From e00320875d0cc5f8099a7227b2f25fbb3231268d Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 14 Feb 2008 08:48:23 +0100 Subject: x86: fix stackprotector canary updates during context switches fix a bug noticed and fixed by pageexec@freemail.hu. if built with -fstack-protector-all then we'll have canary checks built into the __switch_to() function. That does not work well with the canary-switching code there: while we already use the %rsp of the new task, we still call __switch_to() whith the previous task's canary value in the PDA, hence the __switch_to() ssp prologue instructions will store the previous canary. Then we update the PDA and upon return from __switch_to() the canary check triggers and we panic. so update the canary after we have called __switch_to(), where we are at the same stackframe level as the last stackframe of the next (and now freshly current) task. Note: this means that we call __switch_to() [and its sub-functions] still with the old canary, but that is not a problem, both the previous and the next task has a high-quality canary. The only (mostly academic) disadvantage is that the canary of one task may leak onto the stack of another task, increasing the risk of information leaks, were an attacker able to read the stack of specific tasks (but not that of others). To solve this we'll have to reorganize the way we switch tasks, and move the PDA setting into the switch_to() assembly code. That will happen in another patch. Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- include/linux/sched.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index 5395a6176f4b..d6a515158783 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1096,10 +1096,9 @@ struct task_struct { pid_t pid; pid_t tgid; -#ifdef CONFIG_CC_STACKPROTECTOR /* Canary value for the -fstack-protector gcc feature */ unsigned long stack_canary; -#endif + /* * pointers to (original) parent process, youngest child, younger sibling, * older sibling, respectively. (p->father can be replaced with -- cgit v1.2.3-59-g8ed1b From 9b5609fd773e6ac0b1d6d6e1bf68f32cca64e06b Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 14 Feb 2008 09:41:09 +0100 Subject: stackprotector: include files create for core kernel files to include. Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- include/asm-x86/stackprotector.h | 4 ++++ include/linux/stackprotector.h | 8 ++++++++ init/main.c | 1 + 3 files changed, 13 insertions(+) create mode 100644 include/asm-x86/stackprotector.h create mode 100644 include/linux/stackprotector.h (limited to 'include/linux') diff --git a/include/asm-x86/stackprotector.h b/include/asm-x86/stackprotector.h new file mode 100644 index 000000000000..dcac7a6bdba2 --- /dev/null +++ b/include/asm-x86/stackprotector.h @@ -0,0 +1,4 @@ +#ifndef _ASM_STACKPROTECTOR_H +#define _ASM_STACKPROTECTOR_H 1 + +#endif diff --git a/include/linux/stackprotector.h b/include/linux/stackprotector.h new file mode 100644 index 000000000000..d3e8bbe602f8 --- /dev/null +++ b/include/linux/stackprotector.h @@ -0,0 +1,8 @@ +#ifndef _LINUX_STACKPROTECTOR_H +#define _LINUX_STACKPROTECTOR_H 1 + +#ifdef CONFIG_CC_STACKPROTECTOR +# include +#endif + +#endif diff --git a/init/main.c b/init/main.c index f7fb20021d48..a84322ca64a2 100644 --- a/init/main.c +++ b/init/main.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3-59-g8ed1b From 18aa8bb12dcb10adc3d7c9d69714d53667c0ab7f Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 14 Feb 2008 09:42:02 +0100 Subject: stackprotector: add boot_init_stack_canary() add the boot_init_stack_canary() and make the secondary idle threads use it. Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- arch/x86/kernel/process_64.c | 6 ++---- include/asm-x86/stackprotector.h | 20 ++++++++++++++++++++ include/linux/stackprotector.h | 4 ++++ 3 files changed, 26 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index d4c7ac7aa430..5107cb214c7b 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -147,7 +147,6 @@ void cpu_idle(void) { current_thread_info()->status |= TS_POLLING; -#ifdef CONFIG_CC_STACKPROTECTOR /* * If we're the non-boot CPU, nothing set the PDA stack * canary up for us - and if we are the boot CPU we have @@ -156,9 +155,8 @@ void cpu_idle(void) * invalid canaries already on the stack wont ever * trigger): */ - current->stack_canary = get_random_int(); - write_pda(stack_canary, current->stack_canary); -#endif + boot_init_stack_canary(); + /* endless idle loop with no priority at all */ while (1) { tick_nohz_stop_sched_tick(); diff --git a/include/asm-x86/stackprotector.h b/include/asm-x86/stackprotector.h index dcac7a6bdba2..0f91f7a2688c 100644 --- a/include/asm-x86/stackprotector.h +++ b/include/asm-x86/stackprotector.h @@ -1,4 +1,24 @@ #ifndef _ASM_STACKPROTECTOR_H #define _ASM_STACKPROTECTOR_H 1 +/* + * Initialize the stackprotector canary value. + * + * NOTE: this must only be called from functions that never return, + * and it must always be inlined. + */ +static __always_inline void boot_init_stack_canary(void) +{ + /* + * If we're the non-boot CPU, nothing set the PDA stack + * canary up for us - and if we are the boot CPU we have + * a 0 stack canary. This is a good place for updating + * it, as we wont ever return from this function (so the + * invalid canaries already on the stack wont ever + * trigger): + */ + current->stack_canary = get_random_int(); + write_pda(stack_canary, current->stack_canary); +} + #endif diff --git a/include/linux/stackprotector.h b/include/linux/stackprotector.h index d3e8bbe602f8..422e71aafd0b 100644 --- a/include/linux/stackprotector.h +++ b/include/linux/stackprotector.h @@ -3,6 +3,10 @@ #ifdef CONFIG_CC_STACKPROTECTOR # include +#else +static inline void boot_init_stack_canary(void) +{ +} #endif #endif -- cgit v1.2.3-59-g8ed1b From 420594296838fdc9a674470d710cda7d1487f9f4 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 14 Feb 2008 09:44:08 +0100 Subject: x86: fix the stackprotector canary of the boot CPU Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- arch/x86/kernel/process_64.c | 1 + include/linux/stackprotector.h | 4 ++++ init/main.c | 6 ++++++ 3 files changed, 11 insertions(+) (limited to 'include/linux') diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index 5107cb214c7b..cce47f7fbf22 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -16,6 +16,7 @@ #include +#include #include #include #include diff --git a/include/linux/stackprotector.h b/include/linux/stackprotector.h index 422e71aafd0b..6f3e54c704c0 100644 --- a/include/linux/stackprotector.h +++ b/include/linux/stackprotector.h @@ -1,6 +1,10 @@ #ifndef _LINUX_STACKPROTECTOR_H #define _LINUX_STACKPROTECTOR_H 1 +#include +#include +#include + #ifdef CONFIG_CC_STACKPROTECTOR # include #else diff --git a/init/main.c b/init/main.c index a84322ca64a2..b44e4eb0f5e3 100644 --- a/init/main.c +++ b/init/main.c @@ -546,6 +546,12 @@ asmlinkage void __init start_kernel(void) unwind_init(); lockdep_init(); debug_objects_early_init(); + + /* + * Set up the the initial canary ASAP: + */ + boot_init_stack_canary(); + cgroup_init_early(); local_irq_disable(); -- cgit v1.2.3-59-g8ed1b From 7c9f8861e6c9c839f913e49b98c3854daca18f27 Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Tue, 22 Apr 2008 16:38:23 -0500 Subject: stackprotector: use canary at end of stack to indicate overruns at oops time (Updated with a common max-stack-used checker that knows about the canary, as suggested by Joe Perches) Use a canary at the end of the stack to clearly indicate at oops time whether the stack has ever overflowed. This is a very simple implementation with a couple of drawbacks: 1) a thread may legitimately use exactly up to the last word on the stack -- but the chances of doing this and then oopsing later seem slim 2) it's possible that the stack usage isn't dense enough that the canary location could get skipped over -- but the worst that happens is that we don't flag the overrun -- though this happens fairly often in my testing :( With the code in place, an intentionally-bloated stack oops might do: BUG: unable to handle kernel paging request at ffff8103f84cc680 IP: [] update_curr+0x9a/0xa8 PGD 8063 PUD 0 Thread overran stack or stack corrupted Oops: 0000 [1] SMP CPU 0 ... ... unless the stack overrun is so bad that it corrupts some other thread. Signed-off-by: Eric Sandeen Signed-off-by: Ingo Molnar Signed-off-by: Thomas Gleixner --- arch/x86/mm/fault.c | 7 +++++++ include/linux/magic.h | 1 + include/linux/sched.h | 13 +++++++++++++ kernel/exit.c | 5 +---- kernel/fork.c | 5 +++++ kernel/sched.c | 7 +------ 6 files changed, 28 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index fd7e1798c75a..1f524df68b96 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include @@ -581,6 +582,8 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) unsigned long address; int write, si_code; int fault; + unsigned long *stackend; + #ifdef CONFIG_X86_64 unsigned long flags; #endif @@ -850,6 +853,10 @@ no_context: show_fault_oops(regs, error_code, address); + stackend = end_of_stack(tsk); + if (*stackend != STACK_END_MAGIC) + printk(KERN_ALERT "Thread overran stack, or stack corrupted\n"); + tsk->thread.cr2 = address; tsk->thread.trap_no = 14; tsk->thread.error_code = error_code; diff --git a/include/linux/magic.h b/include/linux/magic.h index 1fa0c2ce4dec..74e68e201166 100644 --- a/include/linux/magic.h +++ b/include/linux/magic.h @@ -42,4 +42,5 @@ #define FUTEXFS_SUPER_MAGIC 0xBAD1DEA #define INOTIFYFS_SUPER_MAGIC 0x2BAD1DEA +#define STACK_END_MAGIC 0x57AC6E9D #endif /* __LINUX_MAGIC_H__ */ diff --git a/include/linux/sched.h b/include/linux/sched.h index d6a515158783..c5181e77f305 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1969,6 +1969,19 @@ static inline unsigned long *end_of_stack(struct task_struct *p) extern void thread_info_cache_init(void); +#ifdef CONFIG_DEBUG_STACK_USAGE +static inline unsigned long stack_not_used(struct task_struct *p) +{ + unsigned long *n = end_of_stack(p); + + do { /* Skip over canary */ + n++; + } while (!*n); + + return (unsigned long)n - (unsigned long)end_of_stack(p); +} +#endif + /* set thread flags in other task's structures * - see asm/thread_info.h for TIF_xxxx flags available */ diff --git a/kernel/exit.c b/kernel/exit.c index 8f6185e69b69..fb8de6cbf2c7 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -899,12 +899,9 @@ static void check_stack_usage(void) { static DEFINE_SPINLOCK(low_water_lock); static int lowest_to_date = THREAD_SIZE; - unsigned long *n = end_of_stack(current); unsigned long free; - while (*n == 0) - n++; - free = (unsigned long)n - (unsigned long)end_of_stack(current); + free = stack_not_used(current); if (free >= lowest_to_date) return; diff --git a/kernel/fork.c b/kernel/fork.c index 19908b26cf80..d428336e7aa1 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -54,6 +54,7 @@ #include #include #include +#include #include #include @@ -186,6 +187,8 @@ static struct task_struct *dup_task_struct(struct task_struct *orig) { struct task_struct *tsk; struct thread_info *ti; + unsigned long *stackend; + int err; prepare_to_copy(orig); @@ -211,6 +214,8 @@ static struct task_struct *dup_task_struct(struct task_struct *orig) goto out; setup_thread_stack(tsk, orig); + stackend = end_of_stack(tsk); + *stackend = STACK_END_MAGIC; /* for overflow detection */ #ifdef CONFIG_CC_STACKPROTECTOR tsk->stack_canary = get_random_int(); diff --git a/kernel/sched.c b/kernel/sched.c index cfa222a91539..a964ed945094 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -5748,12 +5748,7 @@ void sched_show_task(struct task_struct *p) printk(KERN_CONT " %016lx ", thread_saved_pc(p)); #endif #ifdef CONFIG_DEBUG_STACK_USAGE - { - unsigned long *n = end_of_stack(p); - while (!*n) - n++; - free = (unsigned long)n - (unsigned long)end_of_stack(p); - } + free = stack_not_used(p); #endif printk(KERN_CONT "%5lu %5d %6d\n", free, task_pid_nr(p), task_pid_nr(p->real_parent)); -- cgit v1.2.3-59-g8ed1b From bc22c17e12c130dc929218a95aa347e0f3fd05dc Mon Sep 17 00:00:00 2001 From: Alain Knaff Date: Sun, 4 Jan 2009 22:46:16 +0100 Subject: bzip2/lzma: library support for gzip, bzip2 and lzma decompression Impact: Replaces inflate.c with a wrapper around zlib_inflate; new library code This is the first part of the bzip2/lzma patch The bzip patch is based on an idea by Christian Ludwig, includes support for compressing the kernel with bzip2 or lzma rather than gzip. Both compressors give smaller sizes than gzip. Lzma's decompresses faster than bzip2. It also supports ramdisks and initramfs' compressed using these two compressors. The functionality has been successfully used for a couple of years by the udpcast project This version applies to "tip" kernel 2.6.28 This part contains: - changed inflate.c to accomodate rest of patch - implementation of bzip2 compression (not used at this stage yet) - implementation of lzma compression (not used at this stage yet) - Makefile routines to support bzip2 and lzma kernel compression Signed-off-by: Alain Knaff Signed-off-by: H. Peter Anvin --- include/linux/decompress/bunzip2.h | 10 + include/linux/decompress/generic.h | 30 ++ include/linux/decompress/inflate.h | 13 + include/linux/decompress/mm.h | 87 +++++ include/linux/decompress/unlzma.h | 12 + lib/decompress_bunzip2.c | 735 +++++++++++++++++++++++++++++++++++++ lib/decompress_inflate.c | 167 +++++++++ lib/decompress_unlzma.c | 647 ++++++++++++++++++++++++++++++++ lib/zlib_inflate/inflate.h | 4 + lib/zlib_inflate/inftrees.h | 4 + scripts/Makefile.lib | 14 + scripts/bin_size | 10 + 12 files changed, 1733 insertions(+) create mode 100644 include/linux/decompress/bunzip2.h create mode 100644 include/linux/decompress/generic.h create mode 100644 include/linux/decompress/inflate.h create mode 100644 include/linux/decompress/mm.h create mode 100644 include/linux/decompress/unlzma.h create mode 100644 lib/decompress_bunzip2.c create mode 100644 lib/decompress_inflate.c create mode 100644 lib/decompress_unlzma.c create mode 100644 scripts/bin_size (limited to 'include/linux') diff --git a/include/linux/decompress/bunzip2.h b/include/linux/decompress/bunzip2.h new file mode 100644 index 000000000000..115272137a9c --- /dev/null +++ b/include/linux/decompress/bunzip2.h @@ -0,0 +1,10 @@ +#ifndef DECOMPRESS_BUNZIP2_H +#define DECOMPRESS_BUNZIP2_H + +int bunzip2(unsigned char *inbuf, int len, + int(*fill)(void*, unsigned int), + int(*flush)(void*, unsigned int), + unsigned char *output, + int *pos, + void(*error)(char *x)); +#endif diff --git a/include/linux/decompress/generic.h b/include/linux/decompress/generic.h new file mode 100644 index 000000000000..f847f514f78e --- /dev/null +++ b/include/linux/decompress/generic.h @@ -0,0 +1,30 @@ +#ifndef DECOMPRESS_GENERIC_H +#define DECOMPRESS_GENERIC_H + +/* Minimal chunksize to be read. + *Bzip2 prefers at least 4096 + *Lzma prefers 0x10000 */ +#define COMPR_IOBUF_SIZE 4096 + +typedef int (*decompress_fn) (unsigned char *inbuf, int len, + int(*fill)(void*, unsigned int), + int(*writebb)(void*, unsigned int), + unsigned char *output, + int *posp, + void(*error)(char *x)); + +/* inbuf - input buffer + *len - len of pre-read data in inbuf + *fill - function to fill inbuf if empty + *writebb - function to write out outbug + *posp - if non-null, input position (number of bytes read) will be + * returned here + * + *If len != 0, the inbuf is initialized (with as much data), and fill + *should not be called + *If len = 0, the inbuf is allocated, but empty. Its size is IOBUF_SIZE + *fill should be called (repeatedly...) to read data, at most IOBUF_SIZE + */ + + +#endif diff --git a/include/linux/decompress/inflate.h b/include/linux/decompress/inflate.h new file mode 100644 index 000000000000..f9b06ccc3e5c --- /dev/null +++ b/include/linux/decompress/inflate.h @@ -0,0 +1,13 @@ +#ifndef INFLATE_H +#define INFLATE_H + +/* Other housekeeping constants */ +#define INBUFSIZ 4096 + +int gunzip(unsigned char *inbuf, int len, + int(*fill)(void*, unsigned int), + int(*flush)(void*, unsigned int), + unsigned char *output, + int *pos, + void(*error_fn)(char *x)); +#endif diff --git a/include/linux/decompress/mm.h b/include/linux/decompress/mm.h new file mode 100644 index 000000000000..12ff8c3f1d05 --- /dev/null +++ b/include/linux/decompress/mm.h @@ -0,0 +1,87 @@ +/* + * linux/compr_mm.h + * + * Memory management for pre-boot and ramdisk uncompressors + * + * Authors: Alain Knaff + * + */ + +#ifndef DECOMPR_MM_H +#define DECOMPR_MM_H + +#ifdef STATIC + +/* Code active when included from pre-boot environment: */ + +/* A trivial malloc implementation, adapted from + * malloc by Hannu Savolainen 1993 and Matthias Urlichs 1994 + */ +static unsigned long malloc_ptr; +static int malloc_count; + +static void *malloc(int size) +{ + void *p; + + if (size < 0) + error("Malloc error"); + if (!malloc_ptr) + malloc_ptr = free_mem_ptr; + + malloc_ptr = (malloc_ptr + 3) & ~3; /* Align */ + + p = (void *)malloc_ptr; + malloc_ptr += size; + + if (free_mem_end_ptr && malloc_ptr >= free_mem_end_ptr) + error("Out of memory"); + + malloc_count++; + return p; +} + +static void free(void *where) +{ + malloc_count--; + if (!malloc_count) + malloc_ptr = free_mem_ptr; +} + +#define large_malloc(a) malloc(a) +#define large_free(a) free(a) + +#define set_error_fn(x) + +#define INIT + +#else /* STATIC */ + +/* Code active when compiled standalone for use when loading ramdisk: */ + +#include +#include +#include +#include + +/* Use defines rather than static inline in order to avoid spurious + * warnings when not needed (indeed large_malloc / large_free are not + * needed by inflate */ + +#define malloc(a) kmalloc(a, GFP_KERNEL) +#define free(a) kfree(a) + +#define large_malloc(a) vmalloc(a) +#define large_free(a) vfree(a) + +static void(*error)(char *m); +#define set_error_fn(x) error = x; + +#define INIT __init +#define STATIC + +#include + +#endif /* STATIC */ + +#endif /* DECOMPR_MM_H */ diff --git a/include/linux/decompress/unlzma.h b/include/linux/decompress/unlzma.h new file mode 100644 index 000000000000..7796538f1bf4 --- /dev/null +++ b/include/linux/decompress/unlzma.h @@ -0,0 +1,12 @@ +#ifndef DECOMPRESS_UNLZMA_H +#define DECOMPRESS_UNLZMA_H + +int unlzma(unsigned char *, int, + int(*fill)(void*, unsigned int), + int(*flush)(void*, unsigned int), + unsigned char *output, + int *posp, + void(*error)(char *x) + ); + +#endif diff --git a/lib/decompress_bunzip2.c b/lib/decompress_bunzip2.c new file mode 100644 index 000000000000..5d3ddb5fcfd9 --- /dev/null +++ b/lib/decompress_bunzip2.c @@ -0,0 +1,735 @@ +/* vi: set sw = 4 ts = 4: */ +/* Small bzip2 deflate implementation, by Rob Landley (rob@landley.net). + + Based on bzip2 decompression code by Julian R Seward (jseward@acm.org), + which also acknowledges contributions by Mike Burrows, David Wheeler, + Peter Fenwick, Alistair Moffat, Radford Neal, Ian H. Witten, + Robert Sedgewick, and Jon L. Bentley. + + This code is licensed under the LGPLv2: + LGPL (http://www.gnu.org/copyleft/lgpl.html +*/ + +/* + Size and speed optimizations by Manuel Novoa III (mjn3@codepoet.org). + + More efficient reading of Huffman codes, a streamlined read_bunzip() + function, and various other tweaks. In (limited) tests, approximately + 20% faster than bzcat on x86 and about 10% faster on arm. + + Note that about 2/3 of the time is spent in read_unzip() reversing + the Burrows-Wheeler transformation. Much of that time is delay + resulting from cache misses. + + I would ask that anyone benefiting from this work, especially those + using it in commercial products, consider making a donation to my local + non-profit hospice organization in the name of the woman I loved, who + passed away Feb. 12, 2003. + + In memory of Toni W. Hagan + + Hospice of Acadiana, Inc. + 2600 Johnston St., Suite 200 + Lafayette, LA 70503-3240 + + Phone (337) 232-1234 or 1-800-738-2226 + Fax (337) 232-1297 + + http://www.hospiceacadiana.com/ + + Manuel + */ + +/* + Made it fit for running in Linux Kernel by Alain Knaff (alain@knaff.lu) +*/ + + +#ifndef STATIC +#include +#endif /* !STATIC */ + +#include + +#ifndef INT_MAX +#define INT_MAX 0x7fffffff +#endif + +/* Constants for Huffman coding */ +#define MAX_GROUPS 6 +#define GROUP_SIZE 50 /* 64 would have been more efficient */ +#define MAX_HUFCODE_BITS 20 /* Longest Huffman code allowed */ +#define MAX_SYMBOLS 258 /* 256 literals + RUNA + RUNB */ +#define SYMBOL_RUNA 0 +#define SYMBOL_RUNB 1 + +/* Status return values */ +#define RETVAL_OK 0 +#define RETVAL_LAST_BLOCK (-1) +#define RETVAL_NOT_BZIP_DATA (-2) +#define RETVAL_UNEXPECTED_INPUT_EOF (-3) +#define RETVAL_UNEXPECTED_OUTPUT_EOF (-4) +#define RETVAL_DATA_ERROR (-5) +#define RETVAL_OUT_OF_MEMORY (-6) +#define RETVAL_OBSOLETE_INPUT (-7) + +/* Other housekeeping constants */ +#define BZIP2_IOBUF_SIZE 4096 + +/* This is what we know about each Huffman coding group */ +struct group_data { + /* We have an extra slot at the end of limit[] for a sentinal value. */ + int limit[MAX_HUFCODE_BITS+1]; + int base[MAX_HUFCODE_BITS]; + int permute[MAX_SYMBOLS]; + int minLen, maxLen; +}; + +/* Structure holding all the housekeeping data, including IO buffers and + memory that persists between calls to bunzip */ +struct bunzip_data { + /* State for interrupting output loop */ + int writeCopies, writePos, writeRunCountdown, writeCount, writeCurrent; + /* I/O tracking data (file handles, buffers, positions, etc.) */ + int (*fill)(void*, unsigned int); + int inbufCount, inbufPos /*, outbufPos*/; + unsigned char *inbuf /*,*outbuf*/; + unsigned int inbufBitCount, inbufBits; + /* The CRC values stored in the block header and calculated from the + data */ + unsigned int crc32Table[256], headerCRC, totalCRC, writeCRC; + /* Intermediate buffer and its size (in bytes) */ + unsigned int *dbuf, dbufSize; + /* These things are a bit too big to go on the stack */ + unsigned char selectors[32768]; /* nSelectors = 15 bits */ + struct group_data groups[MAX_GROUPS]; /* Huffman coding tables */ + int io_error; /* non-zero if we have IO error */ +}; + + +/* Return the next nnn bits of input. All reads from the compressed input + are done through this function. All reads are big endian */ +static unsigned int INIT get_bits(struct bunzip_data *bd, char bits_wanted) +{ + unsigned int bits = 0; + + /* If we need to get more data from the byte buffer, do so. + (Loop getting one byte at a time to enforce endianness and avoid + unaligned access.) */ + while (bd->inbufBitCount < bits_wanted) { + /* If we need to read more data from file into byte buffer, do + so */ + if (bd->inbufPos == bd->inbufCount) { + if (bd->io_error) + return 0; + bd->inbufCount = bd->fill(bd->inbuf, BZIP2_IOBUF_SIZE); + if (bd->inbufCount <= 0) { + bd->io_error = RETVAL_UNEXPECTED_INPUT_EOF; + return 0; + } + bd->inbufPos = 0; + } + /* Avoid 32-bit overflow (dump bit buffer to top of output) */ + if (bd->inbufBitCount >= 24) { + bits = bd->inbufBits&((1 << bd->inbufBitCount)-1); + bits_wanted -= bd->inbufBitCount; + bits <<= bits_wanted; + bd->inbufBitCount = 0; + } + /* Grab next 8 bits of input from buffer. */ + bd->inbufBits = (bd->inbufBits << 8)|bd->inbuf[bd->inbufPos++]; + bd->inbufBitCount += 8; + } + /* Calculate result */ + bd->inbufBitCount -= bits_wanted; + bits |= (bd->inbufBits >> bd->inbufBitCount)&((1 << bits_wanted)-1); + + return bits; +} + +/* Unpacks the next block and sets up for the inverse burrows-wheeler step. */ + +static int INIT get_next_block(struct bunzip_data *bd) +{ + struct group_data *hufGroup = NULL; + int *base = NULL; + int *limit = NULL; + int dbufCount, nextSym, dbufSize, groupCount, selector, + i, j, k, t, runPos, symCount, symTotal, nSelectors, + byteCount[256]; + unsigned char uc, symToByte[256], mtfSymbol[256], *selectors; + unsigned int *dbuf, origPtr; + + dbuf = bd->dbuf; + dbufSize = bd->dbufSize; + selectors = bd->selectors; + + /* Read in header signature and CRC, then validate signature. + (last block signature means CRC is for whole file, return now) */ + i = get_bits(bd, 24); + j = get_bits(bd, 24); + bd->headerCRC = get_bits(bd, 32); + if ((i == 0x177245) && (j == 0x385090)) + return RETVAL_LAST_BLOCK; + if ((i != 0x314159) || (j != 0x265359)) + return RETVAL_NOT_BZIP_DATA; + /* We can add support for blockRandomised if anybody complains. + There was some code for this in busybox 1.0.0-pre3, but nobody ever + noticed that it didn't actually work. */ + if (get_bits(bd, 1)) + return RETVAL_OBSOLETE_INPUT; + origPtr = get_bits(bd, 24); + if (origPtr > dbufSize) + return RETVAL_DATA_ERROR; + /* mapping table: if some byte values are never used (encoding things + like ascii text), the compression code removes the gaps to have fewer + symbols to deal with, and writes a sparse bitfield indicating which + values were present. We make a translation table to convert the + symbols back to the corresponding bytes. */ + t = get_bits(bd, 16); + symTotal = 0; + for (i = 0; i < 16; i++) { + if (t&(1 << (15-i))) { + k = get_bits(bd, 16); + for (j = 0; j < 16; j++) + if (k&(1 << (15-j))) + symToByte[symTotal++] = (16*i)+j; + } + } + /* How many different Huffman coding groups does this block use? */ + groupCount = get_bits(bd, 3); + if (groupCount < 2 || groupCount > MAX_GROUPS) + return RETVAL_DATA_ERROR; + /* nSelectors: Every GROUP_SIZE many symbols we select a new + Huffman coding group. Read in the group selector list, + which is stored as MTF encoded bit runs. (MTF = Move To + Front, as each value is used it's moved to the start of the + list.) */ + nSelectors = get_bits(bd, 15); + if (!nSelectors) + return RETVAL_DATA_ERROR; + for (i = 0; i < groupCount; i++) + mtfSymbol[i] = i; + for (i = 0; i < nSelectors; i++) { + /* Get next value */ + for (j = 0; get_bits(bd, 1); j++) + if (j >= groupCount) + return RETVAL_DATA_ERROR; + /* Decode MTF to get the next selector */ + uc = mtfSymbol[j]; + for (; j; j--) + mtfSymbol[j] = mtfSymbol[j-1]; + mtfSymbol[0] = selectors[i] = uc; + } + /* Read the Huffman coding tables for each group, which code + for symTotal literal symbols, plus two run symbols (RUNA, + RUNB) */ + symCount = symTotal+2; + for (j = 0; j < groupCount; j++) { + unsigned char length[MAX_SYMBOLS], temp[MAX_HUFCODE_BITS+1]; + int minLen, maxLen, pp; + /* Read Huffman code lengths for each symbol. They're + stored in a way similar to mtf; record a starting + value for the first symbol, and an offset from the + previous value for everys symbol after that. + (Subtracting 1 before the loop and then adding it + back at the end is an optimization that makes the + test inside the loop simpler: symbol length 0 + becomes negative, so an unsigned inequality catches + it.) */ + t = get_bits(bd, 5)-1; + for (i = 0; i < symCount; i++) { + for (;;) { + if (((unsigned)t) > (MAX_HUFCODE_BITS-1)) + return RETVAL_DATA_ERROR; + + /* If first bit is 0, stop. Else + second bit indicates whether to + increment or decrement the value. + Optimization: grab 2 bits and unget + the second if the first was 0. */ + + k = get_bits(bd, 2); + if (k < 2) { + bd->inbufBitCount++; + break; + } + /* Add one if second bit 1, else + * subtract 1. Avoids if/else */ + t += (((k+1)&2)-1); + } + /* Correct for the initial -1, to get the + * final symbol length */ + length[i] = t+1; + } + /* Find largest and smallest lengths in this group */ + minLen = maxLen = length[0]; + + for (i = 1; i < symCount; i++) { + if (length[i] > maxLen) + maxLen = length[i]; + else if (length[i] < minLen) + minLen = length[i]; + } + + /* Calculate permute[], base[], and limit[] tables from + * length[]. + * + * permute[] is the lookup table for converting + * Huffman coded symbols into decoded symbols. base[] + * is the amount to subtract from the value of a + * Huffman symbol of a given length when using + * permute[]. + * + * limit[] indicates the largest numerical value a + * symbol with a given number of bits can have. This + * is how the Huffman codes can vary in length: each + * code with a value > limit[length] needs another + * bit. + */ + hufGroup = bd->groups+j; + hufGroup->minLen = minLen; + hufGroup->maxLen = maxLen; + /* Note that minLen can't be smaller than 1, so we + adjust the base and limit array pointers so we're + not always wasting the first entry. We do this + again when using them (during symbol decoding).*/ + base = hufGroup->base-1; + limit = hufGroup->limit-1; + /* Calculate permute[]. Concurently, initialize + * temp[] and limit[]. */ + pp = 0; + for (i = minLen; i <= maxLen; i++) { + temp[i] = limit[i] = 0; + for (t = 0; t < symCount; t++) + if (length[t] == i) + hufGroup->permute[pp++] = t; + } + /* Count symbols coded for at each bit length */ + for (i = 0; i < symCount; i++) + temp[length[i]]++; + /* Calculate limit[] (the largest symbol-coding value + *at each bit length, which is (previous limit << + *1)+symbols at this level), and base[] (number of + *symbols to ignore at each bit length, which is limit + *minus the cumulative count of symbols coded for + *already). */ + pp = t = 0; + for (i = minLen; i < maxLen; i++) { + pp += temp[i]; + /* We read the largest possible symbol size + and then unget bits after determining how + many we need, and those extra bits could be + set to anything. (They're noise from + future symbols.) At each level we're + really only interested in the first few + bits, so here we set all the trailing + to-be-ignored bits to 1 so they don't + affect the value > limit[length] + comparison. */ + limit[i] = (pp << (maxLen - i)) - 1; + pp <<= 1; + base[i+1] = pp-(t += temp[i]); + } + limit[maxLen+1] = INT_MAX; /* Sentinal value for + * reading next sym. */ + limit[maxLen] = pp+temp[maxLen]-1; + base[minLen] = 0; + } + /* We've finished reading and digesting the block header. Now + read this block's Huffman coded symbols from the file and + undo the Huffman coding and run length encoding, saving the + result into dbuf[dbufCount++] = uc */ + + /* Initialize symbol occurrence counters and symbol Move To + * Front table */ + for (i = 0; i < 256; i++) { + byteCount[i] = 0; + mtfSymbol[i] = (unsigned char)i; + } + /* Loop through compressed symbols. */ + runPos = dbufCount = symCount = selector = 0; + for (;;) { + /* Determine which Huffman coding group to use. */ + if (!(symCount--)) { + symCount = GROUP_SIZE-1; + if (selector >= nSelectors) + return RETVAL_DATA_ERROR; + hufGroup = bd->groups+selectors[selector++]; + base = hufGroup->base-1; + limit = hufGroup->limit-1; + } + /* Read next Huffman-coded symbol. */ + /* Note: It is far cheaper to read maxLen bits and + back up than it is to read minLen bits and then an + additional bit at a time, testing as we go. + Because there is a trailing last block (with file + CRC), there is no danger of the overread causing an + unexpected EOF for a valid compressed file. As a + further optimization, we do the read inline + (falling back to a call to get_bits if the buffer + runs dry). The following (up to got_huff_bits:) is + equivalent to j = get_bits(bd, hufGroup->maxLen); + */ + while (bd->inbufBitCount < hufGroup->maxLen) { + if (bd->inbufPos == bd->inbufCount) { + j = get_bits(bd, hufGroup->maxLen); + goto got_huff_bits; + } + bd->inbufBits = + (bd->inbufBits << 8)|bd->inbuf[bd->inbufPos++]; + bd->inbufBitCount += 8; + }; + bd->inbufBitCount -= hufGroup->maxLen; + j = (bd->inbufBits >> bd->inbufBitCount)& + ((1 << hufGroup->maxLen)-1); +got_huff_bits: + /* Figure how how many bits are in next symbol and + * unget extras */ + i = hufGroup->minLen; + while (j > limit[i]) + ++i; + bd->inbufBitCount += (hufGroup->maxLen - i); + /* Huffman decode value to get nextSym (with bounds checking) */ + if ((i > hufGroup->maxLen) + || (((unsigned)(j = (j>>(hufGroup->maxLen-i))-base[i])) + >= MAX_SYMBOLS)) + return RETVAL_DATA_ERROR; + nextSym = hufGroup->permute[j]; + /* We have now decoded the symbol, which indicates + either a new literal byte, or a repeated run of the + most recent literal byte. First, check if nextSym + indicates a repeated run, and if so loop collecting + how many times to repeat the last literal. */ + if (((unsigned)nextSym) <= SYMBOL_RUNB) { /* RUNA or RUNB */ + /* If this is the start of a new run, zero out + * counter */ + if (!runPos) { + runPos = 1; + t = 0; + } + /* Neat trick that saves 1 symbol: instead of + or-ing 0 or 1 at each bit position, add 1 + or 2 instead. For example, 1011 is 1 << 0 + + 1 << 1 + 2 << 2. 1010 is 2 << 0 + 2 << 1 + + 1 << 2. You can make any bit pattern + that way using 1 less symbol than the basic + or 0/1 method (except all bits 0, which + would use no symbols, but a run of length 0 + doesn't mean anything in this context). + Thus space is saved. */ + t += (runPos << nextSym); + /* +runPos if RUNA; +2*runPos if RUNB */ + + runPos <<= 1; + continue; + } + /* When we hit the first non-run symbol after a run, + we now know how many times to repeat the last + literal, so append that many copies to our buffer + of decoded symbols (dbuf) now. (The last literal + used is the one at the head of the mtfSymbol + array.) */ + if (runPos) { + runPos = 0; + if (dbufCount+t >= dbufSize) + return RETVAL_DATA_ERROR; + + uc = symToByte[mtfSymbol[0]]; + byteCount[uc] += t; + while (t--) + dbuf[dbufCount++] = uc; + } + /* Is this the terminating symbol? */ + if (nextSym > symTotal) + break; + /* At this point, nextSym indicates a new literal + character. Subtract one to get the position in the + MTF array at which this literal is currently to be + found. (Note that the result can't be -1 or 0, + because 0 and 1 are RUNA and RUNB. But another + instance of the first symbol in the mtf array, + position 0, would have been handled as part of a + run above. Therefore 1 unused mtf position minus 2 + non-literal nextSym values equals -1.) */ + if (dbufCount >= dbufSize) + return RETVAL_DATA_ERROR; + i = nextSym - 1; + uc = mtfSymbol[i]; + /* Adjust the MTF array. Since we typically expect to + *move only a small number of symbols, and are bound + *by 256 in any case, using memmove here would + *typically be bigger and slower due to function call + *overhead and other assorted setup costs. */ + do { + mtfSymbol[i] = mtfSymbol[i-1]; + } while (--i); + mtfSymbol[0] = uc; + uc = symToByte[uc]; + /* We have our literal byte. Save it into dbuf. */ + byteCount[uc]++; + dbuf[dbufCount++] = (unsigned int)uc; + } + /* At this point, we've read all the Huffman-coded symbols + (and repeated runs) for this block from the input stream, + and decoded them into the intermediate buffer. There are + dbufCount many decoded bytes in dbuf[]. Now undo the + Burrows-Wheeler transform on dbuf. See + http://dogma.net/markn/articles/bwt/bwt.htm + */ + /* Turn byteCount into cumulative occurrence counts of 0 to n-1. */ + j = 0; + for (i = 0; i < 256; i++) { + k = j+byteCount[i]; + byteCount[i] = j; + j = k; + } + /* Figure out what order dbuf would be in if we sorted it. */ + for (i = 0; i < dbufCount; i++) { + uc = (unsigned char)(dbuf[i] & 0xff); + dbuf[byteCount[uc]] |= (i << 8); + byteCount[uc]++; + } + /* Decode first byte by hand to initialize "previous" byte. + Note that it doesn't get output, and if the first three + characters are identical it doesn't qualify as a run (hence + writeRunCountdown = 5). */ + if (dbufCount) { + if (origPtr >= dbufCount) + return RETVAL_DATA_ERROR; + bd->writePos = dbuf[origPtr]; + bd->writeCurrent = (unsigned char)(bd->writePos&0xff); + bd->writePos >>= 8; + bd->writeRunCountdown = 5; + } + bd->writeCount = dbufCount; + + return RETVAL_OK; +} + +/* Undo burrows-wheeler transform on intermediate buffer to produce output. + If start_bunzip was initialized with out_fd =-1, then up to len bytes of + data are written to outbuf. Return value is number of bytes written or + error (all errors are negative numbers). If out_fd!=-1, outbuf and len + are ignored, data is written to out_fd and return is RETVAL_OK or error. +*/ + +static int INIT read_bunzip(struct bunzip_data *bd, char *outbuf, int len) +{ + const unsigned int *dbuf; + int pos, xcurrent, previous, gotcount; + + /* If last read was short due to end of file, return last block now */ + if (bd->writeCount < 0) + return bd->writeCount; + + gotcount = 0; + dbuf = bd->dbuf; + pos = bd->writePos; + xcurrent = bd->writeCurrent; + + /* We will always have pending decoded data to write into the output + buffer unless this is the very first call (in which case we haven't + Huffman-decoded a block into the intermediate buffer yet). */ + + if (bd->writeCopies) { + /* Inside the loop, writeCopies means extra copies (beyond 1) */ + --bd->writeCopies; + /* Loop outputting bytes */ + for (;;) { + /* If the output buffer is full, snapshot + * state and return */ + if (gotcount >= len) { + bd->writePos = pos; + bd->writeCurrent = xcurrent; + bd->writeCopies++; + return len; + } + /* Write next byte into output buffer, updating CRC */ + outbuf[gotcount++] = xcurrent; + bd->writeCRC = (((bd->writeCRC) << 8) + ^bd->crc32Table[((bd->writeCRC) >> 24) + ^xcurrent]); + /* Loop now if we're outputting multiple + * copies of this byte */ + if (bd->writeCopies) { + --bd->writeCopies; + continue; + } +decode_next_byte: + if (!bd->writeCount--) + break; + /* Follow sequence vector to undo + * Burrows-Wheeler transform */ + previous = xcurrent; + pos = dbuf[pos]; + xcurrent = pos&0xff; + pos >>= 8; + /* After 3 consecutive copies of the same + byte, the 4th is a repeat count. We count + down from 4 instead *of counting up because + testing for non-zero is faster */ + if (--bd->writeRunCountdown) { + if (xcurrent != previous) + bd->writeRunCountdown = 4; + } else { + /* We have a repeated run, this byte + * indicates the count */ + bd->writeCopies = xcurrent; + xcurrent = previous; + bd->writeRunCountdown = 5; + /* Sometimes there are just 3 bytes + * (run length 0) */ + if (!bd->writeCopies) + goto decode_next_byte; + /* Subtract the 1 copy we'd output + * anyway to get extras */ + --bd->writeCopies; + } + } + /* Decompression of this block completed successfully */ + bd->writeCRC = ~bd->writeCRC; + bd->totalCRC = ((bd->totalCRC << 1) | + (bd->totalCRC >> 31)) ^ bd->writeCRC; + /* If this block had a CRC error, force file level CRC error. */ + if (bd->writeCRC != bd->headerCRC) { + bd->totalCRC = bd->headerCRC+1; + return RETVAL_LAST_BLOCK; + } + } + + /* Refill the intermediate buffer by Huffman-decoding next + * block of input */ + /* (previous is just a convenient unused temp variable here) */ + previous = get_next_block(bd); + if (previous) { + bd->writeCount = previous; + return (previous != RETVAL_LAST_BLOCK) ? previous : gotcount; + } + bd->writeCRC = 0xffffffffUL; + pos = bd->writePos; + xcurrent = bd->writeCurrent; + goto decode_next_byte; +} + +static int INIT nofill(void *buf, unsigned int len) +{ + return -1; +} + +/* Allocate the structure, read file header. If in_fd ==-1, inbuf must contain + a complete bunzip file (len bytes long). If in_fd!=-1, inbuf and len are + ignored, and data is read from file handle into temporary buffer. */ +static int INIT start_bunzip(struct bunzip_data **bdp, void *inbuf, int len, + int (*fill)(void*, unsigned int)) +{ + struct bunzip_data *bd; + unsigned int i, j, c; + const unsigned int BZh0 = + (((unsigned int)'B') << 24)+(((unsigned int)'Z') << 16) + +(((unsigned int)'h') << 8)+(unsigned int)'0'; + + /* Figure out how much data to allocate */ + i = sizeof(struct bunzip_data); + + /* Allocate bunzip_data. Most fields initialize to zero. */ + bd = *bdp = malloc(i); + memset(bd, 0, sizeof(struct bunzip_data)); + /* Setup input buffer */ + bd->inbuf = inbuf; + bd->inbufCount = len; + if (fill != NULL) + bd->fill = fill; + else + bd->fill = nofill; + + /* Init the CRC32 table (big endian) */ + for (i = 0; i < 256; i++) { + c = i << 24; + for (j = 8; j; j--) + c = c&0x80000000 ? (c << 1)^0x04c11db7 : (c << 1); + bd->crc32Table[i] = c; + } + + /* Ensure that file starts with "BZh['1'-'9']." */ + i = get_bits(bd, 32); + if (((unsigned int)(i-BZh0-1)) >= 9) + return RETVAL_NOT_BZIP_DATA; + + /* Fourth byte (ascii '1'-'9'), indicates block size in units of 100k of + uncompressed data. Allocate intermediate buffer for block. */ + bd->dbufSize = 100000*(i-BZh0); + + bd->dbuf = large_malloc(bd->dbufSize * sizeof(int)); + return RETVAL_OK; +} + +/* Example usage: decompress src_fd to dst_fd. (Stops at end of bzip2 data, + not end of file.) */ +STATIC int INIT bunzip2(unsigned char *buf, int len, + int(*fill)(void*, unsigned int), + int(*flush)(void*, unsigned int), + unsigned char *outbuf, + int *pos, + void(*error_fn)(char *x)) +{ + struct bunzip_data *bd; + int i = -1; + unsigned char *inbuf; + + set_error_fn(error_fn); + if (flush) + outbuf = malloc(BZIP2_IOBUF_SIZE); + else + len -= 4; /* Uncompressed size hack active in pre-boot + environment */ + if (!outbuf) { + error("Could not allocate output bufer"); + return -1; + } + if (buf) + inbuf = buf; + else + inbuf = malloc(BZIP2_IOBUF_SIZE); + if (!inbuf) { + error("Could not allocate input bufer"); + goto exit_0; + } + i = start_bunzip(&bd, inbuf, len, fill); + if (!i) { + for (;;) { + i = read_bunzip(bd, outbuf, BZIP2_IOBUF_SIZE); + if (i <= 0) + break; + if (!flush) + outbuf += i; + else + if (i != flush(outbuf, i)) { + i = RETVAL_UNEXPECTED_OUTPUT_EOF; + break; + } + } + } + /* Check CRC and release memory */ + if (i == RETVAL_LAST_BLOCK) { + if (bd->headerCRC != bd->totalCRC) + error("Data integrity error when decompressing."); + else + i = RETVAL_OK; + } else if (i == RETVAL_UNEXPECTED_OUTPUT_EOF) { + error("Compressed file ends unexpectedly"); + } + if (bd->dbuf) + large_free(bd->dbuf); + if (pos) + *pos = bd->inbufPos; + free(bd); + if (!buf) + free(inbuf); +exit_0: + if (flush) + free(outbuf); + return i; +} + +#define decompress bunzip2 diff --git a/lib/decompress_inflate.c b/lib/decompress_inflate.c new file mode 100644 index 000000000000..163e66aea5f6 --- /dev/null +++ b/lib/decompress_inflate.c @@ -0,0 +1,167 @@ +#ifdef STATIC +/* Pre-boot environment: included */ + +/* prevent inclusion of _LINUX_KERNEL_H in pre-boot environment: lots + * errors about console_printk etc... on ARM */ +#define _LINUX_KERNEL_H + +#include "zlib_inflate/inftrees.c" +#include "zlib_inflate/inffast.c" +#include "zlib_inflate/inflate.c" + +#else /* STATIC */ +/* initramfs et al: linked */ + +#include + +#include "zlib_inflate/inftrees.h" +#include "zlib_inflate/inffast.h" +#include "zlib_inflate/inflate.h" + +#include "zlib_inflate/infutil.h" + +#endif /* STATIC */ + +#include + +#define INBUF_LEN (16*1024) + +/* Included from initramfs et al code */ +STATIC int INIT gunzip(unsigned char *buf, int len, + int(*fill)(void*, unsigned int), + int(*flush)(void*, unsigned int), + unsigned char *out_buf, + int *pos, + void(*error_fn)(char *x)) { + u8 *zbuf; + struct z_stream_s *strm; + int rc; + size_t out_len; + + set_error_fn(error_fn); + rc = -1; + if (flush) { + out_len = 0x8100; /* 32 K */ + out_buf = malloc(out_len); + } else { + out_len = 0x7fffffff; /* no limit */ + } + if (!out_buf) { + error("Out of memory while allocating output buffer"); + goto gunzip_nomem1; + } + + if (buf) + zbuf = buf; + else { + zbuf = malloc(INBUF_LEN); + len = 0; + } + if (!zbuf) { + error("Out of memory while allocating input buffer"); + goto gunzip_nomem2; + } + + strm = malloc(sizeof(*strm)); + if (strm == NULL) { + error("Out of memory while allocating z_stream"); + goto gunzip_nomem3; + } + + strm->workspace = malloc(flush ? zlib_inflate_workspacesize() : + sizeof(struct inflate_state)); + if (strm->workspace == NULL) { + error("Out of memory while allocating workspace"); + goto gunzip_nomem4; + } + + if (len == 0) + len = fill(zbuf, INBUF_LEN); + + /* verify the gzip header */ + if (len < 10 || + zbuf[0] != 0x1f || zbuf[1] != 0x8b || zbuf[2] != 0x08) { + if (pos) + *pos = 0; + error("Not a gzip file"); + goto gunzip_5; + } + + /* skip over gzip header (1f,8b,08... 10 bytes total + + * possible asciz filename) + */ + strm->next_in = zbuf + 10; + /* skip over asciz filename */ + if (zbuf[3] & 0x8) { + while (strm->next_in[0]) + strm->next_in++; + strm->next_in++; + } + strm->avail_in = len - 10; + + strm->next_out = out_buf; + strm->avail_out = out_len; + + rc = zlib_inflateInit2(strm, -MAX_WBITS); + + if (!flush) { + WS(strm)->inflate_state.wsize = 0; + WS(strm)->inflate_state.window = NULL; + } + + while (rc == Z_OK) { + if (strm->avail_in == 0) { + /* TODO: handle case where both pos and fill are set */ + len = fill(zbuf, INBUF_LEN); + if (len < 0) { + rc = -1; + error("read error"); + break; + } + strm->next_in = zbuf; + strm->avail_in = len; + } + rc = zlib_inflate(strm, 0); + + /* Write any data generated */ + if (flush && strm->next_out > out_buf) { + int l = strm->next_out - out_buf; + if (l != flush(out_buf, l)) { + rc = -1; + error("write error"); + break; + } + strm->next_out = out_buf; + strm->avail_out = out_len; + } + + /* after Z_FINISH, only Z_STREAM_END is "we unpacked it all" */ + if (rc == Z_STREAM_END) { + rc = 0; + break; + } else if (rc != Z_OK) { + error("uncompression error"); + rc = -1; + } + } + + zlib_inflateEnd(strm); + if (pos) + /* add + 8 to skip over trailer */ + *pos = strm->next_in - zbuf+8; + +gunzip_5: + free(strm->workspace); +gunzip_nomem4: + free(strm); +gunzip_nomem3: + if (!buf) + free(zbuf); +gunzip_nomem2: + if (flush) + free(out_buf); +gunzip_nomem1: + return rc; /* returns Z_OK (0) if successful */ +} + +#define decompress gunzip diff --git a/lib/decompress_unlzma.c b/lib/decompress_unlzma.c new file mode 100644 index 000000000000..546f2f4c157e --- /dev/null +++ b/lib/decompress_unlzma.c @@ -0,0 +1,647 @@ +/* Lzma decompressor for Linux kernel. Shamelessly snarfed + *from busybox 1.1.1 + * + *Linux kernel adaptation + *Copyright (C) 2006 Alain < alain@knaff.lu > + * + *Based on small lzma deflate implementation/Small range coder + *implementation for lzma. + *Copyright (C) 2006 Aurelien Jacobs < aurel@gnuage.org > + * + *Based on LzmaDecode.c from the LZMA SDK 4.22 (http://www.7-zip.org/) + *Copyright (C) 1999-2005 Igor Pavlov + * + *Copyrights of the parts, see headers below. + * + * + *This program is free software; you can redistribute it and/or + *modify it under the terms of the GNU Lesser General Public + *License as published by the Free Software Foundation; either + *version 2.1 of the License, or (at your option) any later version. + * + *This program is distributed in the hope that it will be useful, + *but WITHOUT ANY WARRANTY; without even the implied warranty of + *MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + *Lesser General Public License for more details. + * + *You should have received a copy of the GNU Lesser General Public + *License along with this library; if not, write to the Free Software + *Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef STATIC +#include +#endif /* STATIC */ + +#include + +#define MIN(a, b) (((a) < (b)) ? (a) : (b)) + +static long long INIT read_int(unsigned char *ptr, int size) +{ + int i; + long long ret = 0; + + for (i = 0; i < size; i++) + ret = (ret << 8) | ptr[size-i-1]; + return ret; +} + +#define ENDIAN_CONVERT(x) \ + x = (typeof(x))read_int((unsigned char *)&x, sizeof(x)) + + +/* Small range coder implementation for lzma. + *Copyright (C) 2006 Aurelien Jacobs < aurel@gnuage.org > + * + *Based on LzmaDecode.c from the LZMA SDK 4.22 (http://www.7-zip.org/) + *Copyright (c) 1999-2005 Igor Pavlov + */ + +#include + +#define LZMA_IOBUF_SIZE 0x10000 + +struct rc { + int (*fill)(void*, unsigned int); + uint8_t *ptr; + uint8_t *buffer; + uint8_t *buffer_end; + int buffer_size; + uint32_t code; + uint32_t range; + uint32_t bound; +}; + + +#define RC_TOP_BITS 24 +#define RC_MOVE_BITS 5 +#define RC_MODEL_TOTAL_BITS 11 + + +/* Called twice: once at startup and once in rc_normalize() */ +static void INIT rc_read(struct rc *rc) +{ + rc->buffer_size = rc->fill((char *)rc->buffer, LZMA_IOBUF_SIZE); + if (rc->buffer_size <= 0) + error("unexpected EOF"); + rc->ptr = rc->buffer; + rc->buffer_end = rc->buffer + rc->buffer_size; +} + +/* Called once */ +static inline void INIT rc_init(struct rc *rc, + int (*fill)(void*, unsigned int), + char *buffer, int buffer_size) +{ + rc->fill = fill; + rc->buffer = (uint8_t *)buffer; + rc->buffer_size = buffer_size; + rc->buffer_end = rc->buffer + rc->buffer_size; + rc->ptr = rc->buffer; + + rc->code = 0; + rc->range = 0xFFFFFFFF; +} + +static inline void INIT rc_init_code(struct rc *rc) +{ + int i; + + for (i = 0; i < 5; i++) { + if (rc->ptr >= rc->buffer_end) + rc_read(rc); + rc->code = (rc->code << 8) | *rc->ptr++; + } +} + + +/* Called once. TODO: bb_maybe_free() */ +static inline void INIT rc_free(struct rc *rc) +{ + free(rc->buffer); +} + +/* Called twice, but one callsite is in inline'd rc_is_bit_0_helper() */ +static void INIT rc_do_normalize(struct rc *rc) +{ + if (rc->ptr >= rc->buffer_end) + rc_read(rc); + rc->range <<= 8; + rc->code = (rc->code << 8) | *rc->ptr++; +} +static inline void INIT rc_normalize(struct rc *rc) +{ + if (rc->range < (1 << RC_TOP_BITS)) + rc_do_normalize(rc); +} + +/* Called 9 times */ +/* Why rc_is_bit_0_helper exists? + *Because we want to always expose (rc->code < rc->bound) to optimizer + */ +static inline uint32_t INIT rc_is_bit_0_helper(struct rc *rc, uint16_t *p) +{ + rc_normalize(rc); + rc->bound = *p * (rc->range >> RC_MODEL_TOTAL_BITS); + return rc->bound; +} +static inline int INIT rc_is_bit_0(struct rc *rc, uint16_t *p) +{ + uint32_t t = rc_is_bit_0_helper(rc, p); + return rc->code < t; +} + +/* Called ~10 times, but very small, thus inlined */ +static inline void INIT rc_update_bit_0(struct rc *rc, uint16_t *p) +{ + rc->range = rc->bound; + *p += ((1 << RC_MODEL_TOTAL_BITS) - *p) >> RC_MOVE_BITS; +} +static inline void rc_update_bit_1(struct rc *rc, uint16_t *p) +{ + rc->range -= rc->bound; + rc->code -= rc->bound; + *p -= *p >> RC_MOVE_BITS; +} + +/* Called 4 times in unlzma loop */ +static int INIT rc_get_bit(struct rc *rc, uint16_t *p, int *symbol) +{ + if (rc_is_bit_0(rc, p)) { + rc_update_bit_0(rc, p); + *symbol *= 2; + return 0; + } else { + rc_update_bit_1(rc, p); + *symbol = *symbol * 2 + 1; + return 1; + } +} + +/* Called once */ +static inline int INIT rc_direct_bit(struct rc *rc) +{ + rc_normalize(rc); + rc->range >>= 1; + if (rc->code >= rc->range) { + rc->code -= rc->range; + return 1; + } + return 0; +} + +/* Called twice */ +static inline void INIT +rc_bit_tree_decode(struct rc *rc, uint16_t *p, int num_levels, int *symbol) +{ + int i = num_levels; + + *symbol = 1; + while (i--) + rc_get_bit(rc, p + *symbol, symbol); + *symbol -= 1 << num_levels; +} + + +/* + * Small lzma deflate implementation. + * Copyright (C) 2006 Aurelien Jacobs < aurel@gnuage.org > + * + * Based on LzmaDecode.c from the LZMA SDK 4.22 (http://www.7-zip.org/) + * Copyright (C) 1999-2005 Igor Pavlov + */ + + +struct lzma_header { + uint8_t pos; + uint32_t dict_size; + uint64_t dst_size; +} __attribute__ ((packed)) ; + + +#define LZMA_BASE_SIZE 1846 +#define LZMA_LIT_SIZE 768 + +#define LZMA_NUM_POS_BITS_MAX 4 + +#define LZMA_LEN_NUM_LOW_BITS 3 +#define LZMA_LEN_NUM_MID_BITS 3 +#define LZMA_LEN_NUM_HIGH_BITS 8 + +#define LZMA_LEN_CHOICE 0 +#define LZMA_LEN_CHOICE_2 (LZMA_LEN_CHOICE + 1) +#define LZMA_LEN_LOW (LZMA_LEN_CHOICE_2 + 1) +#define LZMA_LEN_MID (LZMA_LEN_LOW \ + + (1 << (LZMA_NUM_POS_BITS_MAX + LZMA_LEN_NUM_LOW_BITS))) +#define LZMA_LEN_HIGH (LZMA_LEN_MID \ + +(1 << (LZMA_NUM_POS_BITS_MAX + LZMA_LEN_NUM_MID_BITS))) +#define LZMA_NUM_LEN_PROBS (LZMA_LEN_HIGH + (1 << LZMA_LEN_NUM_HIGH_BITS)) + +#define LZMA_NUM_STATES 12 +#define LZMA_NUM_LIT_STATES 7 + +#define LZMA_START_POS_MODEL_INDEX 4 +#define LZMA_END_POS_MODEL_INDEX 14 +#define LZMA_NUM_FULL_DISTANCES (1 << (LZMA_END_POS_MODEL_INDEX >> 1)) + +#define LZMA_NUM_POS_SLOT_BITS 6 +#define LZMA_NUM_LEN_TO_POS_STATES 4 + +#define LZMA_NUM_ALIGN_BITS 4 + +#define LZMA_MATCH_MIN_LEN 2 + +#define LZMA_IS_MATCH 0 +#define LZMA_IS_REP (LZMA_IS_MATCH + (LZMA_NUM_STATES << LZMA_NUM_POS_BITS_MAX)) +#define LZMA_IS_REP_G0 (LZMA_IS_REP + LZMA_NUM_STATES) +#define LZMA_IS_REP_G1 (LZMA_IS_REP_G0 + LZMA_NUM_STATES) +#define LZMA_IS_REP_G2 (LZMA_IS_REP_G1 + LZMA_NUM_STATES) +#define LZMA_IS_REP_0_LONG (LZMA_IS_REP_G2 + LZMA_NUM_STATES) +#define LZMA_POS_SLOT (LZMA_IS_REP_0_LONG \ + + (LZMA_NUM_STATES << LZMA_NUM_POS_BITS_MAX)) +#define LZMA_SPEC_POS (LZMA_POS_SLOT \ + +(LZMA_NUM_LEN_TO_POS_STATES << LZMA_NUM_POS_SLOT_BITS)) +#define LZMA_ALIGN (LZMA_SPEC_POS \ + + LZMA_NUM_FULL_DISTANCES - LZMA_END_POS_MODEL_INDEX) +#define LZMA_LEN_CODER (LZMA_ALIGN + (1 << LZMA_NUM_ALIGN_BITS)) +#define LZMA_REP_LEN_CODER (LZMA_LEN_CODER + LZMA_NUM_LEN_PROBS) +#define LZMA_LITERAL (LZMA_REP_LEN_CODER + LZMA_NUM_LEN_PROBS) + + +struct writer { + uint8_t *buffer; + uint8_t previous_byte; + size_t buffer_pos; + int bufsize; + size_t global_pos; + int(*flush)(void*, unsigned int); + struct lzma_header *header; +}; + +struct cstate { + int state; + uint32_t rep0, rep1, rep2, rep3; +}; + +static inline size_t INIT get_pos(struct writer *wr) +{ + return + wr->global_pos + wr->buffer_pos; +} + +static inline uint8_t INIT peek_old_byte(struct writer *wr, + uint32_t offs) +{ + if (!wr->flush) { + int32_t pos; + while (offs > wr->header->dict_size) + offs -= wr->header->dict_size; + pos = wr->buffer_pos - offs; + return wr->buffer[pos]; + } else { + uint32_t pos = wr->buffer_pos - offs; + while (pos >= wr->header->dict_size) + pos += wr->header->dict_size; + return wr->buffer[pos]; + } + +} + +static inline void INIT write_byte(struct writer *wr, uint8_t byte) +{ + wr->buffer[wr->buffer_pos++] = wr->previous_byte = byte; + if (wr->flush && wr->buffer_pos == wr->header->dict_size) { + wr->buffer_pos = 0; + wr->global_pos += wr->header->dict_size; + wr->flush((char *)wr->buffer, wr->header->dict_size); + } +} + + +static inline void INIT copy_byte(struct writer *wr, uint32_t offs) +{ + write_byte(wr, peek_old_byte(wr, offs)); +} + +static inline void INIT copy_bytes(struct writer *wr, + uint32_t rep0, int len) +{ + do { + copy_byte(wr, rep0); + len--; + } while (len != 0 && wr->buffer_pos < wr->header->dst_size); +} + +static inline void INIT process_bit0(struct writer *wr, struct rc *rc, + struct cstate *cst, uint16_t *p, + int pos_state, uint16_t *prob, + int lc, uint32_t literal_pos_mask) { + int mi = 1; + rc_update_bit_0(rc, prob); + prob = (p + LZMA_LITERAL + + (LZMA_LIT_SIZE + * (((get_pos(wr) & literal_pos_mask) << lc) + + (wr->previous_byte >> (8 - lc)))) + ); + + if (cst->state >= LZMA_NUM_LIT_STATES) { + int match_byte = peek_old_byte(wr, cst->rep0); + do { + int bit; + uint16_t *prob_lit; + + match_byte <<= 1; + bit = match_byte & 0x100; + prob_lit = prob + 0x100 + bit + mi; + if (rc_get_bit(rc, prob_lit, &mi)) { + if (!bit) + break; + } else { + if (bit) + break; + } + } while (mi < 0x100); + } + while (mi < 0x100) { + uint16_t *prob_lit = prob + mi; + rc_get_bit(rc, prob_lit, &mi); + } + write_byte(wr, mi); + if (cst->state < 4) + cst->state = 0; + else if (cst->state < 10) + cst->state -= 3; + else + cst->state -= 6; +} + +static inline void INIT process_bit1(struct writer *wr, struct rc *rc, + struct cstate *cst, uint16_t *p, + int pos_state, uint16_t *prob) { + int offset; + uint16_t *prob_len; + int num_bits; + int len; + + rc_update_bit_1(rc, prob); + prob = p + LZMA_IS_REP + cst->state; + if (rc_is_bit_0(rc, prob)) { + rc_update_bit_0(rc, prob); + cst->rep3 = cst->rep2; + cst->rep2 = cst->rep1; + cst->rep1 = cst->rep0; + cst->state = cst->state < LZMA_NUM_LIT_STATES ? 0 : 3; + prob = p + LZMA_LEN_CODER; + } else { + rc_update_bit_1(rc, prob); + prob = p + LZMA_IS_REP_G0 + cst->state; + if (rc_is_bit_0(rc, prob)) { + rc_update_bit_0(rc, prob); + prob = (p + LZMA_IS_REP_0_LONG + + (cst->state << + LZMA_NUM_POS_BITS_MAX) + + pos_state); + if (rc_is_bit_0(rc, prob)) { + rc_update_bit_0(rc, prob); + + cst->state = cst->state < LZMA_NUM_LIT_STATES ? + 9 : 11; + copy_byte(wr, cst->rep0); + return; + } else { + rc_update_bit_1(rc, prob); + } + } else { + uint32_t distance; + + rc_update_bit_1(rc, prob); + prob = p + LZMA_IS_REP_G1 + cst->state; + if (rc_is_bit_0(rc, prob)) { + rc_update_bit_0(rc, prob); + distance = cst->rep1; + } else { + rc_update_bit_1(rc, prob); + prob = p + LZMA_IS_REP_G2 + cst->state; + if (rc_is_bit_0(rc, prob)) { + rc_update_bit_0(rc, prob); + distance = cst->rep2; + } else { + rc_update_bit_1(rc, prob); + distance = cst->rep3; + cst->rep3 = cst->rep2; + } + cst->rep2 = cst->rep1; + } + cst->rep1 = cst->rep0; + cst->rep0 = distance; + } + cst->state = cst->state < LZMA_NUM_LIT_STATES ? 8 : 11; + prob = p + LZMA_REP_LEN_CODER; + } + + prob_len = prob + LZMA_LEN_CHOICE; + if (rc_is_bit_0(rc, prob_len)) { + rc_update_bit_0(rc, prob_len); + prob_len = (prob + LZMA_LEN_LOW + + (pos_state << + LZMA_LEN_NUM_LOW_BITS)); + offset = 0; + num_bits = LZMA_LEN_NUM_LOW_BITS; + } else { + rc_update_bit_1(rc, prob_len); + prob_len = prob + LZMA_LEN_CHOICE_2; + if (rc_is_bit_0(rc, prob_len)) { + rc_update_bit_0(rc, prob_len); + prob_len = (prob + LZMA_LEN_MID + + (pos_state << + LZMA_LEN_NUM_MID_BITS)); + offset = 1 << LZMA_LEN_NUM_LOW_BITS; + num_bits = LZMA_LEN_NUM_MID_BITS; + } else { + rc_update_bit_1(rc, prob_len); + prob_len = prob + LZMA_LEN_HIGH; + offset = ((1 << LZMA_LEN_NUM_LOW_BITS) + + (1 << LZMA_LEN_NUM_MID_BITS)); + num_bits = LZMA_LEN_NUM_HIGH_BITS; + } + } + + rc_bit_tree_decode(rc, prob_len, num_bits, &len); + len += offset; + + if (cst->state < 4) { + int pos_slot; + + cst->state += LZMA_NUM_LIT_STATES; + prob = + p + LZMA_POS_SLOT + + ((len < + LZMA_NUM_LEN_TO_POS_STATES ? len : + LZMA_NUM_LEN_TO_POS_STATES - 1) + << LZMA_NUM_POS_SLOT_BITS); + rc_bit_tree_decode(rc, prob, + LZMA_NUM_POS_SLOT_BITS, + &pos_slot); + if (pos_slot >= LZMA_START_POS_MODEL_INDEX) { + int i, mi; + num_bits = (pos_slot >> 1) - 1; + cst->rep0 = 2 | (pos_slot & 1); + if (pos_slot < LZMA_END_POS_MODEL_INDEX) { + cst->rep0 <<= num_bits; + prob = p + LZMA_SPEC_POS + + cst->rep0 - pos_slot - 1; + } else { + num_bits -= LZMA_NUM_ALIGN_BITS; + while (num_bits--) + cst->rep0 = (cst->rep0 << 1) | + rc_direct_bit(rc); + prob = p + LZMA_ALIGN; + cst->rep0 <<= LZMA_NUM_ALIGN_BITS; + num_bits = LZMA_NUM_ALIGN_BITS; + } + i = 1; + mi = 1; + while (num_bits--) { + if (rc_get_bit(rc, prob + mi, &mi)) + cst->rep0 |= i; + i <<= 1; + } + } else + cst->rep0 = pos_slot; + if (++(cst->rep0) == 0) + return; + } + + len += LZMA_MATCH_MIN_LEN; + + copy_bytes(wr, cst->rep0, len); +} + + + +STATIC inline int INIT unlzma(unsigned char *buf, int in_len, + int(*fill)(void*, unsigned int), + int(*flush)(void*, unsigned int), + unsigned char *output, + int *posp, + void(*error_fn)(char *x) + ) +{ + struct lzma_header header; + int lc, pb, lp; + uint32_t pos_state_mask; + uint32_t literal_pos_mask; + uint16_t *p; + int num_probs; + struct rc rc; + int i, mi; + struct writer wr; + struct cstate cst; + unsigned char *inbuf; + int ret = -1; + + set_error_fn(error_fn); + if (!flush) + in_len -= 4; /* Uncompressed size hack active in pre-boot + environment */ + if (buf) + inbuf = buf; + else + inbuf = malloc(LZMA_IOBUF_SIZE); + if (!inbuf) { + error("Could not allocate input bufer"); + goto exit_0; + } + + cst.state = 0; + cst.rep0 = cst.rep1 = cst.rep2 = cst.rep3 = 1; + + wr.header = &header; + wr.flush = flush; + wr.global_pos = 0; + wr.previous_byte = 0; + wr.buffer_pos = 0; + + rc_init(&rc, fill, inbuf, in_len); + + for (i = 0; i < sizeof(header); i++) { + if (rc.ptr >= rc.buffer_end) + rc_read(&rc); + ((unsigned char *)&header)[i] = *rc.ptr++; + } + + if (header.pos >= (9 * 5 * 5)) + error("bad header"); + + mi = 0; + lc = header.pos; + while (lc >= 9) { + mi++; + lc -= 9; + } + pb = 0; + lp = mi; + while (lp >= 5) { + pb++; + lp -= 5; + } + pos_state_mask = (1 << pb) - 1; + literal_pos_mask = (1 << lp) - 1; + + ENDIAN_CONVERT(header.dict_size); + ENDIAN_CONVERT(header.dst_size); + + if (header.dict_size == 0) + header.dict_size = 1; + + if (output) + wr.buffer = output; + else { + wr.bufsize = MIN(header.dst_size, header.dict_size); + wr.buffer = large_malloc(wr.bufsize); + } + if (wr.buffer == NULL) + goto exit_1; + + num_probs = LZMA_BASE_SIZE + (LZMA_LIT_SIZE << (lc + lp)); + p = (uint16_t *) large_malloc(num_probs * sizeof(*p)); + if (p == 0) + goto exit_2; + num_probs = LZMA_LITERAL + (LZMA_LIT_SIZE << (lc + lp)); + for (i = 0; i < num_probs; i++) + p[i] = (1 << RC_MODEL_TOTAL_BITS) >> 1; + + rc_init_code(&rc); + + while (get_pos(&wr) < header.dst_size) { + int pos_state = get_pos(&wr) & pos_state_mask; + uint16_t *prob = p + LZMA_IS_MATCH + + (cst.state << LZMA_NUM_POS_BITS_MAX) + pos_state; + if (rc_is_bit_0(&rc, prob)) + process_bit0(&wr, &rc, &cst, p, pos_state, prob, + lc, literal_pos_mask); + else { + process_bit1(&wr, &rc, &cst, p, pos_state, prob); + if (cst.rep0 == 0) + break; + } + } + + if (posp) + *posp = rc.ptr-rc.buffer; + if (wr.flush) + wr.flush(wr.buffer, wr.buffer_pos); + ret = 0; + large_free(p); +exit_2: + if (!output) + large_free(wr.buffer); +exit_1: + if (!buf) + free(inbuf); +exit_0: + return ret; +} + +#define decompress unlzma diff --git a/lib/zlib_inflate/inflate.h b/lib/zlib_inflate/inflate.h index df8a6c92052d..3d17b3d1b21f 100644 --- a/lib/zlib_inflate/inflate.h +++ b/lib/zlib_inflate/inflate.h @@ -1,3 +1,6 @@ +#ifndef INFLATE_H +#define INFLATE_H + /* inflate.h -- internal inflate state definition * Copyright (C) 1995-2004 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h @@ -105,3 +108,4 @@ struct inflate_state { unsigned short work[288]; /* work area for code table building */ code codes[ENOUGH]; /* space for code tables */ }; +#endif diff --git a/lib/zlib_inflate/inftrees.h b/lib/zlib_inflate/inftrees.h index 5f5219b1240e..b70b4731ac7a 100644 --- a/lib/zlib_inflate/inftrees.h +++ b/lib/zlib_inflate/inftrees.h @@ -1,3 +1,6 @@ +#ifndef INFTREES_H +#define INFTREES_H + /* inftrees.h -- header to use inftrees.c * Copyright (C) 1995-2005 Mark Adler * For conditions of distribution and use, see copyright notice in zlib.h @@ -53,3 +56,4 @@ typedef enum { extern int zlib_inflate_table (codetype type, unsigned short *lens, unsigned codes, code **table, unsigned *bits, unsigned short *work); +#endif diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib index e06365775bdf..70b4676e3b99 100644 --- a/scripts/Makefile.lib +++ b/scripts/Makefile.lib @@ -186,3 +186,17 @@ quiet_cmd_gzip = GZIP $@ cmd_gzip = gzip -f -9 < $< > $@ +# Bzip2 +# --------------------------------------------------------------------------- + +# Bzip2 does not include size in file... so we have to fake that +size_append=$(CONFIG_SHELL) $(srctree)/scripts/bin_size + +quiet_cmd_bzip2 = BZIP2 $@ +cmd_bzip2 = (bzip2 -9 < $< ; $(size_append) $<) > $@ || (rm -f $@ ; false) + +# Lzma +# --------------------------------------------------------------------------- + +quiet_cmd_lzma = LZMA $@ +cmd_lzma = (lzma -9 -c $< ; $(size_append) $<) >$@ || (rm -f $@ ; false) diff --git a/scripts/bin_size b/scripts/bin_size new file mode 100644 index 000000000000..43e1b360cee6 --- /dev/null +++ b/scripts/bin_size @@ -0,0 +1,10 @@ +#!/bin/sh + +if [ $# = 0 ] ; then + echo Usage: $0 file +fi + +size_dec=`stat -c "%s" $1` +size_hex_echo_string=`printf "%08x" $size_dec | + sed 's/\(..\)\(..\)\(..\)\(..\)/\\\\x\4\\\\x\3\\\\x\2\\\\x\1/g'` +/bin/echo -ne $size_hex_echo_string -- cgit v1.2.3-59-g8ed1b From 889c92d21db40be0b7d22a59395060237895bb85 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Thu, 8 Jan 2009 15:14:17 -0800 Subject: bzip2/lzma: centralize format detection Centralize the compression format detection to a common routine in the lib directory, and use it for both initramfs and initrd. Signed-off-by: H. Peter Anvin --- include/linux/decompress/generic.h | 3 +++ init/do_mounts_rd.c | 38 ++++++----------------------- init/initramfs.c | 39 ++++++----------------------- lib/Makefile | 9 ++++--- lib/decompress.c | 50 ++++++++++++++++++++++++++++++++++++++ 5 files changed, 72 insertions(+), 67 deletions(-) create mode 100644 lib/decompress.c (limited to 'include/linux') diff --git a/include/linux/decompress/generic.h b/include/linux/decompress/generic.h index f847f514f78e..6dfb856327bb 100644 --- a/include/linux/decompress/generic.h +++ b/include/linux/decompress/generic.h @@ -26,5 +26,8 @@ typedef int (*decompress_fn) (unsigned char *inbuf, int len, *fill should be called (repeatedly...) to read data, at most IOBUF_SIZE */ +/* Utility routine to detect the decompression method */ +decompress_fn decompress_method(const unsigned char *inbuf, int len, + const char **name); #endif diff --git a/init/do_mounts_rd.c b/init/do_mounts_rd.c index 9c9d7dbcf9ca..a06ed4f92e0e 100644 --- a/init/do_mounts_rd.c +++ b/init/do_mounts_rd.c @@ -12,9 +12,6 @@ #include -#include -#include -#include int __initdata rd_prompt = 1;/* 1 = prompt for RAM disk, 0 = don't prompt */ @@ -49,24 +46,6 @@ static int __init crd_load(int in_fd, int out_fd, decompress_fn deco); * cramfs * gzip */ -static const struct compress_format { - unsigned char magic[2]; - const char *name; - decompress_fn decompressor; -} compressed_formats[] = { -#ifdef CONFIG_RD_GZIP - { {037, 0213}, "gzip", gunzip }, - { {037, 0236}, "gzip", gunzip }, -#endif -#ifdef CONFIG_RD_BZIP2 - { {0x42, 0x5a}, "bzip2", bunzip2 }, -#endif -#ifdef CONFIG_RD_LZMA - { {0x5d, 0x00}, "lzma", unlzma }, -#endif - { {0, 0}, NULL, NULL } -}; - static int __init identify_ramdisk_image(int fd, int start_block, decompress_fn *decompressor) { @@ -77,7 +56,7 @@ identify_ramdisk_image(int fd, int start_block, decompress_fn *decompressor) struct cramfs_super *cramfsb; int nblocks = -1; unsigned char *buf; - const struct compress_format *cf; + const char *compress_name; buf = kmalloc(size, GFP_KERNEL); if (!buf) @@ -95,15 +74,12 @@ identify_ramdisk_image(int fd, int start_block, decompress_fn *decompressor) sys_lseek(fd, start_block * BLOCK_SIZE, 0); sys_read(fd, buf, size); - for (cf = compressed_formats; cf->decompressor; cf++) { - if (buf[0] == cf->magic[0] && buf[1] == cf->magic[1]) { - printk(KERN_NOTICE - "RAMDISK: %s image found at block %d\n", - cf->name, start_block); - *decompressor = cf->decompressor; - nblocks = 0; - goto done; - } + *decompressor = decompress_method(buf, size, &compress_name); + if (*decompressor) { + printk(KERN_NOTICE "RAMDISK: %s image found at block %d\n", + compress_name, start_block); + nblocks = 0; + goto done; } /* romfs is at block zero too */ diff --git a/init/initramfs.c b/init/initramfs.c index a3ba91cdab89..2f42984e5582 100644 --- a/init/initramfs.c +++ b/init/initramfs.c @@ -416,13 +416,13 @@ static int __init flush_buffer(void *bufv, unsigned len) static unsigned my_inptr; /* index of next byte to be processed in inbuf */ -#include -#include -#include +#include static char * __init unpack_to_rootfs(char *buf, unsigned len, int check_only) { int written; + decompress_fn decompress; + dry_run = check_only; header_buf = kmalloc(110, GFP_KERNEL); symlink_buf = kmalloc(PATH_MAX + N_ALIGN(PATH_MAX) + 1, GFP_KERNEL); @@ -450,35 +450,10 @@ static char * __init unpack_to_rootfs(char *buf, unsigned len, int check_only) continue; } this_header = 0; -#ifdef CONFIG_RD_GZIP - if (!gunzip(buf, len, NULL, flush_buffer, NULL, - &my_inptr, error) && - message == NULL) - goto ok; -#endif - -#ifdef CONFIG_RD_BZIP2 - message = NULL; /* Zero out message, or else cpio will - think an error has already occured */ - if (!bunzip2(buf, len, NULL, flush_buffer, NULL, - &my_inptr, error) && - message == NULL) { - goto ok; - } -#endif - -#ifdef CONFIG_RD_LZMA - message = NULL; /* Zero out message, or else cpio will - think an error has already occured */ - if (!unlzma(buf, len, NULL, flush_buffer, NULL, - &my_inptr, error) && - message == NULL) { - goto ok; - } -#endif -#if defined CONFIG_RD_GZIP || defined CONFIG_RD_BZIP2 || defined CONFIG_RD_LZMA -ok: -#endif + decompress = decompress_method(buf, len, NULL); + if (decompress) + decompress(buf, len, NULL, flush_buffer, NULL, + &my_inptr, error); if (state != Reset) error("junk in compressed archive"); this_header = saved_offset + my_inptr; diff --git a/lib/Makefile b/lib/Makefile index d9ac5a414fa7..790de7c25d0d 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -11,7 +11,8 @@ lib-y := ctype.o string.o vsprintf.o cmdline.o \ rbtree.o radix-tree.o dump_stack.o \ idr.o int_sqrt.o extable.o prio_tree.o \ sha1.o irq_regs.o reciprocal_div.o argv_split.o \ - proportions.o prio_heap.o ratelimit.o show_mem.o is_single_threaded.o + proportions.o prio_heap.o ratelimit.o show_mem.o \ + is_single_threaded.o decompress.o lib-$(CONFIG_MMU) += ioremap.o lib-$(CONFIG_SMP) += cpumask.o @@ -65,9 +66,9 @@ obj-$(CONFIG_REED_SOLOMON) += reed_solomon/ obj-$(CONFIG_LZO_COMPRESS) += lzo/ obj-$(CONFIG_LZO_DECOMPRESS) += lzo/ -obj-$(CONFIG_DECOMPRESS_GZIP) += decompress_inflate.o -obj-$(CONFIG_DECOMPRESS_BZIP2) += decompress_bunzip2.o -obj-$(CONFIG_DECOMPRESS_LZMA) += decompress_unlzma.o +lib-$(CONFIG_DECOMPRESS_GZIP) += decompress_inflate.o +lib-$(CONFIG_DECOMPRESS_BZIP2) += decompress_bunzip2.o +lib-$(CONFIG_DECOMPRESS_LZMA) += decompress_unlzma.o obj-$(CONFIG_TEXTSEARCH) += textsearch.o obj-$(CONFIG_TEXTSEARCH_KMP) += ts_kmp.o diff --git a/lib/decompress.c b/lib/decompress.c new file mode 100644 index 000000000000..edac55cc7823 --- /dev/null +++ b/lib/decompress.c @@ -0,0 +1,50 @@ +/* + * decompress.c + * + * Detect the decompression method based on magic number + */ + +#include + +#include +#include +#include + +#include +#include + +static const struct compress_format { + unsigned char magic[2]; + const char *name; + decompress_fn decompressor; +} compressed_formats[] = { +#ifdef CONFIG_DECOMPRESS_GZIP + { {037, 0213}, "gzip", gunzip }, + { {037, 0236}, "gzip", gunzip }, +#endif +#ifdef CONFIG_DECOMPRESS_BZIP2 + { {0x42, 0x5a}, "bzip2", bunzip2 }, +#endif +#ifdef CONFIG_DECOMPRESS_LZMA + { {0x5d, 0x00}, "lzma", unlzma }, +#endif + { {0, 0}, NULL, NULL } +}; + +decompress_fn decompress_method(const unsigned char *inbuf, int len, + const char **name) +{ + const struct compress_format *cf; + + if (len < 2) + return NULL; /* Need at least this much... */ + + for (cf = compressed_formats; cf->decompressor; cf++) { + if (!memcmp(inbuf, cf->magic, 2)) + break; + + } + if (name) + *name = cf->name; + return cf->decompressor; +} -- cgit v1.2.3-59-g8ed1b From 7f7ace0cda64c99599c23785f8979a072e118058 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Sat, 10 Jan 2009 21:58:08 -0800 Subject: cpumask: update irq_desc to use cpumask_var_t Impact: reduce memory usage, use new cpumask API. Replace the affinity and pending_masks with cpumask_var_t's. This adds to the significant size reduction done with the SPARSE_IRQS changes. The added functions (init_alloc_desc_masks & init_copy_desc_masks) are in the include file so they can be inlined (and optimized out for the !CONFIG_CPUMASKS_OFFSTACK case.) [Naming chosen to be consistent with the other init*irq functions, as well as the backwards arg declaration of "from, to" instead of the more common "to, from" standard.] Includes a slight change to the declaration of struct irq_desc to embed the pending_mask within ifdef(CONFIG_SMP) to be consistent with other references, and some small changes to Xen. Tested: sparse/non-sparse/cpumask_offstack/non-cpumask_offstack/nonuma/nosmp on x86_64 Signed-off-by: Mike Travis Cc: Chris Wright Cc: Jeremy Fitzhardinge Cc: KOSAKI Motohiro Cc: Venkatesh Pallipadi Cc: virtualization@lists.osdl.org Cc: xen-devel@lists.xensource.com Cc: Yinghai Lu --- arch/x86/kernel/io_apic.c | 20 ++++++------ arch/x86/kernel/irq_32.c | 2 +- arch/x86/kernel/irq_64.c | 2 +- drivers/xen/events.c | 4 +-- include/linux/irq.h | 81 +++++++++++++++++++++++++++++++++++++++++++++-- kernel/irq/chip.c | 5 ++- kernel/irq/handle.c | 26 ++++++++------- kernel/irq/manage.c | 12 +++---- kernel/irq/migration.c | 12 +++---- kernel/irq/numa_migrate.c | 12 ++++++- kernel/irq/proc.c | 4 +-- 11 files changed, 135 insertions(+), 45 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 1c4a1302536c..1337eab60ecc 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -356,7 +356,7 @@ set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask) if (!cfg->move_in_progress) { /* it means that domain is not changed */ - if (!cpumask_intersects(&desc->affinity, mask)) + if (!cpumask_intersects(desc->affinity, mask)) cfg->move_desc_pending = 1; } } @@ -579,9 +579,9 @@ set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask) if (assign_irq_vector(irq, cfg, mask)) return BAD_APICID; - cpumask_and(&desc->affinity, cfg->domain, mask); + cpumask_and(desc->affinity, cfg->domain, mask); set_extra_move_desc(desc, mask); - return cpu_mask_to_apicid_and(&desc->affinity, cpu_online_mask); + return cpu_mask_to_apicid_and(desc->affinity, cpu_online_mask); } static void @@ -2383,7 +2383,7 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) if (cfg->move_in_progress) send_cleanup_vector(cfg); - cpumask_copy(&desc->affinity, mask); + cpumask_copy(desc->affinity, mask); } static int migrate_irq_remapped_level_desc(struct irq_desc *desc) @@ -2405,11 +2405,11 @@ static int migrate_irq_remapped_level_desc(struct irq_desc *desc) } /* everthing is clear. we have right of way */ - migrate_ioapic_irq_desc(desc, &desc->pending_mask); + migrate_ioapic_irq_desc(desc, desc->pending_mask); ret = 0; desc->status &= ~IRQ_MOVE_PENDING; - cpumask_clear(&desc->pending_mask); + cpumask_clear(desc->pending_mask); unmask: unmask_IO_APIC_irq_desc(desc); @@ -2434,7 +2434,7 @@ static void ir_irq_migration(struct work_struct *work) continue; } - desc->chip->set_affinity(irq, &desc->pending_mask); + desc->chip->set_affinity(irq, desc->pending_mask); spin_unlock_irqrestore(&desc->lock, flags); } } @@ -2448,7 +2448,7 @@ static void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, { if (desc->status & IRQ_LEVEL) { desc->status |= IRQ_MOVE_PENDING; - cpumask_copy(&desc->pending_mask, mask); + cpumask_copy(desc->pending_mask, mask); migrate_irq_remapped_level_desc(desc); return; } @@ -2516,7 +2516,7 @@ static void irq_complete_move(struct irq_desc **descp) /* domain has not changed, but affinity did */ me = smp_processor_id(); - if (cpu_isset(me, desc->affinity)) { + if (cpumask_test_cpu(me, desc->affinity)) { *descp = desc = move_irq_desc(desc, me); /* get the new one */ cfg = desc->chip_data; @@ -4039,7 +4039,7 @@ void __init setup_ioapic_dest(void) */ if (desc->status & (IRQ_NO_BALANCING | IRQ_AFFINITY_SET)) - mask = &desc->affinity; + mask = desc->affinity; else mask = TARGET_CPUS; diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c index 74b9ff7341e9..e0f29be8ab0b 100644 --- a/arch/x86/kernel/irq_32.c +++ b/arch/x86/kernel/irq_32.c @@ -248,7 +248,7 @@ void fixup_irqs(void) if (irq == 2) continue; - affinity = &desc->affinity; + affinity = desc->affinity; if (cpumask_any_and(affinity, cpu_online_mask) >= nr_cpu_ids) { printk("Breaking affinity for irq %i\n", irq); affinity = cpu_all_mask; diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c index 63c88e6ec025..0b21cb1ea11f 100644 --- a/arch/x86/kernel/irq_64.c +++ b/arch/x86/kernel/irq_64.c @@ -100,7 +100,7 @@ void fixup_irqs(void) /* interrupt's are disabled at this point */ spin_lock(&desc->lock); - affinity = &desc->affinity; + affinity = desc->affinity; if (!irq_has_action(irq) || cpumask_equal(affinity, cpu_online_mask)) { spin_unlock(&desc->lock); diff --git a/drivers/xen/events.c b/drivers/xen/events.c index eb0dfdeaa949..e0767ff35d6c 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -125,7 +125,7 @@ static void bind_evtchn_to_cpu(unsigned int chn, unsigned int cpu) BUG_ON(irq == -1); #ifdef CONFIG_SMP - irq_to_desc(irq)->affinity = cpumask_of_cpu(cpu); + cpumask_copy(irq_to_desc(irq)->affinity, cpumask_of(cpu)); #endif __clear_bit(chn, cpu_evtchn_mask[cpu_evtchn[chn]]); @@ -142,7 +142,7 @@ static void init_evtchn_cpu_bindings(void) /* By default all event channels notify CPU#0. */ for_each_irq_desc(i, desc) { - desc->affinity = cpumask_of_cpu(0); + cpumask_copy(desc->affinity, cpumask_of(0)); } #endif diff --git a/include/linux/irq.h b/include/linux/irq.h index f899b502f186..fa27210f1dfd 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -182,11 +182,11 @@ struct irq_desc { unsigned int irqs_unhandled; spinlock_t lock; #ifdef CONFIG_SMP - cpumask_t affinity; + cpumask_var_t affinity; unsigned int cpu; -#endif #ifdef CONFIG_GENERIC_PENDING_IRQ - cpumask_t pending_mask; + cpumask_var_t pending_mask; +#endif #endif #ifdef CONFIG_PROC_FS struct proc_dir_entry *dir; @@ -422,4 +422,79 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry); #endif /* !CONFIG_S390 */ +#ifdef CONFIG_SMP +/** + * init_alloc_desc_masks - allocate cpumasks for irq_desc + * @desc: pointer to irq_desc struct + * @boot: true if need bootmem + * + * Allocates affinity and pending_mask cpumask if required. + * Returns true if successful (or not required). + * Side effect: affinity has all bits set, pending_mask has all bits clear. + */ +static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node, + bool boot) +{ + if (boot) { + alloc_bootmem_cpumask_var(&desc->affinity); + cpumask_setall(desc->affinity); + +#ifdef CONFIG_GENERIC_PENDING_IRQ + alloc_bootmem_cpumask_var(&desc->pending_mask); + cpumask_clear(desc->pending_mask); +#endif + return true; + } + + if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node)) + return false; + cpumask_setall(desc->affinity); + +#ifdef CONFIG_GENERIC_PENDING_IRQ + if (!alloc_cpumask_var_node(&desc->pending_mask, GFP_ATOMIC, node)) { + free_cpumask_var(desc->affinity); + return false; + } + cpumask_clear(desc->pending_mask); +#endif + return true; +} + +/** + * init_copy_desc_masks - copy cpumasks for irq_desc + * @old_desc: pointer to old irq_desc struct + * @new_desc: pointer to new irq_desc struct + * + * Insures affinity and pending_masks are copied to new irq_desc. + * If !CONFIG_CPUMASKS_OFFSTACK the cpumasks are embedded in the + * irq_desc struct so the copy is redundant. + */ + +static inline void init_copy_desc_masks(struct irq_desc *old_desc, + struct irq_desc *new_desc) +{ +#ifdef CONFIG_CPUMASKS_OFFSTACK + cpumask_copy(new_desc->affinity, old_desc->affinity); + +#ifdef CONFIG_GENERIC_PENDING_IRQ + cpumask_copy(new_desc->pending_mask, old_desc->pending_mask); +#endif +#endif +} + +#else /* !CONFIG_SMP */ + +static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node, + bool boot) +{ + return true; +} + +static inline void init_copy_desc_masks(struct irq_desc *old_desc, + struct irq_desc *new_desc) +{ +} + +#endif /* CONFIG_SMP */ + #endif /* _LINUX_IRQ_H */ diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index f63c706d25e1..c248eba98b43 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -46,7 +46,10 @@ void dynamic_irq_init(unsigned int irq) desc->irq_count = 0; desc->irqs_unhandled = 0; #ifdef CONFIG_SMP - cpumask_setall(&desc->affinity); + cpumask_setall(desc->affinity); +#ifdef CONFIG_GENERIC_PENDING_IRQ + cpumask_clear(desc->pending_mask); +#endif #endif spin_unlock_irqrestore(&desc->lock, flags); } diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index c20db0be9173..b8fa1354f01c 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -64,9 +64,6 @@ static struct irq_desc irq_desc_init = { .handle_irq = handle_bad_irq, .depth = 1, .lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock), -#ifdef CONFIG_SMP - .affinity = CPU_MASK_ALL -#endif }; void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) @@ -88,6 +85,8 @@ void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) { + int node = cpu_to_node(cpu); + memcpy(desc, &irq_desc_init, sizeof(struct irq_desc)); spin_lock_init(&desc->lock); @@ -101,6 +100,10 @@ static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) printk(KERN_ERR "can not alloc kstat_irqs\n"); BUG_ON(1); } + if (!init_alloc_desc_masks(desc, node, false)) { + printk(KERN_ERR "can not alloc irq_desc cpumasks\n"); + BUG_ON(1); + } arch_init_chip_data(desc, cpu); } @@ -119,9 +122,6 @@ static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_sm .handle_irq = handle_bad_irq, .depth = 1, .lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock), -#ifdef CONFIG_SMP - .affinity = CPU_MASK_ALL -#endif } }; @@ -141,7 +141,7 @@ int __init early_irq_init(void) desc[i].irq = i; desc[i].kstat_irqs = kstat_irqs_legacy[i]; lockdep_set_class(&desc[i].lock, &irq_desc_lock_class); - + init_alloc_desc_masks(&desc[i], 0, true); irq_desc_ptrs[i] = desc + i; } @@ -188,6 +188,10 @@ struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) printk(KERN_ERR "can not alloc irq_desc\n"); BUG_ON(1); } + if (!init_alloc_desc_masks(desc, node, false)) { + printk(KERN_ERR "can not alloc irq_desc cpumasks\n"); + BUG_ON(1); + } init_one_irq_desc(irq, desc, cpu); irq_desc_ptrs[irq] = desc; @@ -207,9 +211,6 @@ struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { .handle_irq = handle_bad_irq, .depth = 1, .lock = __SPIN_LOCK_UNLOCKED(irq_desc->lock), -#ifdef CONFIG_SMP - .affinity = CPU_MASK_ALL -#endif } }; @@ -222,9 +223,10 @@ int __init early_irq_init(void) desc = irq_desc; count = ARRAY_SIZE(irq_desc); - for (i = 0; i < count; i++) + for (i = 0; i < count; i++) { desc[i].irq = i; - + init_alloc_desc_masks(&desc[i], 0, true); + } return arch_early_irq_init(); } diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index cd0cd8dcb345..b98739af4558 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -98,14 +98,14 @@ int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask) #ifdef CONFIG_GENERIC_PENDING_IRQ if (desc->status & IRQ_MOVE_PCNTXT || desc->status & IRQ_DISABLED) { - cpumask_copy(&desc->affinity, cpumask); + cpumask_copy(desc->affinity, cpumask); desc->chip->set_affinity(irq, cpumask); } else { desc->status |= IRQ_MOVE_PENDING; - cpumask_copy(&desc->pending_mask, cpumask); + cpumask_copy(desc->pending_mask, cpumask); } #else - cpumask_copy(&desc->affinity, cpumask); + cpumask_copy(desc->affinity, cpumask); desc->chip->set_affinity(irq, cpumask); #endif desc->status |= IRQ_AFFINITY_SET; @@ -127,16 +127,16 @@ int do_irq_select_affinity(unsigned int irq, struct irq_desc *desc) * one of the targets is online. */ if (desc->status & (IRQ_AFFINITY_SET | IRQ_NO_BALANCING)) { - if (cpumask_any_and(&desc->affinity, cpu_online_mask) + if (cpumask_any_and(desc->affinity, cpu_online_mask) < nr_cpu_ids) goto set_affinity; else desc->status &= ~IRQ_AFFINITY_SET; } - cpumask_and(&desc->affinity, cpu_online_mask, irq_default_affinity); + cpumask_and(desc->affinity, cpu_online_mask, irq_default_affinity); set_affinity: - desc->chip->set_affinity(irq, &desc->affinity); + desc->chip->set_affinity(irq, desc->affinity); return 0; } diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index bd72329e630c..e05ad9be43b7 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -18,7 +18,7 @@ void move_masked_irq(int irq) desc->status &= ~IRQ_MOVE_PENDING; - if (unlikely(cpumask_empty(&desc->pending_mask))) + if (unlikely(cpumask_empty(desc->pending_mask))) return; if (!desc->chip->set_affinity) @@ -38,13 +38,13 @@ void move_masked_irq(int irq) * For correct operation this depends on the caller * masking the irqs. */ - if (likely(cpumask_any_and(&desc->pending_mask, cpu_online_mask) + if (likely(cpumask_any_and(desc->pending_mask, cpu_online_mask) < nr_cpu_ids)) { - cpumask_and(&desc->affinity, - &desc->pending_mask, cpu_online_mask); - desc->chip->set_affinity(irq, &desc->affinity); + cpumask_and(desc->affinity, + desc->pending_mask, cpu_online_mask); + desc->chip->set_affinity(irq, desc->affinity); } - cpumask_clear(&desc->pending_mask); + cpumask_clear(desc->pending_mask); } void move_native_irq(int irq) diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index ecf765c6a77a..f001a4ea6414 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -46,6 +46,7 @@ static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, desc->cpu = cpu; lockdep_set_class(&desc->lock, &irq_desc_lock_class); init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids); + init_copy_desc_masks(old_desc, desc); arch_init_copy_chip_data(old_desc, desc, cpu); } @@ -76,11 +77,20 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, node = cpu_to_node(cpu); desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); if (!desc) { - printk(KERN_ERR "irq %d: can not get new irq_desc for migration.\n", irq); + printk(KERN_ERR "irq %d: can not get new irq_desc " + "for migration.\n", irq); /* still use old one */ desc = old_desc; goto out_unlock; } + if (!init_alloc_desc_masks(desc, node, false)) { + printk(KERN_ERR "irq %d: can not get new irq_desc cpumask " + "for migration.\n", irq); + /* still use old one */ + kfree(desc); + desc = old_desc; + goto out_unlock; + } init_copy_one_irq_desc(irq, old_desc, desc, cpu); irq_desc_ptrs[irq] = desc; diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index aae3f742bcec..692363dd591f 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -20,11 +20,11 @@ static struct proc_dir_entry *root_irq_dir; static int irq_affinity_proc_show(struct seq_file *m, void *v) { struct irq_desc *desc = irq_to_desc((long)m->private); - const struct cpumask *mask = &desc->affinity; + const struct cpumask *mask = desc->affinity; #ifdef CONFIG_GENERIC_PENDING_IRQ if (desc->status & IRQ_MOVE_PENDING) - mask = &desc->pending_mask; + mask = desc->pending_mask; #endif seq_cpumask(m, mask); seq_putc(m, '\n'); -- cgit v1.2.3-59-g8ed1b From fbd59a8d1f7cf325fdb6828659f1fb76631e87b3 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Sat, 10 Jan 2009 21:58:08 -0800 Subject: cpumask: Use topology_core_cpumask()/topology_thread_cpumask() Impact: reduce stack usage, use new cpumask API. This actually uses topology_core_cpumask() and topology_thread_cpumask(), removing the only users of topology_core_siblings() and topology_thread_siblings() Signed-off-by: Rusty Russell Signed-off-by: Mike Travis Cc: linux-net-drivers@solarflare.com --- Documentation/cputopology.txt | 6 +++--- drivers/base/topology.c | 33 ++++++++++++++++----------------- drivers/net/sfc/efx.c | 4 ++-- include/linux/topology.h | 6 ++++++ 4 files changed, 27 insertions(+), 22 deletions(-) (limited to 'include/linux') diff --git a/Documentation/cputopology.txt b/Documentation/cputopology.txt index 45932ec21cee..b41f3e58aefa 100644 --- a/Documentation/cputopology.txt +++ b/Documentation/cputopology.txt @@ -18,11 +18,11 @@ For an architecture to support this feature, it must define some of these macros in include/asm-XXX/topology.h: #define topology_physical_package_id(cpu) #define topology_core_id(cpu) -#define topology_thread_siblings(cpu) -#define topology_core_siblings(cpu) +#define topology_thread_cpumask(cpu) +#define topology_core_cpumask(cpu) The type of **_id is int. -The type of siblings is cpumask_t. +The type of siblings is (const) struct cpumask *. To be consistent on all architectures, include/linux/topology.h provides default definitions for any of the above macros that are diff --git a/drivers/base/topology.c b/drivers/base/topology.c index a778fb52b11f..bf6b13206d00 100644 --- a/drivers/base/topology.c +++ b/drivers/base/topology.c @@ -31,7 +31,10 @@ #include #include -#define define_one_ro(_name) \ +#define define_one_ro_named(_name, _func) \ +static SYSDEV_ATTR(_name, 0444, _func, NULL) + +#define define_one_ro(_name) \ static SYSDEV_ATTR(_name, 0444, show_##_name, NULL) #define define_id_show_func(name) \ @@ -42,8 +45,8 @@ static ssize_t show_##name(struct sys_device *dev, \ return sprintf(buf, "%d\n", topology_##name(cpu)); \ } -#if defined(topology_thread_siblings) || defined(topology_core_siblings) -static ssize_t show_cpumap(int type, cpumask_t *mask, char *buf) +#if defined(topology_thread_cpumask) || defined(topology_core_cpumask) +static ssize_t show_cpumap(int type, const struct cpumask *mask, char *buf) { ptrdiff_t len = PTR_ALIGN(buf + PAGE_SIZE - 1, PAGE_SIZE) - buf; int n = 0; @@ -65,7 +68,7 @@ static ssize_t show_##name(struct sys_device *dev, \ struct sysdev_attribute *attr, char *buf) \ { \ unsigned int cpu = dev->id; \ - return show_cpumap(0, &(topology_##name(cpu)), buf); \ + return show_cpumap(0, topology_##name(cpu), buf); \ } #define define_siblings_show_list(name) \ @@ -74,7 +77,7 @@ static ssize_t show_##name##_list(struct sys_device *dev, \ char *buf) \ { \ unsigned int cpu = dev->id; \ - return show_cpumap(1, &(topology_##name(cpu)), buf); \ + return show_cpumap(1, topology_##name(cpu), buf); \ } #else @@ -82,9 +85,7 @@ static ssize_t show_##name##_list(struct sys_device *dev, \ static ssize_t show_##name(struct sys_device *dev, \ struct sysdev_attribute *attr, char *buf) \ { \ - unsigned int cpu = dev->id; \ - cpumask_t mask = topology_##name(cpu); \ - return show_cpumap(0, &mask, buf); \ + return show_cpumap(0, topology_##name(dev->id), buf); \ } #define define_siblings_show_list(name) \ @@ -92,9 +93,7 @@ static ssize_t show_##name##_list(struct sys_device *dev, \ struct sysdev_attribute *attr, \ char *buf) \ { \ - unsigned int cpu = dev->id; \ - cpumask_t mask = topology_##name(cpu); \ - return show_cpumap(1, &mask, buf); \ + return show_cpumap(1, topology_##name(dev->id), buf); \ } #endif @@ -107,13 +106,13 @@ define_one_ro(physical_package_id); define_id_show_func(core_id); define_one_ro(core_id); -define_siblings_show_func(thread_siblings); -define_one_ro(thread_siblings); -define_one_ro(thread_siblings_list); +define_siblings_show_func(thread_cpumask); +define_one_ro_named(thread_siblings, show_thread_cpumask); +define_one_ro_named(thread_siblings_list, show_thread_cpumask_list); -define_siblings_show_func(core_siblings); -define_one_ro(core_siblings); -define_one_ro(core_siblings_list); +define_siblings_show_func(core_cpumask); +define_one_ro_named(core_siblings, show_core_cpumask); +define_one_ro_named(core_siblings_list, show_core_cpumask_list); static struct attribute *default_attrs[] = { &attr_physical_package_id.attr, diff --git a/drivers/net/sfc/efx.c b/drivers/net/sfc/efx.c index 7673fd92eaf5..f2e56ceee0ea 100644 --- a/drivers/net/sfc/efx.c +++ b/drivers/net/sfc/efx.c @@ -863,8 +863,8 @@ static int efx_wanted_rx_queues(void) for_each_online_cpu(cpu) { if (!cpu_isset(cpu, core_mask)) { ++count; - cpus_or(core_mask, core_mask, - topology_core_siblings(cpu)); + cpumask_or(&core_mask, &core_mask, + topology_core_cpumask(cpu)); } } diff --git a/include/linux/topology.h b/include/linux/topology.h index e632d29f0544..a16b9e06f2e5 100644 --- a/include/linux/topology.h +++ b/include/linux/topology.h @@ -193,5 +193,11 @@ int arch_update_cpu_topology(void); #ifndef topology_core_siblings #define topology_core_siblings(cpu) cpumask_of_cpu(cpu) #endif +#ifndef topology_thread_cpumask +#define topology_thread_cpumask(cpu) cpumask_of(cpu) +#endif +#ifndef topology_core_cpumask +#define topology_core_cpumask(cpu) cpumask_of(cpu) +#endif #endif /* _LINUX_TOPOLOGY_H */ -- cgit v1.2.3-59-g8ed1b From 802bf931f2688ad125b73db597ce63cc842fb27a Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Sat, 10 Jan 2009 21:58:09 -0800 Subject: cpumask: fix bug in use cpumask_var_t in irq_desc Impact: fix bug where new irq_desc uses old cpumask pointers which are freed. As Yinghai pointed out, init_copy_one_irq_desc() copies the old desc to the new desc overwriting the cpumask pointers. Since the old_desc and the cpumask pointers are freed, then memory corruption will occur if these old pointers are used. Move the allocation of these pointers to after the copy. Signed-off-by: Mike Travis Cc: Yinghai Lu --- include/linux/irq.h | 9 +++++++-- kernel/irq/handle.c | 8 +------- kernel/irq/numa_migrate.c | 13 ++++++++----- 3 files changed, 16 insertions(+), 14 deletions(-) (limited to 'include/linux') diff --git a/include/linux/irq.h b/include/linux/irq.h index fa27210f1dfd..27a67536511e 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -426,15 +426,18 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry); /** * init_alloc_desc_masks - allocate cpumasks for irq_desc * @desc: pointer to irq_desc struct + * @cpu: cpu which will be handling the cpumasks * @boot: true if need bootmem * * Allocates affinity and pending_mask cpumask if required. * Returns true if successful (or not required). * Side effect: affinity has all bits set, pending_mask has all bits clear. */ -static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node, +static inline bool init_alloc_desc_masks(struct irq_desc *desc, int cpu, bool boot) { + int node; + if (boot) { alloc_bootmem_cpumask_var(&desc->affinity); cpumask_setall(desc->affinity); @@ -446,6 +449,8 @@ static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node, return true; } + node = cpu_to_node(cpu); + if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node)) return false; cpumask_setall(desc->affinity); @@ -484,7 +489,7 @@ static inline void init_copy_desc_masks(struct irq_desc *old_desc, #else /* !CONFIG_SMP */ -static inline bool init_alloc_desc_masks(struct irq_desc *desc, int node, +static inline bool init_alloc_desc_masks(struct irq_desc *desc, int cpu, bool boot) { return true; diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index b8fa1354f01c..f01c0a30cb42 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -85,8 +85,6 @@ void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) { - int node = cpu_to_node(cpu); - memcpy(desc, &irq_desc_init, sizeof(struct irq_desc)); spin_lock_init(&desc->lock); @@ -100,7 +98,7 @@ static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) printk(KERN_ERR "can not alloc kstat_irqs\n"); BUG_ON(1); } - if (!init_alloc_desc_masks(desc, node, false)) { + if (!init_alloc_desc_masks(desc, cpu, false)) { printk(KERN_ERR "can not alloc irq_desc cpumasks\n"); BUG_ON(1); } @@ -188,10 +186,6 @@ struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) printk(KERN_ERR "can not alloc irq_desc\n"); BUG_ON(1); } - if (!init_alloc_desc_masks(desc, node, false)) { - printk(KERN_ERR "can not alloc irq_desc cpumasks\n"); - BUG_ON(1); - } init_one_irq_desc(irq, desc, cpu); irq_desc_ptrs[irq] = desc; diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index f001a4ea6414..666260e4c065 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -38,16 +38,22 @@ static void free_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc) old_desc->kstat_irqs = NULL; } -static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, +static bool init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, struct irq_desc *desc, int cpu) { memcpy(desc, old_desc, sizeof(struct irq_desc)); + if (!init_alloc_desc_masks(desc, cpu, false)) { + printk(KERN_ERR "irq %d: can not get new irq_desc cpumask " + "for migration.\n", irq); + return false; + } spin_lock_init(&desc->lock); desc->cpu = cpu; lockdep_set_class(&desc->lock, &irq_desc_lock_class); init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids); init_copy_desc_masks(old_desc, desc); arch_init_copy_chip_data(old_desc, desc, cpu); + return true; } static void free_one_irq_desc(struct irq_desc *old_desc, struct irq_desc *desc) @@ -83,15 +89,12 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, desc = old_desc; goto out_unlock; } - if (!init_alloc_desc_masks(desc, node, false)) { - printk(KERN_ERR "irq %d: can not get new irq_desc cpumask " - "for migration.\n", irq); + if (!init_copy_one_irq_desc(irq, old_desc, desc, cpu)) { /* still use old one */ kfree(desc); desc = old_desc; goto out_unlock; } - init_copy_one_irq_desc(irq, old_desc, desc, cpu); irq_desc_ptrs[irq] = desc; -- cgit v1.2.3-59-g8ed1b From 9332fccdedf8e09448f3b69b624211ae879f6c45 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Sat, 10 Jan 2009 22:24:07 -0800 Subject: irq: initialize nr_irqs based on nr_cpu_ids Impact: Reduce memory usage. This is the second half of the changes to make the irq_desc_ptrs be variable sized based on nr_cpu_ids. This is done by adding a new "max_nr_irqs" macro to irq_vectors.h (and a dummy in irqnr.h) to return a max NR_IRQS value based on NR_CPUS or nr_cpu_ids. This necessitated moving the define of MAX_IO_APICS to a separate file (asm/apicnum.h) so it could be included without the baggage of the other asm/apicdef.h declarations. Signed-off-by: Mike Travis --- arch/x86/include/asm/apicdef.h | 8 ++------ arch/x86/include/asm/apicnum.h | 12 ++++++++++++ arch/x86/include/asm/irq_vectors.h | 16 +++++++++++----- include/linux/irqnr.h | 7 +++++++ kernel/irq/handle.c | 3 +++ 5 files changed, 35 insertions(+), 11 deletions(-) create mode 100644 arch/x86/include/asm/apicnum.h (limited to 'include/linux') diff --git a/arch/x86/include/asm/apicdef.h b/arch/x86/include/asm/apicdef.h index 63134e31e8b9..1a6454ef7f6c 100644 --- a/arch/x86/include/asm/apicdef.h +++ b/arch/x86/include/asm/apicdef.h @@ -132,12 +132,8 @@ #define APIC_BASE_MSR 0x800 #define X2APIC_ENABLE (1UL << 10) -#ifdef CONFIG_X86_32 -# define MAX_IO_APICS 64 -#else -# define MAX_IO_APICS 128 -# define MAX_LOCAL_APIC 32768 -#endif +/* get MAX_IO_APICS */ +#include /* * All x86-64 systems are xAPIC compatible. diff --git a/arch/x86/include/asm/apicnum.h b/arch/x86/include/asm/apicnum.h new file mode 100644 index 000000000000..82f613c607ce --- /dev/null +++ b/arch/x86/include/asm/apicnum.h @@ -0,0 +1,12 @@ +#ifndef _ASM_X86_APICNUM_H +#define _ASM_X86_APICNUM_H + +/* define MAX_IO_APICS */ +#ifdef CONFIG_X86_32 +# define MAX_IO_APICS 64 +#else +# define MAX_IO_APICS 128 +# define MAX_LOCAL_APIC 32768 +#endif + +#endif /* _ASM_X86_APICNUM_H */ diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h index f7ff65032b9d..602361ad0e74 100644 --- a/arch/x86/include/asm/irq_vectors.h +++ b/arch/x86/include/asm/irq_vectors.h @@ -105,6 +105,8 @@ #if defined(CONFIG_X86_IO_APIC) && !defined(CONFIG_X86_VOYAGER) +#include /* need MAX_IO_APICS */ + #ifndef CONFIG_SPARSE_IRQ # if NR_CPUS < MAX_IO_APICS # define NR_IRQS (NR_VECTORS + (32 * NR_CPUS)) @@ -112,11 +114,15 @@ # define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS)) # endif #else -# if (8 * NR_CPUS) > (32 * MAX_IO_APICS) -# define NR_IRQS (NR_VECTORS + (8 * NR_CPUS)) -# else -# define NR_IRQS (NR_VECTORS + (32 * MAX_IO_APICS)) -# endif + +/* defined as a macro so nr_irqs = max_nr_irqs(nr_cpu_ids) can be used */ +# define max_nr_irqs(nr_cpus) \ + ((8 * nr_cpus) > (32 * MAX_IO_APICS) ? \ + (NR_VECTORS + (8 * NR_CPUS)) : \ + (NR_VECTORS + (32 * MAX_IO_APICS))) \ + +# define NR_IRQS max_nr_irqs(NR_CPUS) + #endif #elif defined(CONFIG_X86_VOYAGER) diff --git a/include/linux/irqnr.h b/include/linux/irqnr.h index 86af92e9e84c..de66e4e10406 100644 --- a/include/linux/irqnr.h +++ b/include/linux/irqnr.h @@ -20,11 +20,18 @@ # define for_each_irq_desc_reverse(irq, desc) \ for (irq = nr_irqs - 1; irq >= 0; irq--) + #else /* CONFIG_GENERIC_HARDIRQS */ +#include /* need possible max_nr_irqs() */ + extern int nr_irqs; extern struct irq_desc *irq_to_desc(unsigned int irq); +# ifndef max_nr_irqs +# define max_nr_irqs(nr_cpus) NR_IRQS +# endif + # define for_each_irq_desc(irq, desc) \ for (irq = 0, desc = irq_to_desc(irq); irq < nr_irqs; \ irq++, desc = irq_to_desc(irq)) \ diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index d0b8f7e72790..ebba7a116f14 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -133,6 +133,9 @@ int __init early_irq_init(void) int legacy_count; int i; + /* initialize nr_irqs based on nr_cpu_ids */ + nr_irqs = max_nr_irqs(nr_cpu_ids); + printk(KERN_INFO "NR_IRQS:%d nr_irqs:%d\n", NR_IRQS, nr_irqs); desc = irq_desc_legacy; -- cgit v1.2.3-59-g8ed1b From 92296c6d6e908c35fca287a21af27be814af9c75 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Sun, 11 Jan 2009 09:22:58 -0800 Subject: cpumask, irq: non-x86 build failures Ingo Molnar wrote: > All non-x86 architectures fail to build: > > In file included from /home/mingo/tip/include/linux/random.h:11, > from /home/mingo/tip/include/linux/stackprotector.h:6, > from /home/mingo/tip/init/main.c:17: > /home/mingo/tip/include/linux/irqnr.h:26:63: error: asm/irq_vectors.h: No such file or directory Do not include asm/irq_vectors.h in generic code - it's not available on all architectures. Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apicdef.h | 8 ++++++-- include/linux/irqnr.h | 6 ------ kernel/irq/handle.c | 5 +++++ 3 files changed, 11 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/include/asm/apicdef.h b/arch/x86/include/asm/apicdef.h index 1a6454ef7f6c..63134e31e8b9 100644 --- a/arch/x86/include/asm/apicdef.h +++ b/arch/x86/include/asm/apicdef.h @@ -132,8 +132,12 @@ #define APIC_BASE_MSR 0x800 #define X2APIC_ENABLE (1UL << 10) -/* get MAX_IO_APICS */ -#include +#ifdef CONFIG_X86_32 +# define MAX_IO_APICS 64 +#else +# define MAX_IO_APICS 128 +# define MAX_LOCAL_APIC 32768 +#endif /* * All x86-64 systems are xAPIC compatible. diff --git a/include/linux/irqnr.h b/include/linux/irqnr.h index de66e4e10406..887477bc2ab0 100644 --- a/include/linux/irqnr.h +++ b/include/linux/irqnr.h @@ -23,15 +23,9 @@ #else /* CONFIG_GENERIC_HARDIRQS */ -#include /* need possible max_nr_irqs() */ - extern int nr_irqs; extern struct irq_desc *irq_to_desc(unsigned int irq); -# ifndef max_nr_irqs -# define max_nr_irqs(nr_cpus) NR_IRQS -# endif - # define for_each_irq_desc(irq, desc) \ for (irq = 0, desc = irq_to_desc(irq); irq < nr_irqs; \ irq++, desc = irq_to_desc(irq)) \ diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index b39f32ac8f80..04d3e46031e5 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -58,6 +58,11 @@ int nr_irqs = NR_IRQS; EXPORT_SYMBOL_GPL(nr_irqs); #ifdef CONFIG_SPARSE_IRQ + +#ifndef max_nr_irqs +#define max_nr_irqs(nr_cpus) NR_IRQS +#endif + static struct irq_desc irq_desc_init = { .irq = -1, .status = IRQ_DISABLED, -- cgit v1.2.3-59-g8ed1b From 4a046d1754ee6ebb6f399696805ed61ea0444d4c Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 12 Jan 2009 17:39:24 -0800 Subject: x86: arch_probe_nr_irqs Impact: save RAM with large NR_CPUS, get smaller nr_irqs Signed-off-by: Yinghai Lu Signed-off-by: Mike Travis --- arch/x86/include/asm/irq_vectors.h | 7 ++----- arch/x86/kernel/io_apic.c | 16 ++++++++++++++++ include/linux/interrupt.h | 1 + kernel/irq/handle.c | 9 ++------- kernel/softirq.c | 5 +++++ 5 files changed, 26 insertions(+), 12 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h index 602361ad0e74..a16a2ab2b429 100644 --- a/arch/x86/include/asm/irq_vectors.h +++ b/arch/x86/include/asm/irq_vectors.h @@ -115,14 +115,11 @@ # endif #else -/* defined as a macro so nr_irqs = max_nr_irqs(nr_cpu_ids) can be used */ -# define max_nr_irqs(nr_cpus) \ - ((8 * nr_cpus) > (32 * MAX_IO_APICS) ? \ +# define NR_IRQS \ + ((8 * NR_CPUS) > (32 * MAX_IO_APICS) ? \ (NR_VECTORS + (8 * NR_CPUS)) : \ (NR_VECTORS + (32 * MAX_IO_APICS))) \ -# define NR_IRQS max_nr_irqs(NR_CPUS) - #endif #elif defined(CONFIG_X86_VOYAGER) diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index ae80638012de..157986916cd1 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -3850,6 +3850,22 @@ void __init probe_nr_irqs_gsi(void) nr_irqs_gsi = nr; } +#ifdef CONFIG_SPARSE_IRQ +int __init arch_probe_nr_irqs(void) +{ + int nr; + + nr = ((8 * nr_cpu_ids) > (32 * nr_ioapics) ? + (NR_VECTORS + (8 * nr_cpu_ids)) : + (NR_VECTORS + (32 * nr_ioapics))); + + if (nr < nr_irqs && nr > nr_irqs_gsi) + nr_irqs = nr; + + return 0; +} +#endif + /* -------------------------------------------------------------------------- ACPI-based IOAPIC Configuration -------------------------------------------------------------------------- */ diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index 9127f6b51a39..472f11765f60 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -467,6 +467,7 @@ int show_interrupts(struct seq_file *p, void *v); struct irq_desc; extern int early_irq_init(void); +extern int arch_probe_nr_irqs(void); extern int arch_early_irq_init(void); extern int arch_init_chip_data(struct irq_desc *desc, int cpu); diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 04d3e46031e5..375d68cd5bf0 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -59,10 +59,6 @@ EXPORT_SYMBOL_GPL(nr_irqs); #ifdef CONFIG_SPARSE_IRQ -#ifndef max_nr_irqs -#define max_nr_irqs(nr_cpus) NR_IRQS -#endif - static struct irq_desc irq_desc_init = { .irq = -1, .status = IRQ_DISABLED, @@ -137,9 +133,8 @@ int __init early_irq_init(void) int legacy_count; int i; - /* initialize nr_irqs based on nr_cpu_ids */ - nr_irqs = max_nr_irqs(nr_cpu_ids); - + /* initialize nr_irqs based on nr_cpu_ids */ + arch_probe_nr_irqs(); printk(KERN_INFO "NR_IRQS:%d nr_irqs:%d\n", NR_IRQS, nr_irqs); desc = irq_desc_legacy; diff --git a/kernel/softirq.c b/kernel/softirq.c index bdbe9de9cd8d..0365b4899a3d 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -795,6 +795,11 @@ int __init __weak early_irq_init(void) return 0; } +int __init __weak arch_probe_nr_irqs(void) +{ + return 0; +} + int __init __weak arch_early_irq_init(void) { return 0; -- cgit v1.2.3-59-g8ed1b From 0bd74fa8e29dcad98f7e8ffe01ec05fb3326abaf Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Mon, 19 Jan 2009 12:21:27 +0900 Subject: percpu: refactor percpu.h Impact: cleanup Refactor the DEFINE_PER_CPU_* macros and add .data.percpu.first section. Signed-off-by: Brian Gerst Signed-off-by: Tejun Heo --- include/asm-generic/vmlinux.lds.h | 1 + include/linux/percpu.h | 41 ++++++++++++++++++++++----------------- 2 files changed, 24 insertions(+), 18 deletions(-) (limited to 'include/linux') diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index aa6b9b1b30b5..32bbf50d3055 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -486,6 +486,7 @@ */ #define PERCPU_VADDR(vaddr, phdr) \ PERCPU_PROLOG(vaddr) \ + *(.data.percpu.first) \ *(.data.percpu.page_aligned) \ *(.data.percpu) \ *(.data.percpu.shared_aligned) \ diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 9f2a3751873a..0e24202b5a4e 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -9,34 +9,39 @@ #include #ifdef CONFIG_SMP -#define DEFINE_PER_CPU(type, name) \ - __attribute__((__section__(".data.percpu"))) \ - PER_CPU_ATTRIBUTES __typeof__(type) per_cpu__##name +#define PER_CPU_BASE_SECTION ".data.percpu" #ifdef MODULE -#define SHARED_ALIGNED_SECTION ".data.percpu" +#define PER_CPU_SHARED_ALIGNED_SECTION "" #else -#define SHARED_ALIGNED_SECTION ".data.percpu.shared_aligned" +#define PER_CPU_SHARED_ALIGNED_SECTION ".shared_aligned" #endif +#define PER_CPU_FIRST_SECTION ".first" -#define DEFINE_PER_CPU_SHARED_ALIGNED(type, name) \ - __attribute__((__section__(SHARED_ALIGNED_SECTION))) \ - PER_CPU_ATTRIBUTES __typeof__(type) per_cpu__##name \ - ____cacheline_aligned_in_smp +#else + +#define PER_CPU_BASE_SECTION ".data" +#define PER_CPU_SHARED_ALIGNED_SECTION "" +#define PER_CPU_FIRST_SECTION "" + +#endif -#define DEFINE_PER_CPU_PAGE_ALIGNED(type, name) \ - __attribute__((__section__(".data.percpu.page_aligned"))) \ +#define DEFINE_PER_CPU_SECTION(type, name, section) \ + __attribute__((__section__(PER_CPU_BASE_SECTION section))) \ PER_CPU_ATTRIBUTES __typeof__(type) per_cpu__##name -#else + #define DEFINE_PER_CPU(type, name) \ - PER_CPU_ATTRIBUTES __typeof__(type) per_cpu__##name + DEFINE_PER_CPU_SECTION(type, name, "") -#define DEFINE_PER_CPU_SHARED_ALIGNED(type, name) \ - DEFINE_PER_CPU(type, name) +#define DEFINE_PER_CPU_SHARED_ALIGNED(type, name) \ + DEFINE_PER_CPU_SECTION(type, name, PER_CPU_SHARED_ALIGNED_SECTION) \ + ____cacheline_aligned_in_smp -#define DEFINE_PER_CPU_PAGE_ALIGNED(type, name) \ - DEFINE_PER_CPU(type, name) -#endif +#define DEFINE_PER_CPU_PAGE_ALIGNED(type, name) \ + DEFINE_PER_CPU_SECTION(type, name, ".page_aligned") + +#define DEFINE_PER_CPU_FIRST(type, name) \ + DEFINE_PER_CPU_SECTION(type, name, PER_CPU_FIRST_SECTION) #define EXPORT_PER_CPU_SYMBOL(var) EXPORT_SYMBOL(per_cpu__##var) #define EXPORT_PER_CPU_SYMBOL_GPL(var) EXPORT_SYMBOL_GPL(per_cpu__##var) -- cgit v1.2.3-59-g8ed1b From 65a4e574d2382d83f71b30ea92f86d2e40a6ef8d Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sat, 31 Jan 2009 03:36:17 +0100 Subject: smp, generic: introduce arch_disable_smp_support() instead of disable_ioapic_setup() Impact: cleanup disable_ioapic_setup() in init/main.c is ugly as the function is x86-specific. The #ifdef inline prototype there is ugly too. Replace it with a generic arch_disable_smp_support() function - which has a weak alias for non-x86 architectures and for non-ioapic x86 builds. Signed-off-by: Ingo Molnar --- arch/x86/include/asm/io_apic.h | 9 --------- arch/x86/kernel/apic.c | 4 +--- arch/x86/kernel/io_apic.c | 11 ++++++++++- arch/x86/kernel/smpboot.c | 2 +- include/linux/smp.h | 6 ++++++ init/main.c | 12 ++++++------ 6 files changed, 24 insertions(+), 20 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 08ec793aa043..309d0e23193a 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -143,15 +143,6 @@ extern int noioapicreroute; /* 1 if the timer IRQ uses the '8259A Virtual Wire' mode */ extern int timer_through_8259; -static inline void disable_ioapic_setup(void) -{ -#ifdef CONFIG_PCI - noioapicquirk = 1; - noioapicreroute = -1; -#endif - skip_ioapic_setup = 1; -} - /* * If we use the IO-APIC for IRQ routing, disable automatic * assignment of PCI IRQ's. diff --git a/arch/x86/kernel/apic.c b/arch/x86/kernel/apic.c index 85d8b50d1af7..a04a73a51d20 100644 --- a/arch/x86/kernel/apic.c +++ b/arch/x86/kernel/apic.c @@ -1138,9 +1138,7 @@ void __cpuinit setup_local_APIC(void) int i, j; if (disable_apic) { -#ifdef CONFIG_X86_IO_APIC - disable_ioapic_setup(); -#endif + arch_disable_smp_support(); return; } diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 57d60c741e37..84bccac4619f 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -98,10 +98,19 @@ DECLARE_BITMAP(mp_bus_not_pci, MAX_MP_BUSSES); int skip_ioapic_setup; +void arch_disable_smp_support(void) +{ +#ifdef CONFIG_PCI + noioapicquirk = 1; + noioapicreroute = -1; +#endif + skip_ioapic_setup = 1; +} + static int __init parse_noapic(char *str) { /* disable IO-APIC */ - disable_ioapic_setup(); + arch_disable_smp_support(); return 0; } early_param("noapic", parse_noapic); diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index f40f86fec2fe..96f7d304f5c9 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1071,7 +1071,7 @@ static int __init smp_sanity_check(unsigned max_cpus) printk(KERN_ERR "... forcing use of dummy APIC emulation." "(tell your hw vendor)\n"); smpboot_clear_io_apic(); - disable_ioapic_setup(); + arch_disable_smp_support(); return -1; } diff --git a/include/linux/smp.h b/include/linux/smp.h index 715196b09d67..d41a3a865fe3 100644 --- a/include/linux/smp.h +++ b/include/linux/smp.h @@ -66,6 +66,12 @@ extern int __cpu_up(unsigned int cpunum); */ extern void smp_cpus_done(unsigned int max_cpus); +/* + * Callback to arch code if there's nosmp or maxcpus=0 on the + * boot command line: + */ +extern void arch_disable_smp_support(void); + /* * Call a function on all other processors */ diff --git a/init/main.c b/init/main.c index bfe4fb0c9842..6441083f8273 100644 --- a/init/main.c +++ b/init/main.c @@ -136,14 +136,14 @@ unsigned int __initdata setup_max_cpus = NR_CPUS; * greater than 0, limits the maximum number of CPUs activated in * SMP mode to . */ -#ifndef CONFIG_X86_IO_APIC -static inline void disable_ioapic_setup(void) {}; -#endif + +void __weak arch_disable_smp_support(void) { } static int __init nosmp(char *str) { setup_max_cpus = 0; - disable_ioapic_setup(); + arch_disable_smp_support(); + return 0; } @@ -153,14 +153,14 @@ static int __init maxcpus(char *str) { get_option(&str, &setup_max_cpus); if (setup_max_cpus == 0) - disable_ioapic_setup(); + arch_disable_smp_support(); return 0; } early_param("maxcpus", maxcpus); #else -#define setup_max_cpus NR_CPUS +const unsigned int setup_max_cpus = NR_CPUS; #endif /* -- cgit v1.2.3-59-g8ed1b From a146649bc19d5eba4f5bfac6720c5f252d517a71 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sat, 31 Jan 2009 14:09:06 +0100 Subject: smp, generic: introduce arch_disable_smp_support(), build fix This function should be provided on UP too. Signed-off-by: Ingo Molnar --- include/linux/smp.h | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/smp.h b/include/linux/smp.h index d41a3a865fe3..bbacb7baa446 100644 --- a/include/linux/smp.h +++ b/include/linux/smp.h @@ -66,12 +66,6 @@ extern int __cpu_up(unsigned int cpunum); */ extern void smp_cpus_done(unsigned int max_cpus); -/* - * Callback to arch code if there's nosmp or maxcpus=0 on the - * boot command line: - */ -extern void arch_disable_smp_support(void); - /* * Call a function on all other processors */ @@ -182,6 +176,12 @@ static inline void init_call_single_data(void) #define put_cpu() preempt_enable() #define put_cpu_no_resched() preempt_enable_no_resched() +/* + * Callback to arch code if there's nosmp or maxcpus=0 on the + * boot command line: + */ +extern void arch_disable_smp_support(void); + void smp_setup_processor_id(void); #endif /* __LINUX_SMP_H */ -- cgit v1.2.3-59-g8ed1b From 0fb807c3e573ff9de2965ca38c907605d4735d16 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Sun, 8 Feb 2009 11:00:25 +0530 Subject: unconditionally include asm/types.h from linux/types.h Reported-by: Sam Ravnborg Signed-off-by: Jaswinder Singh Rajput --- include/linux/types.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/types.h b/include/linux/types.h index c30973ace890..fca82ed55f49 100644 --- a/include/linux/types.h +++ b/include/linux/types.h @@ -1,6 +1,8 @@ #ifndef _LINUX_TYPES_H #define _LINUX_TYPES_H +#include + #ifndef __ASSEMBLY__ #ifdef __KERNEL__ @@ -10,7 +12,6 @@ #endif #include -#include #ifndef __KERNEL_STRICT_NAMES -- cgit v1.2.3-59-g8ed1b From d3770449d3cb058b94ca1d050d5ced4a66c75ce4 Mon Sep 17 00:00:00 2001 From: Brian Gerst Date: Sun, 8 Feb 2009 09:58:38 -0500 Subject: percpu: make PER_CPU_BASE_SECTION overridable by arches Impact: bug fix IA-64 needs to put percpu data in the seperate section even on UP. Fixes regression caused by "percpu: refactor percpu.h" Signed-off-by: Brian Gerst Acked-by: Tony Luck Signed-off-by: Ingo Molnar --- arch/ia64/include/asm/percpu.h | 4 ++-- include/linux/percpu.h | 8 +++++++- 2 files changed, 9 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/arch/ia64/include/asm/percpu.h b/arch/ia64/include/asm/percpu.h index 77f30b664b4e..30cf46534dd2 100644 --- a/arch/ia64/include/asm/percpu.h +++ b/arch/ia64/include/asm/percpu.h @@ -27,12 +27,12 @@ extern void *per_cpu_init(void); #else /* ! SMP */ -#define PER_CPU_ATTRIBUTES __attribute__((__section__(".data.percpu"))) - #define per_cpu_init() (__phys_per_cpu_start) #endif /* SMP */ +#define PER_CPU_BASE_SECTION ".data.percpu" + /* * Be extremely careful when taking the address of this variable! Due to virtual * remapping, it is different from the canonical address returned by __get_cpu_var(var)! diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 0e24202b5a4e..3577ffd90d45 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -8,8 +8,15 @@ #include +#ifndef PER_CPU_BASE_SECTION #ifdef CONFIG_SMP #define PER_CPU_BASE_SECTION ".data.percpu" +#else +#define PER_CPU_BASE_SECTION ".data" +#endif +#endif + +#ifdef CONFIG_SMP #ifdef MODULE #define PER_CPU_SHARED_ALIGNED_SECTION "" @@ -20,7 +27,6 @@ #else -#define PER_CPU_BASE_SECTION ".data" #define PER_CPU_SHARED_ALIGNED_SECTION "" #define PER_CPU_FIRST_SECTION "" -- cgit v1.2.3-59-g8ed1b From 7d97277b754d3ee098a5ec69b6aaafb00c94e2f2 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sat, 7 Feb 2009 15:39:41 -0800 Subject: acpi/x86: introduce __apci_map_table, v4 to prevent wrongly overwriting fixmap that still want to use. ACPI used to rely on low mappings being all linearly mapped and grew a habit: it never really unmapped certain kinds of tables after use. This can cause problems - for example the hypothetical case when some spurious access still references it. v2: remove prev_map and prev_size in __apci_map_table v3: let acpi_os_unmap_memory() call early_iounmap too, so remove extral calling to early_acpi_os_unmap_memory v4: fix typo in one acpi_get_table_with_size calling Signed-off-by: Yinghai Lu Acked-by: Len Brown Signed-off-by: Ingo Molnar --- arch/ia64/kernel/acpi.c | 4 ++++ arch/x86/kernel/acpi/boot.c | 17 +++++++---------- drivers/acpi/acpica/tbxface.c | 17 ++++++++++++++--- drivers/acpi/bus.c | 6 ------ drivers/acpi/osl.c | 11 +++++++++-- drivers/acpi/tables.c | 20 ++++++++++++++------ include/acpi/acpiosxf.h | 1 + include/acpi/acpixf.h | 4 ++++ include/linux/acpi.h | 1 + 9 files changed, 54 insertions(+), 27 deletions(-) (limited to 'include/linux') diff --git a/arch/ia64/kernel/acpi.c b/arch/ia64/kernel/acpi.c index d541671caf4a..2363ed173198 100644 --- a/arch/ia64/kernel/acpi.c +++ b/arch/ia64/kernel/acpi.c @@ -199,6 +199,10 @@ char *__init __acpi_map_table(unsigned long phys_addr, unsigned long size) return __va(phys_addr); } +char *__init __acpi_unmap_table(unsigned long virt_addr, unsigned long size) +{ +} + /* -------------------------------------------------------------------------- Boot-time Table Parsing -------------------------------------------------------------------------- */ diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 7217834f6b1d..4c2aaea42930 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -121,21 +121,18 @@ enum acpi_irq_model_id acpi_irq_model = ACPI_IRQ_MODEL_PIC; */ char *__init __acpi_map_table(unsigned long phys, unsigned long size) { - static char *prev_map; - static unsigned long prev_size; - - if (prev_map) { - early_iounmap(prev_map, prev_size); - prev_map = NULL; - } if (!phys || !size) return NULL; - prev_size = size; - prev_map = early_ioremap(phys, size); + return early_ioremap(phys, size); +} +void __init __acpi_unmap_table(char *map, unsigned long size) +{ + if (!map || !size) + return; - return prev_map; + early_iounmap(map, size); } #ifdef CONFIG_PCI_MMCONFIG diff --git a/drivers/acpi/acpica/tbxface.c b/drivers/acpi/acpica/tbxface.c index c3e841f3cde9..ab0aff3c7d6a 100644 --- a/drivers/acpi/acpica/tbxface.c +++ b/drivers/acpi/acpica/tbxface.c @@ -365,7 +365,7 @@ ACPI_EXPORT_SYMBOL(acpi_unload_table_id) /******************************************************************************* * - * FUNCTION: acpi_get_table + * FUNCTION: acpi_get_table_with_size * * PARAMETERS: Signature - ACPI signature of needed table * Instance - Which instance (for SSDTs) @@ -377,8 +377,9 @@ ACPI_EXPORT_SYMBOL(acpi_unload_table_id) * *****************************************************************************/ acpi_status -acpi_get_table(char *signature, - u32 instance, struct acpi_table_header **out_table) +acpi_get_table_with_size(char *signature, + u32 instance, struct acpi_table_header **out_table, + acpi_size *tbl_size) { u32 i; u32 j; @@ -408,6 +409,7 @@ acpi_get_table(char *signature, acpi_tb_verify_table(&acpi_gbl_root_table_list.tables[i]); if (ACPI_SUCCESS(status)) { *out_table = acpi_gbl_root_table_list.tables[i].pointer; + *tbl_size = acpi_gbl_root_table_list.tables[i].length; } if (!acpi_gbl_permanent_mmap) { @@ -420,6 +422,15 @@ acpi_get_table(char *signature, return (AE_NOT_FOUND); } +acpi_status +acpi_get_table(char *signature, + u32 instance, struct acpi_table_header **out_table) +{ + acpi_size tbl_size; + + return acpi_get_table_with_size(signature, + instance, out_table, &tbl_size); +} ACPI_EXPORT_SYMBOL(acpi_get_table) /******************************************************************************* diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c index fb1be7b5dbc1..765fd1c56cd6 100644 --- a/drivers/acpi/bus.c +++ b/drivers/acpi/bus.c @@ -694,12 +694,6 @@ void __init acpi_early_init(void) if (!acpi_strict) acpi_gbl_enable_interpreter_slack = TRUE; - /* - * Doing a zero-sized mapping will clear out the previous - * __acpi_map_table() mapping, if any. - */ - __acpi_map_table(0, 0); - acpi_gbl_permanent_mmap = 1; status = acpi_reallocate_root_table(); diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c index b3193ec0a2ef..d1dd5160daa9 100644 --- a/drivers/acpi/osl.c +++ b/drivers/acpi/osl.c @@ -274,12 +274,19 @@ EXPORT_SYMBOL_GPL(acpi_os_map_memory); void acpi_os_unmap_memory(void __iomem * virt, acpi_size size) { - if (acpi_gbl_permanent_mmap) { + if (acpi_gbl_permanent_mmap) iounmap(virt); - } + else + __acpi_unmap_table(virt, size); } EXPORT_SYMBOL_GPL(acpi_os_unmap_memory); +void early_acpi_os_unmap_memory(void __iomem * virt, acpi_size size) +{ + if (!acpi_gbl_permanent_mmap) + __acpi_unmap_table(virt, size); +} + #ifdef ACPI_FUTURE_USAGE acpi_status acpi_os_get_physical_address(void *virt, acpi_physical_address * phys) diff --git a/drivers/acpi/tables.c b/drivers/acpi/tables.c index a8852952fac4..fec1ae36d431 100644 --- a/drivers/acpi/tables.c +++ b/drivers/acpi/tables.c @@ -181,14 +181,15 @@ acpi_table_parse_entries(char *id, struct acpi_subtable_header *entry; unsigned int count = 0; unsigned long table_end; + acpi_size tbl_size; if (!handler) return -EINVAL; if (strncmp(id, ACPI_SIG_MADT, 4) == 0) - acpi_get_table(id, acpi_apic_instance, &table_header); + acpi_get_table_with_size(id, acpi_apic_instance, &table_header, &tbl_size); else - acpi_get_table(id, 0, &table_header); + acpi_get_table_with_size(id, 0, &table_header, &tbl_size); if (!table_header) { printk(KERN_WARNING PREFIX "%4.4s not present\n", id); @@ -206,8 +207,10 @@ acpi_table_parse_entries(char *id, table_end) { if (entry->type == entry_id && (!max_entries || count++ < max_entries)) - if (handler(entry, table_end)) + if (handler(entry, table_end)) { + early_acpi_os_unmap_memory((char *)table_header, tbl_size); return -EINVAL; + } entry = (struct acpi_subtable_header *) ((unsigned long)entry + entry->length); @@ -217,6 +220,7 @@ acpi_table_parse_entries(char *id, "%i found\n", id, entry_id, count - max_entries, count); } + early_acpi_os_unmap_memory((char *)table_header, tbl_size); return count; } @@ -241,17 +245,19 @@ acpi_table_parse_madt(enum acpi_madt_type id, int __init acpi_table_parse(char *id, acpi_table_handler handler) { struct acpi_table_header *table = NULL; + acpi_size tbl_size; if (!handler) return -EINVAL; if (strncmp(id, ACPI_SIG_MADT, 4) == 0) - acpi_get_table(id, acpi_apic_instance, &table); + acpi_get_table_with_size(id, acpi_apic_instance, &table, &tbl_size); else - acpi_get_table(id, 0, &table); + acpi_get_table_with_size(id, 0, &table, &tbl_size); if (table) { handler(table); + early_acpi_os_unmap_memory(table, tbl_size); return 0; } else return 1; @@ -265,8 +271,9 @@ int __init acpi_table_parse(char *id, acpi_table_handler handler) static void __init check_multiple_madt(void) { struct acpi_table_header *table = NULL; + acpi_size tbl_size; - acpi_get_table(ACPI_SIG_MADT, 2, &table); + acpi_get_table_with_size(ACPI_SIG_MADT, 2, &table, &tbl_size); if (table) { printk(KERN_WARNING PREFIX "BIOS bug: multiple APIC/MADT found," @@ -275,6 +282,7 @@ static void __init check_multiple_madt(void) "If \"acpi_apic_instance=%d\" works better, " "notify linux-acpi@vger.kernel.org\n", acpi_apic_instance ? 0 : 2); + early_acpi_os_unmap_memory(table, tbl_size); } else acpi_apic_instance = 0; diff --git a/include/acpi/acpiosxf.h b/include/acpi/acpiosxf.h index a62720a7edc0..ab0b85cf21f3 100644 --- a/include/acpi/acpiosxf.h +++ b/include/acpi/acpiosxf.h @@ -144,6 +144,7 @@ void __iomem *acpi_os_map_memory(acpi_physical_address where, acpi_size length); void acpi_os_unmap_memory(void __iomem * logical_address, acpi_size size); +void early_acpi_os_unmap_memory(void __iomem * virt, acpi_size size); #ifdef ACPI_FUTURE_USAGE acpi_status diff --git a/include/acpi/acpixf.h b/include/acpi/acpixf.h index c8e8cf45830f..cc40102fe2f3 100644 --- a/include/acpi/acpixf.h +++ b/include/acpi/acpixf.h @@ -130,6 +130,10 @@ acpi_get_table_header(acpi_string signature, struct acpi_table_header *out_table_header); acpi_status +acpi_get_table_with_size(acpi_string signature, + u32 instance, struct acpi_table_header **out_table, + acpi_size *tbl_size); +acpi_status acpi_get_table(acpi_string signature, u32 instance, struct acpi_table_header **out_table); diff --git a/include/linux/acpi.h b/include/linux/acpi.h index 6fce2fc2d124..d59f0fa4d772 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -79,6 +79,7 @@ typedef int (*acpi_table_handler) (struct acpi_table_header *table); typedef int (*acpi_table_entry_handler) (struct acpi_subtable_header *header, const unsigned long end); char * __acpi_map_table (unsigned long phys_addr, unsigned long size); +void __init __acpi_unmap_table(char *map, unsigned long size); int early_acpi_boot_init(void); int acpi_boot_init (void); int acpi_boot_table_init (void); -- cgit v1.2.3-59-g8ed1b From 6cd61c0baa8bce32271226198b46c67a7a05d108 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 9 Feb 2009 22:17:39 +0900 Subject: elf: add ELF_CORE_COPY_KERNEL_REGS() ELF core dump is used for both user land core dump and kernel crash dump. Depending on architecture, register might need to be accessed differently for userland and kernel. Allow architectures to define ELF_CORE_COPY_KERNEL_REGS() and use different operation for kernel register dump. Signed-off-by: Tejun Heo Signed-off-by: Ingo Molnar --- include/linux/elfcore.h | 9 +++++++++ kernel/kexec.c | 2 +- 2 files changed, 10 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/elfcore.h b/include/linux/elfcore.h index 5ca54d77079f..7605c5e9589f 100644 --- a/include/linux/elfcore.h +++ b/include/linux/elfcore.h @@ -111,6 +111,15 @@ static inline void elf_core_copy_regs(elf_gregset_t *elfregs, struct pt_regs *re #endif } +static inline void elf_core_copy_kernel_regs(elf_gregset_t *elfregs, struct pt_regs *regs) +{ +#ifdef ELF_CORE_COPY_KERNEL_REGS + ELF_CORE_COPY_KERNEL_REGS((*elfregs), regs); +#else + elf_core_copy_regs(elfregs, regs); +#endif +} + static inline int elf_core_copy_task_regs(struct task_struct *t, elf_gregset_t* elfregs) { #ifdef ELF_CORE_COPY_TASK_REGS diff --git a/kernel/kexec.c b/kernel/kexec.c index 8a6d7b08864e..795e7b67a228 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -1130,7 +1130,7 @@ void crash_save_cpu(struct pt_regs *regs, int cpu) return; memset(&prstatus, 0, sizeof(prstatus)); prstatus.pr_pid = current->pid; - elf_core_copy_regs(&prstatus.pr_reg, regs); + elf_core_copy_kernel_regs(&prstatus.pr_reg, regs); buf = append_elf_note(buf, KEXEC_CORE_NOTE_NAME, NT_PRSTATUS, &prstatus, sizeof(prstatus)); final_note(buf); -- cgit v1.2.3-59-g8ed1b From 970ec1a8213cd1a1ea29972ebbe4575a8b30bca1 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sun, 15 Feb 2009 14:06:13 -0800 Subject: [IA64] fix __apci_unmap_table Impact: fix build error to fix: tip/arch/ia64/kernel/acpi.c:203: error: conflicting types for '__acpi_unmap_table' tip/include/linux/acpi.h:82: error: previous declaration of '__acpi_unmap_table' was here tip/arch/ia64/kernel/acpi.c:203: error: conflicting types for '__acpi_unmap_table' tip/include/linux/acpi.h:82: error: previous declaration of '__acpi_unmap_table' was here Signed-off-by: Yinghai Lu Cc: Jeremy Fitzhardinge Signed-off-by: Ingo Molnar --- arch/ia64/kernel/acpi.c | 2 +- include/linux/acpi.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/arch/ia64/kernel/acpi.c b/arch/ia64/kernel/acpi.c index 2363ed173198..bdef2ce38c8b 100644 --- a/arch/ia64/kernel/acpi.c +++ b/arch/ia64/kernel/acpi.c @@ -199,7 +199,7 @@ char *__init __acpi_map_table(unsigned long phys_addr, unsigned long size) return __va(phys_addr); } -char *__init __acpi_unmap_table(unsigned long virt_addr, unsigned long size) +void __init __acpi_unmap_table(char *map, unsigned long size) { } diff --git a/include/linux/acpi.h b/include/linux/acpi.h index d59f0fa4d772..78199151c00b 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -79,7 +79,7 @@ typedef int (*acpi_table_handler) (struct acpi_table_header *table); typedef int (*acpi_table_entry_handler) (struct acpi_subtable_header *header, const unsigned long end); char * __acpi_map_table (unsigned long phys_addr, unsigned long size); -void __init __acpi_unmap_table(char *map, unsigned long size); +void __acpi_unmap_table(char *map, unsigned long size); int early_acpi_boot_init(void); int acpi_boot_init (void); int acpi_boot_table_init (void); -- cgit v1.2.3-59-g8ed1b From b36128c830a8f5bd7d4981f5b0b69950f5928ee6 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 20 Feb 2009 16:29:08 +0900 Subject: alloc_percpu: change percpu_ptr to per_cpu_ptr Impact: cleanup There are two allocated per-cpu accessor macros with almost identical spelling. The original and far more popular is per_cpu_ptr (44 files), so change over the other 4 files. tj: kill percpu_ptr() and update UP too Signed-off-by: Rusty Russell Cc: mingo@redhat.com Cc: lenb@kernel.org Cc: cpufreq@vger.kernel.org Signed-off-by: Tejun Heo --- arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c | 2 +- drivers/acpi/processor_perflib.c | 4 ++-- include/linux/percpu.h | 23 +++++++++++------------ kernel/sched.c | 6 +++--- kernel/stop_machine.c | 2 +- 5 files changed, 18 insertions(+), 19 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index 4b1c319d30c3..22590cf688ae 100644 --- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -601,7 +601,7 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy) if (!data) return -ENOMEM; - data->acpi_data = percpu_ptr(acpi_perf_data, cpu); + data->acpi_data = per_cpu_ptr(acpi_perf_data, cpu); per_cpu(drv_data, cpu) = data; if (cpu_has(c, X86_FEATURE_CONSTANT_TSC)) diff --git a/drivers/acpi/processor_perflib.c b/drivers/acpi/processor_perflib.c index 9cc769b587ff..68fd3d292799 100644 --- a/drivers/acpi/processor_perflib.c +++ b/drivers/acpi/processor_perflib.c @@ -516,12 +516,12 @@ int acpi_processor_preregister_performance( continue; } - if (!performance || !percpu_ptr(performance, i)) { + if (!performance || !per_cpu_ptr(performance, i)) { retval = -EINVAL; continue; } - pr->performance = percpu_ptr(performance, i); + pr->performance = per_cpu_ptr(performance, i); cpumask_set_cpu(i, pr->performance->shared_cpu_map); if (acpi_processor_get_psd(pr)) { retval = -EINVAL; diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 3577ffd90d45..c80cfe1260ec 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -81,23 +81,13 @@ struct percpu_data { }; #define __percpu_disguise(pdata) (struct percpu_data *)~(unsigned long)(pdata) -/* - * Use this to get to a cpu's version of the per-cpu object dynamically - * allocated. Non-atomic access to the current CPU's version should - * probably be combined with get_cpu()/put_cpu(). - */ -#define percpu_ptr(ptr, cpu) \ -({ \ - struct percpu_data *__p = __percpu_disguise(ptr); \ - (__typeof__(ptr))__p->ptrs[(cpu)]; \ -}) extern void *__percpu_alloc_mask(size_t size, gfp_t gfp, cpumask_t *mask); extern void percpu_free(void *__pdata); #else /* CONFIG_SMP */ -#define percpu_ptr(ptr, cpu) ({ (void)(cpu); (ptr); }) +#define per_cpu_ptr(ptr, cpu) ({ (void)(cpu); (ptr); }) static __always_inline void *__percpu_alloc_mask(size_t size, gfp_t gfp, cpumask_t *mask) { @@ -122,6 +112,15 @@ static inline void percpu_free(void *__pdata) cpu_possible_map) #define alloc_percpu(type) (type *)__alloc_percpu(sizeof(type)) #define free_percpu(ptr) percpu_free((ptr)) -#define per_cpu_ptr(ptr, cpu) percpu_ptr((ptr), (cpu)) +/* + * Use this to get to a cpu's version of the per-cpu object dynamically + * allocated. Non-atomic access to the current CPU's version should + * probably be combined with get_cpu()/put_cpu(). + */ +#define per_cpu_ptr(ptr, cpu) \ +({ \ + struct percpu_data *__p = __percpu_disguise(ptr); \ + (__typeof__(ptr))__p->ptrs[(cpu)]; \ +}) #endif /* __LINUX_PERCPU_H */ diff --git a/kernel/sched.c b/kernel/sched.c index fc17fd91ab57..9d30ac956328 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -9472,7 +9472,7 @@ cpuacct_destroy(struct cgroup_subsys *ss, struct cgroup *cgrp) static u64 cpuacct_cpuusage_read(struct cpuacct *ca, int cpu) { - u64 *cpuusage = percpu_ptr(ca->cpuusage, cpu); + u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu); u64 data; #ifndef CONFIG_64BIT @@ -9491,7 +9491,7 @@ static u64 cpuacct_cpuusage_read(struct cpuacct *ca, int cpu) static void cpuacct_cpuusage_write(struct cpuacct *ca, int cpu, u64 val) { - u64 *cpuusage = percpu_ptr(ca->cpuusage, cpu); + u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu); #ifndef CONFIG_64BIT /* @@ -9587,7 +9587,7 @@ static void cpuacct_charge(struct task_struct *tsk, u64 cputime) ca = task_ca(tsk); for (; ca; ca = ca->parent) { - u64 *cpuusage = percpu_ptr(ca->cpuusage, cpu); + u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu); *cpuusage += cputime; } } diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c index 0cd415ee62a2..74541ca49536 100644 --- a/kernel/stop_machine.c +++ b/kernel/stop_machine.c @@ -170,7 +170,7 @@ int __stop_machine(int (*fn)(void *), void *data, const struct cpumask *cpus) * doesn't hit this CPU until we're ready. */ get_cpu(); for_each_online_cpu(i) { - sm_work = percpu_ptr(stop_machine_work, i); + sm_work = per_cpu_ptr(stop_machine_work, i); INIT_WORK(sm_work, stop_cpu); queue_work_on(i, stop_machine_wq, sm_work); } -- cgit v1.2.3-59-g8ed1b From 313e458f81ec3852106c5a83830fe0d4f405a71a Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 20 Feb 2009 16:29:08 +0900 Subject: alloc_percpu: add align argument to __alloc_percpu. This prepares for a real __alloc_percpu, by adding an alignment argument. Only one place uses __alloc_percpu directly, and that's for a string. tj: af_inet also uses __alloc_percpu(), update it. Signed-off-by: Rusty Russell Cc: Christoph Lameter Cc: Jens Axboe --- block/blktrace.c | 2 +- include/linux/percpu.h | 5 +++-- net/ipv4/af_inet.c | 4 ++-- 3 files changed, 6 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/block/blktrace.c b/block/blktrace.c index 39cc3bfe56e4..487766237d28 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -363,7 +363,7 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, if (!bt->sequence) goto err; - bt->msg_data = __alloc_percpu(BLK_TN_MAX_MSG); + bt->msg_data = __alloc_percpu(BLK_TN_MAX_MSG, __alignof__(char)); if (!bt->msg_data) goto err; diff --git a/include/linux/percpu.h b/include/linux/percpu.h index c80cfe1260ec..1fdaee93c04d 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -108,9 +108,10 @@ static inline void percpu_free(void *__pdata) /* (legacy) interface for use without CPU hotplug handling */ -#define __alloc_percpu(size) percpu_alloc_mask((size), GFP_KERNEL, \ +#define __alloc_percpu(size, align) percpu_alloc_mask((size), GFP_KERNEL, \ cpu_possible_map) -#define alloc_percpu(type) (type *)__alloc_percpu(sizeof(type)) +#define alloc_percpu(type) (type *)__alloc_percpu(sizeof(type), \ + __alignof__(type)) #define free_percpu(ptr) percpu_free((ptr)) /* * Use this to get to a cpu's version of the per-cpu object dynamically diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 743f5542d65a..3a3dad801354 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1375,10 +1375,10 @@ EXPORT_SYMBOL_GPL(snmp_fold_field); int snmp_mib_init(void *ptr[2], size_t mibsize) { BUG_ON(ptr == NULL); - ptr[0] = __alloc_percpu(mibsize); + ptr[0] = __alloc_percpu(mibsize, __alignof__(unsigned long long)); if (!ptr[0]) goto err0; - ptr[1] = __alloc_percpu(mibsize); + ptr[1] = __alloc_percpu(mibsize, __alignof__(unsigned long long)); if (!ptr[1]) goto err1; return 0; -- cgit v1.2.3-59-g8ed1b From f2a8205c4ef1af917d175c36a4097ae5587791c8 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 20 Feb 2009 16:29:08 +0900 Subject: percpu: kill percpu_alloc() and friends Impact: kill unused functions percpu_alloc() and its friends never saw much action. It was supposed to replace the cpu-mask unaware __alloc_percpu() but it never happened and in fact __percpu_alloc_mask() itself never really grew proper up/down handling interface either (no exported interface for populate/depopulate). percpu allocation is about to go through major reimplementation and there's no reason to carry this unused interface around. Replace it with __alloc_percpu() and free_percpu(). Signed-off-by: Tejun Heo --- include/linux/percpu.h | 47 ++++++++++++++++++++++------------------------- mm/allocpercpu.c | 32 +++++++++++++++++++------------- 2 files changed, 41 insertions(+), 38 deletions(-) (limited to 'include/linux') diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 1fdaee93c04d..d99e24ae1811 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -82,46 +82,43 @@ struct percpu_data { #define __percpu_disguise(pdata) (struct percpu_data *)~(unsigned long)(pdata) -extern void *__percpu_alloc_mask(size_t size, gfp_t gfp, cpumask_t *mask); -extern void percpu_free(void *__pdata); +/* + * Use this to get to a cpu's version of the per-cpu object + * dynamically allocated. Non-atomic access to the current CPU's + * version should probably be combined with get_cpu()/put_cpu(). + */ +#define per_cpu_ptr(ptr, cpu) \ +({ \ + struct percpu_data *__p = __percpu_disguise(ptr); \ + (__typeof__(ptr))__p->ptrs[(cpu)]; \ +}) + +extern void *__alloc_percpu(size_t size, size_t align); +extern void free_percpu(void *__pdata); #else /* CONFIG_SMP */ #define per_cpu_ptr(ptr, cpu) ({ (void)(cpu); (ptr); }) -static __always_inline void *__percpu_alloc_mask(size_t size, gfp_t gfp, cpumask_t *mask) +static inline void *__alloc_percpu(size_t size, size_t align) { + /* + * Can't easily make larger alignment work with kmalloc. WARN + * on it. Larger alignment should only be used for module + * percpu sections on SMP for which this path isn't used. + */ + WARN_ON_ONCE(align > __alignof__(unsigned long long)); return kzalloc(size, gfp); } -static inline void percpu_free(void *__pdata) +static inline void free_percpu(void *p) { - kfree(__pdata); + kfree(p); } #endif /* CONFIG_SMP */ -#define percpu_alloc_mask(size, gfp, mask) \ - __percpu_alloc_mask((size), (gfp), &(mask)) - -#define percpu_alloc(size, gfp) percpu_alloc_mask((size), (gfp), cpu_online_map) - -/* (legacy) interface for use without CPU hotplug handling */ - -#define __alloc_percpu(size, align) percpu_alloc_mask((size), GFP_KERNEL, \ - cpu_possible_map) #define alloc_percpu(type) (type *)__alloc_percpu(sizeof(type), \ __alignof__(type)) -#define free_percpu(ptr) percpu_free((ptr)) -/* - * Use this to get to a cpu's version of the per-cpu object dynamically - * allocated. Non-atomic access to the current CPU's version should - * probably be combined with get_cpu()/put_cpu(). - */ -#define per_cpu_ptr(ptr, cpu) \ -({ \ - struct percpu_data *__p = __percpu_disguise(ptr); \ - (__typeof__(ptr))__p->ptrs[(cpu)]; \ -}) #endif /* __LINUX_PERCPU_H */ diff --git a/mm/allocpercpu.c b/mm/allocpercpu.c index 4297bc41bfd2..3653c570232b 100644 --- a/mm/allocpercpu.c +++ b/mm/allocpercpu.c @@ -99,45 +99,51 @@ static int __percpu_populate_mask(void *__pdata, size_t size, gfp_t gfp, __percpu_populate_mask((__pdata), (size), (gfp), &(mask)) /** - * percpu_alloc_mask - initial setup of per-cpu data + * alloc_percpu - initial setup of per-cpu data * @size: size of per-cpu object - * @gfp: may sleep or not etc. - * @mask: populate per-data for cpu's selected through mask bits + * @align: alignment * - * Populating per-cpu data for all online cpu's would be a typical use case, - * which is simplified by the percpu_alloc() wrapper. - * Per-cpu objects are populated with zeroed buffers. + * Allocate dynamic percpu area. Percpu objects are populated with + * zeroed buffers. */ -void *__percpu_alloc_mask(size_t size, gfp_t gfp, cpumask_t *mask) +void *__alloc_percpu(size_t size, size_t align) { /* * We allocate whole cache lines to avoid false sharing */ size_t sz = roundup(nr_cpu_ids * sizeof(void *), cache_line_size()); - void *pdata = kzalloc(sz, gfp); + void *pdata = kzalloc(sz, GFP_KERNEL); void *__pdata = __percpu_disguise(pdata); + /* + * Can't easily make larger alignment work with kmalloc. WARN + * on it. Larger alignment should only be used for module + * percpu sections on SMP for which this path isn't used. + */ + WARN_ON_ONCE(align > __alignof__(unsigned long long)); + if (unlikely(!pdata)) return NULL; - if (likely(!__percpu_populate_mask(__pdata, size, gfp, mask))) + if (likely(!__percpu_populate_mask(__pdata, size, GFP_KERNEL, + &cpu_possible_map))) return __pdata; kfree(pdata); return NULL; } -EXPORT_SYMBOL_GPL(__percpu_alloc_mask); +EXPORT_SYMBOL_GPL(__alloc_percpu); /** - * percpu_free - final cleanup of per-cpu data + * free_percpu - final cleanup of per-cpu data * @__pdata: object to clean up * * We simply clean up any per-cpu object left. No need for the client to * track and specify through a bis mask which per-cpu objects are to free. */ -void percpu_free(void *__pdata) +void free_percpu(void *__pdata) { if (unlikely(!__pdata)) return; __percpu_depopulate_mask(__pdata, &cpu_possible_map); kfree(__percpu_disguise(__pdata)); } -EXPORT_SYMBOL_GPL(percpu_free); +EXPORT_SYMBOL_GPL(free_percpu); -- cgit v1.2.3-59-g8ed1b From f0aa6617903648077dffe5cfcf7c4458f4610fa7 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 20 Feb 2009 16:29:08 +0900 Subject: vmalloc: implement vm_area_register_early() Impact: allow multiple early vm areas There are places where kernel VM area needs to be allocated before vmalloc is initialized. This is done by allocating static vm_struct, initializing several fields and linking it to vmlist and later vmalloc initialization picking up these from vmlist. This is currently done manually and if there's more than one such areas, there's no defined way to arbitrate who gets which address. This patch implements vm_area_register_early(), which takes vm_area struct with flags and size initialized, assigns address to it and puts it on the vmlist. This way, multiple early vm areas can determine which addresses they should use. The only current user - alpha mm init - is converted to use it. Signed-off-by: Tejun Heo --- arch/alpha/mm/init.c | 20 +++++++++++++------- include/linux/vmalloc.h | 1 + mm/vmalloc.c | 24 ++++++++++++++++++++++++ 3 files changed, 38 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/arch/alpha/mm/init.c b/arch/alpha/mm/init.c index 5d7a16eab312..df6df025ded4 100644 --- a/arch/alpha/mm/init.c +++ b/arch/alpha/mm/init.c @@ -189,9 +189,21 @@ callback_init(void * kernel_end) if (alpha_using_srm) { static struct vm_struct console_remap_vm; - unsigned long vaddr = VMALLOC_START; + unsigned long nr_pages = 0; + unsigned long vaddr; unsigned long i, j; + /* calculate needed size */ + for (i = 0; i < crb->map_entries; ++i) + nr_pages += crb->map[i].count; + + /* register the vm area */ + console_remap_vm.flags = VM_ALLOC; + console_remap_vm.size = nr_pages << PAGE_SHIFT; + vm_area_register_early(&console_remap_vm); + + vaddr = (unsigned long)consle_remap_vm.addr; + /* Set up the third level PTEs and update the virtual addresses of the CRB entries. */ for (i = 0; i < crb->map_entries; ++i) { @@ -213,12 +225,6 @@ callback_init(void * kernel_end) vaddr += PAGE_SIZE; } } - - /* Let vmalloc know that we've allocated some space. */ - console_remap_vm.flags = VM_ALLOC; - console_remap_vm.addr = (void *) VMALLOC_START; - console_remap_vm.size = vaddr - VMALLOC_START; - vmlist = &console_remap_vm; } callback_init_done = 1; diff --git a/include/linux/vmalloc.h b/include/linux/vmalloc.h index 506e7620a986..bbc051392298 100644 --- a/include/linux/vmalloc.h +++ b/include/linux/vmalloc.h @@ -106,5 +106,6 @@ extern long vwrite(char *buf, char *addr, unsigned long count); */ extern rwlock_t vmlist_lock; extern struct vm_struct *vmlist; +extern __init void vm_area_register_early(struct vm_struct *vm); #endif /* _LINUX_VMALLOC_H */ diff --git a/mm/vmalloc.c b/mm/vmalloc.c index c37924a2ee36..d206261ad9ef 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include @@ -982,6 +983,29 @@ void *vm_map_ram(struct page **pages, unsigned int count, int node, pgprot_t pro } EXPORT_SYMBOL(vm_map_ram); +/** + * vm_area_register_early - register vmap area early during boot + * @vm: vm_struct to register + * @size: size of area to register + * + * This function is used to register kernel vm area before + * vmalloc_init() is called. @vm->size and @vm->flags should contain + * proper values on entry and other fields should be zero. On return, + * vm->addr contains the allocated address. + * + * DO NOT USE THIS FUNCTION UNLESS YOU KNOW WHAT YOU'RE DOING. + */ +void __init vm_area_register_early(struct vm_struct *vm) +{ + static size_t vm_init_off __initdata; + + vm->addr = (void *)VMALLOC_START + vm_init_off; + vm_init_off = PFN_ALIGN(vm_init_off + vm->size); + + vm->next = vmlist; + vmlist = vm; +} + void __init vmalloc_init(void) { struct vmap_area *va; -- cgit v1.2.3-59-g8ed1b From 8fc48985006da4ceba24508db64ec77fc0dfe3bb Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 20 Feb 2009 16:29:08 +0900 Subject: vmalloc: add un/map_kernel_range_noflush() Impact: two more public map/unmap functions Implement map_kernel_range_noflush() and unmap_kernel_range_noflush(). These functions respectively map and unmap address range in kernel VM area but doesn't do any vcache or tlb flushing. These will be used by new percpu allocator. Signed-off-by: Tejun Heo Cc: Nick Piggin --- include/linux/vmalloc.h | 3 +++ mm/vmalloc.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++--- 2 files changed, 67 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/include/linux/vmalloc.h b/include/linux/vmalloc.h index bbc051392298..599ba7984310 100644 --- a/include/linux/vmalloc.h +++ b/include/linux/vmalloc.h @@ -91,6 +91,9 @@ extern struct vm_struct *remove_vm_area(const void *addr); extern int map_vm_area(struct vm_struct *area, pgprot_t prot, struct page ***pages); +extern int map_kernel_range_noflush(unsigned long start, unsigned long size, + pgprot_t prot, struct page **pages); +extern void unmap_kernel_range_noflush(unsigned long addr, unsigned long size); extern void unmap_kernel_range(unsigned long addr, unsigned long size); /* Allocate/destroy a 'vmalloc' VM area. */ diff --git a/mm/vmalloc.c b/mm/vmalloc.c index d206261ad9ef..224eca9650a8 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -153,8 +153,8 @@ static int vmap_pud_range(pgd_t *pgd, unsigned long addr, * * Ie. pte at addr+N*PAGE_SIZE shall point to pfn corresponding to pages[N] */ -static int vmap_page_range(unsigned long start, unsigned long end, - pgprot_t prot, struct page **pages) +static int vmap_page_range_noflush(unsigned long start, unsigned long end, + pgprot_t prot, struct page **pages) { pgd_t *pgd; unsigned long next; @@ -170,13 +170,22 @@ static int vmap_page_range(unsigned long start, unsigned long end, if (err) break; } while (pgd++, addr = next, addr != end); - flush_cache_vmap(start, end); if (unlikely(err)) return err; return nr; } +static int vmap_page_range(unsigned long start, unsigned long end, + pgprot_t prot, struct page **pages) +{ + int ret; + + ret = vmap_page_range_noflush(start, end, prot, pages); + flush_cache_vmap(start, end); + return ret; +} + static inline int is_vmalloc_or_module_addr(const void *x) { /* @@ -1033,6 +1042,58 @@ void __init vmalloc_init(void) vmap_initialized = true; } +/** + * map_kernel_range_noflush - map kernel VM area with the specified pages + * @addr: start of the VM area to map + * @size: size of the VM area to map + * @prot: page protection flags to use + * @pages: pages to map + * + * Map PFN_UP(@size) pages at @addr. The VM area @addr and @size + * specify should have been allocated using get_vm_area() and its + * friends. + * + * NOTE: + * This function does NOT do any cache flushing. The caller is + * responsible for calling flush_cache_vmap() on to-be-mapped areas + * before calling this function. + * + * RETURNS: + * The number of pages mapped on success, -errno on failure. + */ +int map_kernel_range_noflush(unsigned long addr, unsigned long size, + pgprot_t prot, struct page **pages) +{ + return vmap_page_range_noflush(addr, addr + size, prot, pages); +} + +/** + * unmap_kernel_range_noflush - unmap kernel VM area + * @addr: start of the VM area to unmap + * @size: size of the VM area to unmap + * + * Unmap PFN_UP(@size) pages at @addr. The VM area @addr and @size + * specify should have been allocated using get_vm_area() and its + * friends. + * + * NOTE: + * This function does NOT do any cache flushing. The caller is + * responsible for calling flush_cache_vunmap() on to-be-mapped areas + * before calling this function and flush_tlb_kernel_range() after. + */ +void unmap_kernel_range_noflush(unsigned long addr, unsigned long size) +{ + vunmap_page_range(addr, addr + size); +} + +/** + * unmap_kernel_range - unmap kernel VM area and flush cache and TLB + * @addr: start of the VM area to unmap + * @size: size of the VM area to unmap + * + * Similar to unmap_kernel_range_noflush() but flushes vcache before + * the unmapping and tlb after. + */ void unmap_kernel_range(unsigned long addr, unsigned long size) { unsigned long end = addr + size; -- cgit v1.2.3-59-g8ed1b From fbf59bc9d74d1fb30b8e0630743aff2806eafcea Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 20 Feb 2009 16:29:08 +0900 Subject: percpu: implement new dynamic percpu allocator Impact: new scalable dynamic percpu allocator which allows dynamic percpu areas to be accessed the same way as static ones Implement scalable dynamic percpu allocator which can be used for both static and dynamic percpu areas. This will allow static and dynamic areas to share faster direct access methods. This feature is optional and enabled only when CONFIG_HAVE_DYNAMIC_PER_CPU_AREA is defined by arch. Please read comment on top of mm/percpu.c for details. Signed-off-by: Tejun Heo Cc: Andrew Morton --- include/linux/percpu.h | 22 +- kernel/module.c | 31 ++ mm/Makefile | 4 + mm/percpu.c | 890 +++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 943 insertions(+), 4 deletions(-) create mode 100644 mm/percpu.c (limited to 'include/linux') diff --git a/include/linux/percpu.h b/include/linux/percpu.h index d99e24ae1811..18080995ff3e 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -76,23 +76,37 @@ #ifdef CONFIG_SMP -struct percpu_data { - void *ptrs[1]; -}; +#ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA -#define __percpu_disguise(pdata) (struct percpu_data *)~(unsigned long)(pdata) +extern void *pcpu_base_addr; +typedef void (*pcpu_populate_pte_fn_t)(unsigned long addr); + +extern size_t __init pcpu_setup_static(pcpu_populate_pte_fn_t populate_pte_fn, + struct page **pages, size_t cpu_size); /* * Use this to get to a cpu's version of the per-cpu object * dynamically allocated. Non-atomic access to the current CPU's * version should probably be combined with get_cpu()/put_cpu(). */ +#define per_cpu_ptr(ptr, cpu) SHIFT_PERCPU_PTR((ptr), per_cpu_offset((cpu))) + +#else /* CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */ + +struct percpu_data { + void *ptrs[1]; +}; + +#define __percpu_disguise(pdata) (struct percpu_data *)~(unsigned long)(pdata) + #define per_cpu_ptr(ptr, cpu) \ ({ \ struct percpu_data *__p = __percpu_disguise(ptr); \ (__typeof__(ptr))__p->ptrs[(cpu)]; \ }) +#endif /* CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */ + extern void *__alloc_percpu(size_t size, size_t align); extern void free_percpu(void *__pdata); diff --git a/kernel/module.c b/kernel/module.c index 52b3497b8748..1f0657ae555b 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -51,6 +51,7 @@ #include #include #include +#include #if 0 #define DEBUGP printk @@ -366,6 +367,34 @@ static struct module *find_module(const char *name) } #ifdef CONFIG_SMP + +#ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA + +static void *percpu_modalloc(unsigned long size, unsigned long align, + const char *name) +{ + void *ptr; + + if (align > PAGE_SIZE) { + printk(KERN_WARNING "%s: per-cpu alignment %li > %li\n", + name, align, PAGE_SIZE); + align = PAGE_SIZE; + } + + ptr = __alloc_percpu(size, align); + if (!ptr) + printk(KERN_WARNING + "Could not allocate %lu bytes percpu data\n", size); + return ptr; +} + +static void percpu_modfree(void *freeme) +{ + free_percpu(freeme); +} + +#else /* ... !CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */ + /* Number of blocks used and allocated. */ static unsigned int pcpu_num_used, pcpu_num_allocated; /* Size of each block. -ve means used. */ @@ -499,6 +528,8 @@ static int percpu_modinit(void) } __initcall(percpu_modinit); +#endif /* CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */ + static unsigned int find_pcpusec(Elf_Ehdr *hdr, Elf_Shdr *sechdrs, const char *secstrings) diff --git a/mm/Makefile b/mm/Makefile index 72255be57f89..818569b68f46 100644 --- a/mm/Makefile +++ b/mm/Makefile @@ -30,6 +30,10 @@ obj-$(CONFIG_FAILSLAB) += failslab.o obj-$(CONFIG_MEMORY_HOTPLUG) += memory_hotplug.o obj-$(CONFIG_FS_XIP) += filemap_xip.o obj-$(CONFIG_MIGRATION) += migrate.o +ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA +obj-$(CONFIG_SMP) += percpu.o +else obj-$(CONFIG_SMP) += allocpercpu.o +endif obj-$(CONFIG_QUICKLIST) += quicklist.o obj-$(CONFIG_CGROUP_MEM_RES_CTLR) += memcontrol.o page_cgroup.o diff --git a/mm/percpu.c b/mm/percpu.c new file mode 100644 index 000000000000..4617d97e877c --- /dev/null +++ b/mm/percpu.c @@ -0,0 +1,890 @@ +/* + * linux/mm/percpu.c - percpu memory allocator + * + * Copyright (C) 2009 SUSE Linux Products GmbH + * Copyright (C) 2009 Tejun Heo + * + * This file is released under the GPLv2. + * + * This is percpu allocator which can handle both static and dynamic + * areas. Percpu areas are allocated in chunks in vmalloc area. Each + * chunk is consisted of num_possible_cpus() units and the first chunk + * is used for static percpu variables in the kernel image (special + * boot time alloc/init handling necessary as these areas need to be + * brought up before allocation services are running). Unit grows as + * necessary and all units grow or shrink in unison. When a chunk is + * filled up, another chunk is allocated. ie. in vmalloc area + * + * c0 c1 c2 + * ------------------- ------------------- ------------ + * | u0 | u1 | u2 | u3 | | u0 | u1 | u2 | u3 | | u0 | u1 | u + * ------------------- ...... ------------------- .... ------------ + * + * Allocation is done in offset-size areas of single unit space. Ie, + * an area of 512 bytes at 6k in c1 occupies 512 bytes at 6k of c1:u0, + * c1:u1, c1:u2 and c1:u3. Percpu access can be done by configuring + * percpu base registers UNIT_SIZE apart. + * + * There are usually many small percpu allocations many of them as + * small as 4 bytes. The allocator organizes chunks into lists + * according to free size and tries to allocate from the fullest one. + * Each chunk keeps the maximum contiguous area size hint which is + * guaranteed to be eqaul to or larger than the maximum contiguous + * area in the chunk. This helps the allocator not to iterate the + * chunk maps unnecessarily. + * + * Allocation state in each chunk is kept using an array of integers + * on chunk->map. A positive value in the map represents a free + * region and negative allocated. Allocation inside a chunk is done + * by scanning this map sequentially and serving the first matching + * entry. This is mostly copied from the percpu_modalloc() allocator. + * Chunks are also linked into a rb tree to ease address to chunk + * mapping during free. + * + * To use this allocator, arch code should do the followings. + * + * - define CONFIG_HAVE_DYNAMIC_PER_CPU_AREA + * + * - define __addr_to_pcpu_ptr() and __pcpu_ptr_to_addr() to translate + * regular address to percpu pointer and back + * + * - use pcpu_setup_static() during percpu area initialization to + * setup kernel static percpu area + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#define PCPU_MIN_UNIT_PAGES_SHIFT 4 /* also max alloc size */ +#define PCPU_SLOT_BASE_SHIFT 5 /* 1-31 shares the same slot */ +#define PCPU_DFL_MAP_ALLOC 16 /* start a map with 16 ents */ + +struct pcpu_chunk { + struct list_head list; /* linked to pcpu_slot lists */ + struct rb_node rb_node; /* key is chunk->vm->addr */ + int free_size; /* free bytes in the chunk */ + int contig_hint; /* max contiguous size hint */ + struct vm_struct *vm; /* mapped vmalloc region */ + int map_used; /* # of map entries used */ + int map_alloc; /* # of map entries allocated */ + int *map; /* allocation map */ + struct page *page[]; /* #cpus * UNIT_PAGES */ +}; + +static int pcpu_unit_pages_shift; +static int pcpu_unit_pages; +static int pcpu_unit_shift; +static int pcpu_unit_size; +static int pcpu_chunk_size; +static int pcpu_nr_slots; +static size_t pcpu_chunk_struct_size; + +/* the address of the first chunk which starts with the kernel static area */ +void *pcpu_base_addr; +EXPORT_SYMBOL_GPL(pcpu_base_addr); + +/* the size of kernel static area */ +static int pcpu_static_size; + +/* + * One mutex to rule them all. + * + * The following mutex is grabbed in the outermost public alloc/free + * interface functions and released only when the operation is + * complete. As such, every function in this file other than the + * outermost functions are called under pcpu_mutex. + * + * It can easily be switched to use spinlock such that only the area + * allocation and page population commit are protected with it doing + * actual [de]allocation without holding any lock. However, given + * what this allocator does, I think it's better to let them run + * sequentially. + */ +static DEFINE_MUTEX(pcpu_mutex); + +static struct list_head *pcpu_slot; /* chunk list slots */ +static struct rb_root pcpu_addr_root = RB_ROOT; /* chunks by address */ + +static int pcpu_size_to_slot(int size) +{ + int highbit = fls(size); + return max(highbit - PCPU_SLOT_BASE_SHIFT + 2, 1); +} + +static int pcpu_chunk_slot(const struct pcpu_chunk *chunk) +{ + if (chunk->free_size < sizeof(int) || chunk->contig_hint < sizeof(int)) + return 0; + + return pcpu_size_to_slot(chunk->free_size); +} + +static int pcpu_page_idx(unsigned int cpu, int page_idx) +{ + return (cpu << pcpu_unit_pages_shift) + page_idx; +} + +static struct page **pcpu_chunk_pagep(struct pcpu_chunk *chunk, + unsigned int cpu, int page_idx) +{ + return &chunk->page[pcpu_page_idx(cpu, page_idx)]; +} + +static unsigned long pcpu_chunk_addr(struct pcpu_chunk *chunk, + unsigned int cpu, int page_idx) +{ + return (unsigned long)chunk->vm->addr + + (pcpu_page_idx(cpu, page_idx) << PAGE_SHIFT); +} + +static bool pcpu_chunk_page_occupied(struct pcpu_chunk *chunk, + int page_idx) +{ + return *pcpu_chunk_pagep(chunk, 0, page_idx) != NULL; +} + +/** + * pcpu_realloc - versatile realloc + * @p: the current pointer (can be NULL for new allocations) + * @size: the current size (can be 0 for new allocations) + * @new_size: the wanted new size (can be 0 for free) + * + * More robust realloc which can be used to allocate, resize or free a + * memory area of arbitrary size. If the needed size goes over + * PAGE_SIZE, kernel VM is used. + * + * RETURNS: + * The new pointer on success, NULL on failure. + */ +static void *pcpu_realloc(void *p, size_t size, size_t new_size) +{ + void *new; + + if (new_size <= PAGE_SIZE) + new = kmalloc(new_size, GFP_KERNEL); + else + new = vmalloc(new_size); + if (new_size && !new) + return NULL; + + memcpy(new, p, min(size, new_size)); + if (new_size > size) + memset(new + size, 0, new_size - size); + + if (size <= PAGE_SIZE) + kfree(p); + else + vfree(p); + + return new; +} + +/** + * pcpu_chunk_relocate - put chunk in the appropriate chunk slot + * @chunk: chunk of interest + * @oslot: the previous slot it was on + * + * This function is called after an allocation or free changed @chunk. + * New slot according to the changed state is determined and @chunk is + * moved to the slot. + */ +static void pcpu_chunk_relocate(struct pcpu_chunk *chunk, int oslot) +{ + int nslot = pcpu_chunk_slot(chunk); + + if (oslot != nslot) { + if (oslot < nslot) + list_move(&chunk->list, &pcpu_slot[nslot]); + else + list_move_tail(&chunk->list, &pcpu_slot[nslot]); + } +} + +static struct rb_node **pcpu_chunk_rb_search(void *addr, + struct rb_node **parentp) +{ + struct rb_node **p = &pcpu_addr_root.rb_node; + struct rb_node *parent = NULL; + struct pcpu_chunk *chunk; + + while (*p) { + parent = *p; + chunk = rb_entry(parent, struct pcpu_chunk, rb_node); + + if (addr < chunk->vm->addr) + p = &(*p)->rb_left; + else if (addr > chunk->vm->addr) + p = &(*p)->rb_right; + else + break; + } + + if (parentp) + *parentp = parent; + return p; +} + +/** + * pcpu_chunk_addr_search - search for chunk containing specified address + * @addr: address to search for + * + * Look for chunk which might contain @addr. More specifically, it + * searchs for the chunk with the highest start address which isn't + * beyond @addr. + * + * RETURNS: + * The address of the found chunk. + */ +static struct pcpu_chunk *pcpu_chunk_addr_search(void *addr) +{ + struct rb_node *n, *parent; + struct pcpu_chunk *chunk; + + n = *pcpu_chunk_rb_search(addr, &parent); + if (!n) { + /* no exactly matching chunk, the parent is the closest */ + n = parent; + BUG_ON(!n); + } + chunk = rb_entry(n, struct pcpu_chunk, rb_node); + + if (addr < chunk->vm->addr) { + /* the parent was the next one, look for the previous one */ + n = rb_prev(n); + BUG_ON(!n); + chunk = rb_entry(n, struct pcpu_chunk, rb_node); + } + + return chunk; +} + +/** + * pcpu_chunk_addr_insert - insert chunk into address rb tree + * @new: chunk to insert + * + * Insert @new into address rb tree. + */ +static void pcpu_chunk_addr_insert(struct pcpu_chunk *new) +{ + struct rb_node **p, *parent; + + p = pcpu_chunk_rb_search(new->vm->addr, &parent); + BUG_ON(*p); + rb_link_node(&new->rb_node, parent, p); + rb_insert_color(&new->rb_node, &pcpu_addr_root); +} + +/** + * pcpu_split_block - split a map block + * @chunk: chunk of interest + * @i: index of map block to split + * @head: head size (can be 0) + * @tail: tail size (can be 0) + * + * Split the @i'th map block into two or three blocks. If @head is + * non-zero, @head bytes block is inserted before block @i moving it + * to @i+1 and reducing its size by @head bytes. + * + * If @tail is non-zero, the target block, which can be @i or @i+1 + * depending on @head, is reduced by @tail bytes and @tail byte block + * is inserted after the target block. + * + * RETURNS: + * 0 on success, -errno on failure. + */ +static int pcpu_split_block(struct pcpu_chunk *chunk, int i, int head, int tail) +{ + int nr_extra = !!head + !!tail; + int target = chunk->map_used + nr_extra; + + /* reallocation required? */ + if (chunk->map_alloc < target) { + int new_alloc = chunk->map_alloc; + int *new; + + while (new_alloc < target) + new_alloc *= 2; + + new = pcpu_realloc(chunk->map, + chunk->map_alloc * sizeof(new[0]), + new_alloc * sizeof(new[0])); + if (!new) + return -ENOMEM; + + chunk->map_alloc = new_alloc; + chunk->map = new; + } + + /* insert a new subblock */ + memmove(&chunk->map[i + nr_extra], &chunk->map[i], + sizeof(chunk->map[0]) * (chunk->map_used - i)); + chunk->map_used += nr_extra; + + if (head) { + chunk->map[i + 1] = chunk->map[i] - head; + chunk->map[i++] = head; + } + if (tail) { + chunk->map[i++] -= tail; + chunk->map[i] = tail; + } + return 0; +} + +/** + * pcpu_alloc_area - allocate area from a pcpu_chunk + * @chunk: chunk of interest + * @size: wanted size + * @align: wanted align + * + * Try to allocate @size bytes area aligned at @align from @chunk. + * Note that this function only allocates the offset. It doesn't + * populate or map the area. + * + * RETURNS: + * Allocated offset in @chunk on success, -errno on failure. + */ +static int pcpu_alloc_area(struct pcpu_chunk *chunk, int size, int align) +{ + int oslot = pcpu_chunk_slot(chunk); + int max_contig = 0; + int i, off; + + /* + * The static chunk initially doesn't have map attached + * because kmalloc wasn't available during init. Give it one. + */ + if (unlikely(!chunk->map)) { + chunk->map = pcpu_realloc(NULL, 0, + PCPU_DFL_MAP_ALLOC * sizeof(chunk->map[0])); + if (!chunk->map) + return -ENOMEM; + + chunk->map_alloc = PCPU_DFL_MAP_ALLOC; + chunk->map[chunk->map_used++] = -pcpu_static_size; + if (chunk->free_size) + chunk->map[chunk->map_used++] = chunk->free_size; + } + + for (i = 0, off = 0; i < chunk->map_used; off += abs(chunk->map[i++])) { + bool is_last = i + 1 == chunk->map_used; + int head, tail; + + /* extra for alignment requirement */ + head = ALIGN(off, align) - off; + BUG_ON(i == 0 && head != 0); + + if (chunk->map[i] < 0) + continue; + if (chunk->map[i] < head + size) { + max_contig = max(chunk->map[i], max_contig); + continue; + } + + /* + * If head is small or the previous block is free, + * merge'em. Note that 'small' is defined as smaller + * than sizeof(int), which is very small but isn't too + * uncommon for percpu allocations. + */ + if (head && (head < sizeof(int) || chunk->map[i - 1] > 0)) { + if (chunk->map[i - 1] > 0) + chunk->map[i - 1] += head; + else { + chunk->map[i - 1] -= head; + chunk->free_size -= head; + } + chunk->map[i] -= head; + off += head; + head = 0; + } + + /* if tail is small, just keep it around */ + tail = chunk->map[i] - head - size; + if (tail < sizeof(int)) + tail = 0; + + /* split if warranted */ + if (head || tail) { + if (pcpu_split_block(chunk, i, head, tail)) + return -ENOMEM; + if (head) { + i++; + off += head; + max_contig = max(chunk->map[i - 1], max_contig); + } + if (tail) + max_contig = max(chunk->map[i + 1], max_contig); + } + + /* update hint and mark allocated */ + if (is_last) + chunk->contig_hint = max_contig; /* fully scanned */ + else + chunk->contig_hint = max(chunk->contig_hint, + max_contig); + + chunk->free_size -= chunk->map[i]; + chunk->map[i] = -chunk->map[i]; + + pcpu_chunk_relocate(chunk, oslot); + return off; + } + + chunk->contig_hint = max_contig; /* fully scanned */ + pcpu_chunk_relocate(chunk, oslot); + + /* + * Tell the upper layer that this chunk has no area left. + * Note that this is not an error condition but a notification + * to upper layer that it needs to look at other chunks. + * -ENOSPC is chosen as it isn't used in memory subsystem and + * matches the meaning in a way. + */ + return -ENOSPC; +} + +/** + * pcpu_free_area - free area to a pcpu_chunk + * @chunk: chunk of interest + * @freeme: offset of area to free + * + * Free area starting from @freeme to @chunk. Note that this function + * only modifies the allocation map. It doesn't depopulate or unmap + * the area. + */ +static void pcpu_free_area(struct pcpu_chunk *chunk, int freeme) +{ + int oslot = pcpu_chunk_slot(chunk); + int i, off; + + for (i = 0, off = 0; i < chunk->map_used; off += abs(chunk->map[i++])) + if (off == freeme) + break; + BUG_ON(off != freeme); + BUG_ON(chunk->map[i] > 0); + + chunk->map[i] = -chunk->map[i]; + chunk->free_size += chunk->map[i]; + + /* merge with previous? */ + if (i > 0 && chunk->map[i - 1] >= 0) { + chunk->map[i - 1] += chunk->map[i]; + chunk->map_used--; + memmove(&chunk->map[i], &chunk->map[i + 1], + (chunk->map_used - i) * sizeof(chunk->map[0])); + i--; + } + /* merge with next? */ + if (i + 1 < chunk->map_used && chunk->map[i + 1] >= 0) { + chunk->map[i] += chunk->map[i + 1]; + chunk->map_used--; + memmove(&chunk->map[i + 1], &chunk->map[i + 2], + (chunk->map_used - (i + 1)) * sizeof(chunk->map[0])); + } + + chunk->contig_hint = max(chunk->map[i], chunk->contig_hint); + pcpu_chunk_relocate(chunk, oslot); +} + +/** + * pcpu_unmap - unmap pages out of a pcpu_chunk + * @chunk: chunk of interest + * @page_start: page index of the first page to unmap + * @page_end: page index of the last page to unmap + 1 + * @flush: whether to flush cache and tlb or not + * + * For each cpu, unmap pages [@page_start,@page_end) out of @chunk. + * If @flush is true, vcache is flushed before unmapping and tlb + * after. + */ +static void pcpu_unmap(struct pcpu_chunk *chunk, int page_start, int page_end, + bool flush) +{ + unsigned int last = num_possible_cpus() - 1; + unsigned int cpu; + + /* + * Each flushing trial can be very expensive, issue flush on + * the whole region at once rather than doing it for each cpu. + * This could be an overkill but is more scalable. + */ + if (flush) + flush_cache_vunmap(pcpu_chunk_addr(chunk, 0, page_start), + pcpu_chunk_addr(chunk, last, page_end)); + + for_each_possible_cpu(cpu) + unmap_kernel_range_noflush( + pcpu_chunk_addr(chunk, cpu, page_start), + (page_end - page_start) << PAGE_SHIFT); + + /* ditto as flush_cache_vunmap() */ + if (flush) + flush_tlb_kernel_range(pcpu_chunk_addr(chunk, 0, page_start), + pcpu_chunk_addr(chunk, last, page_end)); +} + +/** + * pcpu_depopulate_chunk - depopulate and unmap an area of a pcpu_chunk + * @chunk: chunk to depopulate + * @off: offset to the area to depopulate + * @size: size of the area to depopulate + * @flush: whether to flush cache and tlb or not + * + * For each cpu, depopulate and unmap pages [@page_start,@page_end) + * from @chunk. If @flush is true, vcache is flushed before unmapping + * and tlb after. + */ +static void pcpu_depopulate_chunk(struct pcpu_chunk *chunk, size_t off, + size_t size, bool flush) +{ + int page_start = PFN_DOWN(off); + int page_end = PFN_UP(off + size); + int unmap_start = -1; + int uninitialized_var(unmap_end); + unsigned int cpu; + int i; + + for (i = page_start; i < page_end; i++) { + for_each_possible_cpu(cpu) { + struct page **pagep = pcpu_chunk_pagep(chunk, cpu, i); + + if (!*pagep) + continue; + + __free_page(*pagep); + + /* + * If it's partial depopulation, it might get + * populated or depopulated again. Mark the + * page gone. + */ + *pagep = NULL; + + unmap_start = unmap_start < 0 ? i : unmap_start; + unmap_end = i + 1; + } + } + + if (unmap_start >= 0) + pcpu_unmap(chunk, unmap_start, unmap_end, flush); +} + +/** + * pcpu_map - map pages into a pcpu_chunk + * @chunk: chunk of interest + * @page_start: page index of the first page to map + * @page_end: page index of the last page to map + 1 + * + * For each cpu, map pages [@page_start,@page_end) into @chunk. + * vcache is flushed afterwards. + */ +static int pcpu_map(struct pcpu_chunk *chunk, int page_start, int page_end) +{ + unsigned int last = num_possible_cpus() - 1; + unsigned int cpu; + int err; + + for_each_possible_cpu(cpu) { + err = map_kernel_range_noflush( + pcpu_chunk_addr(chunk, cpu, page_start), + (page_end - page_start) << PAGE_SHIFT, + PAGE_KERNEL, + pcpu_chunk_pagep(chunk, cpu, page_start)); + if (err < 0) + return err; + } + + /* flush at once, please read comments in pcpu_unmap() */ + flush_cache_vmap(pcpu_chunk_addr(chunk, 0, page_start), + pcpu_chunk_addr(chunk, last, page_end)); + return 0; +} + +/** + * pcpu_populate_chunk - populate and map an area of a pcpu_chunk + * @chunk: chunk of interest + * @off: offset to the area to populate + * @size: size of the area to populate + * + * For each cpu, populate and map pages [@page_start,@page_end) into + * @chunk. The area is cleared on return. + */ +static int pcpu_populate_chunk(struct pcpu_chunk *chunk, int off, int size) +{ + const gfp_t alloc_mask = GFP_KERNEL | __GFP_HIGHMEM | __GFP_COLD; + int page_start = PFN_DOWN(off); + int page_end = PFN_UP(off + size); + int map_start = -1; + int map_end; + unsigned int cpu; + int i; + + for (i = page_start; i < page_end; i++) { + if (pcpu_chunk_page_occupied(chunk, i)) { + if (map_start >= 0) { + if (pcpu_map(chunk, map_start, map_end)) + goto err; + map_start = -1; + } + continue; + } + + map_start = map_start < 0 ? i : map_start; + map_end = i + 1; + + for_each_possible_cpu(cpu) { + struct page **pagep = pcpu_chunk_pagep(chunk, cpu, i); + + *pagep = alloc_pages_node(cpu_to_node(cpu), + alloc_mask, 0); + if (!*pagep) + goto err; + } + } + + if (map_start >= 0 && pcpu_map(chunk, map_start, map_end)) + goto err; + + for_each_possible_cpu(cpu) + memset(chunk->vm->addr + (cpu << pcpu_unit_shift) + off, 0, + size); + + return 0; +err: + /* likely under heavy memory pressure, give memory back */ + pcpu_depopulate_chunk(chunk, off, size, true); + return -ENOMEM; +} + +static void free_pcpu_chunk(struct pcpu_chunk *chunk) +{ + if (!chunk) + return; + if (chunk->vm) + free_vm_area(chunk->vm); + pcpu_realloc(chunk->map, chunk->map_alloc * sizeof(chunk->map[0]), 0); + kfree(chunk); +} + +static struct pcpu_chunk *alloc_pcpu_chunk(void) +{ + struct pcpu_chunk *chunk; + + chunk = kzalloc(pcpu_chunk_struct_size, GFP_KERNEL); + if (!chunk) + return NULL; + + chunk->map = pcpu_realloc(NULL, 0, + PCPU_DFL_MAP_ALLOC * sizeof(chunk->map[0])); + chunk->map_alloc = PCPU_DFL_MAP_ALLOC; + chunk->map[chunk->map_used++] = pcpu_unit_size; + + chunk->vm = get_vm_area(pcpu_chunk_size, GFP_KERNEL); + if (!chunk->vm) { + free_pcpu_chunk(chunk); + return NULL; + } + + INIT_LIST_HEAD(&chunk->list); + chunk->free_size = pcpu_unit_size; + chunk->contig_hint = pcpu_unit_size; + + return chunk; +} + +/** + * __alloc_percpu - allocate percpu area + * @size: size of area to allocate + * @align: alignment of area (max PAGE_SIZE) + * + * Allocate percpu area of @size bytes aligned at @align. Might + * sleep. Might trigger writeouts. + * + * RETURNS: + * Percpu pointer to the allocated area on success, NULL on failure. + */ +void *__alloc_percpu(size_t size, size_t align) +{ + void *ptr = NULL; + struct pcpu_chunk *chunk; + int slot, off; + + if (unlikely(!size || size > PAGE_SIZE << PCPU_MIN_UNIT_PAGES_SHIFT || + align > PAGE_SIZE)) { + WARN(true, "illegal size (%zu) or align (%zu) for " + "percpu allocation\n", size, align); + return NULL; + } + + mutex_lock(&pcpu_mutex); + + /* allocate area */ + for (slot = pcpu_size_to_slot(size); slot < pcpu_nr_slots; slot++) { + list_for_each_entry(chunk, &pcpu_slot[slot], list) { + if (size > chunk->contig_hint) + continue; + off = pcpu_alloc_area(chunk, size, align); + if (off >= 0) + goto area_found; + if (off != -ENOSPC) + goto out_unlock; + } + } + + /* hmmm... no space left, create a new chunk */ + chunk = alloc_pcpu_chunk(); + if (!chunk) + goto out_unlock; + pcpu_chunk_relocate(chunk, -1); + pcpu_chunk_addr_insert(chunk); + + off = pcpu_alloc_area(chunk, size, align); + if (off < 0) + goto out_unlock; + +area_found: + /* populate, map and clear the area */ + if (pcpu_populate_chunk(chunk, off, size)) { + pcpu_free_area(chunk, off); + goto out_unlock; + } + + ptr = __addr_to_pcpu_ptr(chunk->vm->addr + off); +out_unlock: + mutex_unlock(&pcpu_mutex); + return ptr; +} +EXPORT_SYMBOL_GPL(__alloc_percpu); + +static void pcpu_kill_chunk(struct pcpu_chunk *chunk) +{ + pcpu_depopulate_chunk(chunk, 0, pcpu_unit_size, false); + list_del(&chunk->list); + rb_erase(&chunk->rb_node, &pcpu_addr_root); + free_pcpu_chunk(chunk); +} + +/** + * free_percpu - free percpu area + * @ptr: pointer to area to free + * + * Free percpu area @ptr. Might sleep. + */ +void free_percpu(void *ptr) +{ + void *addr = __pcpu_ptr_to_addr(ptr); + struct pcpu_chunk *chunk; + int off; + + if (!ptr) + return; + + mutex_lock(&pcpu_mutex); + + chunk = pcpu_chunk_addr_search(addr); + off = addr - chunk->vm->addr; + + pcpu_free_area(chunk, off); + + /* the chunk became fully free, kill one if there are other free ones */ + if (chunk->free_size == pcpu_unit_size) { + struct pcpu_chunk *pos; + + list_for_each_entry(pos, + &pcpu_slot[pcpu_chunk_slot(chunk)], list) + if (pos != chunk) { + pcpu_kill_chunk(pos); + break; + } + } + + mutex_unlock(&pcpu_mutex); +} +EXPORT_SYMBOL_GPL(free_percpu); + +/** + * pcpu_setup_static - initialize kernel static percpu area + * @populate_pte_fn: callback to allocate pagetable + * @pages: num_possible_cpus() * PFN_UP(cpu_size) pages + * + * Initialize kernel static percpu area. The caller should allocate + * all the necessary pages and pass them in @pages. + * @populate_pte_fn() is called on each page to be used for percpu + * mapping and is responsible for making sure all the necessary page + * tables for the page is allocated. + * + * RETURNS: + * The determined pcpu_unit_size which can be used to initialize + * percpu access. + */ +size_t __init pcpu_setup_static(pcpu_populate_pte_fn_t populate_pte_fn, + struct page **pages, size_t cpu_size) +{ + static struct vm_struct static_vm; + struct pcpu_chunk *static_chunk; + int nr_cpu_pages = DIV_ROUND_UP(cpu_size, PAGE_SIZE); + unsigned int cpu; + int err, i; + + pcpu_unit_pages_shift = max_t(int, PCPU_MIN_UNIT_PAGES_SHIFT, + order_base_2(cpu_size) - PAGE_SHIFT); + + pcpu_static_size = cpu_size; + pcpu_unit_pages = 1 << pcpu_unit_pages_shift; + pcpu_unit_shift = PAGE_SHIFT + pcpu_unit_pages_shift; + pcpu_unit_size = 1 << pcpu_unit_shift; + pcpu_chunk_size = num_possible_cpus() * pcpu_unit_size; + pcpu_nr_slots = pcpu_size_to_slot(pcpu_unit_size) + 1; + pcpu_chunk_struct_size = sizeof(struct pcpu_chunk) + + (1 << pcpu_unit_pages_shift) * sizeof(struct page *); + + /* allocate chunk slots */ + pcpu_slot = alloc_bootmem(pcpu_nr_slots * sizeof(pcpu_slot[0])); + for (i = 0; i < pcpu_nr_slots; i++) + INIT_LIST_HEAD(&pcpu_slot[i]); + + /* init and register vm area */ + static_vm.flags = VM_ALLOC; + static_vm.size = pcpu_chunk_size; + vm_area_register_early(&static_vm); + + /* init static_chunk */ + static_chunk = alloc_bootmem(pcpu_chunk_struct_size); + INIT_LIST_HEAD(&static_chunk->list); + static_chunk->vm = &static_vm; + static_chunk->free_size = pcpu_unit_size - pcpu_static_size; + static_chunk->contig_hint = static_chunk->free_size; + + /* assign pages and map them */ + for_each_possible_cpu(cpu) { + for (i = 0; i < nr_cpu_pages; i++) { + *pcpu_chunk_pagep(static_chunk, cpu, i) = *pages++; + populate_pte_fn(pcpu_chunk_addr(static_chunk, cpu, i)); + } + } + + err = pcpu_map(static_chunk, 0, nr_cpu_pages); + if (err) + panic("failed to setup static percpu area, err=%d\n", err); + + /* link static_chunk in */ + pcpu_chunk_relocate(static_chunk, -1); + pcpu_chunk_addr_insert(static_chunk); + + /* we're done */ + pcpu_base_addr = (void *)pcpu_chunk_addr(static_chunk, 0, 0); + return pcpu_unit_size; +} -- cgit v1.2.3-59-g8ed1b From b814d41f0987c7648d7ed07471258101c95c026b Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 20 Feb 2009 22:32:10 +0100 Subject: x86, mm: fault.c, simplify kmmio_fault() Impact: cleanup Remove an #ifdef from kmmio_fault() - we can do this by providing default implementations for is_kmmio_active() and kmmio_handler(). The compiler optimizes it all away in the !CONFIG_MMIOTRACE case. Also, while at it, clean up mmiotrace.h a bit: - standard header guards - standard vertical spaces for structure definitions No code changed (both with mmiotrace on and off in the config): text data bss dec hex filename 2947 12 12 2971 b9b fault.o.before 2947 12 12 2971 b9b fault.o.after Cc: Pekka Paalanen Cc: Linus Torvalds Signed-off-by: Ingo Molnar --- arch/x86/mm/fault.c | 5 +-- include/linux/mmiotrace.h | 78 ++++++++++++++++++++++++++++------------------- 2 files changed, 50 insertions(+), 33 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 3e3661462739..fe99af4b86d9 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -55,13 +55,14 @@ enum x86_pf_error_code { PF_INSTR = 1 << 4, }; +/* + * (returns 0 if mmiotrace is disabled) + */ static inline int kmmio_fault(struct pt_regs *regs, unsigned long addr) { -#ifdef CONFIG_MMIOTRACE if (unlikely(is_kmmio_active())) if (kmmio_handler(regs, addr) == 1) return -1; -#endif return 0; } diff --git a/include/linux/mmiotrace.h b/include/linux/mmiotrace.h index 139d7c88d9c9..3d1b7bde1283 100644 --- a/include/linux/mmiotrace.h +++ b/include/linux/mmiotrace.h @@ -1,5 +1,5 @@ -#ifndef MMIOTRACE_H -#define MMIOTRACE_H +#ifndef _LINUX_MMIOTRACE_H +#define _LINUX_MMIOTRACE_H #include #include @@ -13,28 +13,34 @@ typedef void (*kmmio_post_handler_t)(struct kmmio_probe *, unsigned long condition, struct pt_regs *); struct kmmio_probe { - struct list_head list; /* kmmio internal list */ - unsigned long addr; /* start location of the probe point */ - unsigned long len; /* length of the probe region */ - kmmio_pre_handler_t pre_handler; /* Called before addr is executed. */ - kmmio_post_handler_t post_handler; /* Called after addr is executed */ - void *private; + /* kmmio internal list: */ + struct list_head list; + /* start location of the probe point: */ + unsigned long addr; + /* length of the probe region: */ + unsigned long len; + /* Called before addr is executed: */ + kmmio_pre_handler_t pre_handler; + /* Called after addr is executed: */ + kmmio_post_handler_t post_handler; + void *private; }; +extern unsigned int kmmio_count; + +extern int register_kmmio_probe(struct kmmio_probe *p); +extern void unregister_kmmio_probe(struct kmmio_probe *p); + +#ifdef CONFIG_MMIOTRACE /* kmmio is active by some kmmio_probes? */ static inline int is_kmmio_active(void) { - extern unsigned int kmmio_count; return kmmio_count; } -extern int register_kmmio_probe(struct kmmio_probe *p); -extern void unregister_kmmio_probe(struct kmmio_probe *p); - /* Called from page fault handler. */ extern int kmmio_handler(struct pt_regs *regs, unsigned long addr); -#ifdef CONFIG_MMIOTRACE /* Called from ioremap.c */ extern void mmiotrace_ioremap(resource_size_t offset, unsigned long size, void __iomem *addr); @@ -43,7 +49,17 @@ extern void mmiotrace_iounmap(volatile void __iomem *addr); /* For anyone to insert markers. Remember trailing newline. */ extern int mmiotrace_printk(const char *fmt, ...) __attribute__ ((format (printf, 1, 2))); -#else +#else /* !CONFIG_MMIOTRACE: */ +static inline int is_kmmio_active(void) +{ + return 0; +} + +static inline int kmmio_handler(struct pt_regs *regs, unsigned long addr) +{ + return 0; +} + static inline void mmiotrace_ioremap(resource_size_t offset, unsigned long size, void __iomem *addr) { @@ -63,28 +79,28 @@ static inline int mmiotrace_printk(const char *fmt, ...) #endif /* CONFIG_MMIOTRACE */ enum mm_io_opcode { - MMIO_READ = 0x1, /* struct mmiotrace_rw */ - MMIO_WRITE = 0x2, /* struct mmiotrace_rw */ - MMIO_PROBE = 0x3, /* struct mmiotrace_map */ - MMIO_UNPROBE = 0x4, /* struct mmiotrace_map */ - MMIO_UNKNOWN_OP = 0x5, /* struct mmiotrace_rw */ + MMIO_READ = 0x1, /* struct mmiotrace_rw */ + MMIO_WRITE = 0x2, /* struct mmiotrace_rw */ + MMIO_PROBE = 0x3, /* struct mmiotrace_map */ + MMIO_UNPROBE = 0x4, /* struct mmiotrace_map */ + MMIO_UNKNOWN_OP = 0x5, /* struct mmiotrace_rw */ }; struct mmiotrace_rw { - resource_size_t phys; /* PCI address of register */ - unsigned long value; - unsigned long pc; /* optional program counter */ - int map_id; - unsigned char opcode; /* one of MMIO_{READ,WRITE,UNKNOWN_OP} */ - unsigned char width; /* size of register access in bytes */ + resource_size_t phys; /* PCI address of register */ + unsigned long value; + unsigned long pc; /* optional program counter */ + int map_id; + unsigned char opcode; /* one of MMIO_{READ,WRITE,UNKNOWN_OP} */ + unsigned char width; /* size of register access in bytes */ }; struct mmiotrace_map { - resource_size_t phys; /* base address in PCI space */ - unsigned long virt; /* base virtual address */ - unsigned long len; /* mapping size */ - int map_id; - unsigned char opcode; /* MMIO_PROBE or MMIO_UNPROBE */ + resource_size_t phys; /* base address in PCI space */ + unsigned long virt; /* base virtual address */ + unsigned long len; /* mapping size */ + int map_id; + unsigned char opcode; /* MMIO_PROBE or MMIO_UNPROBE */ }; /* in kernel/trace/trace_mmiotrace.c */ @@ -94,4 +110,4 @@ extern void mmio_trace_rw(struct mmiotrace_rw *rw); extern void mmio_trace_mapping(struct mmiotrace_map *map); extern int mmio_trace_printk(const char *fmt, va_list args); -#endif /* MMIOTRACE_H */ +#endif /* _LINUX_MMIOTRACE_H */ -- cgit v1.2.3-59-g8ed1b From b18018126f422f5b706fd750373425e10e84b486 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 20 Feb 2009 22:42:57 +0100 Subject: x86, mm, kprobes: fault.c, simplify notify_page_fault() Impact: cleanup Remove an #ifdef from notify_page_fault(). The function still compiles to nothing in the !CONFIG_KPROBES case. Introduce kprobes_built_in() and kprobe_fault_handler() helpers to allow this - they returns 0 if !CONFIG_KPROBES. No code changed: text data bss dec hex filename 4618 32 24 4674 1242 fault.o.before 4618 32 24 4674 1242 fault.o.after Cc: Masami Hiramatsu Cc: Andrew Morton Cc: Linus Torvalds Signed-off-by: Ingo Molnar --- arch/x86/mm/fault.c | 6 +----- include/linux/kprobes.h | 22 +++++++++++++++++++--- 2 files changed, 20 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index fe99af4b86d9..379beaec6caa 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -68,11 +68,10 @@ static inline int kmmio_fault(struct pt_regs *regs, unsigned long addr) static inline int notify_page_fault(struct pt_regs *regs) { -#ifdef CONFIG_KPROBES int ret = 0; /* kprobe_running() needs smp_processor_id() */ - if (!user_mode_vm(regs)) { + if (kprobes_built_in() && !user_mode_vm(regs)) { preempt_disable(); if (kprobe_running() && kprobe_fault_handler(regs, 14)) ret = 1; @@ -80,9 +79,6 @@ static inline int notify_page_fault(struct pt_regs *regs) } return ret; -#else - return 0; -#endif } /* diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h index 32851eef48f0..2ec6cc14a114 100644 --- a/include/linux/kprobes.h +++ b/include/linux/kprobes.h @@ -182,6 +182,14 @@ struct kprobe_blackpoint { DECLARE_PER_CPU(struct kprobe *, current_kprobe); DECLARE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk); +/* + * For #ifdef avoidance: + */ +static inline int kprobes_built_in(void) +{ + return 1; +} + #ifdef CONFIG_KRETPROBES extern void arch_prepare_kretprobe(struct kretprobe_instance *ri, struct pt_regs *regs); @@ -271,8 +279,16 @@ void unregister_kretprobes(struct kretprobe **rps, int num); void kprobe_flush_task(struct task_struct *tk); void recycle_rp_inst(struct kretprobe_instance *ri, struct hlist_head *head); -#else /* CONFIG_KPROBES */ +#else /* !CONFIG_KPROBES: */ +static inline int kprobes_built_in(void) +{ + return 0; +} +static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) +{ + return 0; +} static inline struct kprobe *get_kprobe(void *addr) { return NULL; @@ -329,5 +345,5 @@ static inline void unregister_kretprobes(struct kretprobe **rps, int num) static inline void kprobe_flush_task(struct task_struct *tk) { } -#endif /* CONFIG_KPROBES */ -#endif /* _LINUX_KPROBES_H */ +#endif /* CONFIG_KPROBES */ +#endif /* _LINUX_KPROBES_H */ -- cgit v1.2.3-59-g8ed1b From c132937556f56ee4b831ef4b23f1846e05fde102 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 24 Feb 2009 11:57:20 +0900 Subject: bootmem: clean up arch-specific bootmem wrapping Impact: cleaner and consistent bootmem wrapping By setting CONFIG_HAVE_ARCH_BOOTMEM_NODE, archs can define arch-specific wrappers for bootmem allocation. However, this is done a bit strangely in that only the high level convenience macros can be changed while lower level, but still exported, interface functions can't be wrapped. This not only is messy but also leads to strange situation where alloc_bootmem() does what the arch wants it to do but the equivalent __alloc_bootmem() call doesn't although they should be able to be used interchangeably. This patch updates bootmem such that archs can override / wrap the backend function - alloc_bootmem_core() instead of the highlevel interface functions to allow simpler and consistent wrapping. Also, HAVE_ARCH_BOOTMEM_NODE is renamed to HAVE_ARCH_BOOTMEM. Signed-off-by: Tejun Heo Cc: Johannes Weiner --- arch/avr32/Kconfig | 2 +- arch/x86/Kconfig | 2 +- arch/x86/include/asm/mmzone_32.h | 43 +++++----------------------------------- include/linux/bootmem.h | 10 ++++------ mm/bootmem.c | 14 ++++++++++--- 5 files changed, 22 insertions(+), 49 deletions(-) (limited to 'include/linux') diff --git a/arch/avr32/Kconfig b/arch/avr32/Kconfig index b189680d18b0..05fe3053dcae 100644 --- a/arch/avr32/Kconfig +++ b/arch/avr32/Kconfig @@ -181,7 +181,7 @@ source "kernel/Kconfig.preempt" config QUICKLIST def_bool y -config HAVE_ARCH_BOOTMEM_NODE +config HAVE_ARCH_BOOTMEM def_bool n config ARCH_HAVE_MEMORY_PRESENT diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index d3f6eadfd4ba..6fd3b2302ed9 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -1111,7 +1111,7 @@ config NODES_SHIFT Specify the maximum number of NUMA Nodes available on the target system. Increases memory reserved to accomodate various tables. -config HAVE_ARCH_BOOTMEM_NODE +config HAVE_ARCH_BOOTMEM def_bool y depends on X86_32 && NUMA diff --git a/arch/x86/include/asm/mmzone_32.h b/arch/x86/include/asm/mmzone_32.h index 07f1af494ca5..1e0fa9e63afa 100644 --- a/arch/x86/include/asm/mmzone_32.h +++ b/arch/x86/include/asm/mmzone_32.h @@ -93,45 +93,12 @@ static inline int pfn_valid(int pfn) #endif /* CONFIG_DISCONTIGMEM */ #ifdef CONFIG_NEED_MULTIPLE_NODES - -/* - * Following are macros that are specific to this numa platform. - */ -#define reserve_bootmem(addr, size, flags) \ - reserve_bootmem_node(NODE_DATA(0), (addr), (size), (flags)) -#define alloc_bootmem(x) \ - __alloc_bootmem_node(NODE_DATA(0), (x), SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS)) -#define alloc_bootmem_nopanic(x) \ - __alloc_bootmem_node_nopanic(NODE_DATA(0), (x), SMP_CACHE_BYTES, \ - __pa(MAX_DMA_ADDRESS)) -#define alloc_bootmem_low(x) \ - __alloc_bootmem_node(NODE_DATA(0), (x), SMP_CACHE_BYTES, 0) -#define alloc_bootmem_pages(x) \ - __alloc_bootmem_node(NODE_DATA(0), (x), PAGE_SIZE, __pa(MAX_DMA_ADDRESS)) -#define alloc_bootmem_pages_nopanic(x) \ - __alloc_bootmem_node_nopanic(NODE_DATA(0), (x), PAGE_SIZE, \ - __pa(MAX_DMA_ADDRESS)) -#define alloc_bootmem_low_pages(x) \ - __alloc_bootmem_node(NODE_DATA(0), (x), PAGE_SIZE, 0) -#define alloc_bootmem_node(pgdat, x) \ -({ \ - struct pglist_data __maybe_unused \ - *__alloc_bootmem_node__pgdat = (pgdat); \ - __alloc_bootmem_node(NODE_DATA(0), (x), SMP_CACHE_BYTES, \ - __pa(MAX_DMA_ADDRESS)); \ -}) -#define alloc_bootmem_pages_node(pgdat, x) \ -({ \ - struct pglist_data __maybe_unused \ - *__alloc_bootmem_node__pgdat = (pgdat); \ - __alloc_bootmem_node(NODE_DATA(0), (x), PAGE_SIZE, \ - __pa(MAX_DMA_ADDRESS)); \ -}) -#define alloc_bootmem_low_pages_node(pgdat, x) \ +/* always use node 0 for bootmem on this numa platform */ +#define alloc_bootmem_core(__bdata, size, align, goal, limit) \ ({ \ - struct pglist_data __maybe_unused \ - *__alloc_bootmem_node__pgdat = (pgdat); \ - __alloc_bootmem_node(NODE_DATA(0), (x), PAGE_SIZE, 0); \ + bootmem_data_t __maybe_unused * __abm_bdata_dummy = (__bdata); \ + __alloc_bootmem_core(NODE_DATA(0)->bdata, \ + (size), (align), (goal), (limit)); \ }) #endif /* CONFIG_NEED_MULTIPLE_NODES */ diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h index 95837bfb5256..3a87f93081ed 100644 --- a/include/linux/bootmem.h +++ b/include/linux/bootmem.h @@ -69,10 +69,9 @@ extern int reserve_bootmem_node(pg_data_t *pgdat, unsigned long physaddr, unsigned long size, int flags); -#ifndef CONFIG_HAVE_ARCH_BOOTMEM_NODE -extern int reserve_bootmem(unsigned long addr, unsigned long size, int flags); -#endif - +extern int reserve_bootmem(unsigned long addr, + unsigned long size, + int flags); extern void *__alloc_bootmem_nopanic(unsigned long size, unsigned long align, unsigned long goal); @@ -94,7 +93,7 @@ extern void *__alloc_bootmem_low_node(pg_data_t *pgdat, unsigned long size, unsigned long align, unsigned long goal); -#ifndef CONFIG_HAVE_ARCH_BOOTMEM_NODE + #define alloc_bootmem(x) \ __alloc_bootmem(x, SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS)) #define alloc_bootmem_nopanic(x) \ @@ -113,7 +112,6 @@ extern void *__alloc_bootmem_low_node(pg_data_t *pgdat, __alloc_bootmem_node(pgdat, x, PAGE_SIZE, __pa(MAX_DMA_ADDRESS)) #define alloc_bootmem_low_pages_node(pgdat, x) \ __alloc_bootmem_low_node(pgdat, x, PAGE_SIZE, 0) -#endif /* !CONFIG_HAVE_ARCH_BOOTMEM_NODE */ extern int reserve_bootmem_generic(unsigned long addr, unsigned long size, int flags); diff --git a/mm/bootmem.c b/mm/bootmem.c index 51a0ccf61e0e..d7140c008ba8 100644 --- a/mm/bootmem.c +++ b/mm/bootmem.c @@ -37,6 +37,16 @@ static struct list_head bdata_list __initdata = LIST_HEAD_INIT(bdata_list); static int bootmem_debug; +/* + * If an arch needs to apply workarounds to bootmem allocation, it can + * set CONFIG_HAVE_ARCH_BOOTMEM and define a wrapper around + * __alloc_bootmem_core(). + */ +#ifndef CONFIG_HAVE_ARCH_BOOTMEM +#define alloc_bootmem_core(bdata, size, align, goal, limit) \ + __alloc_bootmem_core((bdata), (size), (align), (goal), (limit)) +#endif + static int __init bootmem_debug_setup(char *buf) { bootmem_debug = 1; @@ -382,7 +392,6 @@ int __init reserve_bootmem_node(pg_data_t *pgdat, unsigned long physaddr, return mark_bootmem_node(pgdat->bdata, start, end, 1, flags); } -#ifndef CONFIG_HAVE_ARCH_BOOTMEM_NODE /** * reserve_bootmem - mark a page range as usable * @addr: starting address of the range @@ -403,7 +412,6 @@ int __init reserve_bootmem(unsigned long addr, unsigned long size, return mark_bootmem(start, end, 1, flags); } -#endif /* !CONFIG_HAVE_ARCH_BOOTMEM_NODE */ static unsigned long align_idx(struct bootmem_data *bdata, unsigned long idx, unsigned long step) @@ -428,7 +436,7 @@ static unsigned long align_off(struct bootmem_data *bdata, unsigned long off, return ALIGN(base + off, align) - base; } -static void * __init alloc_bootmem_core(struct bootmem_data *bdata, +static void * __init __alloc_bootmem_core(struct bootmem_data *bdata, unsigned long size, unsigned long align, unsigned long goal, unsigned long limit) { -- cgit v1.2.3-59-g8ed1b From 2d0aae41695257603fc281b519677131ab5a752b Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 24 Feb 2009 11:57:21 +0900 Subject: bootmem: reorder interface functions and add a missing one Impact: cleanup and addition of missing interface wrapper The interface functions in bootmem.h was ordered in not so orderly manner. Reorder them such that * functions allocating the same area group together - ie. alloc_bootmem group and alloc_bootmem_low group. * functions w/o node parameter come before the ones w/ node parameter. * nopanic variants are immediately below their panicky counterparts. While at it, add alloc_bootmem_pages_node_nopanic() which was missing. Signed-off-by: Tejun Heo Cc: Johannes Weiner --- include/linux/bootmem.h | 30 +++++++++++++++++------------- 1 file changed, 17 insertions(+), 13 deletions(-) (limited to 'include/linux') diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h index 3a87f93081ed..455d83219fae 100644 --- a/include/linux/bootmem.h +++ b/include/linux/bootmem.h @@ -65,22 +65,20 @@ extern void free_bootmem(unsigned long addr, unsigned long size); #define BOOTMEM_DEFAULT 0 #define BOOTMEM_EXCLUSIVE (1<<0) -extern int reserve_bootmem_node(pg_data_t *pgdat, - unsigned long physaddr, - unsigned long size, - int flags); extern int reserve_bootmem(unsigned long addr, unsigned long size, int flags); -extern void *__alloc_bootmem_nopanic(unsigned long size, +extern int reserve_bootmem_node(pg_data_t *pgdat, + unsigned long physaddr, + unsigned long size, + int flags); + +extern void *__alloc_bootmem(unsigned long size, unsigned long align, unsigned long goal); -extern void *__alloc_bootmem(unsigned long size, +extern void *__alloc_bootmem_nopanic(unsigned long size, unsigned long align, unsigned long goal); -extern void *__alloc_bootmem_low(unsigned long size, - unsigned long align, - unsigned long goal); extern void *__alloc_bootmem_node(pg_data_t *pgdat, unsigned long size, unsigned long align, @@ -89,6 +87,9 @@ extern void *__alloc_bootmem_node_nopanic(pg_data_t *pgdat, unsigned long size, unsigned long align, unsigned long goal); +extern void *__alloc_bootmem_low(unsigned long size, + unsigned long align, + unsigned long goal); extern void *__alloc_bootmem_low_node(pg_data_t *pgdat, unsigned long size, unsigned long align, @@ -98,18 +99,21 @@ extern void *__alloc_bootmem_low_node(pg_data_t *pgdat, __alloc_bootmem(x, SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS)) #define alloc_bootmem_nopanic(x) \ __alloc_bootmem_nopanic(x, SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS)) -#define alloc_bootmem_low(x) \ - __alloc_bootmem_low(x, SMP_CACHE_BYTES, 0) #define alloc_bootmem_pages(x) \ __alloc_bootmem(x, PAGE_SIZE, __pa(MAX_DMA_ADDRESS)) #define alloc_bootmem_pages_nopanic(x) \ __alloc_bootmem_nopanic(x, PAGE_SIZE, __pa(MAX_DMA_ADDRESS)) -#define alloc_bootmem_low_pages(x) \ - __alloc_bootmem_low(x, PAGE_SIZE, 0) #define alloc_bootmem_node(pgdat, x) \ __alloc_bootmem_node(pgdat, x, SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS)) #define alloc_bootmem_pages_node(pgdat, x) \ __alloc_bootmem_node(pgdat, x, PAGE_SIZE, __pa(MAX_DMA_ADDRESS)) +#define alloc_bootmem_pages_node_nopanic(pgdat, x) \ + __alloc_bootmem_node_nopanic(pgdat, x, PAGE_SIZE, __pa(MAX_DMA_ADDRESS)) + +#define alloc_bootmem_low(x) \ + __alloc_bootmem_low(x, SMP_CACHE_BYTES, 0) +#define alloc_bootmem_low_pages(x) \ + __alloc_bootmem_low(x, PAGE_SIZE, 0) #define alloc_bootmem_low_pages_node(pgdat, x) \ __alloc_bootmem_low_node(pgdat, x, PAGE_SIZE, 0) -- cgit v1.2.3-59-g8ed1b From c0c0a29379b5848aec2e8f1c58d853d3cb7118b8 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 24 Feb 2009 11:57:21 +0900 Subject: vmalloc: add @align to vm_area_register_early() Impact: allow larger alignment for early vmalloc area allocation Some early vmalloc users might want larger alignment, for example, for custom large page mapping. Add @align to vm_area_register_early(). While at it, drop docbook comment on non-existent @size. Signed-off-by: Tejun Heo Cc: Nick Piggin Cc: Ivan Kokshaysky --- arch/alpha/mm/init.c | 2 +- include/linux/vmalloc.h | 2 +- mm/percpu.c | 2 +- mm/vmalloc.c | 11 +++++++---- 4 files changed, 10 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/arch/alpha/mm/init.c b/arch/alpha/mm/init.c index df6df025ded4..91eddd8505df 100644 --- a/arch/alpha/mm/init.c +++ b/arch/alpha/mm/init.c @@ -200,7 +200,7 @@ callback_init(void * kernel_end) /* register the vm area */ console_remap_vm.flags = VM_ALLOC; console_remap_vm.size = nr_pages << PAGE_SHIFT; - vm_area_register_early(&console_remap_vm); + vm_area_register_early(&console_remap_vm, PAGE_SIZE); vaddr = (unsigned long)consle_remap_vm.addr; diff --git a/include/linux/vmalloc.h b/include/linux/vmalloc.h index 599ba7984310..2f6994fdf0e0 100644 --- a/include/linux/vmalloc.h +++ b/include/linux/vmalloc.h @@ -109,6 +109,6 @@ extern long vwrite(char *buf, char *addr, unsigned long count); */ extern rwlock_t vmlist_lock; extern struct vm_struct *vmlist; -extern __init void vm_area_register_early(struct vm_struct *vm); +extern __init void vm_area_register_early(struct vm_struct *vm, size_t align); #endif /* _LINUX_VMALLOC_H */ diff --git a/mm/percpu.c b/mm/percpu.c index ed92caa2aa3b..41e7a5f5ab1b 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -860,7 +860,7 @@ size_t __init pcpu_setup_static(pcpu_populate_pte_fn_t populate_pte_fn, /* init and register vm area */ static_vm.flags = VM_ALLOC; static_vm.size = pcpu_chunk_size; - vm_area_register_early(&static_vm); + vm_area_register_early(&static_vm, PAGE_SIZE); /* init static_chunk */ static_chunk = alloc_bootmem(pcpu_chunk_struct_size); diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 224eca9650a8..366ae9ea6af2 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -995,7 +995,7 @@ EXPORT_SYMBOL(vm_map_ram); /** * vm_area_register_early - register vmap area early during boot * @vm: vm_struct to register - * @size: size of area to register + * @align: requested alignment * * This function is used to register kernel vm area before * vmalloc_init() is called. @vm->size and @vm->flags should contain @@ -1004,12 +1004,15 @@ EXPORT_SYMBOL(vm_map_ram); * * DO NOT USE THIS FUNCTION UNLESS YOU KNOW WHAT YOU'RE DOING. */ -void __init vm_area_register_early(struct vm_struct *vm) +void __init vm_area_register_early(struct vm_struct *vm, size_t align) { static size_t vm_init_off __initdata; + unsigned long addr; + + addr = ALIGN(VMALLOC_START + vm_init_off, align); + vm_init_off = PFN_ALIGN(addr + vm->size) - VMALLOC_START; - vm->addr = (void *)VMALLOC_START + vm_init_off; - vm_init_off = PFN_ALIGN(vm_init_off + vm->size); + vm->addr = (void *)addr; vm->next = vmlist; vmlist = vm; -- cgit v1.2.3-59-g8ed1b From 8d408b4be37bc49c9086531f2ebe411cf5731746 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 24 Feb 2009 11:57:21 +0900 Subject: percpu: give more latitude to arch specific first chunk initialization Impact: more latitude for first percpu chunk allocation The first percpu chunk serves the kernel static percpu area and may or may not contain extra room for further dynamic allocation. Initialization of the first chunk needs to be done before normal memory allocation service is up, so it has its own init path - pcpu_setup_static(). It seems archs need more latitude while initializing the first chunk for example to take advantage of large page mapping. This patch makes the following changes to allow this. * Define PERCPU_DYNAMIC_RESERVE to give arch hint about how much space to reserve in the first chunk for further dynamic allocation. * Rename pcpu_setup_static() to pcpu_setup_first_chunk(). * Make pcpu_setup_first_chunk() much more flexible by fetching page pointer by callback and adding optional @unit_size, @free_size and @base_addr arguments which allow archs to selectively part of chunk initialization to their likings. Signed-off-by: Tejun Heo --- arch/x86/kernel/setup_percpu.c | 15 ++++- include/linux/percpu.h | 39 ++++++++++- mm/percpu.c | 149 ++++++++++++++++++++++++++++++++--------- 3 files changed, 167 insertions(+), 36 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 671e6528a82d..d928e8887201 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -41,6 +41,16 @@ unsigned long __per_cpu_offset[NR_CPUS] __read_mostly = { }; EXPORT_SYMBOL(__per_cpu_offset); +static struct page **pcpu4k_pages __initdata; +static int pcpu4k_nr_static_pages __initdata; + +static struct page * __init pcpu4k_get_page(unsigned int cpu, int pageno) +{ + if (pageno < pcpu4k_nr_static_pages) + return pcpu4k_pages[cpu * pcpu4k_nr_static_pages + pageno]; + return NULL; +} + static void __init pcpu4k_populate_pte(unsigned long addr) { populate_extra_pte(addr); @@ -109,7 +119,10 @@ void __init setup_per_cpu_areas(void) } } - pcpu_unit_size = pcpu_setup_static(pcpu4k_populate_pte, pages, size); + pcpu4k_pages = pages; + pcpu4k_nr_static_pages = nr_cpu_pages; + pcpu_unit_size = pcpu_setup_first_chunk(pcpu4k_get_page, size, 0, 0, + NULL, pcpu4k_populate_pte); free_bootmem(__pa(pages), pages_size); diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 18080995ff3e..910beb0abea2 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -78,12 +78,47 @@ #ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA +/* minimum unit size, also is the maximum supported allocation size */ +#define PCPU_MIN_UNIT_SIZE (16UL << PAGE_SHIFT) + +/* + * PERCPU_DYNAMIC_RESERVE indicates the amount of free area to piggy + * back on the first chunk if arch is manually allocating and mapping + * it for faster access (as a part of large page mapping for example). + * Note that dynamic percpu allocator covers both static and dynamic + * areas, so these values are bigger than PERCPU_MODULE_RESERVE. + * + * On typical configuration with modules, the following values leave + * about 8k of free space on the first chunk after boot on both x86_32 + * and 64 when module support is enabled. When module support is + * disabled, it's much tighter. + */ +#ifndef PERCPU_DYNAMIC_RESERVE +# if BITS_PER_LONG > 32 +# ifdef CONFIG_MODULES +# define PERCPU_DYNAMIC_RESERVE (6 << PAGE_SHIFT) +# else +# define PERCPU_DYNAMIC_RESERVE (4 << PAGE_SHIFT) +# endif +# else +# ifdef CONFIG_MODULES +# define PERCPU_DYNAMIC_RESERVE (4 << PAGE_SHIFT) +# else +# define PERCPU_DYNAMIC_RESERVE (2 << PAGE_SHIFT) +# endif +# endif +#endif /* PERCPU_DYNAMIC_RESERVE */ + extern void *pcpu_base_addr; +typedef struct page * (*pcpu_get_page_fn_t)(unsigned int cpu, int pageno); typedef void (*pcpu_populate_pte_fn_t)(unsigned long addr); -extern size_t __init pcpu_setup_static(pcpu_populate_pte_fn_t populate_pte_fn, - struct page **pages, size_t cpu_size); +extern size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, + size_t static_size, size_t unit_size, + size_t free_size, void *base_addr, + pcpu_populate_pte_fn_t populate_pte_fn); + /* * Use this to get to a cpu's version of the per-cpu object * dynamically allocated. Non-atomic access to the current CPU's diff --git a/mm/percpu.c b/mm/percpu.c index d9e6e5d1dbd4..9ac01980cce0 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -48,8 +48,8 @@ * - define __addr_to_pcpu_ptr() and __pcpu_ptr_to_addr() to translate * regular address to percpu pointer and back * - * - use pcpu_setup_static() during percpu area initialization to - * setup kernel static percpu area + * - use pcpu_setup_first_chunk() during percpu area initialization to + * setup the first chunk containing the kernel static percpu area */ #include @@ -67,7 +67,6 @@ #include #include -#define PCPU_MIN_UNIT_PAGES 16 /* max alloc size in pages */ #define PCPU_SLOT_BASE_SHIFT 5 /* 1-31 shares the same slot */ #define PCPU_DFL_MAP_ALLOC 16 /* start a map with 16 ents */ @@ -80,6 +79,7 @@ struct pcpu_chunk { int map_used; /* # of map entries used */ int map_alloc; /* # of map entries allocated */ int *map; /* allocation map */ + bool immutable; /* no [de]population allowed */ struct page *page[]; /* #cpus * UNIT_PAGES */ }; @@ -521,6 +521,9 @@ static void pcpu_unmap(struct pcpu_chunk *chunk, int page_start, int page_end, unsigned int last = num_possible_cpus() - 1; unsigned int cpu; + /* unmap must not be done on immutable chunk */ + WARN_ON(chunk->immutable); + /* * Each flushing trial can be very expensive, issue flush on * the whole region at once rather than doing it for each cpu. @@ -602,6 +605,9 @@ static int pcpu_map(struct pcpu_chunk *chunk, int page_start, int page_end) unsigned int cpu; int err; + /* map must not be done on immutable chunk */ + WARN_ON(chunk->immutable); + for_each_possible_cpu(cpu) { err = map_kernel_range_noflush( pcpu_chunk_addr(chunk, cpu, page_start), @@ -727,8 +733,7 @@ void *__alloc_percpu(size_t size, size_t align) struct pcpu_chunk *chunk; int slot, off; - if (unlikely(!size || size > PCPU_MIN_UNIT_PAGES * PAGE_SIZE || - align > PAGE_SIZE)) { + if (unlikely(!size || size > PCPU_MIN_UNIT_SIZE || align > PAGE_SIZE)) { WARN(true, "illegal size (%zu) or align (%zu) for " "percpu allocation\n", size, align); return NULL; @@ -776,6 +781,7 @@ EXPORT_SYMBOL_GPL(__alloc_percpu); static void pcpu_kill_chunk(struct pcpu_chunk *chunk) { + WARN_ON(chunk->immutable); pcpu_depopulate_chunk(chunk, 0, pcpu_unit_size, false); list_del(&chunk->list); rb_erase(&chunk->rb_node, &pcpu_addr_root); @@ -821,33 +827,73 @@ void free_percpu(void *ptr) EXPORT_SYMBOL_GPL(free_percpu); /** - * pcpu_setup_static - initialize kernel static percpu area - * @populate_pte_fn: callback to allocate pagetable - * @pages: num_possible_cpus() * PFN_UP(cpu_size) pages - * @cpu_size: the size of static percpu area in bytes - * - * Initialize kernel static percpu area. The caller should allocate - * all the necessary pages and pass them in @pages. - * @populate_pte_fn() is called on each page to be used for percpu - * mapping and is responsible for making sure all the necessary page - * tables for the page is allocated. + * pcpu_setup_first_chunk - initialize the first percpu chunk + * @get_page_fn: callback to fetch page pointer + * @static_size: the size of static percpu area in bytes + * @unit_size: unit size in bytes, must be multiple of PAGE_SIZE, 0 for auto + * @free_size: free size in bytes, 0 for auto + * @base_addr: mapped address, NULL for auto + * @populate_pte_fn: callback to allocate pagetable, NULL if unnecessary + * + * Initialize the first percpu chunk which contains the kernel static + * perpcu area. This function is to be called from arch percpu area + * setup path. The first two parameters are mandatory. The rest are + * optional. + * + * @get_page_fn() should return pointer to percpu page given cpu + * number and page number. It should at least return enough pages to + * cover the static area. The returned pages for static area should + * have been initialized with valid data. If @unit_size is specified, + * it can also return pages after the static area. NULL return + * indicates end of pages for the cpu. Note that @get_page_fn() must + * return the same number of pages for all cpus. + * + * @unit_size, if non-zero, determines unit size and must be aligned + * to PAGE_SIZE and equal to or larger than @static_size + @free_size. + * + * @free_size determines the number of free bytes after the static + * area in the first chunk. If zero, whatever left is available. + * Specifying non-zero value make percpu leave the area after + * @static_size + @free_size alone. + * + * Non-null @base_addr means that the caller already allocated virtual + * region for the first chunk and mapped it. percpu must not mess + * with the chunk. Note that @base_addr with 0 @unit_size or non-NULL + * @populate_pte_fn doesn't make any sense. + * + * @populate_pte_fn is used to populate the pagetable. NULL means the + * caller already populated the pagetable. * * RETURNS: * The determined pcpu_unit_size which can be used to initialize * percpu access. */ -size_t __init pcpu_setup_static(pcpu_populate_pte_fn_t populate_pte_fn, - struct page **pages, size_t cpu_size) +size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, + size_t static_size, size_t unit_size, + size_t free_size, void *base_addr, + pcpu_populate_pte_fn_t populate_pte_fn) { static struct vm_struct static_vm; struct pcpu_chunk *static_chunk; - int nr_cpu_pages = DIV_ROUND_UP(cpu_size, PAGE_SIZE); unsigned int cpu; + int nr_pages; int err, i; - pcpu_unit_pages = max_t(int, PCPU_MIN_UNIT_PAGES, PFN_UP(cpu_size)); + /* santiy checks */ + BUG_ON(!static_size); + BUG_ON(!unit_size && free_size); + BUG_ON(unit_size && unit_size < static_size + free_size); + BUG_ON(unit_size & ~PAGE_MASK); + BUG_ON(base_addr && !unit_size); + BUG_ON(base_addr && populate_pte_fn); - pcpu_static_size = cpu_size; + if (unit_size) + pcpu_unit_pages = unit_size >> PAGE_SHIFT; + else + pcpu_unit_pages = max_t(int, PCPU_MIN_UNIT_SIZE >> PAGE_SHIFT, + PFN_UP(static_size)); + + pcpu_static_size = static_size; pcpu_unit_size = pcpu_unit_pages << PAGE_SHIFT; pcpu_chunk_size = num_possible_cpus() * pcpu_unit_size; pcpu_chunk_struct_size = sizeof(struct pcpu_chunk) @@ -862,29 +908,66 @@ size_t __init pcpu_setup_static(pcpu_populate_pte_fn_t populate_pte_fn, for (i = 0; i < pcpu_nr_slots; i++) INIT_LIST_HEAD(&pcpu_slot[i]); - /* init and register vm area */ - static_vm.flags = VM_ALLOC; - static_vm.size = pcpu_chunk_size; - vm_area_register_early(&static_vm, PAGE_SIZE); - /* init static_chunk */ static_chunk = alloc_bootmem(pcpu_chunk_struct_size); INIT_LIST_HEAD(&static_chunk->list); static_chunk->vm = &static_vm; - static_chunk->free_size = pcpu_unit_size - pcpu_static_size; + + if (free_size) + static_chunk->free_size = free_size; + else + static_chunk->free_size = pcpu_unit_size - pcpu_static_size; + static_chunk->contig_hint = static_chunk->free_size; - /* assign pages and map them */ + /* allocate vm address */ + static_vm.flags = VM_ALLOC; + static_vm.size = pcpu_chunk_size; + + if (!base_addr) + vm_area_register_early(&static_vm, PAGE_SIZE); + else { + /* + * Pages already mapped. No need to remap into + * vmalloc area. In this case the static chunk can't + * be mapped or unmapped by percpu and is marked + * immutable. + */ + static_vm.addr = base_addr; + static_chunk->immutable = true; + } + + /* assign pages */ + nr_pages = -1; for_each_possible_cpu(cpu) { - for (i = 0; i < nr_cpu_pages; i++) { - *pcpu_chunk_pagep(static_chunk, cpu, i) = *pages++; - populate_pte_fn(pcpu_chunk_addr(static_chunk, cpu, i)); + for (i = 0; i < pcpu_unit_pages; i++) { + struct page *page = get_page_fn(cpu, i); + + if (!page) + break; + *pcpu_chunk_pagep(static_chunk, cpu, i) = page; } + + BUG_ON(i < PFN_UP(pcpu_static_size)); + + if (nr_pages < 0) + nr_pages = i; + else + BUG_ON(nr_pages != i); } - err = pcpu_map(static_chunk, 0, nr_cpu_pages); - if (err) - panic("failed to setup static percpu area, err=%d\n", err); + /* map them */ + if (populate_pte_fn) { + for_each_possible_cpu(cpu) + for (i = 0; i < nr_pages; i++) + populate_pte_fn(pcpu_chunk_addr(static_chunk, + cpu, i)); + + err = pcpu_map(static_chunk, 0, nr_pages); + if (err) + panic("failed to setup static percpu area, err=%d\n", + err); + } /* link static_chunk in */ pcpu_chunk_relocate(static_chunk, -1); -- cgit v1.2.3-59-g8ed1b From 3255aa2eb636a508fc82a73fabbb8aaf2ff23c0f Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 25 Feb 2009 08:21:52 +0100 Subject: x86, mm: pass in 'total' to __copy_from_user_*nocache() Impact: cleanup, enable future change Add a 'total bytes copied' parameter to __copy_from_user_*nocache(), and update all the callsites. The parameter is not used yet - architecture code can use it to more intelligently decide whether the copy should be cached or non-temporal. Cc: Salman Qazi Cc: Nick Piggin Cc: Linus Torvalds Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uaccess_32.h | 4 ++-- arch/x86/include/asm/uaccess_64.h | 5 ++--- drivers/gpu/drm/i915/i915_gem.c | 2 +- include/linux/uaccess.h | 4 ++-- mm/filemap.c | 10 ++++++---- mm/filemap_xip.c | 2 +- 6 files changed, 14 insertions(+), 13 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/include/asm/uaccess_32.h b/arch/x86/include/asm/uaccess_32.h index 5e06259e90e5..a0ba61386972 100644 --- a/arch/x86/include/asm/uaccess_32.h +++ b/arch/x86/include/asm/uaccess_32.h @@ -157,7 +157,7 @@ __copy_from_user(void *to, const void __user *from, unsigned long n) } static __always_inline unsigned long __copy_from_user_nocache(void *to, - const void __user *from, unsigned long n) + const void __user *from, unsigned long n, unsigned long total) { might_fault(); if (__builtin_constant_p(n)) { @@ -180,7 +180,7 @@ static __always_inline unsigned long __copy_from_user_nocache(void *to, static __always_inline unsigned long __copy_from_user_inatomic_nocache(void *to, const void __user *from, - unsigned long n) + unsigned long n, unsigned long total) { return __copy_from_user_ll_nocache_nozero(to, from, n); } diff --git a/arch/x86/include/asm/uaccess_64.h b/arch/x86/include/asm/uaccess_64.h index 987a2c10fe20..a748253db0c9 100644 --- a/arch/x86/include/asm/uaccess_64.h +++ b/arch/x86/include/asm/uaccess_64.h @@ -189,7 +189,7 @@ extern long __copy_user_nocache(void *dst, const void __user *src, unsigned size, int zerorest); static inline int __copy_from_user_nocache(void *dst, const void __user *src, - unsigned size) + unsigned size, unsigned long total) { might_sleep(); /* @@ -205,8 +205,7 @@ static inline int __copy_from_user_nocache(void *dst, const void __user *src, } static inline int __copy_from_user_inatomic_nocache(void *dst, - const void __user *src, - unsigned size) + const void __user *src, unsigned size, unsigned total) { if (likely(size >= PAGE_SIZE)) return __copy_user_nocache(dst, src, size, 0); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 818576654092..6b209db8370d 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -215,7 +215,7 @@ fast_user_write(struct io_mapping *mapping, vaddr_atomic = io_mapping_map_atomic_wc(mapping, page_base); unwritten = __copy_from_user_inatomic_nocache(vaddr_atomic + page_offset, - user_data, length); + user_data, length, length); io_mapping_unmap_atomic(vaddr_atomic); if (unwritten) return -EFAULT; diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h index 6b58367d145e..6f3c603b0d67 100644 --- a/include/linux/uaccess.h +++ b/include/linux/uaccess.h @@ -41,13 +41,13 @@ static inline void pagefault_enable(void) #ifndef ARCH_HAS_NOCACHE_UACCESS static inline unsigned long __copy_from_user_inatomic_nocache(void *to, - const void __user *from, unsigned long n) + const void __user *from, unsigned long n, unsigned long total) { return __copy_from_user_inatomic(to, from, n); } static inline unsigned long __copy_from_user_nocache(void *to, - const void __user *from, unsigned long n) + const void __user *from, unsigned long n, unsigned long total) { return __copy_from_user(to, from, n); } diff --git a/mm/filemap.c b/mm/filemap.c index 23acefe51808..60fd56772cc6 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -1816,14 +1816,14 @@ EXPORT_SYMBOL(file_remove_suid); static size_t __iovec_copy_from_user_inatomic(char *vaddr, const struct iovec *iov, size_t base, size_t bytes) { - size_t copied = 0, left = 0; + size_t copied = 0, left = 0, total = bytes; while (bytes) { char __user *buf = iov->iov_base + base; int copy = min(bytes, iov->iov_len - base); base = 0; - left = __copy_from_user_inatomic_nocache(vaddr, buf, copy); + left = __copy_from_user_inatomic_nocache(vaddr, buf, copy, total); copied += copy; bytes -= copy; vaddr += copy; @@ -1851,8 +1851,9 @@ size_t iov_iter_copy_from_user_atomic(struct page *page, if (likely(i->nr_segs == 1)) { int left; char __user *buf = i->iov->iov_base + i->iov_offset; + left = __copy_from_user_inatomic_nocache(kaddr + offset, - buf, bytes); + buf, bytes, bytes); copied = bytes - left; } else { copied = __iovec_copy_from_user_inatomic(kaddr + offset, @@ -1880,7 +1881,8 @@ size_t iov_iter_copy_from_user(struct page *page, if (likely(i->nr_segs == 1)) { int left; char __user *buf = i->iov->iov_base + i->iov_offset; - left = __copy_from_user_nocache(kaddr + offset, buf, bytes); + + left = __copy_from_user_nocache(kaddr + offset, buf, bytes, bytes); copied = bytes - left; } else { copied = __iovec_copy_from_user_inatomic(kaddr + offset, diff --git a/mm/filemap_xip.c b/mm/filemap_xip.c index 0c04615651b7..bf54f8a2cf1d 100644 --- a/mm/filemap_xip.c +++ b/mm/filemap_xip.c @@ -354,7 +354,7 @@ __xip_file_write(struct file *filp, const char __user *buf, break; copied = bytes - - __copy_from_user_nocache(xip_mem + offset, buf, bytes); + __copy_from_user_nocache(xip_mem + offset, buf, bytes, bytes); if (likely(copied > 0)) { status = copied; -- cgit v1.2.3-59-g8ed1b From 17581ad812a9abb0182260374ef2e52d4a808a64 Mon Sep 17 00:00:00 2001 From: Venkatesh Pallipadi Date: Tue, 24 Feb 2009 17:35:14 -0800 Subject: gpu/drm, x86, PAT: PAT support for io_mapping_* Make io_mapping_create_wc and io_mapping_free go through PAT to make sure that there are no memory type aliases. Signed-off-by: Venkatesh Pallipadi Signed-off-by: Suresh Siddha Cc: Dave Airlie Cc: Jesse Barnes Cc: Eric Anholt Cc: Keith Packard Signed-off-by: Ingo Molnar --- arch/x86/include/asm/iomap.h | 5 ++++- arch/x86/mm/iomap_32.c | 42 ++++++++++++++++++++++++++++++++++++++++-- include/linux/io-mapping.h | 6 ++++-- 3 files changed, 48 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/include/asm/iomap.h b/arch/x86/include/asm/iomap.h index 86af26091d6c..bd46495ff7de 100644 --- a/arch/x86/include/asm/iomap.h +++ b/arch/x86/include/asm/iomap.h @@ -24,7 +24,10 @@ #include int -is_io_mapping_possible(resource_size_t base, unsigned long size); +reserve_io_memtype_wc(u64 base, unsigned long size, pgprot_t *prot); + +void +free_io_memtype(u64 base, unsigned long size); void * iomap_atomic_prot_pfn(unsigned long pfn, enum km_type type, pgprot_t prot); diff --git a/arch/x86/mm/iomap_32.c b/arch/x86/mm/iomap_32.c index 6c2b1af16926..94596f794b1b 100644 --- a/arch/x86/mm/iomap_32.c +++ b/arch/x86/mm/iomap_32.c @@ -21,13 +21,13 @@ #include #ifdef CONFIG_X86_PAE -int +static int is_io_mapping_possible(resource_size_t base, unsigned long size) { return 1; } #else -int +static int is_io_mapping_possible(resource_size_t base, unsigned long size) { /* There is no way to map greater than 1 << 32 address without PAE */ @@ -38,6 +38,44 @@ is_io_mapping_possible(resource_size_t base, unsigned long size) } #endif +int +reserve_io_memtype_wc(u64 base, unsigned long size, pgprot_t *prot) +{ + unsigned long ret_flag; + + if (!is_io_mapping_possible(base, size)) + goto out_err; + + if (!pat_enabled) { + *prot = pgprot_noncached(PAGE_KERNEL); + return 0; + } + + if (reserve_memtype(base, base + size, _PAGE_CACHE_WC, &ret_flag)) + goto out_err; + + if (ret_flag == _PAGE_CACHE_WB) + goto out_free; + + if (kernel_map_sync_memtype(base, size, ret_flag)) + goto out_free; + + *prot = __pgprot(__PAGE_KERNEL | ret_flag); + return 0; + +out_free: + free_memtype(base, base + size); +out_err: + return -EINVAL; +} + +void +free_io_memtype(u64 base, unsigned long size) +{ + if (pat_enabled) + free_memtype(base, base + size); +} + /* Map 'pfn' using fixed map 'type' and protections 'prot' */ void * diff --git a/include/linux/io-mapping.h b/include/linux/io-mapping.h index cbc2f0cd631b..f1ed66c43787 100644 --- a/include/linux/io-mapping.h +++ b/include/linux/io-mapping.h @@ -49,8 +49,9 @@ static inline struct io_mapping * io_mapping_create_wc(resource_size_t base, unsigned long size) { struct io_mapping *iomap; + pgprot_t prot; - if (!is_io_mapping_possible(base, size)) + if (!reserve_io_memtype_wc(base, size, &prot)) return NULL; iomap = kmalloc(sizeof(*iomap), GFP_KERNEL); @@ -59,13 +60,14 @@ io_mapping_create_wc(resource_size_t base, unsigned long size) iomap->base = base; iomap->size = size; - iomap->prot = pgprot_writecombine(__pgprot(__PAGE_KERNEL)); + iomap->prot = prot; return iomap; } static inline void io_mapping_free(struct io_mapping *mapping) { + free_io_memtype(mapping->base, mapping->size); kfree(mapping); } -- cgit v1.2.3-59-g8ed1b From d2b0261506602bd969164879206027b30358ffdf Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 25 Feb 2009 14:36:45 +0100 Subject: alloc_percpu: fix UP build Impact: build fix the !SMP branch had a 'gfp' leftover: include/linux/percpu.h: In function '__alloc_percpu': include/linux/percpu.h:160: error: 'gfp' undeclared (first use in this function) include/linux/percpu.h:160: error: (Each undeclared identifier is reported only once include/linux/percpu.h:160: error: for each function it appears in.) Use GFP_KERNEL like the SMP version does. Cc: Rusty Russell Cc: Tejun Heo Signed-off-by: Ingo Molnar --- include/linux/percpu.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 910beb0abea2..d8e5a9abbce0 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -157,7 +157,7 @@ static inline void *__alloc_percpu(size_t size, size_t align) * percpu sections on SMP for which this path isn't used. */ WARN_ON_ONCE(align > __alignof__(unsigned long long)); - return kzalloc(size, gfp); + return kzalloc(size, GFP_KERNEL); } static inline void free_percpu(void *p) -- cgit v1.2.3-59-g8ed1b From e317603694bfd17b28a40de9d65e1a4ec12f816e Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 26 Feb 2009 10:54:17 +0900 Subject: percpu: fix too low alignment restriction on UP UP __alloc_percpu() triggered WARN_ON_ONCE() if the requested alignment is larger than that of unsigned long long, which is too small for all the cacheline aligned allocations. Bump it up to SMP_CACHE_BYTES which kmalloc allocations generally guarantee. Signed-off-by: Tejun Heo Reported-by: Ingo Molnar --- include/linux/percpu.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/percpu.h b/include/linux/percpu.h index d8e5a9abbce0..545b068bcb70 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -156,7 +156,7 @@ static inline void *__alloc_percpu(size_t size, size_t align) * on it. Larger alignment should only be used for module * percpu sections on SMP for which this path isn't used. */ - WARN_ON_ONCE(align > __alignof__(unsigned long long)); + WARN_ON_ONCE(align > SMP_CACHE_BYTES); return kzalloc(size, GFP_KERNEL); } -- cgit v1.2.3-59-g8ed1b From f5c1aa1537be39d8b9bb5279b5881d81898fd3cd Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sun, 1 Mar 2009 12:32:08 +0100 Subject: Revert "gpu/drm, x86, PAT: PAT support for io_mapping_*" This reverts commit 17581ad812a9abb0182260374ef2e52d4a808a64. Sitsofe Wheeler reported that /dev/dri/card0 is MIA on his EeePC 900 and bisected it to this commit. Graphics card is an i915 in an EeePC 900: 00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller [8086:2592] (rev 04) ( Most likely the ioremap() of the driver failed and hence the card did not initialize. ) Reported-by: Sitsofe Wheeler Bisected-by: Sitsofe Wheeler Cc: Venkatesh Pallipadi Cc: Suresh Siddha Signed-off-by: Ingo Molnar --- arch/x86/include/asm/iomap.h | 5 +---- arch/x86/mm/iomap_32.c | 44 ++------------------------------------------ include/linux/io-mapping.h | 6 ++---- 3 files changed, 5 insertions(+), 50 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/include/asm/iomap.h b/arch/x86/include/asm/iomap.h index bd46495ff7de..86af26091d6c 100644 --- a/arch/x86/include/asm/iomap.h +++ b/arch/x86/include/asm/iomap.h @@ -24,10 +24,7 @@ #include int -reserve_io_memtype_wc(u64 base, unsigned long size, pgprot_t *prot); - -void -free_io_memtype(u64 base, unsigned long size); +is_io_mapping_possible(resource_size_t base, unsigned long size); void * iomap_atomic_prot_pfn(unsigned long pfn, enum km_type type, pgprot_t prot); diff --git a/arch/x86/mm/iomap_32.c b/arch/x86/mm/iomap_32.c index d5e28424622c..6c2b1af16926 100644 --- a/arch/x86/mm/iomap_32.c +++ b/arch/x86/mm/iomap_32.c @@ -21,13 +21,13 @@ #include #ifdef CONFIG_X86_PAE -static int +int is_io_mapping_possible(resource_size_t base, unsigned long size) { return 1; } #else -static int +int is_io_mapping_possible(resource_size_t base, unsigned long size) { /* There is no way to map greater than 1 << 32 address without PAE */ @@ -38,46 +38,6 @@ is_io_mapping_possible(resource_size_t base, unsigned long size) } #endif -int -reserve_io_memtype_wc(u64 base, unsigned long size, pgprot_t *prot) -{ - unsigned long ret_flag; - - if (!is_io_mapping_possible(base, size)) - goto out_err; - - if (!pat_enabled) { - *prot = pgprot_noncached(PAGE_KERNEL); - return 0; - } - - if (reserve_memtype(base, base + size, _PAGE_CACHE_WC, &ret_flag)) - goto out_err; - - if (ret_flag == _PAGE_CACHE_WB) - goto out_free; - - if (kernel_map_sync_memtype(base, size, ret_flag)) - goto out_free; - - *prot = __pgprot(__PAGE_KERNEL | ret_flag); - return 0; - -out_free: - free_memtype(base, base + size); -out_err: - return -EINVAL; -} -EXPORT_SYMBOL_GPL(reserve_io_memtype_wc); - -void -free_io_memtype(u64 base, unsigned long size) -{ - if (pat_enabled) - free_memtype(base, base + size); -} -EXPORT_SYMBOL_GPL(free_io_memtype); - /* Map 'pfn' using fixed map 'type' and protections 'prot' */ void * diff --git a/include/linux/io-mapping.h b/include/linux/io-mapping.h index f1ed66c43787..cbc2f0cd631b 100644 --- a/include/linux/io-mapping.h +++ b/include/linux/io-mapping.h @@ -49,9 +49,8 @@ static inline struct io_mapping * io_mapping_create_wc(resource_size_t base, unsigned long size) { struct io_mapping *iomap; - pgprot_t prot; - if (!reserve_io_memtype_wc(base, size, &prot)) + if (!is_io_mapping_possible(base, size)) return NULL; iomap = kmalloc(sizeof(*iomap), GFP_KERNEL); @@ -60,14 +59,13 @@ io_mapping_create_wc(resource_size_t base, unsigned long size) iomap->base = base; iomap->size = size; - iomap->prot = prot; + iomap->prot = pgprot_writecombine(__pgprot(__PAGE_KERNEL)); return iomap; } static inline void io_mapping_free(struct io_mapping *mapping) { - free_io_memtype(mapping->base, mapping->size); kfree(mapping); } -- cgit v1.2.3-59-g8ed1b From f180053694b43d5714bf56cb95499a3c32ff155c Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 2 Mar 2009 11:00:57 +0100 Subject: x86, mm: dont use non-temporal stores in pagecache accesses Impact: standardize IO on cached ops On modern CPUs it is almost always a bad idea to use non-temporal stores, as the regression in this commit has shown it: 30d697f: x86: fix performance regression in write() syscall The kernel simply has no good information about whether using non-temporal stores is a good idea or not - and trying to add heuristics only increases complexity and inserts fragility. The regression on cached write()s took very long to be found - over two years. So dont take any chances and let the hardware decide how it makes use of its caches. The only exception is drivers/gpu/drm/i915/i915_gem.c: there were we are absolutely sure that another entity (the GPU) will pick up the dirty data immediately and that the CPU will not touch that data before the GPU will. Also, keep the _nocache() primitives to make it easier for people to experiment with these details. There may be more clear-cut cases where non-cached copies can be used, outside of filemap.c. Cc: Salman Qazi Cc: Nick Piggin Cc: Linus Torvalds Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uaccess_32.h | 4 ++-- arch/x86/include/asm/uaccess_64.h | 25 +++++++------------------ drivers/gpu/drm/i915/i915_gem.c | 2 +- include/linux/uaccess.h | 4 ++-- mm/filemap.c | 11 ++++------- mm/filemap_xip.c | 2 +- 6 files changed, 17 insertions(+), 31 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/include/asm/uaccess_32.h b/arch/x86/include/asm/uaccess_32.h index a0ba61386972..5e06259e90e5 100644 --- a/arch/x86/include/asm/uaccess_32.h +++ b/arch/x86/include/asm/uaccess_32.h @@ -157,7 +157,7 @@ __copy_from_user(void *to, const void __user *from, unsigned long n) } static __always_inline unsigned long __copy_from_user_nocache(void *to, - const void __user *from, unsigned long n, unsigned long total) + const void __user *from, unsigned long n) { might_fault(); if (__builtin_constant_p(n)) { @@ -180,7 +180,7 @@ static __always_inline unsigned long __copy_from_user_nocache(void *to, static __always_inline unsigned long __copy_from_user_inatomic_nocache(void *to, const void __user *from, - unsigned long n, unsigned long total) + unsigned long n) { return __copy_from_user_ll_nocache_nozero(to, from, n); } diff --git a/arch/x86/include/asm/uaccess_64.h b/arch/x86/include/asm/uaccess_64.h index dcaa0404cf7b..8cc687326eb8 100644 --- a/arch/x86/include/asm/uaccess_64.h +++ b/arch/x86/include/asm/uaccess_64.h @@ -188,29 +188,18 @@ __copy_to_user_inatomic(void __user *dst, const void *src, unsigned size) extern long __copy_user_nocache(void *dst, const void __user *src, unsigned size, int zerorest); -static inline int __copy_from_user_nocache(void *dst, const void __user *src, - unsigned size, unsigned long total) +static inline int +__copy_from_user_nocache(void *dst, const void __user *src, unsigned size) { might_sleep(); - /* - * In practice this limit means that large file write()s - * which get chunked to 4K copies get handled via - * non-temporal stores here. Smaller writes get handled - * via regular __copy_from_user(): - */ - if (likely(total >= PAGE_SIZE)) - return __copy_user_nocache(dst, src, size, 1); - else - return __copy_from_user(dst, src, size); + return __copy_user_nocache(dst, src, size, 1); } -static inline int __copy_from_user_inatomic_nocache(void *dst, - const void __user *src, unsigned size, unsigned total) +static inline int +__copy_from_user_inatomic_nocache(void *dst, const void __user *src, + unsigned size) { - if (likely(total >= PAGE_SIZE)) - return __copy_user_nocache(dst, src, size, 0); - else - return __copy_from_user_inatomic(dst, src, size); + return __copy_user_nocache(dst, src, size, 0); } unsigned long diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 6b209db8370d..818576654092 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -215,7 +215,7 @@ fast_user_write(struct io_mapping *mapping, vaddr_atomic = io_mapping_map_atomic_wc(mapping, page_base); unwritten = __copy_from_user_inatomic_nocache(vaddr_atomic + page_offset, - user_data, length, length); + user_data, length); io_mapping_unmap_atomic(vaddr_atomic); if (unwritten) return -EFAULT; diff --git a/include/linux/uaccess.h b/include/linux/uaccess.h index 6f3c603b0d67..6b58367d145e 100644 --- a/include/linux/uaccess.h +++ b/include/linux/uaccess.h @@ -41,13 +41,13 @@ static inline void pagefault_enable(void) #ifndef ARCH_HAS_NOCACHE_UACCESS static inline unsigned long __copy_from_user_inatomic_nocache(void *to, - const void __user *from, unsigned long n, unsigned long total) + const void __user *from, unsigned long n) { return __copy_from_user_inatomic(to, from, n); } static inline unsigned long __copy_from_user_nocache(void *to, - const void __user *from, unsigned long n, unsigned long total) + const void __user *from, unsigned long n) { return __copy_from_user(to, from, n); } diff --git a/mm/filemap.c b/mm/filemap.c index 60fd56772cc6..126d3973b3d1 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -1816,14 +1816,14 @@ EXPORT_SYMBOL(file_remove_suid); static size_t __iovec_copy_from_user_inatomic(char *vaddr, const struct iovec *iov, size_t base, size_t bytes) { - size_t copied = 0, left = 0, total = bytes; + size_t copied = 0, left = 0; while (bytes) { char __user *buf = iov->iov_base + base; int copy = min(bytes, iov->iov_len - base); base = 0; - left = __copy_from_user_inatomic_nocache(vaddr, buf, copy, total); + left = __copy_from_user_inatomic(vaddr, buf, copy); copied += copy; bytes -= copy; vaddr += copy; @@ -1851,9 +1851,7 @@ size_t iov_iter_copy_from_user_atomic(struct page *page, if (likely(i->nr_segs == 1)) { int left; char __user *buf = i->iov->iov_base + i->iov_offset; - - left = __copy_from_user_inatomic_nocache(kaddr + offset, - buf, bytes, bytes); + left = __copy_from_user_inatomic(kaddr + offset, buf, bytes); copied = bytes - left; } else { copied = __iovec_copy_from_user_inatomic(kaddr + offset, @@ -1881,8 +1879,7 @@ size_t iov_iter_copy_from_user(struct page *page, if (likely(i->nr_segs == 1)) { int left; char __user *buf = i->iov->iov_base + i->iov_offset; - - left = __copy_from_user_nocache(kaddr + offset, buf, bytes, bytes); + left = __copy_from_user(kaddr + offset, buf, bytes); copied = bytes - left; } else { copied = __iovec_copy_from_user_inatomic(kaddr + offset, diff --git a/mm/filemap_xip.c b/mm/filemap_xip.c index bf54f8a2cf1d..0c04615651b7 100644 --- a/mm/filemap_xip.c +++ b/mm/filemap_xip.c @@ -354,7 +354,7 @@ __xip_file_write(struct file *filp, const char __user *buf, break; copied = bytes - - __copy_from_user_nocache(xip_mem + offset, buf, bytes, bytes); + __copy_from_user_nocache(xip_mem + offset, buf, bytes); if (likely(copied > 0)) { status = copied; -- cgit v1.2.3-59-g8ed1b From 6a242909b01120f6f3d571c0b75e20ec61f0d8d3 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 6 Mar 2009 14:33:58 +0900 Subject: percpu: clean up percpu constants Impact: cleaup Make the following cleanups. * There isn't much arch-specific about PERCPU_MODULE_RESERVE. Always define it whether arch overrides PERCPU_ENOUGH_ROOM or not. * blackfin overrides PERCPU_ENOUGH_ROOM to align static area size. Do it by default. * percpu allocation sizes doesn't have much to do with the page size. Don't use PAGE_SHIFT in their definition. Signed-off-by: Tejun Heo Cc: Bryan Wu --- arch/blackfin/include/asm/percpu.h | 10 ---------- include/linux/percpu.h | 24 +++++++++++++----------- 2 files changed, 13 insertions(+), 21 deletions(-) (limited to 'include/linux') diff --git a/arch/blackfin/include/asm/percpu.h b/arch/blackfin/include/asm/percpu.h index 797c0c165069..c94c7bc88c71 100644 --- a/arch/blackfin/include/asm/percpu.h +++ b/arch/blackfin/include/asm/percpu.h @@ -3,14 +3,4 @@ #include -#ifdef CONFIG_MODULES -#define PERCPU_MODULE_RESERVE 8192 -#else -#define PERCPU_MODULE_RESERVE 0 -#endif - -#define PERCPU_ENOUGH_ROOM \ - (ALIGN(__per_cpu_end - __per_cpu_start, SMP_CACHE_BYTES) + \ - PERCPU_MODULE_RESERVE) - #endif /* __ARCH_BLACKFIN_PERCPU__ */ diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 545b068bcb70..2d34b038fe70 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -5,6 +5,7 @@ #include /* For kmalloc() */ #include #include +#include #include @@ -52,17 +53,18 @@ #define EXPORT_PER_CPU_SYMBOL(var) EXPORT_SYMBOL(per_cpu__##var) #define EXPORT_PER_CPU_SYMBOL_GPL(var) EXPORT_SYMBOL_GPL(per_cpu__##var) -/* Enough to cover all DEFINE_PER_CPUs in kernel, including modules. */ -#ifndef PERCPU_ENOUGH_ROOM +/* enough to cover all DEFINE_PER_CPUs in modules */ #ifdef CONFIG_MODULES -#define PERCPU_MODULE_RESERVE 8192 +#define PERCPU_MODULE_RESERVE (8 << 10) #else -#define PERCPU_MODULE_RESERVE 0 +#define PERCPU_MODULE_RESERVE 0 #endif +#ifndef PERCPU_ENOUGH_ROOM #define PERCPU_ENOUGH_ROOM \ - (__per_cpu_end - __per_cpu_start + PERCPU_MODULE_RESERVE) -#endif /* PERCPU_ENOUGH_ROOM */ + (ALIGN(__per_cpu_end - __per_cpu_start, SMP_CACHE_BYTES) + \ + PERCPU_MODULE_RESERVE) +#endif /* * Must be an lvalue. Since @var must be a simple identifier, @@ -79,7 +81,7 @@ #ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA /* minimum unit size, also is the maximum supported allocation size */ -#define PCPU_MIN_UNIT_SIZE (16UL << PAGE_SHIFT) +#define PCPU_MIN_UNIT_SIZE PFN_ALIGN(64 << 10) /* * PERCPU_DYNAMIC_RESERVE indicates the amount of free area to piggy @@ -96,15 +98,15 @@ #ifndef PERCPU_DYNAMIC_RESERVE # if BITS_PER_LONG > 32 # ifdef CONFIG_MODULES -# define PERCPU_DYNAMIC_RESERVE (6 << PAGE_SHIFT) +# define PERCPU_DYNAMIC_RESERVE (24 << 10) # else -# define PERCPU_DYNAMIC_RESERVE (4 << PAGE_SHIFT) +# define PERCPU_DYNAMIC_RESERVE (16 << 10) # endif # else # ifdef CONFIG_MODULES -# define PERCPU_DYNAMIC_RESERVE (4 << PAGE_SHIFT) +# define PERCPU_DYNAMIC_RESERVE (16 << 10) # else -# define PERCPU_DYNAMIC_RESERVE (2 << PAGE_SHIFT) +# define PERCPU_DYNAMIC_RESERVE (8 << 10) # endif # endif #endif /* PERCPU_DYNAMIC_RESERVE */ -- cgit v1.2.3-59-g8ed1b From 2441d15c97d498b18f03ae9fba262ffeae42a08b Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 6 Mar 2009 14:33:59 +0900 Subject: percpu: cosmetic renames in pcpu_setup_first_chunk() Impact: cosmetic, preparation for future changes Make the following renames in pcpur_setup_first_chunk() in preparation for future changes. * s/free_size/dyn_size/ * s/static_vm/first_vm/ * s/static_chunk/schunk/ Signed-off-by: Tejun Heo --- include/linux/percpu.h | 2 +- mm/percpu.c | 58 +++++++++++++++++++++++++------------------------- 2 files changed, 30 insertions(+), 30 deletions(-) (limited to 'include/linux') diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 2d34b038fe70..a0b4ea2a3354 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -118,7 +118,7 @@ typedef void (*pcpu_populate_pte_fn_t)(unsigned long addr); extern size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, size_t static_size, size_t unit_size, - size_t free_size, void *base_addr, + size_t dyn_size, void *base_addr, pcpu_populate_pte_fn_t populate_pte_fn); /* diff --git a/mm/percpu.c b/mm/percpu.c index 3d0f5456827c..9531590e6b69 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -831,7 +831,7 @@ EXPORT_SYMBOL_GPL(free_percpu); * @get_page_fn: callback to fetch page pointer * @static_size: the size of static percpu area in bytes * @unit_size: unit size in bytes, must be multiple of PAGE_SIZE, 0 for auto - * @free_size: free size in bytes, 0 for auto + * @dyn_size: free size for dynamic allocation in bytes, 0 for auto * @base_addr: mapped address, NULL for auto * @populate_pte_fn: callback to allocate pagetable, NULL if unnecessary * @@ -849,12 +849,12 @@ EXPORT_SYMBOL_GPL(free_percpu); * return the same number of pages for all cpus. * * @unit_size, if non-zero, determines unit size and must be aligned - * to PAGE_SIZE and equal to or larger than @static_size + @free_size. + * to PAGE_SIZE and equal to or larger than @static_size + @dyn_size. * - * @free_size determines the number of free bytes after the static + * @dyn_size determines the number of free bytes after the static * area in the first chunk. If zero, whatever left is available. * Specifying non-zero value make percpu leave the area after - * @static_size + @free_size alone. + * @static_size + @dyn_size alone. * * Non-null @base_addr means that the caller already allocated virtual * region for the first chunk and mapped it. percpu must not mess @@ -870,19 +870,19 @@ EXPORT_SYMBOL_GPL(free_percpu); */ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, size_t static_size, size_t unit_size, - size_t free_size, void *base_addr, + size_t dyn_size, void *base_addr, pcpu_populate_pte_fn_t populate_pte_fn) { - static struct vm_struct static_vm; - struct pcpu_chunk *static_chunk; + static struct vm_struct first_vm; + struct pcpu_chunk *schunk; unsigned int cpu; int nr_pages; int err, i; /* santiy checks */ BUG_ON(!static_size); - BUG_ON(!unit_size && free_size); - BUG_ON(unit_size && unit_size < static_size + free_size); + BUG_ON(!unit_size && dyn_size); + BUG_ON(unit_size && unit_size < static_size + dyn_size); BUG_ON(unit_size & ~PAGE_MASK); BUG_ON(base_addr && !unit_size); BUG_ON(base_addr && populate_pte_fn); @@ -908,24 +908,24 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, for (i = 0; i < pcpu_nr_slots; i++) INIT_LIST_HEAD(&pcpu_slot[i]); - /* init static_chunk */ - static_chunk = alloc_bootmem(pcpu_chunk_struct_size); - INIT_LIST_HEAD(&static_chunk->list); - static_chunk->vm = &static_vm; + /* init static chunk */ + schunk = alloc_bootmem(pcpu_chunk_struct_size); + INIT_LIST_HEAD(&schunk->list); + schunk->vm = &first_vm; - if (free_size) - static_chunk->free_size = free_size; + if (dyn_size) + schunk->free_size = dyn_size; else - static_chunk->free_size = pcpu_unit_size - pcpu_static_size; + schunk->free_size = pcpu_unit_size - pcpu_static_size; - static_chunk->contig_hint = static_chunk->free_size; + schunk->contig_hint = schunk->free_size; /* allocate vm address */ - static_vm.flags = VM_ALLOC; - static_vm.size = pcpu_chunk_size; + first_vm.flags = VM_ALLOC; + first_vm.size = pcpu_chunk_size; if (!base_addr) - vm_area_register_early(&static_vm, PAGE_SIZE); + vm_area_register_early(&first_vm, PAGE_SIZE); else { /* * Pages already mapped. No need to remap into @@ -933,8 +933,8 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, * be mapped or unmapped by percpu and is marked * immutable. */ - static_vm.addr = base_addr; - static_chunk->immutable = true; + first_vm.addr = base_addr; + schunk->immutable = true; } /* assign pages */ @@ -945,7 +945,7 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, if (!page) break; - *pcpu_chunk_pagep(static_chunk, cpu, i) = page; + *pcpu_chunk_pagep(schunk, cpu, i) = page; } BUG_ON(i < PFN_UP(pcpu_static_size)); @@ -960,20 +960,20 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, if (populate_pte_fn) { for_each_possible_cpu(cpu) for (i = 0; i < nr_pages; i++) - populate_pte_fn(pcpu_chunk_addr(static_chunk, + populate_pte_fn(pcpu_chunk_addr(schunk, cpu, i)); - err = pcpu_map(static_chunk, 0, nr_pages); + err = pcpu_map(schunk, 0, nr_pages); if (err) panic("failed to setup static percpu area, err=%d\n", err); } - /* link static_chunk in */ - pcpu_chunk_relocate(static_chunk, -1); - pcpu_chunk_addr_insert(static_chunk); + /* link the first chunk in */ + pcpu_chunk_relocate(schunk, -1); + pcpu_chunk_addr_insert(schunk); /* we're done */ - pcpu_base_addr = (void *)pcpu_chunk_addr(static_chunk, 0, 0); + pcpu_base_addr = (void *)pcpu_chunk_addr(schunk, 0, 0); return pcpu_unit_size; } -- cgit v1.2.3-59-g8ed1b From cafe8816b217b98dc3f268d3b77445da498beb4f Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 6 Mar 2009 14:33:59 +0900 Subject: percpu: use negative for auto for pcpu_setup_first_chunk() arguments Impact: argument semantic cleanup In pcpu_setup_first_chunk(), zero @unit_size and @dyn_size meant auto-sizing. It's okay for @unit_size as 0 doesn't make sense but 0 dynamic reserve size is valid. Alos, if arch @dyn_size is calculated from other parameters, it might end up passing in 0 @dyn_size and malfunction when the size is automatically adjusted. This patch makes both @unit_size and @dyn_size ssize_t and use -1 for auto sizing. Signed-off-by: Tejun Heo --- arch/x86/kernel/setup_percpu.c | 2 +- include/linux/percpu.h | 5 +++-- mm/percpu.c | 46 +++++++++++++++++++++++------------------- 3 files changed, 29 insertions(+), 24 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index c29f301d3885..ef3a2cd3fe64 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -344,7 +344,7 @@ static ssize_t __init setup_pcpu_4k(size_t static_size) pr_info("PERCPU: Allocated %d 4k pages, static data %zu bytes\n", pcpu4k_nr_static_pages, static_size); - ret = pcpu_setup_first_chunk(pcpu4k_get_page, static_size, 0, 0, NULL, + ret = pcpu_setup_first_chunk(pcpu4k_get_page, static_size, -1, -1, NULL, pcpu4k_populate_pte); goto out_free_ar; diff --git a/include/linux/percpu.h b/include/linux/percpu.h index a0b4ea2a3354..a96fc53bbd62 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -117,8 +117,9 @@ typedef struct page * (*pcpu_get_page_fn_t)(unsigned int cpu, int pageno); typedef void (*pcpu_populate_pte_fn_t)(unsigned long addr); extern size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, - size_t static_size, size_t unit_size, - size_t dyn_size, void *base_addr, + size_t static_size, + ssize_t unit_size, ssize_t dyn_size, + void *base_addr, pcpu_populate_pte_fn_t populate_pte_fn); /* diff --git a/mm/percpu.c b/mm/percpu.c index 503ccad091af..a84cf9977faf 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -824,8 +824,8 @@ EXPORT_SYMBOL_GPL(free_percpu); * pcpu_setup_first_chunk - initialize the first percpu chunk * @get_page_fn: callback to fetch page pointer * @static_size: the size of static percpu area in bytes - * @unit_size: unit size in bytes, must be multiple of PAGE_SIZE, 0 for auto - * @dyn_size: free size for dynamic allocation in bytes, 0 for auto + * @unit_size: unit size in bytes, must be multiple of PAGE_SIZE, -1 for auto + * @dyn_size: free size for dynamic allocation in bytes, -1 for auto * @base_addr: mapped address, NULL for auto * @populate_pte_fn: callback to allocate pagetable, NULL if unnecessary * @@ -842,13 +842,14 @@ EXPORT_SYMBOL_GPL(free_percpu); * indicates end of pages for the cpu. Note that @get_page_fn() must * return the same number of pages for all cpus. * - * @unit_size, if non-zero, determines unit size and must be aligned - * to PAGE_SIZE and equal to or larger than @static_size + @dyn_size. + * @unit_size, if non-negative, specifies unit size and must be + * aligned to PAGE_SIZE and equal to or larger than @static_size + + * @dyn_size. * - * @dyn_size determines the number of free bytes after the static - * area in the first chunk. If zero, whatever left is available. - * Specifying non-zero value make percpu leave the area after - * @static_size + @dyn_size alone. + * @dyn_size, if non-negative, limits the number of bytes available + * for dynamic allocation in the first chunk. Specifying non-negative + * value make percpu leave alone the area beyond @static_size + + * @dyn_size. * * Non-null @base_addr means that the caller already allocated virtual * region for the first chunk and mapped it. percpu must not mess @@ -863,8 +864,9 @@ EXPORT_SYMBOL_GPL(free_percpu); * percpu access. */ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, - size_t static_size, size_t unit_size, - size_t dyn_size, void *base_addr, + size_t static_size, + ssize_t unit_size, ssize_t dyn_size, + void *base_addr, pcpu_populate_pte_fn_t populate_pte_fn) { static struct vm_struct first_vm; @@ -877,13 +879,17 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, /* santiy checks */ BUILD_BUG_ON(ARRAY_SIZE(smap) >= PCPU_DFL_MAP_ALLOC); BUG_ON(!static_size); - BUG_ON(!unit_size && dyn_size); - BUG_ON(unit_size && unit_size < static_size + dyn_size); - BUG_ON(unit_size & ~PAGE_MASK); - BUG_ON(base_addr && !unit_size); + if (unit_size >= 0) { + BUG_ON(unit_size < static_size + + (dyn_size >= 0 ? dyn_size : 0)); + BUG_ON(unit_size & ~PAGE_MASK); + } else { + BUG_ON(dyn_size >= 0); + BUG_ON(base_addr); + } BUG_ON(base_addr && populate_pte_fn); - if (unit_size) + if (unit_size >= 0) pcpu_unit_pages = unit_size >> PAGE_SHIFT; else pcpu_unit_pages = max_t(int, PCPU_MIN_UNIT_SIZE >> PAGE_SHIFT, @@ -894,6 +900,9 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, pcpu_chunk_struct_size = sizeof(struct pcpu_chunk) + num_possible_cpus() * pcpu_unit_pages * sizeof(struct page *); + if (dyn_size < 0) + dyn_size = pcpu_unit_size - static_size; + /* * Allocate chunk slots. The additional last slot is for * empty chunks. @@ -909,12 +918,7 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, schunk->vm = &first_vm; schunk->map = smap; schunk->map_alloc = ARRAY_SIZE(smap); - - if (dyn_size) - schunk->free_size = dyn_size; - else - schunk->free_size = pcpu_unit_size - static_size; - + schunk->free_size = dyn_size; schunk->contig_hint = schunk->free_size; schunk->map[schunk->map_used++] = -static_size; -- cgit v1.2.3-59-g8ed1b From edcb463997ed7b2ffa3bac76e3e75957318f2e01 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 6 Mar 2009 14:33:59 +0900 Subject: percpu, module: implement reserved allocation and use it for module percpu variables Impact: add reserved allocation functionality and use it for module percpu variables This patch implements reserved allocation from the first chunk. When setting up the first chunk, arch can ask to set aside certain number of bytes right after the core static area which is available only through a separate reserved allocator. This will be used primarily for module static percpu variables on architectures with limited relocation range to ensure that the module perpcu symbols are inside the relocatable range. If reserved area is requested, the first chunk becomes reserved and isn't available for regular allocation. If the first chunk also includes piggy-back dynamic allocation area, a separate chunk mapping the same region is created to serve dynamic allocation. The first one is called static first chunk and the second dynamic first chunk. Although they share the page map, their different area map initializations guarantee they serve disjoint areas according to their purposes. If arch doesn't setup reserved area, reserved allocation is handled like any other allocation. Signed-off-by: Tejun Heo --- arch/x86/kernel/setup_percpu.c | 8 +-- include/linux/percpu.h | 10 +-- kernel/module.c | 2 +- mm/percpu.c | 153 +++++++++++++++++++++++++++++++++++------ 4 files changed, 144 insertions(+), 29 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 38e2b2a470a5..dd4eabc747c8 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -217,7 +217,7 @@ proceed: pr_info("PERCPU: Remapped at %p with large pages, static data " "%zu bytes\n", vm.addr, static_size); - ret = pcpu_setup_first_chunk(pcpur_get_page, static_size, PMD_SIZE, + ret = pcpu_setup_first_chunk(pcpur_get_page, static_size, 0, PMD_SIZE, pcpur_size - static_size, vm.addr, NULL); goto out_free_ar; @@ -297,7 +297,7 @@ static ssize_t __init setup_pcpu_embed(size_t static_size) pr_info("PERCPU: Embedded %zu pages at %p, static data %zu bytes\n", pcpue_size >> PAGE_SHIFT, pcpue_ptr, static_size); - return pcpu_setup_first_chunk(pcpue_get_page, static_size, + return pcpu_setup_first_chunk(pcpue_get_page, static_size, 0, pcpue_unit_size, dyn_size, pcpue_ptr, NULL); } @@ -356,8 +356,8 @@ static ssize_t __init setup_pcpu_4k(size_t static_size) pr_info("PERCPU: Allocated %d 4k pages, static data %zu bytes\n", pcpu4k_nr_static_pages, static_size); - ret = pcpu_setup_first_chunk(pcpu4k_get_page, static_size, -1, -1, NULL, - pcpu4k_populate_pte); + ret = pcpu_setup_first_chunk(pcpu4k_get_page, static_size, 0, -1, -1, + NULL, pcpu4k_populate_pte); goto out_free_ar; enomem: diff --git a/include/linux/percpu.h b/include/linux/percpu.h index a96fc53bbd62..8ff15153ae20 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -117,10 +117,10 @@ typedef struct page * (*pcpu_get_page_fn_t)(unsigned int cpu, int pageno); typedef void (*pcpu_populate_pte_fn_t)(unsigned long addr); extern size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, - size_t static_size, - ssize_t unit_size, ssize_t dyn_size, - void *base_addr, - pcpu_populate_pte_fn_t populate_pte_fn); + size_t static_size, size_t reserved_size, + ssize_t unit_size, ssize_t dyn_size, + void *base_addr, + pcpu_populate_pte_fn_t populate_pte_fn); /* * Use this to get to a cpu's version of the per-cpu object @@ -129,6 +129,8 @@ extern size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, */ #define per_cpu_ptr(ptr, cpu) SHIFT_PERCPU_PTR((ptr), per_cpu_offset((cpu))) +extern void *__alloc_reserved_percpu(size_t size, size_t align); + #else /* CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */ struct percpu_data { diff --git a/kernel/module.c b/kernel/module.c index 1f0657ae555b..f0e04d6b67d8 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -381,7 +381,7 @@ static void *percpu_modalloc(unsigned long size, unsigned long align, align = PAGE_SIZE; } - ptr = __alloc_percpu(size, align); + ptr = __alloc_reserved_percpu(size, align); if (!ptr) printk(KERN_WARNING "Could not allocate %lu bytes percpu data\n", size); diff --git a/mm/percpu.c b/mm/percpu.c index 5b47d9fe65f5..ef8e169b7731 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -94,6 +94,11 @@ static size_t pcpu_chunk_struct_size __read_mostly; void *pcpu_base_addr __read_mostly; EXPORT_SYMBOL_GPL(pcpu_base_addr); +/* optional reserved chunk, only accessible for reserved allocations */ +static struct pcpu_chunk *pcpu_reserved_chunk; +/* offset limit of the reserved chunk */ +static int pcpu_reserved_chunk_limit; + /* * One mutex to rule them all. * @@ -201,13 +206,14 @@ static void *pcpu_realloc(void *p, size_t size, size_t new_size) * * This function is called after an allocation or free changed @chunk. * New slot according to the changed state is determined and @chunk is - * moved to the slot. + * moved to the slot. Note that the reserved chunk is never put on + * chunk slots. */ static void pcpu_chunk_relocate(struct pcpu_chunk *chunk, int oslot) { int nslot = pcpu_chunk_slot(chunk); - if (oslot != nslot) { + if (chunk != pcpu_reserved_chunk && oslot != nslot) { if (oslot < nslot) list_move(&chunk->list, &pcpu_slot[nslot]); else @@ -255,6 +261,15 @@ static struct pcpu_chunk *pcpu_chunk_addr_search(void *addr) struct rb_node *n, *parent; struct pcpu_chunk *chunk; + /* is it in the reserved chunk? */ + if (pcpu_reserved_chunk) { + void *start = pcpu_reserved_chunk->vm->addr; + + if (addr >= start && addr < start + pcpu_reserved_chunk_limit) + return pcpu_reserved_chunk; + } + + /* nah... search the regular ones */ n = *pcpu_chunk_rb_search(addr, &parent); if (!n) { /* no exactly matching chunk, the parent is the closest */ @@ -713,9 +728,10 @@ static struct pcpu_chunk *alloc_pcpu_chunk(void) } /** - * __alloc_percpu - allocate percpu area + * pcpu_alloc - the percpu allocator * @size: size of area to allocate in bytes * @align: alignment of area (max PAGE_SIZE) + * @reserved: allocate from the reserved chunk if available * * Allocate percpu area of @size bytes aligned at @align. Might * sleep. Might trigger writeouts. @@ -723,7 +739,7 @@ static struct pcpu_chunk *alloc_pcpu_chunk(void) * RETURNS: * Percpu pointer to the allocated area on success, NULL on failure. */ -void *__alloc_percpu(size_t size, size_t align) +static void *pcpu_alloc(size_t size, size_t align, bool reserved) { void *ptr = NULL; struct pcpu_chunk *chunk; @@ -737,7 +753,18 @@ void *__alloc_percpu(size_t size, size_t align) mutex_lock(&pcpu_mutex); - /* allocate area */ + /* serve reserved allocations from the reserved chunk if available */ + if (reserved && pcpu_reserved_chunk) { + chunk = pcpu_reserved_chunk; + if (size > chunk->contig_hint) + goto out_unlock; + off = pcpu_alloc_area(chunk, size, align); + if (off >= 0) + goto area_found; + goto out_unlock; + } + + /* search through normal chunks */ for (slot = pcpu_size_to_slot(size); slot < pcpu_nr_slots; slot++) { list_for_each_entry(chunk, &pcpu_slot[slot], list) { if (size > chunk->contig_hint) @@ -773,8 +800,41 @@ out_unlock: mutex_unlock(&pcpu_mutex); return ptr; } + +/** + * __alloc_percpu - allocate dynamic percpu area + * @size: size of area to allocate in bytes + * @align: alignment of area (max PAGE_SIZE) + * + * Allocate percpu area of @size bytes aligned at @align. Might + * sleep. Might trigger writeouts. + * + * RETURNS: + * Percpu pointer to the allocated area on success, NULL on failure. + */ +void *__alloc_percpu(size_t size, size_t align) +{ + return pcpu_alloc(size, align, false); +} EXPORT_SYMBOL_GPL(__alloc_percpu); +/** + * __alloc_reserved_percpu - allocate reserved percpu area + * @size: size of area to allocate in bytes + * @align: alignment of area (max PAGE_SIZE) + * + * Allocate percpu area of @size bytes aligned at @align from reserved + * percpu area if arch has set it up; otherwise, allocation is served + * from the same dynamic area. Might sleep. Might trigger writeouts. + * + * RETURNS: + * Percpu pointer to the allocated area on success, NULL on failure. + */ +void *__alloc_reserved_percpu(size_t size, size_t align) +{ + return pcpu_alloc(size, align, true); +} + static void pcpu_kill_chunk(struct pcpu_chunk *chunk) { WARN_ON(chunk->immutable); @@ -826,6 +886,7 @@ EXPORT_SYMBOL_GPL(free_percpu); * pcpu_setup_first_chunk - initialize the first percpu chunk * @get_page_fn: callback to fetch page pointer * @static_size: the size of static percpu area in bytes + * @reserved_size: the size of reserved percpu area in bytes * @unit_size: unit size in bytes, must be multiple of PAGE_SIZE, -1 for auto * @dyn_size: free size for dynamic allocation in bytes, -1 for auto * @base_addr: mapped address, NULL for auto @@ -844,14 +905,22 @@ EXPORT_SYMBOL_GPL(free_percpu); * indicates end of pages for the cpu. Note that @get_page_fn() must * return the same number of pages for all cpus. * + * @reserved_size, if non-zero, specifies the amount of bytes to + * reserve after the static area in the first chunk. This reserves + * the first chunk such that it's available only through reserved + * percpu allocation. This is primarily used to serve module percpu + * static areas on architectures where the addressing model has + * limited offset range for symbol relocations to guarantee module + * percpu symbols fall inside the relocatable range. + * * @unit_size, if non-negative, specifies unit size and must be * aligned to PAGE_SIZE and equal to or larger than @static_size + - * @dyn_size. + * @reserved_size + @dyn_size. * * @dyn_size, if non-negative, limits the number of bytes available * for dynamic allocation in the first chunk. Specifying non-negative * value make percpu leave alone the area beyond @static_size + - * @dyn_size. + * @reserved_size + @dyn_size. * * Non-null @base_addr means that the caller already allocated virtual * region for the first chunk and mapped it. percpu must not mess @@ -861,28 +930,36 @@ EXPORT_SYMBOL_GPL(free_percpu); * @populate_pte_fn is used to populate the pagetable. NULL means the * caller already populated the pagetable. * + * If the first chunk ends up with both reserved and dynamic areas, it + * is served by two chunks - one to serve the core static and reserved + * areas and the other for the dynamic area. They share the same vm + * and page map but uses different area allocation map to stay away + * from each other. The latter chunk is circulated in the chunk slots + * and available for dynamic allocation like any other chunks. + * * RETURNS: * The determined pcpu_unit_size which can be used to initialize * percpu access. */ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, - size_t static_size, + size_t static_size, size_t reserved_size, ssize_t unit_size, ssize_t dyn_size, void *base_addr, pcpu_populate_pte_fn_t populate_pte_fn) { static struct vm_struct first_vm; - static int smap[2]; - struct pcpu_chunk *schunk; + static int smap[2], dmap[2]; + struct pcpu_chunk *schunk, *dchunk = NULL; unsigned int cpu; int nr_pages; int err, i; /* santiy checks */ - BUILD_BUG_ON(ARRAY_SIZE(smap) >= PCPU_DFL_MAP_ALLOC); + BUILD_BUG_ON(ARRAY_SIZE(smap) >= PCPU_DFL_MAP_ALLOC || + ARRAY_SIZE(dmap) >= PCPU_DFL_MAP_ALLOC); BUG_ON(!static_size); if (unit_size >= 0) { - BUG_ON(unit_size < static_size + + BUG_ON(unit_size < static_size + reserved_size + (dyn_size >= 0 ? dyn_size : 0)); BUG_ON(unit_size & ~PAGE_MASK); } else { @@ -895,7 +972,7 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, pcpu_unit_pages = unit_size >> PAGE_SHIFT; else pcpu_unit_pages = max_t(int, PCPU_MIN_UNIT_SIZE >> PAGE_SHIFT, - PFN_UP(static_size)); + PFN_UP(static_size + reserved_size)); pcpu_unit_size = pcpu_unit_pages << PAGE_SHIFT; pcpu_chunk_size = num_possible_cpus() * pcpu_unit_size; @@ -903,7 +980,7 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, + num_possible_cpus() * pcpu_unit_pages * sizeof(struct page *); if (dyn_size < 0) - dyn_size = pcpu_unit_size - static_size; + dyn_size = pcpu_unit_size - static_size - reserved_size; /* * Allocate chunk slots. The additional last slot is for @@ -914,20 +991,49 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, for (i = 0; i < pcpu_nr_slots; i++) INIT_LIST_HEAD(&pcpu_slot[i]); - /* init static chunk */ + /* + * Initialize static chunk. If reserved_size is zero, the + * static chunk covers static area + dynamic allocation area + * in the first chunk. If reserved_size is not zero, it + * covers static area + reserved area (mostly used for module + * static percpu allocation). + */ schunk = alloc_bootmem(pcpu_chunk_struct_size); INIT_LIST_HEAD(&schunk->list); schunk->vm = &first_vm; schunk->map = smap; schunk->map_alloc = ARRAY_SIZE(smap); schunk->page = schunk->page_ar; - schunk->free_size = dyn_size; + + if (reserved_size) { + schunk->free_size = reserved_size; + pcpu_reserved_chunk = schunk; /* not for dynamic alloc */ + } else { + schunk->free_size = dyn_size; + dyn_size = 0; /* dynamic area covered */ + } schunk->contig_hint = schunk->free_size; schunk->map[schunk->map_used++] = -static_size; if (schunk->free_size) schunk->map[schunk->map_used++] = schunk->free_size; + pcpu_reserved_chunk_limit = static_size + schunk->free_size; + + /* init dynamic chunk if necessary */ + if (dyn_size) { + dchunk = alloc_bootmem(sizeof(struct pcpu_chunk)); + INIT_LIST_HEAD(&dchunk->list); + dchunk->vm = &first_vm; + dchunk->map = dmap; + dchunk->map_alloc = ARRAY_SIZE(dmap); + dchunk->page = schunk->page_ar; /* share page map with schunk */ + + dchunk->contig_hint = dchunk->free_size = dyn_size; + dchunk->map[dchunk->map_used++] = -pcpu_reserved_chunk_limit; + dchunk->map[dchunk->map_used++] = dchunk->free_size; + } + /* allocate vm address */ first_vm.flags = VM_ALLOC; first_vm.size = pcpu_chunk_size; @@ -937,12 +1043,14 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, else { /* * Pages already mapped. No need to remap into - * vmalloc area. In this case the static chunk can't - * be mapped or unmapped by percpu and is marked + * vmalloc area. In this case the first chunks can't + * be mapped or unmapped by percpu and are marked * immutable. */ first_vm.addr = base_addr; schunk->immutable = true; + if (dchunk) + dchunk->immutable = true; } /* assign pages */ @@ -978,8 +1086,13 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, } /* link the first chunk in */ - pcpu_chunk_relocate(schunk, -1); - pcpu_chunk_addr_insert(schunk); + if (!dchunk) { + pcpu_chunk_relocate(schunk, -1); + pcpu_chunk_addr_insert(schunk); + } else { + pcpu_chunk_relocate(dchunk, -1); + pcpu_chunk_addr_insert(dchunk); + } /* we're done */ pcpu_base_addr = (void *)pcpu_chunk_addr(schunk, 0, 0); -- cgit v1.2.3-59-g8ed1b From 6b19b0c2400437a3c10059ede0e59b517092e1bd Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 6 Mar 2009 14:33:59 +0900 Subject: x86, percpu: setup reserved percpu area for x86_64 Impact: fix relocation overflow during module load x86_64 uses 32bit relocations for symbol access and static percpu symbols whether in core or modules must be inside 2GB of the percpu segement base which the dynamic percpu allocator doesn't guarantee. This patch makes x86_64 reserve PERCPU_MODULE_RESERVE bytes in the first chunk so that module percpu areas are always allocated from the first chunk which is always inside the relocatable range. This problem exists for any percpu allocator but is easily triggered when using the embedding allocator because the second chunk is located beyond 2GB on it. This patch also changes the meaning of PERCPU_DYNAMIC_RESERVE such that it only indicates the size of the area to reserve for dynamic allocation as static and dynamic areas can be separate. New PERCPU_DYNAMIC_RESERVED is increased by 4k for both 32 and 64bits as the reserved area separation eats away some allocatable space and having slightly more headroom (currently between 4 and 8k after minimal boot sans module area) makes sense for common case performance. x86_32 can address anywhere from anywhere and doesn't need reserving. Mike Galbraith first reported the problem first and bisected it to the embedding percpu allocator commit. Signed-off-by: Tejun Heo Reported-by: Mike Galbraith Reported-by: Jaswinder Singh Rajput --- arch/x86/kernel/setup_percpu.c | 37 ++++++++++++++++++++++++++++--------- include/linux/percpu.h | 35 ++++++++++++----------------------- 2 files changed, 40 insertions(+), 32 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index dd4eabc747c8..efa615f2bf43 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -42,6 +42,19 @@ unsigned long __per_cpu_offset[NR_CPUS] __read_mostly = { }; EXPORT_SYMBOL(__per_cpu_offset); +/* + * On x86_64 symbols referenced from code should be reachable using + * 32bit relocations. Reserve space for static percpu variables in + * modules so that they are always served from the first chunk which + * is located at the percpu segment base. On x86_32, anything can + * address anywhere. No need to reserve space in the first chunk. + */ +#ifdef CONFIG_X86_64 +#define PERCPU_FIRST_CHUNK_RESERVE PERCPU_MODULE_RESERVE +#else +#define PERCPU_FIRST_CHUNK_RESERVE 0 +#endif + /** * pcpu_need_numa - determine percpu allocation needs to consider NUMA * @@ -141,7 +154,7 @@ static ssize_t __init setup_pcpu_remap(size_t static_size) { static struct vm_struct vm; pg_data_t *last; - size_t ptrs_size; + size_t ptrs_size, dyn_size; unsigned int cpu; ssize_t ret; @@ -169,12 +182,14 @@ proceed: * Currently supports only single page. Supporting multiple * pages won't be too difficult if it ever becomes necessary. */ - pcpur_size = PFN_ALIGN(static_size + PERCPU_DYNAMIC_RESERVE); + pcpur_size = PFN_ALIGN(static_size + PERCPU_MODULE_RESERVE + + PERCPU_DYNAMIC_RESERVE); if (pcpur_size > PMD_SIZE) { pr_warning("PERCPU: static data is larger than large page, " "can't use large page\n"); return -EINVAL; } + dyn_size = pcpur_size - static_size - PERCPU_FIRST_CHUNK_RESERVE; /* allocate pointer array and alloc large pages */ ptrs_size = PFN_ALIGN(num_possible_cpus() * sizeof(pcpur_ptrs[0])); @@ -217,8 +232,9 @@ proceed: pr_info("PERCPU: Remapped at %p with large pages, static data " "%zu bytes\n", vm.addr, static_size); - ret = pcpu_setup_first_chunk(pcpur_get_page, static_size, 0, PMD_SIZE, - pcpur_size - static_size, vm.addr, NULL); + ret = pcpu_setup_first_chunk(pcpur_get_page, static_size, + PERCPU_FIRST_CHUNK_RESERVE, + PMD_SIZE, dyn_size, vm.addr, NULL); goto out_free_ar; enomem: @@ -276,9 +292,10 @@ static ssize_t __init setup_pcpu_embed(size_t static_size) return -EINVAL; /* allocate and copy */ - pcpue_size = PFN_ALIGN(static_size + PERCPU_DYNAMIC_RESERVE); + pcpue_size = PFN_ALIGN(static_size + PERCPU_MODULE_RESERVE + + PERCPU_DYNAMIC_RESERVE); pcpue_unit_size = max_t(size_t, pcpue_size, PCPU_MIN_UNIT_SIZE); - dyn_size = pcpue_size - static_size; + dyn_size = pcpue_size - static_size - PERCPU_FIRST_CHUNK_RESERVE; pcpue_ptr = pcpu_alloc_bootmem(0, num_possible_cpus() * pcpue_unit_size, PAGE_SIZE); @@ -297,7 +314,8 @@ static ssize_t __init setup_pcpu_embed(size_t static_size) pr_info("PERCPU: Embedded %zu pages at %p, static data %zu bytes\n", pcpue_size >> PAGE_SHIFT, pcpue_ptr, static_size); - return pcpu_setup_first_chunk(pcpue_get_page, static_size, 0, + return pcpu_setup_first_chunk(pcpue_get_page, static_size, + PERCPU_FIRST_CHUNK_RESERVE, pcpue_unit_size, dyn_size, pcpue_ptr, NULL); } @@ -356,8 +374,9 @@ static ssize_t __init setup_pcpu_4k(size_t static_size) pr_info("PERCPU: Allocated %d 4k pages, static data %zu bytes\n", pcpu4k_nr_static_pages, static_size); - ret = pcpu_setup_first_chunk(pcpu4k_get_page, static_size, 0, -1, -1, - NULL, pcpu4k_populate_pte); + ret = pcpu_setup_first_chunk(pcpu4k_get_page, static_size, + PERCPU_FIRST_CHUNK_RESERVE, -1, -1, NULL, + pcpu4k_populate_pte); goto out_free_ar; enomem: diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 8ff15153ae20..54a968b4b924 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -85,31 +85,20 @@ /* * PERCPU_DYNAMIC_RESERVE indicates the amount of free area to piggy - * back on the first chunk if arch is manually allocating and mapping - * it for faster access (as a part of large page mapping for example). - * Note that dynamic percpu allocator covers both static and dynamic - * areas, so these values are bigger than PERCPU_MODULE_RESERVE. + * back on the first chunk for dynamic percpu allocation if arch is + * manually allocating and mapping it for faster access (as a part of + * large page mapping for example). * - * On typical configuration with modules, the following values leave - * about 8k of free space on the first chunk after boot on both x86_32 - * and 64 when module support is enabled. When module support is - * disabled, it's much tighter. + * The following values give between one and two pages of free space + * after typical minimal boot (2-way SMP, single disk and NIC) with + * both defconfig and a distro config on x86_64 and 32. More + * intelligent way to determine this would be nice. */ -#ifndef PERCPU_DYNAMIC_RESERVE -# if BITS_PER_LONG > 32 -# ifdef CONFIG_MODULES -# define PERCPU_DYNAMIC_RESERVE (24 << 10) -# else -# define PERCPU_DYNAMIC_RESERVE (16 << 10) -# endif -# else -# ifdef CONFIG_MODULES -# define PERCPU_DYNAMIC_RESERVE (16 << 10) -# else -# define PERCPU_DYNAMIC_RESERVE (8 << 10) -# endif -# endif -#endif /* PERCPU_DYNAMIC_RESERVE */ +#if BITS_PER_LONG > 32 +#define PERCPU_DYNAMIC_RESERVE (20 << 10) +#else +#define PERCPU_DYNAMIC_RESERVE (12 << 10) +#endif extern void *pcpu_base_addr; -- cgit v1.2.3-59-g8ed1b From 6074d5b0a319fe8400ff079a3c289406ca024321 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 10 Mar 2009 16:27:48 +0900 Subject: percpu: more flexibility for @dyn_size of pcpu_setup_first_chunk() Impact: cleanup, more flexibility for first chunk init Non-negative @dyn_size used to be allowed iff @unit_size wasn't auto. This restriction stemmed from implementation detail and made things a bit less intuitive. This patch allows @dyn_size to be specified regardless of @unit_size and swaps the positions of @dyn_size and @unit_size so that the parameter order makes more sense (static, reserved and dyn sizes followed by enclosing unit_size). While at it, add @unit_size >= PCPU_MIN_UNIT_SIZE sanity check. Signed-off-by: Tejun Heo --- arch/x86/kernel/setup_percpu.c | 13 ++++++------- include/linux/percpu.h | 2 +- mm/percpu.c | 28 ++++++++++++++-------------- 3 files changed, 21 insertions(+), 22 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index efa615f2bf43..e41c51f6ada1 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -233,8 +233,8 @@ proceed: "%zu bytes\n", vm.addr, static_size); ret = pcpu_setup_first_chunk(pcpur_get_page, static_size, - PERCPU_FIRST_CHUNK_RESERVE, - PMD_SIZE, dyn_size, vm.addr, NULL); + PERCPU_FIRST_CHUNK_RESERVE, dyn_size, + PMD_SIZE, vm.addr, NULL); goto out_free_ar; enomem: @@ -315,9 +315,8 @@ static ssize_t __init setup_pcpu_embed(size_t static_size) pcpue_size >> PAGE_SHIFT, pcpue_ptr, static_size); return pcpu_setup_first_chunk(pcpue_get_page, static_size, - PERCPU_FIRST_CHUNK_RESERVE, - pcpue_unit_size, dyn_size, - pcpue_ptr, NULL); + PERCPU_FIRST_CHUNK_RESERVE, dyn_size, + pcpue_unit_size, pcpue_ptr, NULL); } /* @@ -375,8 +374,8 @@ static ssize_t __init setup_pcpu_4k(size_t static_size) pcpu4k_nr_static_pages, static_size); ret = pcpu_setup_first_chunk(pcpu4k_get_page, static_size, - PERCPU_FIRST_CHUNK_RESERVE, -1, -1, NULL, - pcpu4k_populate_pte); + PERCPU_FIRST_CHUNK_RESERVE, -1, + -1, NULL, pcpu4k_populate_pte); goto out_free_ar; enomem: diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 54a968b4b924..fb455dcc59c7 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -107,7 +107,7 @@ typedef void (*pcpu_populate_pte_fn_t)(unsigned long addr); extern size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, size_t static_size, size_t reserved_size, - ssize_t unit_size, ssize_t dyn_size, + ssize_t dyn_size, ssize_t unit_size, void *base_addr, pcpu_populate_pte_fn_t populate_pte_fn); diff --git a/mm/percpu.c b/mm/percpu.c index c6f38a2aface..2f94661d3e36 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -1027,8 +1027,8 @@ EXPORT_SYMBOL_GPL(free_percpu); * @get_page_fn: callback to fetch page pointer * @static_size: the size of static percpu area in bytes * @reserved_size: the size of reserved percpu area in bytes - * @unit_size: unit size in bytes, must be multiple of PAGE_SIZE, -1 for auto * @dyn_size: free size for dynamic allocation in bytes, -1 for auto + * @unit_size: unit size in bytes, must be multiple of PAGE_SIZE, -1 for auto * @base_addr: mapped address, NULL for auto * @populate_pte_fn: callback to allocate pagetable, NULL if unnecessary * @@ -1053,14 +1053,14 @@ EXPORT_SYMBOL_GPL(free_percpu); * limited offset range for symbol relocations to guarantee module * percpu symbols fall inside the relocatable range. * + * @dyn_size, if non-negative, determines the number of bytes + * available for dynamic allocation in the first chunk. Specifying + * non-negative value makes percpu leave alone the area beyond + * @static_size + @reserved_size + @dyn_size. + * * @unit_size, if non-negative, specifies unit size and must be * aligned to PAGE_SIZE and equal to or larger than @static_size + - * @reserved_size + @dyn_size. - * - * @dyn_size, if non-negative, limits the number of bytes available - * for dynamic allocation in the first chunk. Specifying non-negative - * value make percpu leave alone the area beyond @static_size + - * @reserved_size + @dyn_size. + * @reserved_size + if non-negative, @dyn_size. * * Non-null @base_addr means that the caller already allocated virtual * region for the first chunk and mapped it. percpu must not mess @@ -1083,12 +1083,14 @@ EXPORT_SYMBOL_GPL(free_percpu); */ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, size_t static_size, size_t reserved_size, - ssize_t unit_size, ssize_t dyn_size, + ssize_t dyn_size, ssize_t unit_size, void *base_addr, pcpu_populate_pte_fn_t populate_pte_fn) { static struct vm_struct first_vm; static int smap[2], dmap[2]; + size_t size_sum = static_size + reserved_size + + (dyn_size >= 0 ? dyn_size : 0); struct pcpu_chunk *schunk, *dchunk = NULL; unsigned int cpu; int nr_pages; @@ -1099,20 +1101,18 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, ARRAY_SIZE(dmap) >= PCPU_DFL_MAP_ALLOC); BUG_ON(!static_size); if (unit_size >= 0) { - BUG_ON(unit_size < static_size + reserved_size + - (dyn_size >= 0 ? dyn_size : 0)); + BUG_ON(unit_size < size_sum); BUG_ON(unit_size & ~PAGE_MASK); - } else { - BUG_ON(dyn_size >= 0); + BUG_ON(unit_size < PCPU_MIN_UNIT_SIZE); + } else BUG_ON(base_addr); - } BUG_ON(base_addr && populate_pte_fn); if (unit_size >= 0) pcpu_unit_pages = unit_size >> PAGE_SHIFT; else pcpu_unit_pages = max_t(int, PCPU_MIN_UNIT_SIZE >> PAGE_SHIFT, - PFN_UP(static_size + reserved_size)); + PFN_UP(size_sum)); pcpu_unit_size = pcpu_unit_pages << PAGE_SHIFT; pcpu_chunk_size = num_possible_cpus() * pcpu_unit_size; -- cgit v1.2.3-59-g8ed1b From 66c3a75772247c31feabefb724e082220a1ab060 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 10 Mar 2009 16:27:48 +0900 Subject: percpu: generalize embedding first chunk setup helper Impact: code reorganization Separate out embedding first chunk setup helper from x86 embedding first chunk allocator and put it in mm/percpu.c. This will be used by the default percpu first chunk allocator and possibly by other archs. Signed-off-by: Tejun Heo --- arch/x86/kernel/setup_percpu.c | 54 +++----------------------- include/linux/percpu.h | 4 ++ mm/percpu.c | 86 ++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 96 insertions(+), 48 deletions(-) (limited to 'include/linux') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index e41c51f6ada1..400331b50a53 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -257,31 +257,13 @@ static ssize_t __init setup_pcpu_remap(size_t static_size) * Embedding allocator * * The first chunk is sized to just contain the static area plus - * module and dynamic reserves, and allocated as a contiguous area - * using bootmem allocator and used as-is without being mapped into - * vmalloc area. This enables the first chunk to piggy back on the - * linear physical PMD mapping and doesn't add any additional pressure - * to TLB. Note that if the needed size is smaller than the minimum - * unit size, the leftover is returned to the bootmem allocator. + * module and dynamic reserves and embedded into linear physical + * mapping so that it can use PMD mapping without additional TLB + * pressure. */ -static void *pcpue_ptr __initdata; -static size_t pcpue_size __initdata; -static size_t pcpue_unit_size __initdata; - -static struct page * __init pcpue_get_page(unsigned int cpu, int pageno) -{ - size_t off = (size_t)pageno << PAGE_SHIFT; - - if (off >= pcpue_size) - return NULL; - - return virt_to_page(pcpue_ptr + cpu * pcpue_unit_size + off); -} - static ssize_t __init setup_pcpu_embed(size_t static_size) { - unsigned int cpu; - size_t dyn_size; + size_t reserve = PERCPU_MODULE_RESERVE + PERCPU_DYNAMIC_RESERVE; /* * If large page isn't supported, there's no benefit in doing @@ -291,32 +273,8 @@ static ssize_t __init setup_pcpu_embed(size_t static_size) if (!cpu_has_pse || pcpu_need_numa()) return -EINVAL; - /* allocate and copy */ - pcpue_size = PFN_ALIGN(static_size + PERCPU_MODULE_RESERVE + - PERCPU_DYNAMIC_RESERVE); - pcpue_unit_size = max_t(size_t, pcpue_size, PCPU_MIN_UNIT_SIZE); - dyn_size = pcpue_size - static_size - PERCPU_FIRST_CHUNK_RESERVE; - - pcpue_ptr = pcpu_alloc_bootmem(0, num_possible_cpus() * pcpue_unit_size, - PAGE_SIZE); - if (!pcpue_ptr) - return -ENOMEM; - - for_each_possible_cpu(cpu) { - void *ptr = pcpue_ptr + cpu * pcpue_unit_size; - - free_bootmem(__pa(ptr + pcpue_size), - pcpue_unit_size - pcpue_size); - memcpy(ptr, __per_cpu_load, static_size); - } - - /* we're ready, commit */ - pr_info("PERCPU: Embedded %zu pages at %p, static data %zu bytes\n", - pcpue_size >> PAGE_SHIFT, pcpue_ptr, static_size); - - return pcpu_setup_first_chunk(pcpue_get_page, static_size, - PERCPU_FIRST_CHUNK_RESERVE, dyn_size, - pcpue_unit_size, pcpue_ptr, NULL); + return pcpu_embed_first_chunk(static_size, PERCPU_FIRST_CHUNK_RESERVE, + reserve - PERCPU_FIRST_CHUNK_RESERVE, -1); } /* diff --git a/include/linux/percpu.h b/include/linux/percpu.h index fb455dcc59c7..ee5615d65211 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -111,6 +111,10 @@ extern size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, void *base_addr, pcpu_populate_pte_fn_t populate_pte_fn); +extern ssize_t __init pcpu_embed_first_chunk( + size_t static_size, size_t reserved_size, + ssize_t dyn_size, ssize_t unit_size); + /* * Use this to get to a cpu's version of the per-cpu object * dynamically allocated. Non-atomic access to the current CPU's diff --git a/mm/percpu.c b/mm/percpu.c index 2f94661d3e36..1aa5d8fbca12 100644 --- a/mm/percpu.c +++ b/mm/percpu.c @@ -1238,3 +1238,89 @@ size_t __init pcpu_setup_first_chunk(pcpu_get_page_fn_t get_page_fn, pcpu_base_addr = (void *)pcpu_chunk_addr(schunk, 0, 0); return pcpu_unit_size; } + +/* + * Embedding first chunk setup helper. + */ +static void *pcpue_ptr __initdata; +static size_t pcpue_size __initdata; +static size_t pcpue_unit_size __initdata; + +static struct page * __init pcpue_get_page(unsigned int cpu, int pageno) +{ + size_t off = (size_t)pageno << PAGE_SHIFT; + + if (off >= pcpue_size) + return NULL; + + return virt_to_page(pcpue_ptr + cpu * pcpue_unit_size + off); +} + +/** + * pcpu_embed_first_chunk - embed the first percpu chunk into bootmem + * @static_size: the size of static percpu area in bytes + * @reserved_size: the size of reserved percpu area in bytes + * @dyn_size: free size for dynamic allocation in bytes, -1 for auto + * @unit_size: unit size in bytes, must be multiple of PAGE_SIZE, -1 for auto + * + * This is a helper to ease setting up embedded first percpu chunk and + * can be called where pcpu_setup_first_chunk() is expected. + * + * If this function is used to setup the first chunk, it is allocated + * as a contiguous area using bootmem allocator and used as-is without + * being mapped into vmalloc area. This enables the first chunk to + * piggy back on the linear physical mapping which often uses larger + * page size. + * + * When @dyn_size is positive, dynamic area might be larger than + * specified to fill page alignment. Also, when @dyn_size is auto, + * @dyn_size does not fill the whole first chunk but only what's + * necessary for page alignment after static and reserved areas. + * + * If the needed size is smaller than the minimum or specified unit + * size, the leftover is returned to the bootmem allocator. + * + * RETURNS: + * The determined pcpu_unit_size which can be used to initialize + * percpu access on success, -errno on failure. + */ +ssize_t __init pcpu_embed_first_chunk(size_t static_size, size_t reserved_size, + ssize_t dyn_size, ssize_t unit_size) +{ + unsigned int cpu; + + /* determine parameters and allocate */ + pcpue_size = PFN_ALIGN(static_size + reserved_size + + (dyn_size >= 0 ? dyn_size : 0)); + if (dyn_size != 0) + dyn_size = pcpue_size - static_size - reserved_size; + + if (unit_size >= 0) { + BUG_ON(unit_size < pcpue_size); + pcpue_unit_size = unit_size; + } else + pcpue_unit_size = max_t(size_t, pcpue_size, PCPU_MIN_UNIT_SIZE); + + pcpue_ptr = __alloc_bootmem_nopanic( + num_possible_cpus() * pcpue_unit_size, + PAGE_SIZE, __pa(MAX_DMA_ADDRESS)); + if (!pcpue_ptr) + return -ENOMEM; + + /* return the leftover and copy */ + for_each_possible_cpu(cpu) { + void *ptr = pcpue_ptr + cpu * pcpue_unit_size; + + free_bootmem(__pa(ptr + pcpue_size), + pcpue_unit_size - pcpue_size); + memcpy(ptr, __per_cpu_load, static_size); + } + + /* we're ready, commit */ + pr_info("PERCPU: Embedded %zu pages at %p, static data %zu bytes\n", + pcpue_size >> PAGE_SHIFT, pcpue_ptr, static_size); + + return pcpu_setup_first_chunk(pcpue_get_page, static_size, + reserved_size, dyn_size, + pcpue_unit_size, pcpue_ptr, NULL); +} -- cgit v1.2.3-59-g8ed1b From 71969fd9e2c523d22bf1742eb31f1562247710eb Mon Sep 17 00:00:00 2001 From: Boaz Harrosh Date: Sun, 25 Jan 2009 16:50:02 +0200 Subject: [SCSI] major.h: char-major number for OSD device driver Allocate major 260 for osd. Signed-off-by: Boaz Harrosh CC: Torben Mathiasen Signed-off-by: James Bottomley --- Documentation/devices.txt | 6 ++++++ include/linux/major.h | 1 + 2 files changed, 7 insertions(+) (limited to 'include/linux') diff --git a/Documentation/devices.txt b/Documentation/devices.txt index 2be08240ee80..62254d4510c6 100644 --- a/Documentation/devices.txt +++ b/Documentation/devices.txt @@ -3145,6 +3145,12 @@ Your cooperation is appreciated. 1 = /dev/blockrom1 Second ROM card's translation layer interface ... +260 char OSD (Object-based-device) SCSI Device + 0 = /dev/osd0 First OSD Device + 1 = /dev/osd1 Second OSD Device + ... + 255 = /dev/osd255 256th OSD Device + **** ADDITIONAL /dev DIRECTORY ENTRIES This section details additional entries that should or may exist in diff --git a/include/linux/major.h b/include/linux/major.h index 88249452b935..058ec15dd060 100644 --- a/include/linux/major.h +++ b/include/linux/major.h @@ -171,5 +171,6 @@ #define VIOTAPE_MAJOR 230 #define BLOCK_EXT_MAJOR 259 +#define SCSI_OSD_MAJOR 260 /* open-osd's OSD scsi device */ #endif -- cgit v1.2.3-59-g8ed1b From 211c738d86f3f423f1b218ab3a356c9538e38047 Mon Sep 17 00:00:00 2001 From: Yi Zou Date: Fri, 27 Feb 2009 14:06:37 -0800 Subject: [SCSI] net, fcoe: add ETH_P_FCOE for Fibre Channel over Ethernet (FCoE) This adds eth type ETH_P_FCOE for Fibre Channel over Ethernet (FCoE), consequently, the ETH_P_FCOE from fc_fcoe.h and fcoe skb->protocol is not set as ETH_P_FCOE. Signed-off-by: Yi Zou Acked-by: David Miller Signed-off-by: James Bottomley --- drivers/scsi/fcoe/libfcoe.c | 2 +- include/linux/if_ether.h | 1 + include/scsi/fc/fc_fcoe.h | 7 ------- 3 files changed, 2 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/drivers/scsi/fcoe/libfcoe.c b/drivers/scsi/fcoe/libfcoe.c index 5548bf3bb58b..a99a42807b38 100644 --- a/drivers/scsi/fcoe/libfcoe.c +++ b/drivers/scsi/fcoe/libfcoe.c @@ -460,7 +460,7 @@ int fcoe_xmit(struct fc_lport *lp, struct fc_frame *fp) skb_reset_mac_header(skb); skb_reset_network_header(skb); skb->mac_len = elen; - skb->protocol = htons(ETH_P_802_3); + skb->protocol = htons(ETH_P_FCOE); skb->dev = fc->real_dev; /* fill up mac and fcoe headers */ diff --git a/include/linux/if_ether.h b/include/linux/if_ether.h index 7f3c735f422b..59d197cb4851 100644 --- a/include/linux/if_ether.h +++ b/include/linux/if_ether.h @@ -78,6 +78,7 @@ #define ETH_P_PAE 0x888E /* Port Access Entity (IEEE 802.1X) */ #define ETH_P_AOE 0x88A2 /* ATA over Ethernet */ #define ETH_P_TIPC 0x88CA /* TIPC */ +#define ETH_P_FCOE 0x8906 /* Fibre Channel over Ethernet */ #define ETH_P_EDSA 0xDADA /* Ethertype DSA [ NOT AN OFFICIALLY REGISTERED ID ] */ /* diff --git a/include/scsi/fc/fc_fcoe.h b/include/scsi/fc/fc_fcoe.h index f271d9cc0fc2..ccb3dbe90463 100644 --- a/include/scsi/fc/fc_fcoe.h +++ b/include/scsi/fc/fc_fcoe.h @@ -24,13 +24,6 @@ * FCoE - Fibre Channel over Ethernet. */ -/* - * The FCoE ethertype eventually goes in net/if_ether.h. - */ -#ifndef ETH_P_FCOE -#define ETH_P_FCOE 0x8906 /* FCOE ether type */ -#endif - /* * FC_FCOE_OUI hasn't been standardized yet. XXX TBD. */ -- cgit v1.2.3-59-g8ed1b From 43eb99c5b349b188f82725652f3d1018c619d682 Mon Sep 17 00:00:00 2001 From: Chris Leech Date: Fri, 27 Feb 2009 14:06:43 -0800 Subject: [SCSI] net: reclaim 8 upper bits of the netdev->features from GSO Reclaim 8 upper bits of netdev->features from GSO. Signed-off-by: Chris Leech Signed-off-by: Yi Zou Acked-by: David Miller Signed-off-by: James Bottomley --- drivers/net/xen-netfront.c | 2 +- include/linux/netdevice.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c index cd6184ee08ee..2ce536fcd209 100644 --- a/drivers/net/xen-netfront.c +++ b/drivers/net/xen-netfront.c @@ -1511,7 +1511,7 @@ static int xennet_set_tso(struct net_device *dev, u32 data) static void xennet_set_features(struct net_device *dev) { /* Turn off all GSO bits except ROBUST. */ - dev->features &= (1 << NETIF_F_GSO_SHIFT) - 1; + dev->features &= ~NETIF_F_GSO_MASK; dev->features |= NETIF_F_GSO_ROBUST; xennet_set_sg(dev, 0); diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index ec54785d34f9..c8238d9ba376 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -652,7 +652,7 @@ struct net_device /* Segmentation offload features */ #define NETIF_F_GSO_SHIFT 16 -#define NETIF_F_GSO_MASK 0xffff0000 +#define NETIF_F_GSO_MASK 0x00ff0000 #define NETIF_F_TSO (SKB_GSO_TCPV4 << NETIF_F_GSO_SHIFT) #define NETIF_F_UFO (SKB_GSO_UDP << NETIF_F_GSO_SHIFT) #define NETIF_F_GSO_ROBUST (SKB_GSO_DODGY << NETIF_F_GSO_SHIFT) -- cgit v1.2.3-59-g8ed1b From 01d5b2fca1fa58ed5039239fd531e9f658971ace Mon Sep 17 00:00:00 2001 From: Chris Leech Date: Fri, 27 Feb 2009 14:06:49 -0800 Subject: [SCSI] net: define feature flags for FCoE offloads Define feature flags for FCoE offloads. Signed-off-by: Chris Leech Signed-off-by: Yi Zou Acked-by: David Miller Signed-off-by: James Bottomley --- include/linux/netdevice.h | 3 +++ include/linux/skbuff.h | 2 ++ 2 files changed, 5 insertions(+) (limited to 'include/linux') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index c8238d9ba376..5c405571cb60 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -650,6 +650,8 @@ struct net_device #define NETIF_F_GRO 16384 /* Generic receive offload */ #define NETIF_F_LRO 32768 /* large receive offload */ +#define NETIF_F_FCOE_CRC (1 << 24) /* FCoE CRC32 */ + /* Segmentation offload features */ #define NETIF_F_GSO_SHIFT 16 #define NETIF_F_GSO_MASK 0x00ff0000 @@ -658,6 +660,7 @@ struct net_device #define NETIF_F_GSO_ROBUST (SKB_GSO_DODGY << NETIF_F_GSO_SHIFT) #define NETIF_F_TSO_ECN (SKB_GSO_TCP_ECN << NETIF_F_GSO_SHIFT) #define NETIF_F_TSO6 (SKB_GSO_TCPV6 << NETIF_F_GSO_SHIFT) +#define NETIF_F_FSO (SKB_GSO_FCOE << NETIF_F_GSO_SHIFT) /* List of features with software fallbacks. */ #define NETIF_F_GSO_SOFTWARE (NETIF_F_TSO | NETIF_F_TSO_ECN | NETIF_F_TSO6) diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 9dcf956ad18a..02adea2099a7 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -188,6 +188,8 @@ enum { SKB_GSO_TCP_ECN = 1 << 3, SKB_GSO_TCPV6 = 1 << 4, + + SKB_GSO_FCOE = 1 << 5, }; #if BITS_PER_LONG > 32 -- cgit v1.2.3-59-g8ed1b From 4d288d5767f853bfca25adc7b6030dc95518cb2e Mon Sep 17 00:00:00 2001 From: Yi Zou Date: Fri, 27 Feb 2009 14:06:59 -0800 Subject: [SCSI] net: add FCoE offload support through net_device This adds support to provide Fiber Channel over Ethernet (FCoE) offload through net_device's net_device_ops struct. The offload through net_device for FCoE is enabled in kernel as built-in or module driver. Signed-off-by: Yi Zou Acked-by: David Miller Signed-off-by: James Bottomley --- include/linux/netdevice.h | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'include/linux') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 5c405571cb60..7ed49f5335b1 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -582,6 +582,14 @@ struct net_device_ops { #define HAVE_NETDEV_POLL void (*ndo_poll_controller)(struct net_device *dev); #endif +#if defined(CONFIG_FCOE) || defined(CONFIG_FCOE_MODULE) + int (*ndo_fcoe_ddp_setup)(struct net_device *dev, + u16 xid, + struct scatterlist *sgl, + unsigned int sgc); + int (*ndo_fcoe_ddp_done)(struct net_device *dev, + u16 xid); +#endif }; /* @@ -843,6 +851,11 @@ struct net_device struct dcbnl_rtnl_ops *dcbnl_ops; #endif +#if defined(CONFIG_FCOE) || defined(CONFIG_FCOE_MODULE) + /* max exchange id for FCoE LRO by ddp */ + unsigned int fcoe_ddp_xid; +#endif + #ifdef CONFIG_COMPAT_NET_DEV_OPS struct { int (*init)(struct net_device *dev); -- cgit v1.2.3-59-g8ed1b From 4ab3b73f85ca2e99d9dbdb55ac13e57327a7e915 Mon Sep 17 00:00:00 2001 From: Douglas Gilbert Date: Mon, 9 Mar 2009 10:51:38 -0400 Subject: [SCSI] bsg: add linux/types.h include to bsg.h Since bsg.h has recently been added to the list of kernel headers that should be exported to the user space, this attachment makes bsg.h more user space "friendly". Specifically autotools dislike headers that don't compile freestanding and bsg.h's use of __u32 types (and friends) are not standard C (C90 or C99). The inclusion of linux/types.h fixes that. Signed-off-by: Douglas Gilbert Signed-off-by: James Bottomley --- include/linux/bsg.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/linux') diff --git a/include/linux/bsg.h b/include/linux/bsg.h index cf0303a60611..6c0a00dfa90c 100644 --- a/include/linux/bsg.h +++ b/include/linux/bsg.h @@ -1,6 +1,8 @@ #ifndef BSG_H #define BSG_H +#include + #define BSG_PROTOCOL_SCSI 0 #define BSG_SUB_PROTOCOL_SCSI_CMD 0 -- cgit v1.2.3-59-g8ed1b From dec3f95959bff957f5bcbf16c2a2823f7e33d1e7 Mon Sep 17 00:00:00 2001 From: Eric Moore Date: Mon, 9 Mar 2009 01:27:49 -0600 Subject: [SCSI] mpt2sas: add MPT2SAS_MINOR(221) to miscdevice.h Signed-off-by: Eric Moore Signed-off-by: James Bottomley --- include/linux/miscdevice.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include/linux') diff --git a/include/linux/miscdevice.h b/include/linux/miscdevice.h index a820f816a49e..beb6ec99cfef 100644 --- a/include/linux/miscdevice.h +++ b/include/linux/miscdevice.h @@ -26,6 +26,7 @@ #define TUN_MINOR 200 #define MWAVE_MINOR 219 /* ACP/Mwave Modem */ #define MPT_MINOR 220 +#define MPT2SAS_MINOR 221 #define HPET_MINOR 228 #define FUSE_MINOR 229 #define KVM_MINOR 232 -- cgit v1.2.3-59-g8ed1b From bf8e3355ec8f4e472f9841e94203cd759b45226e Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Fri, 5 Dec 2008 22:43:41 +0100 Subject: firewire: cdev: documentation fixlet Reported-by: Jay Fenlason Signed-off-by: Stefan Richter --- include/linux/firewire-cdev.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 4d078e99c017..899ef279f5be 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -229,7 +229,7 @@ struct fw_cdev_get_info { * Send a request to the device. This ioctl implements all outgoing requests. * Both quadlet and block request specify the payload as a pointer to the data * in the @data field. Once the transaction completes, the kernel writes an - * &fw_cdev_event_request event back. The @closure field is passed back to + * &fw_cdev_event_response event back. The @closure field is passed back to * user space in the response event. */ struct fw_cdev_send_request { -- cgit v1.2.3-59-g8ed1b From 632321ecd99bf85c982a75f8329b4ecbb95b3a8f Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Fri, 2 Jan 2009 12:47:13 +0100 Subject: firewire: cdev: fix documentation of FW_CDEV_IOC_GET_INFO The FW_CDEV_IOC_GET_INFO ioctl looks at client->device->config_rom, not at the local node's config ROM. We could fix the implementation or the documentation. I believe the way how it is currently implemented is more useful than the way how it is currently documented. In fact, libdc1394 uses the ABI already as implemented, not as documented. Hence let's change the documentation. Signed-off-by: Stefan Richter --- include/linux/firewire-cdev.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 899ef279f5be..86c8ff5326f9 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -201,7 +201,7 @@ union fw_cdev_event { * case, @rom_length is updated with the actual length of the * configuration ROM. * @rom: If non-zero, address of a buffer to be filled by a copy of the - * local node's configuration ROM + * device's configuration ROM * @bus_reset: If non-zero, address of a buffer to be filled by a * &struct fw_cdev_event_bus_reset with the current state * of the bus. This does not cause a bus reset to happen. -- cgit v1.2.3-59-g8ed1b From b1bda4cdc2037447bd66753bf5ccab66d91b0b59 Mon Sep 17 00:00:00 2001 From: "Jay Fenlason, Stefan Richter" Date: Sun, 4 Jan 2009 16:23:29 +0100 Subject: firewire: cdev: add ioctls for isochronous resource management Based on Date: Tue, 18 Nov 2008 11:41:27 -0500 From: Jay Fenlason Subject: [Patch V4] Add ISO resource management support with several changes to the ABI and implementation. Only the part of the ABI which enables auto-reallocation and auto-deallocation is included here. This implements ioctls for kernel-assisted allocation of isochronous channels and isochronous bandwidth. The benefits are: - The client does not have to have write access to the /dev/fw* device corresponding to the IRM. - The client does not have to perform reallocation after bus resets. - Channel and bandwidth are deallocated by the kernel if the file is closed before the client deallocated the resources. Thus resources are released even if the client crashes. It is anticipated that future in-kernel code (firewire-core IRM code; the firewire port of firedtv), will use the fw-iso.c portions of this code too. Signed-off-by: Stefan Richter Tested-by: David Moore --- drivers/firewire/fw-cdev.c | 215 +++++++++++++++++++++++++++++++++++++- drivers/firewire/fw-iso.c | 176 +++++++++++++++++++++++++++++-- drivers/firewire/fw-transaction.h | 4 + include/linux/firewire-cdev.h | 100 +++++++++++++++--- 4 files changed, 475 insertions(+), 20 deletions(-) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index 4c33b51b735a..a227853aa1e2 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include #include @@ -35,6 +36,7 @@ #include #include #include +#include #include #include @@ -114,6 +116,21 @@ struct descriptor_resource { u32 data[0]; }; +struct iso_resource { + struct client_resource resource; + struct client *client; + /* Schedule work and access todo only with client->lock held. */ + struct delayed_work work; + enum {ISO_RES_ALLOC, ISO_RES_REALLOC, ISO_RES_DEALLOC,} todo; + int generation; + u64 channels; + s32 bandwidth; + struct iso_resource_event *e_alloc, *e_dealloc; +}; + +static void schedule_iso_resource(struct iso_resource *); +static void release_iso_resource(struct client *, struct client_resource *); + /* * dequeue_event() just kfree()'s the event, so the event has to be * the first field in a struct XYZ_event. @@ -145,6 +162,11 @@ struct iso_interrupt_event { struct fw_cdev_event_iso_interrupt interrupt; }; +struct iso_resource_event { + struct event event; + struct fw_cdev_event_iso_resource resource; +}; + static inline void __user *u64_to_uptr(__u64 value) { return (void __user *)(unsigned long)value; @@ -290,6 +312,16 @@ static void for_each_client(struct fw_device *device, mutex_unlock(&device->client_list_mutex); } +static int schedule_reallocations(int id, void *p, void *data) +{ + struct client_resource *r = p; + + if (r->release == release_iso_resource) + schedule_iso_resource(container_of(r, + struct iso_resource, resource)); + return 0; +} + static void queue_bus_reset_event(struct client *client) { struct bus_reset_event *e; @@ -304,6 +336,10 @@ static void queue_bus_reset_event(struct client *client) queue_event(client, &e->event, &e->reset, sizeof(e->reset), NULL, 0); + + spin_lock_irq(&client->lock); + idr_for_each(&client->resource_idr, schedule_reallocations, client); + spin_unlock_irq(&client->lock); } void fw_device_cdev_update(struct fw_device *device) @@ -376,8 +412,12 @@ static int add_client_resource(struct client *client, else ret = idr_get_new(&client->resource_idr, resource, &resource->handle); - if (ret >= 0) + if (ret >= 0) { client_get(client); + if (resource->release == release_iso_resource) + schedule_iso_resource(container_of(resource, + struct iso_resource, resource)); + } spin_unlock_irqrestore(&client->lock, flags); if (ret == -EAGAIN) @@ -970,6 +1010,177 @@ static int ioctl_get_cycle_timer(struct client *client, void *buffer) return 0; } +static void iso_resource_work(struct work_struct *work) +{ + struct iso_resource_event *e; + struct iso_resource *r = + container_of(work, struct iso_resource, work.work); + struct client *client = r->client; + int generation, channel, bandwidth, todo; + bool skip, free, success; + + spin_lock_irq(&client->lock); + generation = client->device->generation; + todo = r->todo; + /* Allow 1000ms grace period for other reallocations. */ + if (todo == ISO_RES_ALLOC && + time_is_after_jiffies(client->device->card->reset_jiffies + HZ)) { + if (schedule_delayed_work(&r->work, DIV_ROUND_UP(HZ, 3))) + client_get(client); + skip = true; + } else { + /* We could be called twice within the same generation. */ + skip = todo == ISO_RES_REALLOC && + r->generation == generation; + } + free = todo == ISO_RES_DEALLOC; + r->generation = generation; + spin_unlock_irq(&client->lock); + + if (skip) + goto out; + + bandwidth = r->bandwidth; + + fw_iso_resource_manage(client->device->card, generation, + r->channels, &channel, &bandwidth, + todo == ISO_RES_ALLOC || todo == ISO_RES_REALLOC); + /* + * Is this generation outdated already? As long as this resource sticks + * in the idr, it will be scheduled again for a newer generation or at + * shutdown. + */ + if (channel == -EAGAIN && + (todo == ISO_RES_ALLOC || todo == ISO_RES_REALLOC)) + goto out; + + success = channel >= 0 || bandwidth > 0; + + spin_lock_irq(&client->lock); + /* + * Transit from allocation to reallocation, except if the client + * requested deallocation in the meantime. + */ + if (r->todo == ISO_RES_ALLOC) + r->todo = ISO_RES_REALLOC; + /* + * Allocation or reallocation failure? Pull this resource out of the + * idr and prepare for deletion, unless the client is shutting down. + */ + if (r->todo == ISO_RES_REALLOC && !success && + !client->in_shutdown && + idr_find(&client->resource_idr, r->resource.handle)) { + idr_remove(&client->resource_idr, r->resource.handle); + client_put(client); + free = true; + } + spin_unlock_irq(&client->lock); + + if (todo == ISO_RES_ALLOC && channel >= 0) + r->channels = 1ULL << (63 - channel); + + if (todo == ISO_RES_REALLOC && success) + goto out; + + if (todo == ISO_RES_ALLOC) { + e = r->e_alloc; + r->e_alloc = NULL; + } else { + e = r->e_dealloc; + r->e_dealloc = NULL; + } + e->resource.handle = r->resource.handle; + e->resource.channel = channel; + e->resource.bandwidth = bandwidth; + + queue_event(client, &e->event, + &e->resource, sizeof(e->resource), NULL, 0); + + if (free) { + cancel_delayed_work(&r->work); + kfree(r->e_alloc); + kfree(r->e_dealloc); + kfree(r); + } + out: + client_put(client); +} + +static void schedule_iso_resource(struct iso_resource *r) +{ + if (schedule_delayed_work(&r->work, 0)) + client_get(r->client); +} + +static void release_iso_resource(struct client *client, + struct client_resource *resource) +{ + struct iso_resource *r = + container_of(resource, struct iso_resource, resource); + + spin_lock_irq(&client->lock); + r->todo = ISO_RES_DEALLOC; + schedule_iso_resource(r); + spin_unlock_irq(&client->lock); +} + +static int ioctl_allocate_iso_resource(struct client *client, void *buffer) +{ + struct fw_cdev_allocate_iso_resource *request = buffer; + struct iso_resource_event *e1, *e2; + struct iso_resource *r; + int ret; + + if ((request->channels == 0 && request->bandwidth == 0) || + request->bandwidth > BANDWIDTH_AVAILABLE_INITIAL || + request->bandwidth < 0) + return -EINVAL; + + r = kmalloc(sizeof(*r), GFP_KERNEL); + e1 = kmalloc(sizeof(*e1), GFP_KERNEL); + e2 = kmalloc(sizeof(*e2), GFP_KERNEL); + if (r == NULL || e1 == NULL || e2 == NULL) { + ret = -ENOMEM; + goto fail; + } + + INIT_DELAYED_WORK(&r->work, iso_resource_work); + r->client = client; + r->todo = ISO_RES_ALLOC; + r->generation = -1; + r->channels = request->channels; + r->bandwidth = request->bandwidth; + r->e_alloc = e1; + r->e_dealloc = e2; + + e1->resource.closure = request->closure; + e1->resource.type = FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED; + e2->resource.closure = request->closure; + e2->resource.type = FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED; + + r->resource.release = release_iso_resource; + ret = add_client_resource(client, &r->resource, GFP_KERNEL); + if (ret < 0) + goto fail; + request->handle = r->resource.handle; + + return 0; + fail: + kfree(r); + kfree(e1); + kfree(e2); + + return ret; +} + +static int ioctl_deallocate_iso_resource(struct client *client, void *buffer) +{ + struct fw_cdev_deallocate *request = buffer; + + return release_client_resource(client, request->handle, + release_iso_resource, NULL); +} + static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_get_info, ioctl_send_request, @@ -984,6 +1195,8 @@ static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_start_iso, ioctl_stop_iso, ioctl_get_cycle_timer, + ioctl_allocate_iso_resource, + ioctl_deallocate_iso_resource, }; static int dispatch_ioctl(struct client *client, diff --git a/drivers/firewire/fw-iso.c b/drivers/firewire/fw-iso.c index 39f3bacee404..a7b57b253b06 100644 --- a/drivers/firewire/fw-iso.c +++ b/drivers/firewire/fw-iso.c @@ -1,5 +1,7 @@ /* - * Isochronous IO functionality + * Isochronous I/O functionality: + * - Isochronous DMA context management + * - Isochronous bus resource management (channels, bandwidth), client side * * Copyright (C) 2006 Kristian Hoegsberg * @@ -18,15 +20,20 @@ * Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ -#include -#include #include -#include +#include +#include +#include #include +#include +#include -#include "fw-transaction.h" #include "fw-topology.h" -#include "fw-device.h" +#include "fw-transaction.h" + +/* + * Isochronous DMA context management + */ int fw_iso_buffer_init(struct fw_iso_buffer *buffer, struct fw_card *card, int page_count, enum dma_data_direction direction) @@ -153,3 +160,160 @@ int fw_iso_context_stop(struct fw_iso_context *ctx) { return ctx->card->driver->stop_iso(ctx); } + +/* + * Isochronous bus resource management (channels, bandwidth), client side + */ + +static int manage_bandwidth(struct fw_card *card, int irm_id, int generation, + int bandwidth, bool allocate) +{ + __be32 data[2]; + int try, new, old = allocate ? BANDWIDTH_AVAILABLE_INITIAL : 0; + + /* + * On a 1394a IRM with low contention, try < 1 is enough. + * On a 1394-1995 IRM, we need at least try < 2. + * Let's just do try < 5. + */ + for (try = 0; try < 5; try++) { + new = allocate ? old - bandwidth : old + bandwidth; + if (new < 0 || new > BANDWIDTH_AVAILABLE_INITIAL) + break; + + data[0] = cpu_to_be32(old); + data[1] = cpu_to_be32(new); + switch (fw_run_transaction(card, TCODE_LOCK_COMPARE_SWAP, + irm_id, generation, SCODE_100, + CSR_REGISTER_BASE + CSR_BANDWIDTH_AVAILABLE, + data, sizeof(data))) { + case RCODE_GENERATION: + /* A generation change frees all bandwidth. */ + return allocate ? -EAGAIN : bandwidth; + + case RCODE_COMPLETE: + if (be32_to_cpup(data) == old) + return bandwidth; + + old = be32_to_cpup(data); + /* Fall through. */ + } + } + + return -EIO; +} + +static int manage_channel(struct fw_card *card, int irm_id, int generation, + __be32 channels_mask, u64 offset, bool allocate) +{ + __be32 data[2], c, old = allocate ? cpu_to_be32(~0) : 0; + int i, retry = 5; + + for (i = 0; i < 32; i++) { + c = cpu_to_be32(1 << (31 - i)); + if (!(channels_mask & c)) + continue; + + if (allocate == !(old & c)) + continue; + + data[0] = old; + data[1] = old ^ c; + switch (fw_run_transaction(card, TCODE_LOCK_COMPARE_SWAP, + irm_id, generation, SCODE_100, + offset, data, sizeof(data))) { + case RCODE_GENERATION: + /* A generation change frees all channels. */ + return allocate ? -EAGAIN : i; + + case RCODE_COMPLETE: + if (data[0] == old) + return i; + + old = data[0]; + + /* Is the IRM 1394a-2000 compliant? */ + if ((data[0] & c) != (data[1] & c)) + continue; + + /* 1394-1995 IRM, fall through to retry. */ + default: + if (retry--) + i--; + } + } + + return -EIO; +} + +static void deallocate_channel(struct fw_card *card, int irm_id, + int generation, int channel) +{ + __be32 mask; + u64 offset; + + mask = channel < 32 ? cpu_to_be32(1 << (31 - channel)) : + cpu_to_be32(1 << (63 - channel)); + offset = channel < 32 ? CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_HI : + CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_LO; + + manage_channel(card, irm_id, generation, mask, offset, false); +} + +/** + * fw_iso_resource_manage - Allocate or deallocate a channel and/or bandwidth + * + * In parameters: card, generation, channels_mask, bandwidth, allocate + * Out parameters: channel, bandwidth + * This function blocks (sleeps) during communication with the IRM. + * Allocates or deallocates at most one channel out of channels_mask. + * + * Returns channel < 0 if no channel was allocated or deallocated. + * Returns bandwidth = 0 if no bandwidth was allocated or deallocated. + * + * If generation is stale, deallocations succeed but allocations fail with + * channel = -EAGAIN. + * + * If channel (de)allocation fails, bandwidth (de)allocation fails too. + * If bandwidth allocation fails, no channel will be allocated either. + * If bandwidth deallocation fails, channel deallocation may still have been + * successful. + */ +void fw_iso_resource_manage(struct fw_card *card, int generation, + u64 channels_mask, int *channel, int *bandwidth, + bool allocate) +{ + __be32 channels_hi = cpu_to_be32(channels_mask >> 32); + __be32 channels_lo = cpu_to_be32(channels_mask); + int irm_id, ret, c = -EINVAL; + + spin_lock_irq(&card->lock); + irm_id = card->irm_node->node_id; + spin_unlock_irq(&card->lock); + + if (channels_hi) + c = manage_channel(card, irm_id, generation, channels_hi, + CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_HI, allocate); + if (channels_lo && c < 0) { + c = manage_channel(card, irm_id, generation, channels_lo, + CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_LO, allocate); + if (c >= 0) + c += 32; + } + *channel = c; + + if (channels_mask != 0 && c < 0) + *bandwidth = 0; + + if (*bandwidth == 0) + return; + + ret = manage_bandwidth(card, irm_id, generation, *bandwidth, allocate); + if (ret < 0) + *bandwidth = 0; + + if (ret < 0 && c >= 0 && allocate) { + deallocate_channel(card, irm_id, generation, c); + *channel = ret; + } +} diff --git a/drivers/firewire/fw-transaction.h b/drivers/firewire/fw-transaction.h index 48e88d53998b..212a10293828 100644 --- a/drivers/firewire/fw-transaction.h +++ b/drivers/firewire/fw-transaction.h @@ -82,6 +82,7 @@ #define CSR_SPEED_MAP 0x2000 #define CSR_SPEED_MAP_END 0x3000 +#define BANDWIDTH_AVAILABLE_INITIAL 4915 #define BROADCAST_CHANNEL_INITIAL (1 << 31 | 31) #define BROADCAST_CHANNEL_VALID (1 << 30) @@ -343,6 +344,9 @@ int fw_iso_context_start(struct fw_iso_context *ctx, int fw_iso_context_stop(struct fw_iso_context *ctx); void fw_iso_context_destroy(struct fw_iso_context *ctx); +void fw_iso_resource_manage(struct fw_card *card, int generation, + u64 channels_mask, int *channel, int *bandwidth, bool allocate); + struct fw_card_driver { /* * Enable the given card with the given initial config rom. diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 86c8ff5326f9..25b96dd0574f 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -25,10 +25,12 @@ #include #include -#define FW_CDEV_EVENT_BUS_RESET 0x00 -#define FW_CDEV_EVENT_RESPONSE 0x01 -#define FW_CDEV_EVENT_REQUEST 0x02 -#define FW_CDEV_EVENT_ISO_INTERRUPT 0x03 +#define FW_CDEV_EVENT_BUS_RESET 0x00 +#define FW_CDEV_EVENT_RESPONSE 0x01 +#define FW_CDEV_EVENT_REQUEST 0x02 +#define FW_CDEV_EVENT_ISO_INTERRUPT 0x03 +#define FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED 0x04 +#define FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED 0x05 /** * struct fw_cdev_event_common - Common part of all fw_cdev_event_ types @@ -146,6 +148,37 @@ struct fw_cdev_event_iso_interrupt { __u32 header[0]; }; +/** + * struct fw_cdev_event_iso_resource - Iso resources were allocated or freed + * @closure: See &fw_cdev_event_common; + * set by %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE ioctl + * @type: %FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED or + * %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED + * @handle: Reference by which an allocated resource can be deallocated + * @channel: Isochronous channel which was (de)allocated, if any + * @bandwidth: Bandwidth allocation units which were (de)allocated, if any + * @channels_available: Last known availability of channels + * @bandwidth_available: Last known availability of bandwidth + * + * An %FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED event is sent after an isochronous + * resource was allocated at the IRM. The client has to check @channel and + * @bandwidth for whether the allocation actually succeeded. + * + * @channel is <0 if no channel was allocated. + * @bandwidth is 0 if no bandwidth was allocated. + * + * An %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED event is sent after an isochronous + * resource was deallocated at the IRM. It is also sent when automatic + * reallocation after a bus reset failed. + */ +struct fw_cdev_event_iso_resource { + __u64 closure; + __u32 type; + __u32 handle; + __s32 channel; + __s32 bandwidth; +}; + /** * union fw_cdev_event - Convenience union of fw_cdev_event_ types * @common: Valid for all types @@ -153,6 +186,9 @@ struct fw_cdev_event_iso_interrupt { * @response: Valid if @common.type == %FW_CDEV_EVENT_RESPONSE * @request: Valid if @common.type == %FW_CDEV_EVENT_REQUEST * @iso_interrupt: Valid if @common.type == %FW_CDEV_EVENT_ISO_INTERRUPT + * @iso_resource: Valid if @common.type == + * %FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED or + * %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED * * Convenience union for userspace use. Events could be read(2) into an * appropriately aligned char buffer and then cast to this union for further @@ -163,13 +199,15 @@ struct fw_cdev_event_iso_interrupt { * not fit will be discarded so that the next read(2) will return a new event. */ union fw_cdev_event { - struct fw_cdev_event_common common; - struct fw_cdev_event_bus_reset bus_reset; - struct fw_cdev_event_response response; - struct fw_cdev_event_request request; - struct fw_cdev_event_iso_interrupt iso_interrupt; + struct fw_cdev_event_common common; + struct fw_cdev_event_bus_reset bus_reset; + struct fw_cdev_event_response response; + struct fw_cdev_event_request request; + struct fw_cdev_event_iso_interrupt iso_interrupt; + struct fw_cdev_event_iso_resource iso_resource; }; +/* available since kernel version 2.6.22 */ #define FW_CDEV_IOC_GET_INFO _IOWR('#', 0x00, struct fw_cdev_get_info) #define FW_CDEV_IOC_SEND_REQUEST _IOW('#', 0x01, struct fw_cdev_send_request) #define FW_CDEV_IOC_ALLOCATE _IOWR('#', 0x02, struct fw_cdev_allocate) @@ -178,13 +216,18 @@ union fw_cdev_event { #define FW_CDEV_IOC_INITIATE_BUS_RESET _IOW('#', 0x05, struct fw_cdev_initiate_bus_reset) #define FW_CDEV_IOC_ADD_DESCRIPTOR _IOWR('#', 0x06, struct fw_cdev_add_descriptor) #define FW_CDEV_IOC_REMOVE_DESCRIPTOR _IOW('#', 0x07, struct fw_cdev_remove_descriptor) - #define FW_CDEV_IOC_CREATE_ISO_CONTEXT _IOWR('#', 0x08, struct fw_cdev_create_iso_context) #define FW_CDEV_IOC_QUEUE_ISO _IOWR('#', 0x09, struct fw_cdev_queue_iso) #define FW_CDEV_IOC_START_ISO _IOW('#', 0x0a, struct fw_cdev_start_iso) #define FW_CDEV_IOC_STOP_ISO _IOW('#', 0x0b, struct fw_cdev_stop_iso) + +/* available since kernel version 2.6.24 */ #define FW_CDEV_IOC_GET_CYCLE_TIMER _IOR('#', 0x0c, struct fw_cdev_get_cycle_timer) +/* available since kernel version 2.6.30 */ +#define FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE _IOWR('#', 0x0d, struct fw_cdev_allocate_iso_resource) +#define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE _IOW('#', 0x0e, struct fw_cdev_deallocate) + /* FW_CDEV_VERSION History * * 1 Feb 18, 2007: Initial version. @@ -284,9 +327,9 @@ struct fw_cdev_allocate { }; /** - * struct fw_cdev_deallocate - Free an address range allocation - * @handle: Handle to the address range, as returned by the kernel when the - * range was allocated + * struct fw_cdev_deallocate - Free a CSR address range or isochronous resource + * @handle: Handle to the address range or iso resource, as returned by the + * kernel when the range or resource was allocated */ struct fw_cdev_deallocate { __u32 handle; @@ -479,4 +522,35 @@ struct fw_cdev_get_cycle_timer { __u32 cycle_timer; }; +/** + * struct fw_cdev_allocate_iso_resource - Allocate a channel or bandwidth + * @closure: Passed back to userspace in correponding iso resource events + * @channels: Isochronous channels of which one is to be allocated + * @bandwidth: Isochronous bandwidth units to be allocated + * @handle: Handle to the allocation, written by the kernel + * + * The %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE ioctl initiates allocation of an + * isochronous channel and/or of isochronous bandwidth at the isochronous + * resource manager (IRM). Only one of the channels specified in @channels is + * allocated. An %FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED is sent after + * communication with the IRM, indicating success or failure in the event data. + * The kernel will automatically reallocate the resources after bus resets. + * Should a reallocation fail, an %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED event + * will be sent. The kernel will also automatically deallocate the resources + * when the file descriptor is closed. + * + * @channels is a host-endian bitfield with the most significant bit + * representing channel 0 and the least significant bit representing channel 63: + * 1ULL << (63 - c) + * + * @bandwidth is expressed in bandwidth allocation units, i.e. the time to send + * one quadlet of data (payload or header data) at speed S1600. + */ +struct fw_cdev_allocate_iso_resource { + __u64 closure; + __u64 channels; + __u32 bandwidth; + __u32 handle; +}; + #endif /* _LINUX_FIREWIRE_CDEV_H */ -- cgit v1.2.3-59-g8ed1b From 1ec3c0269d7196118cc7c403654ca5f19ef4d584 Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Sun, 4 Jan 2009 16:23:29 +0100 Subject: firewire: cdev: add ioctls for manual iso resource management This adds ioctls for allocation and deallocation of a channel or/and bandwidth without auto-reallocation and without auto-deallocation. The benefit of these ioctls is that libraw1394-style isochronous resource management can be implemented without write access to the IRM's character device file. Signed-off-by: Stefan Richter --- drivers/firewire/fw-cdev.c | 67 ++++++++++++++++++++++++++++++++++--------- include/linux/firewire-cdev.h | 42 ++++++++++++++++++++------- 2 files changed, 86 insertions(+), 23 deletions(-) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index a227853aa1e2..08fe68d34f32 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -121,14 +121,15 @@ struct iso_resource { struct client *client; /* Schedule work and access todo only with client->lock held. */ struct delayed_work work; - enum {ISO_RES_ALLOC, ISO_RES_REALLOC, ISO_RES_DEALLOC,} todo; + enum {ISO_RES_ALLOC, ISO_RES_REALLOC, ISO_RES_DEALLOC, + ISO_RES_ALLOC_ONCE, ISO_RES_DEALLOC_ONCE,} todo; int generation; u64 channels; s32 bandwidth; struct iso_resource_event *e_alloc, *e_dealloc; }; -static void schedule_iso_resource(struct iso_resource *); +static int schedule_iso_resource(struct iso_resource *); static void release_iso_resource(struct client *, struct client_resource *); /* @@ -1033,7 +1034,9 @@ static void iso_resource_work(struct work_struct *work) skip = todo == ISO_RES_REALLOC && r->generation == generation; } - free = todo == ISO_RES_DEALLOC; + free = todo == ISO_RES_DEALLOC || + todo == ISO_RES_ALLOC_ONCE || + todo == ISO_RES_DEALLOC_ONCE; r->generation = generation; spin_unlock_irq(&client->lock); @@ -1044,7 +1047,9 @@ static void iso_resource_work(struct work_struct *work) fw_iso_resource_manage(client->device->card, generation, r->channels, &channel, &bandwidth, - todo == ISO_RES_ALLOC || todo == ISO_RES_REALLOC); + todo == ISO_RES_ALLOC || + todo == ISO_RES_REALLOC || + todo == ISO_RES_ALLOC_ONCE); /* * Is this generation outdated already? As long as this resource sticks * in the idr, it will be scheduled again for a newer generation or at @@ -1082,7 +1087,7 @@ static void iso_resource_work(struct work_struct *work) if (todo == ISO_RES_REALLOC && success) goto out; - if (todo == ISO_RES_ALLOC) { + if (todo == ISO_RES_ALLOC || todo == ISO_RES_ALLOC_ONCE) { e = r->e_alloc; r->e_alloc = NULL; } else { @@ -1106,10 +1111,17 @@ static void iso_resource_work(struct work_struct *work) client_put(client); } -static void schedule_iso_resource(struct iso_resource *r) +static int schedule_iso_resource(struct iso_resource *r) { - if (schedule_delayed_work(&r->work, 0)) - client_get(r->client); + int scheduled; + + client_get(r->client); + + scheduled = schedule_delayed_work(&r->work, 0); + if (!scheduled) + client_put(r->client); + + return scheduled; } static void release_iso_resource(struct client *client, @@ -1124,9 +1136,9 @@ static void release_iso_resource(struct client *client, spin_unlock_irq(&client->lock); } -static int ioctl_allocate_iso_resource(struct client *client, void *buffer) +static int init_iso_resource(struct client *client, + struct fw_cdev_allocate_iso_resource *request, int todo) { - struct fw_cdev_allocate_iso_resource *request = buffer; struct iso_resource_event *e1, *e2; struct iso_resource *r; int ret; @@ -1146,7 +1158,7 @@ static int ioctl_allocate_iso_resource(struct client *client, void *buffer) INIT_DELAYED_WORK(&r->work, iso_resource_work); r->client = client; - r->todo = ISO_RES_ALLOC; + r->todo = todo; r->generation = -1; r->channels = request->channels; r->bandwidth = request->bandwidth; @@ -1158,8 +1170,14 @@ static int ioctl_allocate_iso_resource(struct client *client, void *buffer) e2->resource.closure = request->closure; e2->resource.type = FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED; - r->resource.release = release_iso_resource; - ret = add_client_resource(client, &r->resource, GFP_KERNEL); + if (todo == ISO_RES_ALLOC) { + r->resource.release = release_iso_resource; + ret = add_client_resource(client, &r->resource, GFP_KERNEL); + } else { + r->resource.release = NULL; + r->resource.handle = -1; + ret = schedule_iso_resource(r) ? 0 : -ENOMEM; + } if (ret < 0) goto fail; request->handle = r->resource.handle; @@ -1173,6 +1191,13 @@ static int ioctl_allocate_iso_resource(struct client *client, void *buffer) return ret; } +static int ioctl_allocate_iso_resource(struct client *client, void *buffer) +{ + struct fw_cdev_allocate_iso_resource *request = buffer; + + return init_iso_resource(client, request, ISO_RES_ALLOC); +} + static int ioctl_deallocate_iso_resource(struct client *client, void *buffer) { struct fw_cdev_deallocate *request = buffer; @@ -1181,6 +1206,20 @@ static int ioctl_deallocate_iso_resource(struct client *client, void *buffer) release_iso_resource, NULL); } +static int ioctl_allocate_iso_resource_once(struct client *client, void *buffer) +{ + struct fw_cdev_allocate_iso_resource *request = buffer; + + return init_iso_resource(client, request, ISO_RES_ALLOC_ONCE); +} + +static int ioctl_deallocate_iso_resource_once(struct client *client, void *buffer) +{ + struct fw_cdev_allocate_iso_resource *request = buffer; + + return init_iso_resource(client, request, ISO_RES_DEALLOC_ONCE); +} + static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_get_info, ioctl_send_request, @@ -1197,6 +1236,8 @@ static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_get_cycle_timer, ioctl_allocate_iso_resource, ioctl_deallocate_iso_resource, + ioctl_allocate_iso_resource_once, + ioctl_deallocate_iso_resource_once, }; static int dispatch_ioctl(struct client *client, diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 25b96dd0574f..08ca838a727b 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -151,7 +151,7 @@ struct fw_cdev_event_iso_interrupt { /** * struct fw_cdev_event_iso_resource - Iso resources were allocated or freed * @closure: See &fw_cdev_event_common; - * set by %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE ioctl + * set by %FW_CDEV_IOC_(DE)ALLOCATE_ISO_RESOURCE(_ONCE) ioctl * @type: %FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED or * %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED * @handle: Reference by which an allocated resource can be deallocated @@ -164,12 +164,12 @@ struct fw_cdev_event_iso_interrupt { * resource was allocated at the IRM. The client has to check @channel and * @bandwidth for whether the allocation actually succeeded. * - * @channel is <0 if no channel was allocated. - * @bandwidth is 0 if no bandwidth was allocated. - * * An %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED event is sent after an isochronous * resource was deallocated at the IRM. It is also sent when automatic * reallocation after a bus reset failed. + * + * @channel is <0 if no channel was (de)allocated or if reallocation failed. + * @bandwidth is 0 if no bandwidth was (de)allocated or if reallocation failed. */ struct fw_cdev_event_iso_resource { __u64 closure; @@ -225,8 +225,10 @@ union fw_cdev_event { #define FW_CDEV_IOC_GET_CYCLE_TIMER _IOR('#', 0x0c, struct fw_cdev_get_cycle_timer) /* available since kernel version 2.6.30 */ -#define FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE _IOWR('#', 0x0d, struct fw_cdev_allocate_iso_resource) -#define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE _IOW('#', 0x0e, struct fw_cdev_deallocate) +#define FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE _IOWR('#', 0x0d, struct fw_cdev_allocate_iso_resource) +#define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE _IOW('#', 0x0e, struct fw_cdev_deallocate) +#define FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x0f, struct fw_cdev_allocate_iso_resource) +#define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x10, struct fw_cdev_allocate_iso_resource) /* FW_CDEV_VERSION History * @@ -523,11 +525,12 @@ struct fw_cdev_get_cycle_timer { }; /** - * struct fw_cdev_allocate_iso_resource - Allocate a channel or bandwidth + * struct fw_cdev_allocate_iso_resource - (De)allocate a channel or bandwidth * @closure: Passed back to userspace in correponding iso resource events - * @channels: Isochronous channels of which one is to be allocated - * @bandwidth: Isochronous bandwidth units to be allocated - * @handle: Handle to the allocation, written by the kernel + * @channels: Isochronous channels of which one is to be (de)allocated + * @bandwidth: Isochronous bandwidth units to be (de)allocated + * @handle: Handle to the allocation, written by the kernel (only valid in + * case of %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE ioctls) * * The %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE ioctl initiates allocation of an * isochronous channel and/or of isochronous bandwidth at the isochronous @@ -539,6 +542,25 @@ struct fw_cdev_get_cycle_timer { * will be sent. The kernel will also automatically deallocate the resources * when the file descriptor is closed. * + * The %FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE ioctl can be used to initiate + * deallocation of resources which were allocated as described above. + * An %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED event concludes this operation. + * + * The %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE ioctl is a variant of allocation + * without automatic re- or deallocation. + * An %FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED event concludes this operation, + * indicating success or failure in its data. + * + * The %FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE_ONCE ioctl works like + * %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE except that resources are freed + * instead of allocated. At most one channel may be specified in this ioctl. + * An %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED event concludes this operation. + * + * To summarize, %FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE allocates iso resources + * for the lifetime of the fd or handle. + * In contrast, %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE allocates iso resources + * for the duration of a bus generation. + * * @channels is a host-endian bitfield with the most significant bit * representing channel 0 and the least significant bit representing channel 63: * 1ULL << (63 - c) -- cgit v1.2.3-59-g8ed1b From 33580a3ef5ba3bc0ee1b520df82a24bb37ce28f0 Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Sun, 4 Jan 2009 16:23:29 +0100 Subject: firewire: cdev: add ioctl to query maximum transmission speed While the speed of asynchronous transactions is automatically chosen by the kernel, the speed of isochronous streams has to be chosen by the initiating client. In case of 1394a bus topologies, the maximum possible speed could be figured out with some effort by evaluation of the remote node's link speed field in the config ROM, the local node's link speed field, and the PHY speeds and topologic information in the local node's or IRM's topology map CSR. However, this does not work in case of 1394b buses. Hence add an ioctl to export the maximum speed which the kernel already determined. Signed-off-by: Stefan Richter --- drivers/firewire/fw-cdev.c | 10 ++++++++++ include/linux/firewire-cdev.h | 10 ++++++++++ 2 files changed, 20 insertions(+) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index 08fe68d34f32..05ad2a8f286c 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -1220,6 +1220,15 @@ static int ioctl_deallocate_iso_resource_once(struct client *client, void *buffe return init_iso_resource(client, request, ISO_RES_DEALLOC_ONCE); } +static int ioctl_get_speed(struct client *client, void *buffer) +{ + struct fw_cdev_get_speed *request = buffer; + + request->max_speed = client->device->max_speed; + + return 0; +} + static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_get_info, ioctl_send_request, @@ -1238,6 +1247,7 @@ static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_deallocate_iso_resource, ioctl_allocate_iso_resource_once, ioctl_deallocate_iso_resource_once, + ioctl_get_speed, }; static int dispatch_ioctl(struct client *client, diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 08ca838a727b..f819c1026958 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -229,6 +229,7 @@ union fw_cdev_event { #define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE _IOW('#', 0x0e, struct fw_cdev_deallocate) #define FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x0f, struct fw_cdev_allocate_iso_resource) #define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x10, struct fw_cdev_allocate_iso_resource) +#define FW_CDEV_IOC_GET_SPEED _IOR('#', 0x11, struct fw_cdev_get_speed) /* FW_CDEV_VERSION History * @@ -575,4 +576,13 @@ struct fw_cdev_allocate_iso_resource { __u32 handle; }; +/** + * struct fw_cdev_get_speed - Query maximum speed to or from this device + * @max_speed: Speed code; minimum of the device's link speed, the local node's + * link speed, and all PHY port speeds between the two links + */ +struct fw_cdev_get_speed { + __u32 max_speed; +}; + #endif /* _LINUX_FIREWIRE_CDEV_H */ -- cgit v1.2.3-59-g8ed1b From acfe8333572cad5dc70fce18ac966be0446548d7 Mon Sep 17 00:00:00 2001 From: "Jay Fenlason, Stefan Richter" Date: Sun, 4 Jan 2009 16:23:29 +0100 Subject: firewire: cdev: add ioctl for broadcast write requests Write transactions to the broadcast node ID are a convenient way to trigger functions of multiple nodes at once. IIDC is a protocol which can make use of this if multiple cameras with same command_regs_base are connected at the same bus. Based on Date: Wed, 10 Sep 2008 11:32:16 -0400 From: Jay Fenlason Subject: [patch] SEND_BROADCAST_REQUEST Changes: ioctl_send_request() and ioctl_send_broadcast_request() now share code. Broadcast speed corrected to S100. Check for proper tcode. Signed-off-by: Stefan Richter --- drivers/firewire/fw-cdev.c | 74 +++++++++++++++++++++++++++---------------- include/linux/firewire-cdev.h | 1 + 2 files changed, 48 insertions(+), 27 deletions(-) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index 05ad2a8f286c..a1637a86da3d 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -518,10 +518,10 @@ static void complete_transaction(struct fw_card *card, int rcode, client_put(client); } -static int ioctl_send_request(struct client *client, void *buffer) +static int init_request(struct client *client, + struct fw_cdev_send_request *request, + int destination_id, int speed) { - struct fw_device *device = client->device; - struct fw_cdev_send_request *request = buffer; struct outbound_transaction_event *e; int ret; @@ -544,24 +544,6 @@ static int ioctl_send_request(struct client *client, void *buffer) goto failed; } - switch (request->tcode) { - case TCODE_WRITE_QUADLET_REQUEST: - case TCODE_WRITE_BLOCK_REQUEST: - case TCODE_READ_QUADLET_REQUEST: - case TCODE_READ_BLOCK_REQUEST: - case TCODE_LOCK_MASK_SWAP: - case TCODE_LOCK_COMPARE_SWAP: - case TCODE_LOCK_FETCH_ADD: - case TCODE_LOCK_LITTLE_ADD: - case TCODE_LOCK_BOUNDED_ADD: - case TCODE_LOCK_WRAP_ADD: - case TCODE_LOCK_VENDOR_DEPENDENT: - break; - default: - ret = -EINVAL; - goto failed; - } - e->r.resource.release = release_transaction; ret = add_client_resource(client, &e->r.resource, GFP_KERNEL); if (ret < 0) @@ -570,12 +552,9 @@ static int ioctl_send_request(struct client *client, void *buffer) /* Get a reference for the transaction callback */ client_get(client); - fw_send_request(device->card, &e->r.transaction, - request->tcode & 0x1f, - device->node->node_id, - request->generation, - device->max_speed, - request->offset, + fw_send_request(client->device->card, &e->r.transaction, + request->tcode & 0x1f, destination_id, + request->generation, speed, request->offset, e->response.data, request->length, complete_transaction, e); @@ -589,6 +568,31 @@ static int ioctl_send_request(struct client *client, void *buffer) return ret; } +static int ioctl_send_request(struct client *client, void *buffer) +{ + struct fw_cdev_send_request *request = buffer; + + switch (request->tcode) { + case TCODE_WRITE_QUADLET_REQUEST: + case TCODE_WRITE_BLOCK_REQUEST: + case TCODE_READ_QUADLET_REQUEST: + case TCODE_READ_BLOCK_REQUEST: + case TCODE_LOCK_MASK_SWAP: + case TCODE_LOCK_COMPARE_SWAP: + case TCODE_LOCK_FETCH_ADD: + case TCODE_LOCK_LITTLE_ADD: + case TCODE_LOCK_BOUNDED_ADD: + case TCODE_LOCK_WRAP_ADD: + case TCODE_LOCK_VENDOR_DEPENDENT: + break; + default: + return -EINVAL; + } + + return init_request(client, request, client->device->node->node_id, + client->device->max_speed); +} + static void release_request(struct client *client, struct client_resource *resource) { @@ -1229,6 +1233,21 @@ static int ioctl_get_speed(struct client *client, void *buffer) return 0; } +static int ioctl_send_broadcast_request(struct client *client, void *buffer) +{ + struct fw_cdev_send_request *request = buffer; + + switch (request->tcode) { + case TCODE_WRITE_QUADLET_REQUEST: + case TCODE_WRITE_BLOCK_REQUEST: + break; + default: + return -EINVAL; + } + + return init_request(client, request, LOCAL_BUS | 0x3f, SCODE_100); +} + static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_get_info, ioctl_send_request, @@ -1248,6 +1267,7 @@ static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_allocate_iso_resource_once, ioctl_deallocate_iso_resource_once, ioctl_get_speed, + ioctl_send_broadcast_request, }; static int dispatch_ioctl(struct client *client, diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index f819c1026958..340a78502bca 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -230,6 +230,7 @@ union fw_cdev_event { #define FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x0f, struct fw_cdev_allocate_iso_resource) #define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x10, struct fw_cdev_allocate_iso_resource) #define FW_CDEV_IOC_GET_SPEED _IOR('#', 0x11, struct fw_cdev_get_speed) +#define FW_CDEV_IOC_SEND_BROADCAST_REQUEST _IOW('#', 0x12, struct fw_cdev_send_request) /* FW_CDEV_VERSION History * -- cgit v1.2.3-59-g8ed1b From 77258da403be4cfce84b6abcdb515ad0bd1f92f1 Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Wed, 7 Jan 2009 20:14:53 +0100 Subject: firewire: cdev: increment fw_cdev_version, update documentation Necessary due to Date: Tue, 22 Jul 2008 23:23:40 -0700 From: David Moore Subject: firewire: Include iso timestamp in headers when header_size > 4 Side note: The lack of upwards compatibility sounds worse than it is. All existing client implementations, libraw1394 and libdc1394, set header_size = 4. And since the ABI v1 behaviour does not offer any advantages over the new behaviour, we deliberately do not provide the old behaviour anymore. Also add documentation about the format of fw_cdev_get_cycle_timer which may be used in conjunction with the timestamp of iso packets but has a different format. Signed-off-by: Stefan Richter --- include/linux/firewire-cdev.h | 35 ++++++++++++++++++++++++++++++----- 1 file changed, 30 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 340a78502bca..6ed9127680fd 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -138,7 +138,24 @@ struct fw_cdev_event_request { * This event is sent when the controller has completed an &fw_cdev_iso_packet * with the %FW_CDEV_ISO_INTERRUPT bit set. In the receive case, the headers * stripped of all packets up until and including the interrupt packet are - * returned in the @header field. + * returned in the @header field. The amount of header data per packet is as + * specified at iso context creation by &fw_cdev_create_iso_context.header_size. + * + * In version 1 of this ABI, header data consisted of the 1394 isochronous + * packet header, followed by quadlets from the packet payload if + * &fw_cdev_create_iso_context.header_size > 4. + * + * In version 2 of this ABI, header data consist of the 1394 isochronous + * packet header, followed by a timestamp quadlet if + * &fw_cdev_create_iso_context.header_size > 4, followed by quadlets from the + * packet payload if &fw_cdev_create_iso_context.header_size > 8. + * + * Behaviour of ver. 1 of this ABI is no longer available since ABI ver. 2. + * + * Format of 1394 iso packet header: 16 bits len, 2 bits tag, 6 bits channel, + * 4 bits tcode, 4 bits sy, in big endian byte order. Format of timestamp: + * 16 bits invalid, 3 bits cycleSeconds, 13 bits cycleCount, in big endian byte + * order. */ struct fw_cdev_event_iso_interrupt { __u64 closure; @@ -232,11 +249,13 @@ union fw_cdev_event { #define FW_CDEV_IOC_GET_SPEED _IOR('#', 0x11, struct fw_cdev_get_speed) #define FW_CDEV_IOC_SEND_BROADCAST_REQUEST _IOW('#', 0x12, struct fw_cdev_send_request) -/* FW_CDEV_VERSION History - * - * 1 Feb 18, 2007: Initial version. +/* + * FW_CDEV_VERSION History + * 1 (2.6.22) - initial version + * 2 (2.6.30) - changed &fw_cdev_event_iso_interrupt.header if + * &fw_cdev_create_iso_context.header_size is 8 or more */ -#define FW_CDEV_VERSION 1 +#define FW_CDEV_VERSION 2 /** * struct fw_cdev_get_info - General purpose information ioctl @@ -417,6 +436,9 @@ struct fw_cdev_remove_descriptor { * * If a context was successfully created, the kernel writes back a handle to the * context, which must be passed in for subsequent operations on that context. + * + * Note that the effect of a @header_size > 4 depends on + * &fw_cdev_get_info.version, as documented at &fw_cdev_event_iso_interrupt. */ struct fw_cdev_create_iso_context { __u32 type; @@ -520,6 +542,9 @@ struct fw_cdev_stop_iso { * The %FW_CDEV_IOC_GET_CYCLE_TIMER ioctl reads the isochronous cycle timer * and also the system clock. This allows to express the receive time of an * isochronous packet as a system time with microsecond accuracy. + * + * @cycle_timer consists of 7 bits cycleSeconds, 13 bits cycleCount, and + * 12 bits cycleOffset, in host byte order. */ struct fw_cdev_get_cycle_timer { __u64 local_time; -- cgit v1.2.3-59-g8ed1b From 5d9cb7d276a9c465fef5a771792eac2cf1929f2b Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Thu, 8 Jan 2009 23:07:40 +0100 Subject: firewire: cdev: add ioctls for iso resource management, amendment Some fixes: - Remove stale documentation. - Fix a != vs. == thinko that got in the way of channel management. - Try bandwidth deallocation even if channel deallocation failed. A simplification: - fw_cdev_allocate_iso_resource.channels is now ordered like libdc1394's dc1394_iso_allocate_channel() channels_allowed argument. By the way, I looked closer at cards from NEC, TI, and VIA, and noticed that they all don't implement IEEE 1394a behaviour which is meant to deviate from IEEE 1212's notion of lock compare-swap. This means that we have to do two lock transactions instead of one in many cases where one transaction would already succeed on a fully 1394a compliant IRM. Signed-off-by: Stefan Richter --- drivers/firewire/fw-cdev.c | 2 +- drivers/firewire/fw-iso.c | 38 ++++++++++++++++++++++---------------- include/linux/firewire-cdev.h | 10 ++++------ 3 files changed, 27 insertions(+), 23 deletions(-) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index b93ad9c0a0d0..257b0c709a8b 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -1082,7 +1082,7 @@ static void iso_resource_work(struct work_struct *work) spin_unlock_irq(&client->lock); if (todo == ISO_RES_ALLOC && channel >= 0) - r->channels = 1ULL << (63 - channel); + r->channels = 1ULL << channel; if (todo == ISO_RES_REALLOC && success) goto out; diff --git a/drivers/firewire/fw-iso.c b/drivers/firewire/fw-iso.c index a7b57b253b06..f511d16efaee 100644 --- a/drivers/firewire/fw-iso.c +++ b/drivers/firewire/fw-iso.c @@ -204,17 +204,19 @@ static int manage_bandwidth(struct fw_card *card, int irm_id, int generation, } static int manage_channel(struct fw_card *card, int irm_id, int generation, - __be32 channels_mask, u64 offset, bool allocate) + u32 channels_mask, u64 offset, bool allocate) { - __be32 data[2], c, old = allocate ? cpu_to_be32(~0) : 0; + __be32 data[2], c, all, old; int i, retry = 5; + old = all = allocate ? cpu_to_be32(~0) : 0; + for (i = 0; i < 32; i++) { - c = cpu_to_be32(1 << (31 - i)); - if (!(channels_mask & c)) + if (!(channels_mask & 1 << i)) continue; - if (allocate == !(old & c)) + c = cpu_to_be32(1 << (31 - i)); + if ((old & c) != (all & c)) continue; data[0] = old; @@ -233,7 +235,7 @@ static int manage_channel(struct fw_card *card, int irm_id, int generation, old = data[0]; /* Is the IRM 1394a-2000 compliant? */ - if ((data[0] & c) != (data[1] & c)) + if ((data[0] & c) == (data[1] & c)) continue; /* 1394-1995 IRM, fall through to retry. */ @@ -249,11 +251,10 @@ static int manage_channel(struct fw_card *card, int irm_id, int generation, static void deallocate_channel(struct fw_card *card, int irm_id, int generation, int channel) { - __be32 mask; + u32 mask; u64 offset; - mask = channel < 32 ? cpu_to_be32(1 << (31 - channel)) : - cpu_to_be32(1 << (63 - channel)); + mask = channel < 32 ? 1 << channel : 1 << (channel - 32); offset = channel < 32 ? CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_HI : CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_LO; @@ -266,7 +267,12 @@ static void deallocate_channel(struct fw_card *card, int irm_id, * In parameters: card, generation, channels_mask, bandwidth, allocate * Out parameters: channel, bandwidth * This function blocks (sleeps) during communication with the IRM. + * * Allocates or deallocates at most one channel out of channels_mask. + * channels_mask is a bitfield with MSB for channel 63 and LSB for channel 0. + * (Note, the IRM's CHANNELS_AVAILABLE is a big-endian bitfield with MSB for + * channel 0 and LSB for channel 63.) + * Allocates or deallocates as many bandwidth allocation units as specified. * * Returns channel < 0 if no channel was allocated or deallocated. * Returns bandwidth = 0 if no bandwidth was allocated or deallocated. @@ -274,17 +280,17 @@ static void deallocate_channel(struct fw_card *card, int irm_id, * If generation is stale, deallocations succeed but allocations fail with * channel = -EAGAIN. * - * If channel (de)allocation fails, bandwidth (de)allocation fails too. + * If channel allocation fails, no bandwidth will be allocated either. * If bandwidth allocation fails, no channel will be allocated either. - * If bandwidth deallocation fails, channel deallocation may still have been - * successful. + * But deallocations of channel and bandwidth are tried independently + * of each other's success. */ void fw_iso_resource_manage(struct fw_card *card, int generation, u64 channels_mask, int *channel, int *bandwidth, bool allocate) { - __be32 channels_hi = cpu_to_be32(channels_mask >> 32); - __be32 channels_lo = cpu_to_be32(channels_mask); + u32 channels_hi = channels_mask; /* channels 31...0 */ + u32 channels_lo = channels_mask >> 32; /* channels 63...32 */ int irm_id, ret, c = -EINVAL; spin_lock_irq(&card->lock); @@ -302,7 +308,7 @@ void fw_iso_resource_manage(struct fw_card *card, int generation, } *channel = c; - if (channels_mask != 0 && c < 0) + if (allocate && channels_mask != 0 && c < 0) *bandwidth = 0; if (*bandwidth == 0) @@ -312,7 +318,7 @@ void fw_iso_resource_manage(struct fw_card *card, int generation, if (ret < 0) *bandwidth = 0; - if (ret < 0 && c >= 0 && allocate) { + if (allocate && ret < 0 && c >= 0) { deallocate_channel(card, irm_id, generation, c); *channel = ret; } diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 6ed9127680fd..2e35379bf96c 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -174,8 +174,6 @@ struct fw_cdev_event_iso_interrupt { * @handle: Reference by which an allocated resource can be deallocated * @channel: Isochronous channel which was (de)allocated, if any * @bandwidth: Bandwidth allocation units which were (de)allocated, if any - * @channels_available: Last known availability of channels - * @bandwidth_available: Last known availability of bandwidth * * An %FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED event is sent after an isochronous * resource was allocated at the IRM. The client has to check @channel and @@ -580,7 +578,7 @@ struct fw_cdev_get_cycle_timer { * * The %FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE_ONCE ioctl works like * %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE except that resources are freed - * instead of allocated. At most one channel may be specified in this ioctl. + * instead of allocated. * An %FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED event concludes this operation. * * To summarize, %FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE allocates iso resources @@ -588,9 +586,9 @@ struct fw_cdev_get_cycle_timer { * In contrast, %FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE allocates iso resources * for the duration of a bus generation. * - * @channels is a host-endian bitfield with the most significant bit - * representing channel 0 and the least significant bit representing channel 63: - * 1ULL << (63 - c) + * @channels is a host-endian bitfield with the least significant bit + * representing channel 0 and the most significant bit representing channel 63: + * 1ULL << c for each channel c that is a candidate for (de)allocation. * * @bandwidth is expressed in bandwidth allocation units, i.e. the time to send * one quadlet of data (payload or header data) at speed S1600. -- cgit v1.2.3-59-g8ed1b From f8c2287c65f8f72000102fc058232669e4540bc4 Mon Sep 17 00:00:00 2001 From: Jay Fenlason Date: Thu, 5 Mar 2009 19:08:40 +0100 Subject: firewire: implement asynchronous stream transmission Allow userspace and other firewire drivers (fw-ipv4 I'm looking at you!) to send Asynchronous Transmit Streams as described in 7.8.3 of release 1.1 of the 1394 Open Host Controller Interface Specification. Signed-off-by: Jay Fenlason Signed-off-by: Stefan Richter (tweaks) --- drivers/firewire/fw-cdev.c | 33 +++++++++++++++++++++++++++++++++ drivers/firewire/fw-ohci.c | 21 +++++++++++++++++++-- drivers/firewire/fw-transaction.c | 25 +++++++++++++++++++++++++ drivers/firewire/fw-transaction.h | 4 ++++ include/linux/firewire-cdev.h | 27 +++++++++++++++++++++++++++ 5 files changed, 108 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index 214e534efee5..539dae5eb5b2 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -1242,6 +1242,38 @@ static int ioctl_send_broadcast_request(struct client *client, void *buffer) return init_request(client, request, LOCAL_BUS | 0x3f, SCODE_100); } +struct stream_packet { + struct fw_packet packet; + u8 data[0]; +}; + +static void send_stream_packet_done(struct fw_packet *packet, + struct fw_card *card, int status) +{ + kfree(container_of(packet, struct stream_packet, packet)); +} + +static int ioctl_send_stream_packet(struct client *client, void *buffer) +{ + struct fw_cdev_send_stream_packet *request = buffer; + struct stream_packet *p; + + p = kmalloc(sizeof(*p) + request->size, GFP_KERNEL); + if (p == NULL) + return -ENOMEM; + + if (request->data && + copy_from_user(p->data, u64_to_uptr(request->data), request->size)) { + kfree(p); + return -EFAULT; + } + fw_send_stream_packet(client->device->card, &p->packet, + request->generation, request->speed, + request->channel, request->sy, request->tag, + p->data, request->size, send_stream_packet_done); + return 0; +} + static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_get_info, ioctl_send_request, @@ -1262,6 +1294,7 @@ static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { ioctl_deallocate_iso_resource_once, ioctl_get_speed, ioctl_send_broadcast_request, + ioctl_send_stream_packet, }; static int dispatch_ioctl(struct client *client, diff --git a/drivers/firewire/fw-ohci.c b/drivers/firewire/fw-ohci.c index c92278374658..1180d0be0bb4 100644 --- a/drivers/firewire/fw-ohci.c +++ b/drivers/firewire/fw-ohci.c @@ -936,7 +936,9 @@ static int at_context_queue_packet(struct context *ctx, */ header = (__le32 *) &d[1]; - if (packet->header_length > 8) { + switch (packet->header_length) { + case 16: + case 12: header[0] = cpu_to_le32((packet->header[0] & 0xffff) | (packet->speed << 16)); header[1] = cpu_to_le32((packet->header[1] & 0xffff) | @@ -950,12 +952,27 @@ static int at_context_queue_packet(struct context *ctx, header[3] = (__force __le32) packet->header[3]; d[0].req_count = cpu_to_le16(packet->header_length); - } else { + break; + + case 8: header[0] = cpu_to_le32((OHCI1394_phy_tcode << 4) | (packet->speed << 16)); header[1] = cpu_to_le32(packet->header[0]); header[2] = cpu_to_le32(packet->header[1]); d[0].req_count = cpu_to_le16(12); + break; + + case 4: + header[0] = cpu_to_le32((packet->header[0] & 0xffff) | + (packet->speed << 16)); + header[1] = cpu_to_le32(packet->header[0] & 0xffff0000); + d[0].req_count = cpu_to_le16(8); + break; + + default: + /* BUG(); */ + packet->ack = RCODE_SEND_ERROR; + return -1; } driver_data = (struct driver_data *) &d[3]; diff --git a/drivers/firewire/fw-transaction.c b/drivers/firewire/fw-transaction.c index 76938fe432a0..e3da58991960 100644 --- a/drivers/firewire/fw-transaction.c +++ b/drivers/firewire/fw-transaction.c @@ -37,6 +37,10 @@ #include "fw-topology.h" #include "fw-device.h" +#define HEADER_TAG(tag) ((tag) << 14) +#define HEADER_CHANNEL(ch) ((ch) << 8) +#define HEADER_SY(sy) ((sy) << 0) + #define HEADER_PRI(pri) ((pri) << 0) #define HEADER_TCODE(tcode) ((tcode) << 4) #define HEADER_RETRY(retry) ((retry) << 8) @@ -293,6 +297,27 @@ void fw_send_request(struct fw_card *card, struct fw_transaction *t, int tcode, } EXPORT_SYMBOL(fw_send_request); +void fw_send_stream_packet(struct fw_card *card, struct fw_packet *p, + int generation, int speed, int channel, int sy, int tag, + void *payload, size_t length, fw_packet_callback_t callback) +{ + p->callback = callback; + p->header[0] = + HEADER_DATA_LENGTH(length) + | HEADER_TAG(tag) + | HEADER_CHANNEL(channel) + | HEADER_TCODE(TCODE_STREAM_DATA) + | HEADER_SY(sy); + p->header_length = 4; + p->payload = payload; + p->payload_length = length; + p->speed = speed; + p->generation = generation; + p->ack = 0; + + card->driver->send_request(card, p); +} + struct transaction_callback_data { struct completion done; void *payload; diff --git a/drivers/firewire/fw-transaction.h b/drivers/firewire/fw-transaction.h index 35d0a4bb6d5c..eed2e295eb3c 100644 --- a/drivers/firewire/fw-transaction.h +++ b/drivers/firewire/fw-transaction.h @@ -407,6 +407,10 @@ void fw_send_request(struct fw_card *card, struct fw_transaction *t, int tcode, int destination_id, int generation, int speed, unsigned long long offset, void *payload, size_t length, fw_transaction_callback_t callback, void *callback_data); +void fw_send_stream_packet(struct fw_card *card, struct fw_packet *p, + int generation, int speed, int channel, int sy, int tag, + void *payload, size_t length, fw_packet_callback_t callback); + int fw_cancel_transaction(struct fw_card *card, struct fw_transaction *transaction); void fw_flush_transactions(struct fw_card *card); diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 2e35379bf96c..4dfc84d0ac76 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -246,6 +246,7 @@ union fw_cdev_event { #define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x10, struct fw_cdev_allocate_iso_resource) #define FW_CDEV_IOC_GET_SPEED _IOR('#', 0x11, struct fw_cdev_get_speed) #define FW_CDEV_IOC_SEND_BROADCAST_REQUEST _IOW('#', 0x12, struct fw_cdev_send_request) +#define FW_CDEV_IOC_SEND_STREAM_PACKET _IOW('#', 0x13, struct fw_cdev_send_stream_packet) /* * FW_CDEV_VERSION History @@ -609,4 +610,30 @@ struct fw_cdev_get_speed { __u32 max_speed; }; +/** + * struct fw_cdev_send_stream_packet - send an asynchronous stream packet + * @generation: Bus generation where the packet is valid + * @speed: Speed code to send the packet at + * @channel: Channel to send the packet on + * @sy: Four-bit sy code for the packet + * @tag: Two-bit tag field to use for the packet + * @size: Size of the packet's data payload + * @data: Userspace pointer to the payload + * + * The %FW_CDEV_IOC_SEND_STREAM_PACKET ioctl sends an asynchronous stream packet + * to every device (that is listening to the specified channel) on the + * firewire bus. It is the applications's job to ensure + * that the intended device(s) will be able to receive the packet at the chosen + * transmit speed. + */ +struct fw_cdev_send_stream_packet { + __u32 generation; + __u32 speed; + __u32 channel; + __u32 sy; + __u32 tag; + __u32 size; + __u64 data; +}; + #endif /* _LINUX_FIREWIRE_CDEV_H */ -- cgit v1.2.3-59-g8ed1b From c8a25900f35e575938c791507894c036c0f2ca7d Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Tue, 10 Mar 2009 20:59:16 +0100 Subject: firewire: cdev: amendment to "add ioctl to query maximum transmission speed" The as yet unreleased FW_CDEV_IOC_GET_SPEED ioctl puts only a single integer into the parameter buffer. We can use ioctl()'s return value instead. (Also: Some whitespace change in firewire-cdev.h.) Signed-off-by: Stefan Richter --- drivers/firewire/fw-cdev.c | 11 ++++++----- include/linux/firewire-cdev.h | 37 ++++++++++++++----------------------- 2 files changed, 20 insertions(+), 28 deletions(-) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index 539dae5eb5b2..2784f91896db 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -1214,13 +1214,14 @@ static int ioctl_deallocate_iso_resource_once(struct client *client, void *buffe return init_iso_resource(client, request, ISO_RES_DEALLOC_ONCE); } +/* + * Returns a speed code: Maximum speed to or from this device, + * limited by the device's link speed, the local node's link speed, + * and all PHY port speeds between the two links. + */ static int ioctl_get_speed(struct client *client, void *buffer) { - struct fw_cdev_get_speed *request = buffer; - - request->max_speed = client->device->max_speed; - - return 0; + return client->device->max_speed; } static int ioctl_send_broadcast_request(struct client *client, void *buffer) diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 4dfc84d0ac76..de4035792f70 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -223,28 +223,28 @@ union fw_cdev_event { }; /* available since kernel version 2.6.22 */ -#define FW_CDEV_IOC_GET_INFO _IOWR('#', 0x00, struct fw_cdev_get_info) -#define FW_CDEV_IOC_SEND_REQUEST _IOW('#', 0x01, struct fw_cdev_send_request) -#define FW_CDEV_IOC_ALLOCATE _IOWR('#', 0x02, struct fw_cdev_allocate) -#define FW_CDEV_IOC_DEALLOCATE _IOW('#', 0x03, struct fw_cdev_deallocate) -#define FW_CDEV_IOC_SEND_RESPONSE _IOW('#', 0x04, struct fw_cdev_send_response) -#define FW_CDEV_IOC_INITIATE_BUS_RESET _IOW('#', 0x05, struct fw_cdev_initiate_bus_reset) -#define FW_CDEV_IOC_ADD_DESCRIPTOR _IOWR('#', 0x06, struct fw_cdev_add_descriptor) -#define FW_CDEV_IOC_REMOVE_DESCRIPTOR _IOW('#', 0x07, struct fw_cdev_remove_descriptor) -#define FW_CDEV_IOC_CREATE_ISO_CONTEXT _IOWR('#', 0x08, struct fw_cdev_create_iso_context) -#define FW_CDEV_IOC_QUEUE_ISO _IOWR('#', 0x09, struct fw_cdev_queue_iso) -#define FW_CDEV_IOC_START_ISO _IOW('#', 0x0a, struct fw_cdev_start_iso) -#define FW_CDEV_IOC_STOP_ISO _IOW('#', 0x0b, struct fw_cdev_stop_iso) +#define FW_CDEV_IOC_GET_INFO _IOWR('#', 0x00, struct fw_cdev_get_info) +#define FW_CDEV_IOC_SEND_REQUEST _IOW('#', 0x01, struct fw_cdev_send_request) +#define FW_CDEV_IOC_ALLOCATE _IOWR('#', 0x02, struct fw_cdev_allocate) +#define FW_CDEV_IOC_DEALLOCATE _IOW('#', 0x03, struct fw_cdev_deallocate) +#define FW_CDEV_IOC_SEND_RESPONSE _IOW('#', 0x04, struct fw_cdev_send_response) +#define FW_CDEV_IOC_INITIATE_BUS_RESET _IOW('#', 0x05, struct fw_cdev_initiate_bus_reset) +#define FW_CDEV_IOC_ADD_DESCRIPTOR _IOWR('#', 0x06, struct fw_cdev_add_descriptor) +#define FW_CDEV_IOC_REMOVE_DESCRIPTOR _IOW('#', 0x07, struct fw_cdev_remove_descriptor) +#define FW_CDEV_IOC_CREATE_ISO_CONTEXT _IOWR('#', 0x08, struct fw_cdev_create_iso_context) +#define FW_CDEV_IOC_QUEUE_ISO _IOWR('#', 0x09, struct fw_cdev_queue_iso) +#define FW_CDEV_IOC_START_ISO _IOW('#', 0x0a, struct fw_cdev_start_iso) +#define FW_CDEV_IOC_STOP_ISO _IOW('#', 0x0b, struct fw_cdev_stop_iso) /* available since kernel version 2.6.24 */ -#define FW_CDEV_IOC_GET_CYCLE_TIMER _IOR('#', 0x0c, struct fw_cdev_get_cycle_timer) +#define FW_CDEV_IOC_GET_CYCLE_TIMER _IOR('#', 0x0c, struct fw_cdev_get_cycle_timer) /* available since kernel version 2.6.30 */ #define FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE _IOWR('#', 0x0d, struct fw_cdev_allocate_iso_resource) #define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE _IOW('#', 0x0e, struct fw_cdev_deallocate) #define FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x0f, struct fw_cdev_allocate_iso_resource) #define FW_CDEV_IOC_DEALLOCATE_ISO_RESOURCE_ONCE _IOW('#', 0x10, struct fw_cdev_allocate_iso_resource) -#define FW_CDEV_IOC_GET_SPEED _IOR('#', 0x11, struct fw_cdev_get_speed) +#define FW_CDEV_IOC_GET_SPEED _IO('#', 0x11) /* returns speed code */ #define FW_CDEV_IOC_SEND_BROADCAST_REQUEST _IOW('#', 0x12, struct fw_cdev_send_request) #define FW_CDEV_IOC_SEND_STREAM_PACKET _IOW('#', 0x13, struct fw_cdev_send_stream_packet) @@ -601,15 +601,6 @@ struct fw_cdev_allocate_iso_resource { __u32 handle; }; -/** - * struct fw_cdev_get_speed - Query maximum speed to or from this device - * @max_speed: Speed code; minimum of the device's link speed, the local node's - * link speed, and all PHY port speeds between the two links - */ -struct fw_cdev_get_speed { - __u32 max_speed; -}; - /** * struct fw_cdev_send_stream_packet - send an asynchronous stream packet * @generation: Bus generation where the packet is valid -- cgit v1.2.3-59-g8ed1b From de487da8ca5839d057e1f4b57ee3f387e180b800 Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Tue, 10 Mar 2009 21:00:23 +0100 Subject: firewire: cdev: secure add_descriptor ioctl The access permissions and ownership or ACL of /dev/fw* character device files will typically be set based on the device type of the respective nodes, as obtained by firewire-core from descriptors in the device's configuration ROM. An example policy is to deny write permission by default but grant write permission to files of AV/C video and audio devices and IIDC video devices. The FW_CDEV_IOC_ADD_DESCRIPTOR ioctl could be used to partly subvert such a policy: Find a device file with relaxed permissions, use the ioctl to add a descriptor with AV/C marker to the local node's ROM, thus gain access to the local node's character device file. (This is only possible if there are udev scripts installed which actively relax permissions for known device types and if there is a device of such a type connected.) Accessibility of the local node's device file is relevant to host security if the host contains two or more IEEE 1394 link layer controllers which are plugged into a single bus. Therefore change the ABI to deny FW_CDEV_IOC_ADD_DESCRIPTOR if the file belongs to a remote node. (This change has no impact on known implementers of the ABI: None of them uses the ioctl yet.) Also clarify the documentation: The ioctl affects all local nodes, not just one local node. Cc: stable@kernel.org Signed-off-by: Stefan Richter --- drivers/firewire/fw-cdev.c | 8 ++++++++ include/linux/firewire-cdev.h | 5 ++++- 2 files changed, 12 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index 2784f91896db..160cb27e120c 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -742,9 +742,17 @@ static void release_descriptor(struct client *client, static int ioctl_add_descriptor(struct client *client, void *buffer) { struct fw_cdev_add_descriptor *request = buffer; + struct fw_card *card = client->device->card; struct descriptor_resource *r; int ret; + /* Access policy: Allow this ioctl only on local nodes' device files. */ + spin_lock_irq(&card->lock); + ret = client->device->node_id != card->local_node->node_id; + spin_unlock_irq(&card->lock); + if (ret) + return -ENOSYS; + if (request->length > 256) return -EINVAL; diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index de4035792f70..25bc82726ef7 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -394,6 +394,9 @@ struct fw_cdev_initiate_bus_reset { * If successful, the kernel adds the descriptor and writes back a handle to the * kernel-side object to be used for later removal of the descriptor block and * immediate key. + * + * This ioctl affects the configuration ROMs of all local nodes. + * The ioctl only succeeds on device files which represent a local node. */ struct fw_cdev_add_descriptor { __u32 immediate; @@ -409,7 +412,7 @@ struct fw_cdev_add_descriptor { * descriptor was added * * Remove a descriptor block and accompanying immediate key from the local - * node's configuration ROM. + * nodes' configuration ROMs. */ struct fw_cdev_remove_descriptor { __u32 handle; -- cgit v1.2.3-59-g8ed1b From 18e9b10fcdc090d3a38606958167d5923c7099b7 Mon Sep 17 00:00:00 2001 From: Stefan Richter Date: Tue, 10 Mar 2009 21:02:21 +0100 Subject: firewire: cdev: add closure to async stream ioctl This changes the as yet unreleased FW_CDEV_IOC_SEND_STREAM_PACKET ioctl to generate an fw_cdev_event_response event just like the other two ioctls for asynchronous request transmission do. This way, clients get feedback on successful or unsuccessful transmission. This also adds input validation for length, tag, channel, sy, speed. Signed-off-by: Stefan Richter --- drivers/firewire/fw-cdev.c | 46 ++++++++++++++++----------------------- drivers/firewire/fw-transaction.c | 42 ++++++++++++++--------------------- drivers/firewire/fw-transaction.h | 9 ++++---- include/linux/firewire-cdev.h | 31 +++++++++++++------------- 4 files changed, 56 insertions(+), 72 deletions(-) (limited to 'include/linux') diff --git a/drivers/firewire/fw-cdev.c b/drivers/firewire/fw-cdev.c index 95a207545eb3..7eb6594cc3e5 100644 --- a/drivers/firewire/fw-cdev.c +++ b/drivers/firewire/fw-cdev.c @@ -522,7 +522,8 @@ static int init_request(struct client *client, struct outbound_transaction_event *e; int ret; - if (request->length > 4096 || request->length > 512 << speed) + if (request->tcode != TCODE_STREAM_DATA && + (request->length > 4096 || request->length > 512 << speed)) return -EIO; e = kmalloc(sizeof(*e) + request->length, GFP_KERNEL); @@ -1247,36 +1248,27 @@ static int ioctl_send_broadcast_request(struct client *client, void *buffer) return init_request(client, request, LOCAL_BUS | 0x3f, SCODE_100); } -struct stream_packet { - struct fw_packet packet; - u8 data[0]; -}; - -static void send_stream_packet_done(struct fw_packet *packet, - struct fw_card *card, int status) -{ - kfree(container_of(packet, struct stream_packet, packet)); -} - static int ioctl_send_stream_packet(struct client *client, void *buffer) { - struct fw_cdev_send_stream_packet *request = buffer; - struct stream_packet *p; + struct fw_cdev_send_stream_packet *p = buffer; + struct fw_cdev_send_request request; + int dest; - p = kmalloc(sizeof(*p) + request->size, GFP_KERNEL); - if (p == NULL) - return -ENOMEM; + if (p->speed > client->device->card->link_speed || + p->length > 1024 << p->speed) + return -EIO; - if (request->data && - copy_from_user(p->data, u64_to_uptr(request->data), request->size)) { - kfree(p); - return -EFAULT; - } - fw_send_stream_packet(client->device->card, &p->packet, - request->generation, request->speed, - request->channel, request->sy, request->tag, - p->data, request->size, send_stream_packet_done); - return 0; + if (p->tag > 3 || p->channel > 63 || p->sy > 15) + return -EINVAL; + + dest = fw_stream_packet_destination_id(p->tag, p->channel, p->sy); + request.tcode = TCODE_STREAM_DATA; + request.length = p->length; + request.closure = p->closure; + request.data = p->data; + request.generation = p->generation; + + return init_request(client, &request, dest, p->speed); } static int (* const ioctl_handlers[])(struct client *client, void *buffer) = { diff --git a/drivers/firewire/fw-transaction.c b/drivers/firewire/fw-transaction.c index e3da58991960..4a9b37461c26 100644 --- a/drivers/firewire/fw-transaction.c +++ b/drivers/firewire/fw-transaction.c @@ -37,10 +37,6 @@ #include "fw-topology.h" #include "fw-device.h" -#define HEADER_TAG(tag) ((tag) << 14) -#define HEADER_CHANNEL(ch) ((ch) << 8) -#define HEADER_SY(sy) ((sy) << 0) - #define HEADER_PRI(pri) ((pri) << 0) #define HEADER_TCODE(tcode) ((tcode) << 4) #define HEADER_RETRY(retry) ((retry) << 8) @@ -158,6 +154,18 @@ static void fw_fill_request(struct fw_packet *packet, int tcode, int tlabel, { int ext_tcode; + if (tcode == TCODE_STREAM_DATA) { + packet->header[0] = + HEADER_DATA_LENGTH(length) | + destination_id | + HEADER_TCODE(TCODE_STREAM_DATA); + packet->header_length = 4; + packet->payload = payload; + packet->payload_length = length; + + goto common; + } + if (tcode > 0x10) { ext_tcode = tcode & ~0x10; tcode = TCODE_LOCK_REQUEST; @@ -204,7 +212,7 @@ static void fw_fill_request(struct fw_packet *packet, int tcode, int tlabel, packet->payload_length = 0; break; } - + common: packet->speed = speed; packet->generation = generation; packet->ack = 0; @@ -246,6 +254,9 @@ static void fw_fill_request(struct fw_packet *packet, int tcode, int tlabel, * @param callback function to be called when the transaction is completed * @param callback_data pointer to arbitrary data, which will be * passed to the callback + * + * In case of asynchronous stream packets i.e. TCODE_STREAM_DATA, the caller + * needs to synthesize @destination_id with fw_stream_packet_destination_id(). */ void fw_send_request(struct fw_card *card, struct fw_transaction *t, int tcode, int destination_id, int generation, int speed, @@ -297,27 +308,6 @@ void fw_send_request(struct fw_card *card, struct fw_transaction *t, int tcode, } EXPORT_SYMBOL(fw_send_request); -void fw_send_stream_packet(struct fw_card *card, struct fw_packet *p, - int generation, int speed, int channel, int sy, int tag, - void *payload, size_t length, fw_packet_callback_t callback) -{ - p->callback = callback; - p->header[0] = - HEADER_DATA_LENGTH(length) - | HEADER_TAG(tag) - | HEADER_CHANNEL(channel) - | HEADER_TCODE(TCODE_STREAM_DATA) - | HEADER_SY(sy); - p->header_length = 4; - p->payload = payload; - p->payload_length = length; - p->speed = speed; - p->generation = generation; - p->ack = 0; - - card->driver->send_request(card, p); -} - struct transaction_callback_data { struct completion done; void *payload; diff --git a/drivers/firewire/fw-transaction.h b/drivers/firewire/fw-transaction.h index f90f09c05833..d4f42cecbdfa 100644 --- a/drivers/firewire/fw-transaction.h +++ b/drivers/firewire/fw-transaction.h @@ -412,10 +412,6 @@ void fw_send_request(struct fw_card *card, struct fw_transaction *t, int tcode, int destination_id, int generation, int speed, unsigned long long offset, void *payload, size_t length, fw_transaction_callback_t callback, void *callback_data); -void fw_send_stream_packet(struct fw_card *card, struct fw_packet *p, - int generation, int speed, int channel, int sy, int tag, - void *payload, size_t length, fw_packet_callback_t callback); - int fw_cancel_transaction(struct fw_card *card, struct fw_transaction *transaction); void fw_flush_transactions(struct fw_card *card); @@ -425,6 +421,11 @@ int fw_run_transaction(struct fw_card *card, int tcode, int destination_id, void fw_send_phy_config(struct fw_card *card, int node_id, int generation, int gap_count); +static inline int fw_stream_packet_destination_id(int tag, int channel, int sy) +{ + return tag << 14 | channel << 8 | sy; +} + /* * Called by the topology code to inform the device code of node * activity; found, lost, or updated nodes. diff --git a/include/linux/firewire-cdev.h b/include/linux/firewire-cdev.h index 25bc82726ef7..c6b3ca3af6df 100644 --- a/include/linux/firewire-cdev.h +++ b/include/linux/firewire-cdev.h @@ -606,28 +606,29 @@ struct fw_cdev_allocate_iso_resource { /** * struct fw_cdev_send_stream_packet - send an asynchronous stream packet - * @generation: Bus generation where the packet is valid - * @speed: Speed code to send the packet at - * @channel: Channel to send the packet on - * @sy: Four-bit sy code for the packet - * @tag: Two-bit tag field to use for the packet - * @size: Size of the packet's data payload - * @data: Userspace pointer to the payload + * @length: Length of outgoing payload, in bytes + * @tag: Data format tag + * @channel: Isochronous channel to transmit to + * @sy: Synchronization code + * @closure: Passed back to userspace in the response event + * @data: Userspace pointer to payload + * @generation: The bus generation where packet is valid + * @speed: Speed to transmit at * * The %FW_CDEV_IOC_SEND_STREAM_PACKET ioctl sends an asynchronous stream packet - * to every device (that is listening to the specified channel) on the - * firewire bus. It is the applications's job to ensure - * that the intended device(s) will be able to receive the packet at the chosen - * transmit speed. + * to every device which is listening to the specified channel. The kernel + * writes an &fw_cdev_event_response event which indicates success or failure of + * the transmission. */ struct fw_cdev_send_stream_packet { - __u32 generation; - __u32 speed; + __u32 length; + __u32 tag; __u32 channel; __u32 sy; - __u32 tag; - __u32 size; + __u64 closure; __u64 data; + __u32 generation; + __u32 speed; }; #endif /* _LINUX_FIREWIRE_CDEV_H */ -- cgit v1.2.3-59-g8ed1b From f18df228997fb716990590d248663981a15f17d4 Mon Sep 17 00:00:00 2001 From: Mingming Cao Date: Tue, 13 Jan 2009 16:43:09 +0100 Subject: quota: Add quota reservation support Delayed allocation defers the block allocation at the dirty pages flush-out time, doing quota charge/check at that time is too late. But we can't charge the quota blocks until blocks are really allocated, otherwise users could get overcharged after reboot from system crash. This patch adds quota reservation for delayed allocation. Quota blocks are reserved in memory, inode and quota won't gets dirtied until later block allocation time. Signed-off-by: Mingming Cao Signed-off-by: Jan Kara --- fs/dquot.c | 117 ++++++++++++++++++++++++++++++++++------------- include/linux/quota.h | 3 ++ include/linux/quotaops.h | 21 +++++++++ 3 files changed, 110 insertions(+), 31 deletions(-) (limited to 'include/linux') diff --git a/fs/dquot.c b/fs/dquot.c index bca3cac4bee7..9b1c4d3c9d83 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -899,6 +899,11 @@ static inline void dquot_incr_space(struct dquot *dquot, qsize_t number) dquot->dq_dqb.dqb_curspace += number; } +static inline void dquot_resv_space(struct dquot *dquot, qsize_t number) +{ + dquot->dq_dqb.dqb_rsvspace += number; +} + static inline void dquot_decr_inodes(struct dquot *dquot, qsize_t number) { if (sb_dqopt(dquot->dq_sb)->flags & DQUOT_NEGATIVE_USAGE || @@ -1068,7 +1073,11 @@ err_out: kfree_skb(skb); } #endif - +/* + * Write warnings to the console and send warning messages over netlink. + * + * Note that this function can sleep. + */ static inline void flush_warnings(struct dquot * const *dquots, char *warntype) { int i; @@ -1129,13 +1138,18 @@ static int check_idq(struct dquot *dquot, qsize_t inodes, char *warntype) /* needs dq_data_lock */ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *warntype) { + qsize_t tspace; + *warntype = QUOTA_NL_NOWARN; if (!sb_has_quota_limits_enabled(dquot->dq_sb, dquot->dq_type) || test_bit(DQ_FAKE_B, &dquot->dq_flags)) return QUOTA_OK; + tspace = dquot->dq_dqb.dqb_curspace + dquot->dq_dqb.dqb_rsvspace + + space; + if (dquot->dq_dqb.dqb_bhardlimit && - dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bhardlimit && + tspace > dquot->dq_dqb.dqb_bhardlimit && !ignore_hardlimit(dquot)) { if (!prealloc) *warntype = QUOTA_NL_BHARDWARN; @@ -1143,7 +1157,7 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war } if (dquot->dq_dqb.dqb_bsoftlimit && - dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bsoftlimit && + tspace > dquot->dq_dqb.dqb_bsoftlimit && dquot->dq_dqb.dqb_btime && get_seconds() >= dquot->dq_dqb.dqb_btime && !ignore_hardlimit(dquot)) { if (!prealloc) @@ -1152,7 +1166,7 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war } if (dquot->dq_dqb.dqb_bsoftlimit && - dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bsoftlimit && + tspace > dquot->dq_dqb.dqb_bsoftlimit && dquot->dq_dqb.dqb_btime == 0) { if (!prealloc) { *warntype = QUOTA_NL_BSOFTWARN; @@ -1306,51 +1320,92 @@ void vfs_dq_drop(struct inode *inode) /* * This operation can block, but only after everything is updated */ -int dquot_alloc_space(struct inode *inode, qsize_t number, int warn) +int __dquot_alloc_space(struct inode *inode, qsize_t number, + int warn, int reserve) { - int cnt, ret = NO_QUOTA; + int cnt, ret = QUOTA_OK; char warntype[MAXQUOTAS]; - /* First test before acquiring mutex - solves deadlocks when we - * re-enter the quota code and are already holding the mutex */ - if (IS_NOQUOTA(inode)) { -out_add: - inode_add_bytes(inode, number); - return QUOTA_OK; - } for (cnt = 0; cnt < MAXQUOTAS; cnt++) warntype[cnt] = QUOTA_NL_NOWARN; - down_read(&sb_dqopt(inode->i_sb)->dqptr_sem); - if (IS_NOQUOTA(inode)) { /* Now we can do reliable test... */ - up_read(&sb_dqopt(inode->i_sb)->dqptr_sem); - goto out_add; - } spin_lock(&dq_data_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (inode->i_dquot[cnt] == NODQUOT) continue; - if (check_bdq(inode->i_dquot[cnt], number, warn, warntype+cnt) == NO_QUOTA) - goto warn_put_all; + if (check_bdq(inode->i_dquot[cnt], number, warn, warntype+cnt) + == NO_QUOTA) { + ret = NO_QUOTA; + goto out_unlock; + } } for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (inode->i_dquot[cnt] == NODQUOT) continue; - dquot_incr_space(inode->i_dquot[cnt], number); + if (reserve) + dquot_resv_space(inode->i_dquot[cnt], number); + else + dquot_incr_space(inode->i_dquot[cnt], number); } - inode_add_bytes(inode, number); - ret = QUOTA_OK; -warn_put_all: + if (!reserve) + inode_add_bytes(inode, number); +out_unlock: spin_unlock(&dq_data_lock); - if (ret == QUOTA_OK) - /* Dirtify all the dquots - this can block when journalling */ - for (cnt = 0; cnt < MAXQUOTAS; cnt++) - if (inode->i_dquot[cnt]) - mark_dquot_dirty(inode->i_dquot[cnt]); flush_warnings(inode->i_dquot, warntype); + return ret; +} + +int dquot_alloc_space(struct inode *inode, qsize_t number, int warn) +{ + int cnt, ret = QUOTA_OK; + + /* + * First test before acquiring mutex - solves deadlocks when we + * re-enter the quota code and are already holding the mutex + */ + if (IS_NOQUOTA(inode)) { + inode_add_bytes(inode, number); + goto out; + } + + down_read(&sb_dqopt(inode->i_sb)->dqptr_sem); + if (IS_NOQUOTA(inode)) { + inode_add_bytes(inode, number); + goto out_unlock; + } + + ret = __dquot_alloc_space(inode, number, warn, 0); + if (ret == NO_QUOTA) + goto out_unlock; + + /* Dirtify all the dquots - this can block when journalling */ + for (cnt = 0; cnt < MAXQUOTAS; cnt++) + if (inode->i_dquot[cnt]) + mark_dquot_dirty(inode->i_dquot[cnt]); +out_unlock: up_read(&sb_dqopt(inode->i_sb)->dqptr_sem); +out: + return ret; +} + +int dquot_reserve_space(struct inode *inode, qsize_t number, int warn) +{ + int ret = QUOTA_OK; + + if (IS_NOQUOTA(inode)) + goto out; + + down_read(&sb_dqopt(inode->i_sb)->dqptr_sem); + if (IS_NOQUOTA(inode)) + goto out_unlock; + + ret = __dquot_alloc_space(inode, number, warn, 1); +out_unlock: + up_read(&sb_dqopt(inode->i_sb)->dqptr_sem); +out: return ret; } +EXPORT_SYMBOL(dquot_reserve_space); /* * This operation can block, but only after everything is updated @@ -2057,7 +2112,7 @@ static void do_get_dqblk(struct dquot *dquot, struct if_dqblk *di) spin_lock(&dq_data_lock); di->dqb_bhardlimit = stoqb(dm->dqb_bhardlimit); di->dqb_bsoftlimit = stoqb(dm->dqb_bsoftlimit); - di->dqb_curspace = dm->dqb_curspace; + di->dqb_curspace = dm->dqb_curspace + dm->dqb_rsvspace; di->dqb_ihardlimit = dm->dqb_ihardlimit; di->dqb_isoftlimit = dm->dqb_isoftlimit; di->dqb_curinodes = dm->dqb_curinodes; @@ -2097,7 +2152,7 @@ static int do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) spin_lock(&dq_data_lock); if (di->dqb_valid & QIF_SPACE) { - dm->dqb_curspace = di->dqb_curspace; + dm->dqb_curspace = di->dqb_curspace - dm->dqb_rsvspace; check_blim = 1; __set_bit(DQ_LASTSET_B + QIF_SPACE_B, &dquot->dq_flags); } diff --git a/include/linux/quota.h b/include/linux/quota.h index d72d5d84fde5..54b837fa64f2 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -198,6 +198,7 @@ struct mem_dqblk { qsize_t dqb_bhardlimit; /* absolute limit on disk blks alloc */ qsize_t dqb_bsoftlimit; /* preferred limit on disk blks */ qsize_t dqb_curspace; /* current used space */ + qsize_t dqb_rsvspace; /* current reserved space for delalloc*/ qsize_t dqb_ihardlimit; /* absolute limit on allocated inodes */ qsize_t dqb_isoftlimit; /* preferred inode limit */ qsize_t dqb_curinodes; /* current # allocated inodes */ @@ -308,6 +309,8 @@ struct dquot_operations { int (*release_dquot) (struct dquot *); /* Quota is going to be deleted from disk */ int (*mark_dirty) (struct dquot *); /* Dquot is marked dirty */ int (*write_info) (struct super_block *, int); /* Write of quota "superblock" */ + /* reserve quota for delayed block allocation */ + int (*reserve_space) (struct inode *, qsize_t, int); }; /* Operations handling requests from userspace */ diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 0b35b3a1be05..3e3a0d2874d9 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -183,6 +183,16 @@ static inline int vfs_dq_alloc_space(struct inode *inode, qsize_t nr) return ret; } +static inline int vfs_dq_reserve_space(struct inode *inode, qsize_t nr) +{ + if (sb_any_quota_active(inode->i_sb)) { + /* Used space is updated in alloc_space() */ + if (inode->i_sb->dq_op->reserve_space(inode, nr, 0) == NO_QUOTA) + return 1; + } + return 0; +} + static inline int vfs_dq_alloc_inode(struct inode *inode) { if (sb_any_quota_active(inode->i_sb)) { @@ -339,6 +349,11 @@ static inline int vfs_dq_alloc_space(struct inode *inode, qsize_t nr) return 0; } +static inline int vfs_dq_reserve_space(struct inode *inode, qsize_t nr) +{ + return 0; +} + static inline void vfs_dq_free_space_nodirty(struct inode *inode, qsize_t nr) { inode_sub_bytes(inode, nr); @@ -376,6 +391,12 @@ static inline int vfs_dq_alloc_block(struct inode *inode, qsize_t nr) nr << inode->i_sb->s_blocksize_bits); } +static inline int vfs_dq_reserve_block(struct inode *inode, qsize_t nr) +{ + return vfs_dq_reserve_space(inode, + nr << inode->i_blkbits); +} + static inline void vfs_dq_free_block_nodirty(struct inode *inode, qsize_t nr) { vfs_dq_free_space_nodirty(inode, nr << inode->i_sb->s_blocksize_bits); -- cgit v1.2.3-59-g8ed1b From 740d9dcd949a986c88886a591054a0cdb89ef669 Mon Sep 17 00:00:00 2001 From: Mingming Cao Date: Tue, 13 Jan 2009 16:43:14 +0100 Subject: quota: Add quota reservation claim and released operations Reserved quota will be claimed at the block allocation time. Over-booked quota could be returned back with the release callback function. Signed-off-by: Mingming Cao Signed-off-by: Jan Kara --- fs/dquot.c | 110 +++++++++++++++++++++++++++++++++++++++++++++-- include/linux/quota.h | 6 +++ include/linux/quotaops.h | 53 +++++++++++++++++++++++ 3 files changed, 165 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/fs/dquot.c b/fs/dquot.c index 9b1c4d3c9d83..2916f91ca40c 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -904,6 +904,23 @@ static inline void dquot_resv_space(struct dquot *dquot, qsize_t number) dquot->dq_dqb.dqb_rsvspace += number; } +/* + * Claim reserved quota space + */ +static void dquot_claim_reserved_space(struct dquot *dquot, + qsize_t number) +{ + WARN_ON(dquot->dq_dqb.dqb_rsvspace < number); + dquot->dq_dqb.dqb_curspace += number; + dquot->dq_dqb.dqb_rsvspace -= number; +} + +static inline +void dquot_free_reserved_space(struct dquot *dquot, qsize_t number) +{ + dquot->dq_dqb.dqb_rsvspace -= number; +} + static inline void dquot_decr_inodes(struct dquot *dquot, qsize_t number) { if (sb_dqopt(dquot->dq_sb)->flags & DQUOT_NEGATIVE_USAGE || @@ -1452,6 +1469,72 @@ warn_put_all: return ret; } +int dquot_claim_space(struct inode *inode, qsize_t number) +{ + int cnt; + int ret = QUOTA_OK; + + if (IS_NOQUOTA(inode)) { + inode_add_bytes(inode, number); + goto out; + } + + down_read(&sb_dqopt(inode->i_sb)->dqptr_sem); + if (IS_NOQUOTA(inode)) { + up_read(&sb_dqopt(inode->i_sb)->dqptr_sem); + inode_add_bytes(inode, number); + goto out; + } + + spin_lock(&dq_data_lock); + /* Claim reserved quotas to allocated quotas */ + for (cnt = 0; cnt < MAXQUOTAS; cnt++) { + if (inode->i_dquot[cnt] != NODQUOT) + dquot_claim_reserved_space(inode->i_dquot[cnt], + number); + } + /* Update inode bytes */ + inode_add_bytes(inode, number); + spin_unlock(&dq_data_lock); + /* Dirtify all the dquots - this can block when journalling */ + for (cnt = 0; cnt < MAXQUOTAS; cnt++) + if (inode->i_dquot[cnt]) + mark_dquot_dirty(inode->i_dquot[cnt]); + up_read(&sb_dqopt(inode->i_sb)->dqptr_sem); +out: + return ret; +} +EXPORT_SYMBOL(dquot_claim_space); + +/* + * Release reserved quota space + */ +void dquot_release_reserved_space(struct inode *inode, qsize_t number) +{ + int cnt; + + if (IS_NOQUOTA(inode)) + goto out; + + down_read(&sb_dqopt(inode->i_sb)->dqptr_sem); + if (IS_NOQUOTA(inode)) + goto out_unlock; + + spin_lock(&dq_data_lock); + /* Release reserved dquots */ + for (cnt = 0; cnt < MAXQUOTAS; cnt++) { + if (inode->i_dquot[cnt] != NODQUOT) + dquot_free_reserved_space(inode->i_dquot[cnt], number); + } + spin_unlock(&dq_data_lock); + +out_unlock: + up_read(&sb_dqopt(inode->i_sb)->dqptr_sem); +out: + return; +} +EXPORT_SYMBOL(dquot_release_reserved_space); + /* * This operation can block, but only after everything is updated */ @@ -1528,6 +1611,19 @@ int dquot_free_inode(const struct inode *inode, qsize_t number) return QUOTA_OK; } +/* + * call back function, get reserved quota space from underlying fs + */ +qsize_t dquot_get_reserved_space(struct inode *inode) +{ + qsize_t reserved_space = 0; + + if (sb_any_quota_active(inode->i_sb) && + inode->i_sb->dq_op->get_reserved_space) + reserved_space = inode->i_sb->dq_op->get_reserved_space(inode); + return reserved_space; +} + /* * Transfer the number of inode and blocks from one diskquota to an other. * @@ -1536,7 +1632,8 @@ int dquot_free_inode(const struct inode *inode, qsize_t number) */ int dquot_transfer(struct inode *inode, struct iattr *iattr) { - qsize_t space; + qsize_t space, cur_space; + qsize_t rsv_space = 0; struct dquot *transfer_from[MAXQUOTAS]; struct dquot *transfer_to[MAXQUOTAS]; int cnt, ret = QUOTA_OK; @@ -1575,7 +1672,9 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) goto put_all; } spin_lock(&dq_data_lock); - space = inode_get_bytes(inode); + cur_space = inode_get_bytes(inode); + rsv_space = dquot_get_reserved_space(inode); + space = cur_space + rsv_space; /* Build the transfer_from list and check the limits */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (transfer_to[cnt] == NODQUOT) @@ -1604,11 +1703,14 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) warntype_from_space[cnt] = info_bdq_free(transfer_from[cnt], space); dquot_decr_inodes(transfer_from[cnt], 1); - dquot_decr_space(transfer_from[cnt], space); + dquot_decr_space(transfer_from[cnt], cur_space); + dquot_free_reserved_space(transfer_from[cnt], + rsv_space); } dquot_incr_inodes(transfer_to[cnt], 1); - dquot_incr_space(transfer_to[cnt], space); + dquot_incr_space(transfer_to[cnt], cur_space); + dquot_resv_space(transfer_to[cnt], rsv_space); inode->i_dquot[cnt] = transfer_to[cnt]; } diff --git a/include/linux/quota.h b/include/linux/quota.h index 54b837fa64f2..a510d91561f4 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -311,6 +311,12 @@ struct dquot_operations { int (*write_info) (struct super_block *, int); /* Write of quota "superblock" */ /* reserve quota for delayed block allocation */ int (*reserve_space) (struct inode *, qsize_t, int); + /* claim reserved quota for delayed alloc */ + int (*claim_space) (struct inode *, qsize_t); + /* release rsved quota for delayed alloc */ + void (*release_rsv) (struct inode *, qsize_t); + /* get reserved quota for delayed alloc */ + qsize_t (*get_reserved_space) (struct inode *); }; /* Operations handling requests from userspace */ diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 3e3a0d2874d9..7369d04e0a86 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -35,6 +35,11 @@ void dquot_destroy(struct dquot *dquot); int dquot_alloc_space(struct inode *inode, qsize_t number, int prealloc); int dquot_alloc_inode(const struct inode *inode, qsize_t number); +int dquot_reserve_space(struct inode *inode, qsize_t number, int prealloc); +int dquot_claim_space(struct inode *inode, qsize_t number); +void dquot_release_reserved_space(struct inode *inode, qsize_t number); +qsize_t dquot_get_reserved_space(struct inode *inode); + int dquot_free_space(struct inode *inode, qsize_t number); int dquot_free_inode(const struct inode *inode, qsize_t number); @@ -203,6 +208,31 @@ static inline int vfs_dq_alloc_inode(struct inode *inode) return 0; } +/* + * Convert in-memory reserved quotas to real consumed quotas + */ +static inline int vfs_dq_claim_space(struct inode *inode, qsize_t nr) +{ + if (sb_any_quota_active(inode->i_sb)) { + if (inode->i_sb->dq_op->claim_space(inode, nr) == NO_QUOTA) + return 1; + } else + inode_add_bytes(inode, nr); + + mark_inode_dirty(inode); + return 0; +} + +/* + * Release reserved (in-memory) quotas + */ +static inline +void vfs_dq_release_reservation_space(struct inode *inode, qsize_t nr) +{ + if (sb_any_quota_active(inode->i_sb)) + inode->i_sb->dq_op->release_rsv(inode, nr); +} + static inline void vfs_dq_free_space_nodirty(struct inode *inode, qsize_t nr) { if (sb_any_quota_active(inode->i_sb)) @@ -354,6 +384,17 @@ static inline int vfs_dq_reserve_space(struct inode *inode, qsize_t nr) return 0; } +static inline int vfs_dq_claim_space(struct inode *inode, qsize_t nr) +{ + return vfs_dq_alloc_space(inode, nr); +} + +static inline +int vfs_dq_release_reservation_space(struct inode *inode, qsize_t nr) +{ + return 0; +} + static inline void vfs_dq_free_space_nodirty(struct inode *inode, qsize_t nr) { inode_sub_bytes(inode, nr); @@ -397,6 +438,18 @@ static inline int vfs_dq_reserve_block(struct inode *inode, qsize_t nr) nr << inode->i_blkbits); } +static inline int vfs_dq_claim_block(struct inode *inode, qsize_t nr) +{ + return vfs_dq_claim_space(inode, + nr << inode->i_blkbits); +} + +static inline +void vfs_dq_release_reservation_block(struct inode *inode, qsize_t nr) +{ + vfs_dq_release_reservation_space(inode, nr << inode->i_blkbits); +} + static inline void vfs_dq_free_block_nodirty(struct inode *inode, qsize_t nr) { vfs_dq_free_space_nodirty(inode, nr << inode->i_sb->s_blocksize_bits); -- cgit v1.2.3-59-g8ed1b From 9900ba3487f9ba392db30e12d210f768a90abb13 Mon Sep 17 00:00:00 2001 From: Mingming Cao Date: Wed, 14 Jan 2009 16:18:57 +0100 Subject: quota: Use inode->i_blkbits to get block bits Andrew has suggested to use inode->i_blkbits to get the block bits info, rather than use super block's blockbits. That should be faster and emit less code. Signed-off-by: Mingming Cao Signed-off-by: Jan Kara --- include/linux/quotaops.h | 22 ++++++++-------------- 1 file changed, 8 insertions(+), 14 deletions(-) (limited to 'include/linux') diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 7369d04e0a86..69b502e5eba0 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -410,38 +410,32 @@ static inline void vfs_dq_free_space(struct inode *inode, qsize_t nr) static inline int vfs_dq_prealloc_block_nodirty(struct inode *inode, qsize_t nr) { - return vfs_dq_prealloc_space_nodirty(inode, - nr << inode->i_sb->s_blocksize_bits); + return vfs_dq_prealloc_space_nodirty(inode, nr << inode->i_blkbits); } static inline int vfs_dq_prealloc_block(struct inode *inode, qsize_t nr) { - return vfs_dq_prealloc_space(inode, - nr << inode->i_sb->s_blocksize_bits); + return vfs_dq_prealloc_space(inode, nr << inode->i_blkbits); } static inline int vfs_dq_alloc_block_nodirty(struct inode *inode, qsize_t nr) { - return vfs_dq_alloc_space_nodirty(inode, - nr << inode->i_sb->s_blocksize_bits); + return vfs_dq_alloc_space_nodirty(inode, nr << inode->i_blkbits); } static inline int vfs_dq_alloc_block(struct inode *inode, qsize_t nr) { - return vfs_dq_alloc_space(inode, - nr << inode->i_sb->s_blocksize_bits); + return vfs_dq_alloc_space(inode, nr << inode->i_blkbits); } static inline int vfs_dq_reserve_block(struct inode *inode, qsize_t nr) { - return vfs_dq_reserve_space(inode, - nr << inode->i_blkbits); + return vfs_dq_reserve_space(inode, nr << inode->i_blkbits); } static inline int vfs_dq_claim_block(struct inode *inode, qsize_t nr) { - return vfs_dq_claim_space(inode, - nr << inode->i_blkbits); + return vfs_dq_claim_space(inode, nr << inode->i_blkbits); } static inline @@ -452,12 +446,12 @@ void vfs_dq_release_reservation_block(struct inode *inode, qsize_t nr) static inline void vfs_dq_free_block_nodirty(struct inode *inode, qsize_t nr) { - vfs_dq_free_space_nodirty(inode, nr << inode->i_sb->s_blocksize_bits); + vfs_dq_free_space_nodirty(inode, nr << inode->i_blkbits); } static inline void vfs_dq_free_block(struct inode *inode, qsize_t nr) { - vfs_dq_free_space(inode, nr << inode->i_sb->s_blocksize_bits); + vfs_dq_free_space(inode, nr << inode->i_blkbits); } /* -- cgit v1.2.3-59-g8ed1b From dd6f3c6d5a26a282521f15a183fdc2d6f35cfa0f Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 26 Jan 2009 16:01:43 +0100 Subject: quota: Remove NODQUOT macro Remove this macro which is just a definition of NULL. Fix a few coding style issues along the way. Signed-off-by: Jan Kara --- fs/quota/dquot.c | 70 ++++++++++++++++++++++++++------------------------- include/linux/quota.h | 2 -- 2 files changed, 36 insertions(+), 36 deletions(-) (limited to 'include/linux') diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c index e840fa2b112e..4881db32e56d 100644 --- a/fs/quota/dquot.c +++ b/fs/quota/dquot.c @@ -253,7 +253,7 @@ static inline struct dquot *find_dquot(unsigned int hashent, struct super_block if (dquot->dq_sb == sb && dquot->dq_id == id && dquot->dq_type == type) return dquot; } - return NODQUOT; + return NULL; } /* Add a dquot to the tail of the free list */ @@ -696,7 +696,7 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type) dquot = sb->dq_op->alloc_dquot(sb, type); if(!dquot) - return NODQUOT; + return NULL; mutex_init(&dquot->dq_lock); INIT_LIST_HEAD(&dquot->dq_free); @@ -722,10 +722,10 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type) struct dquot *dqget(struct super_block *sb, unsigned int id, int type) { unsigned int hashent = hashfn(sb, id, type); - struct dquot *dquot = NODQUOT, *empty = NODQUOT; + struct dquot *dquot = NULL, *empty = NULL; if (!sb_has_quota_active(sb, type)) - return NODQUOT; + return NULL; we_slept: spin_lock(&dq_list_lock); spin_lock(&dq_state_lock); @@ -736,15 +736,17 @@ we_slept: } spin_unlock(&dq_state_lock); - if ((dquot = find_dquot(hashent, sb, id, type)) == NODQUOT) { - if (empty == NODQUOT) { + dquot = find_dquot(hashent, sb, id, type); + if (!dquot) { + if (!empty) { spin_unlock(&dq_list_lock); - if ((empty = get_empty_dquot(sb, type)) == NODQUOT) + empty = get_empty_dquot(sb, type); + if (!empty) schedule(); /* Try to wait for a moment... */ goto we_slept; } dquot = empty; - empty = NODQUOT; + empty = NULL; dquot->dq_id = id; /* all dquots go on the inuse_list */ put_inuse(dquot); @@ -766,7 +768,7 @@ we_slept: /* Read the dquot and instantiate it (everything done only if needed) */ if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags) && sb->dq_op->acquire_dquot(dquot) < 0) { dqput(dquot); - dquot = NODQUOT; + dquot = NULL; goto out; } #ifdef __DQUOT_PARANOIA @@ -787,9 +789,9 @@ static int dqinit_needed(struct inode *inode, int type) if (IS_NOQUOTA(inode)) return 0; if (type != -1) - return inode->i_dquot[type] == NODQUOT; + return !inode->i_dquot[type]; for (cnt = 0; cnt < MAXQUOTAS; cnt++) - if (inode->i_dquot[cnt] == NODQUOT) + if (!inode->i_dquot[cnt]) return 1; return 0; } @@ -840,8 +842,8 @@ static int remove_inode_dquot_ref(struct inode *inode, int type, { struct dquot *dquot = inode->i_dquot[type]; - inode->i_dquot[type] = NODQUOT; - if (dquot != NODQUOT) { + inode->i_dquot[type] = NULL; + if (dquot) { if (dqput_blocks(dquot)) { #ifdef __DQUOT_PARANOIA if (atomic_read(&dquot->dq_count) != 1) @@ -1112,7 +1114,7 @@ static inline void flush_warnings(struct dquot * const *dquots, char *warntype) int i; for (i = 0; i < MAXQUOTAS; i++) - if (dquots[i] != NODQUOT && warntype[i] != QUOTA_NL_NOWARN && + if (dquots[i] && warntype[i] != QUOTA_NL_NOWARN && !warning_issued(dquots[i], warntype[i])) { #ifdef CONFIG_PRINT_QUOTA_WARNING print_warning(dquots[i], warntype[i]); @@ -1249,7 +1251,7 @@ int dquot_initialize(struct inode *inode, int type) { unsigned int id = 0; int cnt, ret = 0; - struct dquot *got[MAXQUOTAS] = { NODQUOT, NODQUOT }; + struct dquot *got[MAXQUOTAS] = { NULL, NULL }; struct super_block *sb = inode->i_sb; /* First test before acquiring mutex - solves deadlocks when we @@ -1282,9 +1284,9 @@ int dquot_initialize(struct inode *inode, int type) /* Avoid races with quotaoff() */ if (!sb_has_quota_active(sb, cnt)) continue; - if (inode->i_dquot[cnt] == NODQUOT) { + if (!inode->i_dquot[cnt]) { inode->i_dquot[cnt] = got[cnt]; - got[cnt] = NODQUOT; + got[cnt] = NULL; } } out_err: @@ -1307,7 +1309,7 @@ int dquot_drop(struct inode *inode) down_write(&sb_dqopt(inode->i_sb)->dqptr_sem); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { put[cnt] = inode->i_dquot[cnt]; - inode->i_dquot[cnt] = NODQUOT; + inode->i_dquot[cnt] = NULL; } up_write(&sb_dqopt(inode->i_sb)->dqptr_sem); @@ -1332,7 +1334,7 @@ void vfs_dq_drop(struct inode *inode) * must assure that nobody can come after the DQUOT_DROP and * add quota pointers back anyway */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) - if (inode->i_dquot[cnt] != NODQUOT) + if (inode->i_dquot[cnt]) break; if (cnt < MAXQUOTAS) inode->i_sb->dq_op->drop(inode); @@ -1363,7 +1365,7 @@ int __dquot_alloc_space(struct inode *inode, qsize_t number, spin_lock(&dq_data_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] == NODQUOT) + if (!inode->i_dquot[cnt]) continue; if (check_bdq(inode->i_dquot[cnt], number, warn, warntype+cnt) == NO_QUOTA) { @@ -1372,7 +1374,7 @@ int __dquot_alloc_space(struct inode *inode, qsize_t number, } } for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] == NODQUOT) + if (!inode->i_dquot[cnt]) continue; if (reserve) dquot_resv_space(inode->i_dquot[cnt], number); @@ -1461,14 +1463,14 @@ int dquot_alloc_inode(const struct inode *inode, qsize_t number) } spin_lock(&dq_data_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] == NODQUOT) + if (!inode->i_dquot[cnt]) continue; if (check_idq(inode->i_dquot[cnt], number, warntype+cnt) == NO_QUOTA) goto warn_put_all; } for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] == NODQUOT) + if (!inode->i_dquot[cnt]) continue; dquot_incr_inodes(inode->i_dquot[cnt], number); } @@ -1506,7 +1508,7 @@ int dquot_claim_space(struct inode *inode, qsize_t number) spin_lock(&dq_data_lock); /* Claim reserved quotas to allocated quotas */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] != NODQUOT) + if (inode->i_dquot[cnt]) dquot_claim_reserved_space(inode->i_dquot[cnt], number); } @@ -1540,7 +1542,7 @@ void dquot_release_reserved_space(struct inode *inode, qsize_t number) spin_lock(&dq_data_lock); /* Release reserved dquots */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] != NODQUOT) + if (inode->i_dquot[cnt]) dquot_free_reserved_space(inode->i_dquot[cnt], number); } spin_unlock(&dq_data_lock); @@ -1576,7 +1578,7 @@ out_sub: } spin_lock(&dq_data_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] == NODQUOT) + if (!inode->i_dquot[cnt]) continue; warntype[cnt] = info_bdq_free(inode->i_dquot[cnt], number); dquot_decr_space(inode->i_dquot[cnt], number); @@ -1614,7 +1616,7 @@ int dquot_free_inode(const struct inode *inode, qsize_t number) } spin_lock(&dq_data_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (inode->i_dquot[cnt] == NODQUOT) + if (!inode->i_dquot[cnt]) continue; warntype[cnt] = info_idq_free(inode->i_dquot[cnt], number); dquot_decr_inodes(inode->i_dquot[cnt], number); @@ -1667,8 +1669,8 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) return QUOTA_OK; /* Initialize the arrays */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - transfer_from[cnt] = NODQUOT; - transfer_to[cnt] = NODQUOT; + transfer_from[cnt] = NULL; + transfer_to[cnt] = NULL; warntype_to[cnt] = QUOTA_NL_NOWARN; switch (cnt) { case USRQUOTA: @@ -1696,7 +1698,7 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) space = cur_space + rsv_space; /* Build the transfer_from list and check the limits */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (transfer_to[cnt] == NODQUOT) + if (!transfer_to[cnt]) continue; transfer_from[cnt] = inode->i_dquot[cnt]; if (check_idq(transfer_to[cnt], 1, warntype_to + cnt) == @@ -1712,7 +1714,7 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) /* * Skip changes for same uid or gid or for turned off quota-type. */ - if (transfer_to[cnt] == NODQUOT) + if (!transfer_to[cnt]) continue; /* Due to IO error we might not have transfer_from[] structure */ @@ -1743,7 +1745,7 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr) if (transfer_to[cnt]) { mark_dquot_dirty(transfer_to[cnt]); /* The reference we got is transferred to the inode */ - transfer_to[cnt] = NODQUOT; + transfer_to[cnt] = NULL; } } warn_put_all: @@ -1761,7 +1763,7 @@ over_quota: up_write(&sb_dqopt(inode->i_sb)->dqptr_sem); /* Clear dquot pointers we don't want to dqput() */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) - transfer_from[cnt] = NODQUOT; + transfer_from[cnt] = NULL; ret = NO_QUOTA; goto warn_put_all; } @@ -2256,7 +2258,7 @@ int vfs_get_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *d struct dquot *dquot; dquot = dqget(sb, id, type); - if (dquot == NODQUOT) + if (!dquot) return -ESRCH; do_get_dqblk(dquot, di); dqput(dquot); diff --git a/include/linux/quota.h b/include/linux/quota.h index a510d91561f4..78c48895b12a 100644 --- a/include/linux/quota.h +++ b/include/linux/quota.h @@ -277,8 +277,6 @@ struct dquot { struct mem_dqblk dq_dqb; /* Diskquota usage */ }; -#define NODQUOT (struct dquot *)NULL - #define QUOTA_OK 0 #define NO_QUOTA 1 -- cgit v1.2.3-59-g8ed1b From bf84c82d000b9820b01f516d13d328f354f8a8ee Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 26 Jan 2009 17:37:01 +0100 Subject: quota: Remove uppercase aliases for quota functions. Since all users have been converted, remove uppercase names of quota functions. Signed-off-by: Jan Kara --- include/linux/quotaops.h | 31 ------------------------------- 1 file changed, 31 deletions(-) (limited to 'include/linux') diff --git a/include/linux/quotaops.h b/include/linux/quotaops.h index 69b502e5eba0..36353d95c8db 100644 --- a/include/linux/quotaops.h +++ b/include/linux/quotaops.h @@ -454,35 +454,4 @@ static inline void vfs_dq_free_block(struct inode *inode, qsize_t nr) vfs_dq_free_space(inode, nr << inode->i_blkbits); } -/* - * Define uppercase equivalents for compatibility with old function names - * Can go away when we think all users have been converted (15/04/2008) - */ -#define DQUOT_INIT(inode) vfs_dq_init(inode) -#define DQUOT_DROP(inode) vfs_dq_drop(inode) -#define DQUOT_PREALLOC_SPACE_NODIRTY(inode, nr) \ - vfs_dq_prealloc_space_nodirty(inode, nr) -#define DQUOT_PREALLOC_SPACE(inode, nr) vfs_dq_prealloc_space(inode, nr) -#define DQUOT_ALLOC_SPACE_NODIRTY(inode, nr) \ - vfs_dq_alloc_space_nodirty(inode, nr) -#define DQUOT_ALLOC_SPACE(inode, nr) vfs_dq_alloc_space(inode, nr) -#define DQUOT_PREALLOC_BLOCK_NODIRTY(inode, nr) \ - vfs_dq_prealloc_block_nodirty(inode, nr) -#define DQUOT_PREALLOC_BLOCK(inode, nr) vfs_dq_prealloc_block(inode, nr) -#define DQUOT_ALLOC_BLOCK_NODIRTY(inode, nr) \ - vfs_dq_alloc_block_nodirty(inode, nr) -#define DQUOT_ALLOC_BLOCK(inode, nr) vfs_dq_alloc_block(inode, nr) -#define DQUOT_ALLOC_INODE(inode) vfs_dq_alloc_inode(inode) -#define DQUOT_FREE_SPACE_NODIRTY(inode, nr) \ - vfs_dq_free_space_nodirty(inode, nr) -#define DQUOT_FREE_SPACE(inode, nr) vfs_dq_free_space(inode, nr) -#define DQUOT_FREE_BLOCK_NODIRTY(inode, nr) \ - vfs_dq_free_block_nodirty(inode, nr) -#define DQUOT_FREE_BLOCK(inode, nr) vfs_dq_free_block(inode, nr) -#define DQUOT_FREE_INODE(inode) vfs_dq_free_inode(inode) -#define DQUOT_TRANSFER(inode, iattr) vfs_dq_transfer(inode, iattr) -#define DQUOT_SYNC(sb) vfs_dq_sync(sb) -#define DQUOT_OFF(sb, remount) vfs_dq_off(sb, remount) -#define DQUOT_ON_REMOUNT(sb) vfs_dq_quota_on_remount(sb) - #endif /* _LINUX_QUOTAOPS_ */ -- cgit v1.2.3-59-g8ed1b From 05fd8e73e1357feaea9c48938d937eae76b4aef4 Mon Sep 17 00:00:00 2001 From: Sascha Hauer Date: Sat, 7 Mar 2009 12:55:49 +0100 Subject: clkdev: add possibility to get a clock based on the device name This adds clk_get_sys to get a clock without the associated struct device. Signed-off-by: Sascha Hauer --- arch/arm/common/clkdev.c | 11 +++++++++-- include/linux/clk.h | 17 +++++++++++++++++ 2 files changed, 26 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/arch/arm/common/clkdev.c b/arch/arm/common/clkdev.c index 1037bba18329..5589444ff437 100644 --- a/arch/arm/common/clkdev.c +++ b/arch/arm/common/clkdev.c @@ -62,9 +62,8 @@ static struct clk *clk_find(const char *dev_id, const char *con_id) return clk; } -struct clk *clk_get(struct device *dev, const char *con_id) +struct clk *clk_get_sys(const char *dev_id, const char *con_id) { - const char *dev_id = dev ? dev_name(dev) : NULL; struct clk *clk; mutex_lock(&clocks_mutex); @@ -75,6 +74,14 @@ struct clk *clk_get(struct device *dev, const char *con_id) return clk ? clk : ERR_PTR(-ENOENT); } +EXPORT_SYMBOL(clk_get_sys); + +struct clk *clk_get(struct device *dev, const char *con_id) +{ + const char *dev_id = dev ? dev_name(dev) : NULL; + + return clk_get_sys(dev_id, con_id); +} EXPORT_SYMBOL(clk_get); void clk_put(struct clk *clk) diff --git a/include/linux/clk.h b/include/linux/clk.h index 778777316ea4..1db9bbf444a3 100644 --- a/include/linux/clk.h +++ b/include/linux/clk.h @@ -125,4 +125,21 @@ int clk_set_parent(struct clk *clk, struct clk *parent); */ struct clk *clk_get_parent(struct clk *clk); +/** + * clk_get_sys - get a clock based upon the device name + * @dev_id: device name + * @con_id: connection ID + * + * Returns a struct clk corresponding to the clock producer, or + * valid IS_ERR() condition containing errno. The implementation + * uses @dev_id and @con_id to determine the clock consumer, and + * thereby the clock producer. In contrast to clk_get() this function + * takes the device name instead of the device itself for identification. + * + * Drivers must assume that the clock source is not enabled. + * + * clk_get_sys should not be called from within interrupt context. + */ +struct clk *clk_get_sys(const char *dev_id, const char *con_id); + #endif -- cgit v1.2.3-59-g8ed1b From 2b1c6bd77d4e6a727ffac8630cd154b2144b751a Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 28 Nov 2008 10:09:09 +0100 Subject: generic compat_sys_ustat Due to a different size of ino_t ustat needs a compat handler, but currently only x86 and mips provide one. Add a generic compat_sys_ustat and switch all architectures over to it. Instead of doing various user copy hacks compat_sys_ustat just reimplements sys_ustat as it's trivial. This was suggested by Arnd Bergmann. Found by Eric Sandeen when running xfstests/017 on ppc64, which causes stack smashing warnings on RHEL/Fedora due to the too large amount of data writen by the syscall. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- arch/ia64/ia32/ia32_entry.S | 2 +- arch/mips/kernel/linux32.c | 34 ---------------------------------- arch/mips/kernel/scall64-n32.S | 2 +- arch/mips/kernel/scall64-o32.S | 2 +- arch/parisc/kernel/syscall_table.S | 2 +- arch/powerpc/include/asm/systbl.h | 2 +- arch/s390/kernel/compat_wrapper.S | 2 +- arch/x86/ia32/ia32entry.S | 2 +- arch/x86/ia32/sys_ia32.c | 22 ---------------------- arch/x86/include/asm/ia32.h | 7 ------- arch/x86/include/asm/sys_ia32.h | 2 -- fs/compat.c | 28 ++++++++++++++++++++++++++++ include/linux/compat.h | 8 ++++++++ 13 files changed, 43 insertions(+), 72 deletions(-) (limited to 'include/linux') diff --git a/arch/ia64/ia32/ia32_entry.S b/arch/ia64/ia32/ia32_entry.S index a46f8395e9a5..af9405cd70e5 100644 --- a/arch/ia64/ia32/ia32_entry.S +++ b/arch/ia64/ia32/ia32_entry.S @@ -240,7 +240,7 @@ ia32_syscall_table: data8 sys_ni_syscall data8 sys_umask /* 60 */ data8 sys_chroot - data8 sys_ustat + data8 compat_sys_ustat data8 sys_dup2 data8 sys_getppid data8 sys_getpgrp /* 65 */ diff --git a/arch/mips/kernel/linux32.c b/arch/mips/kernel/linux32.c index 1a86f84fa947..784859cedef7 100644 --- a/arch/mips/kernel/linux32.c +++ b/arch/mips/kernel/linux32.c @@ -356,40 +356,6 @@ SYSCALL_DEFINE1(32_personality, unsigned long, personality) return ret; } -/* ustat compatibility */ -struct ustat32 { - compat_daddr_t f_tfree; - compat_ino_t f_tinode; - char f_fname[6]; - char f_fpack[6]; -}; - -extern asmlinkage long sys_ustat(dev_t dev, struct ustat __user * ubuf); - -SYSCALL_DEFINE2(32_ustat, dev_t, dev, struct ustat32 __user *, ubuf32) -{ - int err; - struct ustat tmp; - struct ustat32 tmp32; - mm_segment_t old_fs = get_fs(); - - set_fs(KERNEL_DS); - err = sys_ustat(dev, (struct ustat __user *)&tmp); - set_fs(old_fs); - - if (err) - goto out; - - memset(&tmp32, 0, sizeof(struct ustat32)); - tmp32.f_tfree = tmp.f_tfree; - tmp32.f_tinode = tmp.f_tinode; - - err = copy_to_user(ubuf32, &tmp32, sizeof(struct ustat32)) ? -EFAULT : 0; - -out: - return err; -} - SYSCALL_DEFINE4(32_sendfile, long, out_fd, long, in_fd, compat_off_t __user *, offset, s32, count) { diff --git a/arch/mips/kernel/scall64-n32.S b/arch/mips/kernel/scall64-n32.S index 7438e92f8a01..f61d6b0e5731 100644 --- a/arch/mips/kernel/scall64-n32.S +++ b/arch/mips/kernel/scall64-n32.S @@ -253,7 +253,7 @@ EXPORT(sysn32_call_table) PTR compat_sys_utime /* 6130 */ PTR sys_mknod PTR sys_32_personality - PTR sys_32_ustat + PTR compat_sys_ustat PTR compat_sys_statfs PTR compat_sys_fstatfs /* 6135 */ PTR sys_sysfs diff --git a/arch/mips/kernel/scall64-o32.S b/arch/mips/kernel/scall64-o32.S index b0fef4ff9827..60997f1f69d4 100644 --- a/arch/mips/kernel/scall64-o32.S +++ b/arch/mips/kernel/scall64-o32.S @@ -265,7 +265,7 @@ sys_call_table: PTR sys_olduname PTR sys_umask /* 4060 */ PTR sys_chroot - PTR sys_32_ustat + PTR compat_sys_ustat PTR sys_dup2 PTR sys_getppid PTR sys_getpgrp /* 4065 */ diff --git a/arch/parisc/kernel/syscall_table.S b/arch/parisc/kernel/syscall_table.S index 303d2b647e41..03b9a01bc16c 100644 --- a/arch/parisc/kernel/syscall_table.S +++ b/arch/parisc/kernel/syscall_table.S @@ -130,7 +130,7 @@ ENTRY_OURS(newuname) ENTRY_SAME(umask) /* 60 */ ENTRY_SAME(chroot) - ENTRY_SAME(ustat) + ENTRY_COMP(ustat) ENTRY_SAME(dup2) ENTRY_SAME(getppid) ENTRY_SAME(getpgrp) /* 65 */ diff --git a/arch/powerpc/include/asm/systbl.h b/arch/powerpc/include/asm/systbl.h index 72353f6070a4..fe166491e9dc 100644 --- a/arch/powerpc/include/asm/systbl.h +++ b/arch/powerpc/include/asm/systbl.h @@ -65,7 +65,7 @@ SYSCALL(ni_syscall) SYSX(sys_ni_syscall,sys_olduname, sys_olduname) COMPAT_SYS_SPU(umask) SYSCALL_SPU(chroot) -SYSCALL(ustat) +COMPAT_SYS(ustat) SYSCALL_SPU(dup2) SYSCALL_SPU(getppid) SYSCALL_SPU(getpgrp) diff --git a/arch/s390/kernel/compat_wrapper.S b/arch/s390/kernel/compat_wrapper.S index 62c706eb0de6..87cf5a79a351 100644 --- a/arch/s390/kernel/compat_wrapper.S +++ b/arch/s390/kernel/compat_wrapper.S @@ -252,7 +252,7 @@ sys32_chroot_wrapper: sys32_ustat_wrapper: llgfr %r2,%r2 # dev_t llgtr %r3,%r3 # struct ustat * - jg sys_ustat + jg compat_sys_ustat .globl sys32_dup2_wrapper sys32_dup2_wrapper: diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S index 5a0d76dc56a4..8ef8876666b2 100644 --- a/arch/x86/ia32/ia32entry.S +++ b/arch/x86/ia32/ia32entry.S @@ -557,7 +557,7 @@ ia32_sys_call_table: .quad sys32_olduname .quad sys_umask /* 60 */ .quad sys_chroot - .quad sys32_ustat + .quad compat_sys_ustat .quad sys_dup2 .quad sys_getppid .quad sys_getpgrp /* 65 */ diff --git a/arch/x86/ia32/sys_ia32.c b/arch/x86/ia32/sys_ia32.c index 6c0d7f6231af..efac92fd1efb 100644 --- a/arch/x86/ia32/sys_ia32.c +++ b/arch/x86/ia32/sys_ia32.c @@ -638,28 +638,6 @@ long sys32_uname(struct old_utsname __user *name) return err ? -EFAULT : 0; } -long sys32_ustat(unsigned dev, struct ustat32 __user *u32p) -{ - struct ustat u; - mm_segment_t seg; - int ret; - - seg = get_fs(); - set_fs(KERNEL_DS); - ret = sys_ustat(dev, (struct ustat __user *)&u); - set_fs(seg); - if (ret < 0) - return ret; - - if (!access_ok(VERIFY_WRITE, u32p, sizeof(struct ustat32)) || - __put_user((__u32) u.f_tfree, &u32p->f_tfree) || - __put_user((__u32) u.f_tinode, &u32p->f_tfree) || - __copy_to_user(&u32p->f_fname, u.f_fname, sizeof(u.f_fname)) || - __copy_to_user(&u32p->f_fpack, u.f_fpack, sizeof(u.f_fpack))) - ret = -EFAULT; - return ret; -} - asmlinkage long sys32_execve(char __user *name, compat_uptr_t __user *argv, compat_uptr_t __user *envp, struct pt_regs *regs) { diff --git a/arch/x86/include/asm/ia32.h b/arch/x86/include/asm/ia32.h index 50ca486fd88c..1f7e62517284 100644 --- a/arch/x86/include/asm/ia32.h +++ b/arch/x86/include/asm/ia32.h @@ -129,13 +129,6 @@ typedef struct compat_siginfo { } _sifields; } compat_siginfo_t; -struct ustat32 { - __u32 f_tfree; - compat_ino_t f_tinode; - char f_fname[6]; - char f_fpack[6]; -}; - #define IA32_STACK_TOP IA32_PAGE_OFFSET #ifdef __KERNEL__ diff --git a/arch/x86/include/asm/sys_ia32.h b/arch/x86/include/asm/sys_ia32.h index ffb08be2a530..72a6dcd1299b 100644 --- a/arch/x86/include/asm/sys_ia32.h +++ b/arch/x86/include/asm/sys_ia32.h @@ -70,8 +70,6 @@ struct old_utsname; asmlinkage long sys32_olduname(struct oldold_utsname __user *); long sys32_uname(struct old_utsname __user *); -long sys32_ustat(unsigned, struct ustat32 __user *); - asmlinkage long sys32_execve(char __user *, compat_uptr_t __user *, compat_uptr_t __user *, struct pt_regs *); asmlinkage long sys32_clone(unsigned int, unsigned int, struct pt_regs *); diff --git a/fs/compat.c b/fs/compat.c index d0145ca27572..4e0db94b5353 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -378,6 +378,34 @@ out: return error; } +/* + * This is a copy of sys_ustat, just dealing with a structure layout. + * Given how simple this syscall is that apporach is more maintainable + * than the various conversion hacks. + */ +asmlinkage long compat_sys_ustat(unsigned dev, struct compat_ustat __user *u) +{ + struct super_block *sb; + struct compat_ustat tmp; + struct kstatfs sbuf; + int err; + + sb = user_get_super(new_decode_dev(dev)); + if (!sb) + return -EINVAL; + err = vfs_statfs(sb->s_root, &sbuf); + drop_super(sb); + if (err) + return err; + + memset(&tmp, 0, sizeof(struct compat_ustat)); + tmp.f_tfree = sbuf.f_bfree; + tmp.f_tinode = sbuf.f_ffree; + if (copy_to_user(u, &tmp, sizeof(struct compat_ustat))) + return -EFAULT; + return 0; +} + static int get_compat_flock(struct flock *kfl, struct compat_flock __user *ufl) { if (!access_ok(VERIFY_READ, ufl, sizeof(*ufl)) || diff --git a/include/linux/compat.h b/include/linux/compat.h index 3fd2194ff573..b880864672de 100644 --- a/include/linux/compat.h +++ b/include/linux/compat.h @@ -125,6 +125,13 @@ struct compat_dirent { char d_name[256]; }; +struct compat_ustat { + compat_daddr_t f_tfree; + compat_ino_t f_tinode; + char f_fname[6]; + char f_fpack[6]; +}; + typedef union compat_sigval { compat_int_t sival_int; compat_uptr_t sival_ptr; @@ -178,6 +185,7 @@ long compat_sys_semtimedop(int semid, struct sembuf __user *tsems, unsigned nsems, const struct compat_timespec __user *timeout); asmlinkage long compat_sys_keyctl(u32 option, u32 arg2, u32 arg3, u32 arg4, u32 arg5); +asmlinkage long compat_sys_ustat(unsigned dev, struct compat_ustat __user *u32); asmlinkage ssize_t compat_sys_readv(unsigned long fd, const struct compat_iovec __user *vec, unsigned long vlen); -- cgit v1.2.3-59-g8ed1b From 10f303ae1e5e77a9f7cb053e6329906afb132c67 Mon Sep 17 00:00:00 2001 From: Cheng Renquan Date: Wed, 14 Jan 2009 17:01:33 +0800 Subject: do_pipe cleanup: drop its last user in arch/alpha/ The last user of do_pipe is in arch/alpha/, after replacing it with do_pipe_flags, the do_pipe can be totally dropped. Signed-off-by: Cheng Renquan Acked-by: Richard Henderson Signed-off-by: Al Viro --- arch/alpha/kernel/entry.S | 3 ++- arch/alpha/kernel/osf_sys.c | 2 -- fs/pipe.c | 5 ----- include/linux/fs.h | 1 - 4 files changed, 2 insertions(+), 9 deletions(-) (limited to 'include/linux') diff --git a/arch/alpha/kernel/entry.S b/arch/alpha/kernel/entry.S index e4a54b615894..b45d913a51c3 100644 --- a/arch/alpha/kernel/entry.S +++ b/arch/alpha/kernel/entry.S @@ -903,8 +903,9 @@ sys_alpha_pipe: stq $26, 0($sp) .prologue 0 + mov $31, $17 lda $16, 8($sp) - jsr $26, do_pipe + jsr $26, do_pipe_flags ldq $26, 0($sp) bne $0, 1f diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c index ae41f097864b..42ee05981e71 100644 --- a/arch/alpha/kernel/osf_sys.c +++ b/arch/alpha/kernel/osf_sys.c @@ -46,8 +46,6 @@ #include #include -extern int do_pipe(int *); - /* * Brk needs to return an error. Still support Linux's brk(0) query idiom, * which OSF programs just shouldn't be doing. We're still not quite diff --git a/fs/pipe.c b/fs/pipe.c index 14f502b89cf5..df3719562fc1 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -1034,11 +1034,6 @@ int do_pipe_flags(int *fd, int flags) return error; } -int do_pipe(int *fd) -{ - return do_pipe_flags(fd, 0); -} - /* * sys_pipe() is the normal C calling standard for creating * a pipe. It's not the way Unix traditionally does this, though. diff --git a/include/linux/fs.h b/include/linux/fs.h index 92734c0012e6..51de83bd8a87 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1881,7 +1881,6 @@ static inline void allow_write_access(struct file *file) if (file) atomic_inc(&file->f_path.dentry->d_inode->i_writecount); } -extern int do_pipe(int *); extern int do_pipe_flags(int *, int); extern struct file *create_read_pipe(struct file *f, int flags); extern struct file *create_write_pipe(int flags); -- cgit v1.2.3-59-g8ed1b From c2aca5e529a2499d454c41e01f59f1d5fe4a1364 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 20 Jan 2009 10:29:45 +0000 Subject: vfs: Update fs.h to use inline functions when no file locking set This avoids various issues which might give rise to compiler warnings about missing functions and/or unused variable with the previous macros. This also fixes a bug where one of the macros was returning 0, but it should have been void. Reported-by: Randy Dunlap Signed-off-by: Steven Whitehouse Tested-by: Randy Dunlap Signed-off-by: Al Viro --- include/linux/fs.h | 165 ++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 139 insertions(+), 26 deletions(-) (limited to 'include/linux') diff --git a/include/linux/fs.h b/include/linux/fs.h index 51de83bd8a87..d84020b7e676 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1063,34 +1063,147 @@ extern int lease_modify(struct file_lock **, int); extern int lock_may_read(struct inode *, loff_t start, unsigned long count); extern int lock_may_write(struct inode *, loff_t start, unsigned long count); #else /* !CONFIG_FILE_LOCKING */ -#define fcntl_getlk(a, b) ({ -EINVAL; }) -#define fcntl_setlk(a, b, c, d) ({ -EACCES; }) +static inline int fcntl_getlk(struct file *file, struct flock __user *user) +{ + return -EINVAL; +} + +static inline int fcntl_setlk(unsigned int fd, struct file *file, + unsigned int cmd, struct flock __user *user) +{ + return -EACCES; +} + #if BITS_PER_LONG == 32 -#define fcntl_getlk64(a, b) ({ -EINVAL; }) -#define fcntl_setlk64(a, b, c, d) ({ -EACCES; }) +static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user) +{ + return -EINVAL; +} + +static inline int fcntl_setlk64(unsigned int fd, struct file *file, + unsigned int cmd, struct flock64 __user *user) +{ + return -EACCES; +} #endif -#define fcntl_setlease(a, b, c) ({ 0; }) -#define fcntl_getlease(a) ({ 0; }) -#define locks_init_lock(a) ({ }) -#define __locks_copy_lock(a, b) ({ }) -#define locks_copy_lock(a, b) ({ }) -#define locks_remove_posix(a, b) ({ }) -#define locks_remove_flock(a) ({ }) -#define posix_test_lock(a, b) ({ 0; }) -#define posix_lock_file(a, b, c) ({ -ENOLCK; }) -#define posix_lock_file_wait(a, b) ({ -ENOLCK; }) -#define posix_unblock_lock(a, b) (-ENOENT) -#define vfs_test_lock(a, b) ({ 0; }) -#define vfs_lock_file(a, b, c, d) (-ENOLCK) -#define vfs_cancel_lock(a, b) ({ 0; }) -#define flock_lock_file_wait(a, b) ({ -ENOLCK; }) -#define __break_lease(a, b) ({ 0; }) -#define lease_get_mtime(a, b) ({ }) -#define generic_setlease(a, b, c) ({ -EINVAL; }) -#define vfs_setlease(a, b, c) ({ -EINVAL; }) -#define lease_modify(a, b) ({ -EINVAL; }) -#define lock_may_read(a, b, c) ({ 1; }) -#define lock_may_write(a, b, c) ({ 1; }) +static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg) +{ + return 0; +} + +static inline int fcntl_getlease(struct file *filp) +{ + return 0; +} + +static inline void locks_init_lock(struct file_lock *fl) +{ + return; +} + +static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl) +{ + return; +} + +static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl) +{ + return; +} + +static inline void locks_remove_posix(struct file *filp, fl_owner_t owner) +{ + return; +} + +static inline void locks_remove_flock(struct file *filp) +{ + return; +} + +static inline void posix_test_lock(struct file *filp, struct file_lock *fl) +{ + return; +} + +static inline int posix_lock_file(struct file *filp, struct file_lock *fl, + struct file_lock *conflock) +{ + return -ENOLCK; +} + +static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl) +{ + return -ENOLCK; +} + +static inline int posix_unblock_lock(struct file *filp, + struct file_lock *waiter) +{ + return -ENOENT; +} + +static inline int vfs_test_lock(struct file *filp, struct file_lock *fl) +{ + return 0; +} + +static inline int vfs_lock_file(struct file *filp, unsigned int cmd, + struct file_lock *fl, struct file_lock *conf) +{ + return -ENOLCK; +} + +static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl) +{ + return 0; +} + +static inline int flock_lock_file_wait(struct file *filp, + struct file_lock *request) +{ + return -ENOLCK; +} + +static inline int __break_lease(struct inode *inode, unsigned int mode) +{ + return 0; +} + +static inline void lease_get_mtime(struct inode *inode, struct timespec *time) +{ + return; +} + +static inline int generic_setlease(struct file *filp, long arg, + struct file_lock **flp) +{ + return -EINVAL; +} + +static inline int vfs_setlease(struct file *filp, long arg, + struct file_lock **lease) +{ + return -EINVAL; +} + +static inline int lease_modify(struct file_lock **before, int arg) +{ + return -EINVAL; +} + +static inline int lock_may_read(struct inode *inode, loff_t start, + unsigned long len) +{ + return 1; +} + +static inline int lock_may_write(struct inode *inode, loff_t start, + unsigned long len) +{ + return 1; +} + #endif /* !CONFIG_FILE_LOCKING */ -- cgit v1.2.3-59-g8ed1b From af5df56688acfb75c1b15b4e000ec5e82a9cdc29 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 20 Jan 2009 10:29:46 +0000 Subject: vfs: Further changes from macro to inline function in fs.h There is a second set of macros for when CONFIG_FILE_LOCKING is not set. This patch updates those to become inline functions as well. Signed-off-by: Steven Whitehouse Signed-off-by: Al Viro --- include/linux/fs.h | 45 ++++++++++++++++++++++++++++++++++++++------- 1 file changed, 38 insertions(+), 7 deletions(-) (limited to 'include/linux') diff --git a/include/linux/fs.h b/include/linux/fs.h index d84020b7e676..5f74d616cd7d 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1800,13 +1800,44 @@ static inline int break_lease(struct inode *inode, unsigned int mode) return 0; } #else /* !CONFIG_FILE_LOCKING */ -#define locks_mandatory_locked(a) ({ 0; }) -#define locks_mandatory_area(a, b, c, d, e) ({ 0; }) -#define __mandatory_lock(a) ({ 0; }) -#define mandatory_lock(a) ({ 0; }) -#define locks_verify_locked(a) ({ 0; }) -#define locks_verify_truncate(a, b, c) ({ 0; }) -#define break_lease(a, b) ({ 0; }) +static inline int locks_mandatory_locked(struct inode *inode) +{ + return 0; +} + +static inline int locks_mandatory_area(int rw, struct inode *inode, + struct file *filp, loff_t offset, + size_t count) +{ + return 0; +} + +static inline int __mandatory_lock(struct inode *inode) +{ + return 0; +} + +static inline int mandatory_lock(struct inode *inode) +{ + return 0; +} + +static inline int locks_verify_locked(struct inode *inode) +{ + return 0; +} + +static inline int locks_verify_truncate(struct inode *inode, struct file *filp, + size_t size) +{ + return 0; +} + +static inline int break_lease(struct inode *inode, unsigned int mode) +{ + return 0; +} + #endif /* CONFIG_FILE_LOCKING */ /* fs/open.c */ -- cgit v1.2.3-59-g8ed1b From e56980d451904b623573ef4966cbab768e433c79 Mon Sep 17 00:00:00 2001 From: Jan Engelhardt Date: Wed, 11 Feb 2009 13:14:54 -0800 Subject: fs: make struct dentry->d_op const This change will allow for tagging many dentry_operations const in the source tree. Signed-off-by: Jan Engelhardt Cc: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- include/linux/dcache.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/dcache.h b/include/linux/dcache.h index c66d22487bf8..15156364d196 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -112,7 +112,7 @@ struct dentry { struct list_head d_subdirs; /* our children */ struct list_head d_alias; /* inode alias list */ unsigned long d_time; /* used by d_revalidate */ - struct dentry_operations *d_op; + const struct dentry_operations *d_op; struct super_block *d_sb; /* The root of the dentry tree */ void *d_fsdata; /* fs-specific data */ -- cgit v1.2.3-59-g8ed1b From f786aa90e026f2174bb0c26d49f338c5c46ede55 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 20 Feb 2009 05:51:22 +0000 Subject: constify dentry_operations: NFS Signed-off-by: Al Viro --- fs/nfs/dir.c | 4 ++-- fs/nfs/nfs4_fs.h | 2 +- include/linux/nfs_fs.h | 2 +- include/linux/nfs_xdr.h | 2 +- 4 files changed, 5 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 672368f865ca..78bf72fc1db3 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -899,7 +899,7 @@ static void nfs_dentry_iput(struct dentry *dentry, struct inode *inode) iput(inode); } -struct dentry_operations nfs_dentry_operations = { +const struct dentry_operations nfs_dentry_operations = { .d_revalidate = nfs_lookup_revalidate, .d_delete = nfs_dentry_delete, .d_iput = nfs_dentry_iput, @@ -967,7 +967,7 @@ out: #ifdef CONFIG_NFS_V4 static int nfs_open_revalidate(struct dentry *, struct nameidata *); -struct dentry_operations nfs4_dentry_operations = { +const struct dentry_operations nfs4_dentry_operations = { .d_revalidate = nfs_open_revalidate, .d_delete = nfs_dentry_delete, .d_iput = nfs_dentry_iput, diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 4e4d33204376..84345deab26f 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -179,7 +179,7 @@ struct nfs4_state_recovery_ops { int (*recover_lock)(struct nfs4_state *, struct file_lock *); }; -extern struct dentry_operations nfs4_dentry_operations; +extern const struct dentry_operations nfs4_dentry_operations; extern const struct inode_operations nfs4_dir_inode_operations; /* inode.c */ diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index db867b04ac3c..8cc8807f77d6 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -415,7 +415,7 @@ extern const struct inode_operations nfs_dir_inode_operations; extern const struct inode_operations nfs3_dir_inode_operations; #endif /* CONFIG_NFS_V3 */ extern const struct file_operations nfs_dir_operations; -extern struct dentry_operations nfs_dentry_operations; +extern const struct dentry_operations nfs_dentry_operations; extern void nfs_force_lookup_revalidate(struct inode *dir); extern int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fh, struct nfs_fattr *fattr); diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 2e5f00066afd..43a713fce11c 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -785,7 +785,7 @@ struct nfs_access_entry; */ struct nfs_rpc_ops { u32 version; /* Protocol version */ - struct dentry_operations *dentry_ops; + const struct dentry_operations *dentry_ops; const struct inode_operations *dir_inode_ops; const struct inode_operations *file_inode_ops; -- cgit v1.2.3-59-g8ed1b From e16404ed0f3f330dc3e99b95cef69bb60bcd27f7 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 20 Feb 2009 05:55:13 +0000 Subject: constify dentry_operations: misc filesystems Signed-off-by: Al Viro --- fs/adfs/adfs.h | 2 +- fs/adfs/dir.c | 2 +- fs/affs/affs.h | 3 +-- fs/affs/namei.c | 4 ++-- fs/coda/dir.c | 2 +- fs/hfs/hfs_fs.h | 2 +- fs/hfs/sysdep.c | 2 +- fs/hfsplus/hfsplus_fs.h | 2 +- fs/hfsplus/inode.c | 2 +- fs/hostfs/hostfs_kern.c | 4 ++-- fs/hpfs/dentry.c | 2 +- fs/isofs/inode.c | 2 +- fs/ncpfs/dir.c | 4 ++-- fs/reiserfs/xattr.c | 2 +- fs/smbfs/dir.c | 4 ++-- fs/sysv/namei.c | 2 +- fs/sysv/sysv.h | 2 +- include/linux/ncp_fs.h | 2 +- 18 files changed, 22 insertions(+), 23 deletions(-) (limited to 'include/linux') diff --git a/fs/adfs/adfs.h b/fs/adfs/adfs.h index 831157502d5a..e0a85dbeeb88 100644 --- a/fs/adfs/adfs.h +++ b/fs/adfs/adfs.h @@ -86,7 +86,7 @@ void __adfs_error(struct super_block *sb, const char *function, /* dir_*.c */ extern const struct inode_operations adfs_dir_inode_operations; extern const struct file_operations adfs_dir_operations; -extern struct dentry_operations adfs_dentry_operations; +extern const struct dentry_operations adfs_dentry_operations; extern struct adfs_dir_ops adfs_f_dir_ops; extern struct adfs_dir_ops adfs_fplus_dir_ops; diff --git a/fs/adfs/dir.c b/fs/adfs/dir.c index 85a30e929800..e867ccf37246 100644 --- a/fs/adfs/dir.c +++ b/fs/adfs/dir.c @@ -263,7 +263,7 @@ adfs_compare(struct dentry *parent, struct qstr *entry, struct qstr *name) return 0; } -struct dentry_operations adfs_dentry_operations = { +const struct dentry_operations adfs_dentry_operations = { .d_hash = adfs_hash, .d_compare = adfs_compare, }; diff --git a/fs/affs/affs.h b/fs/affs/affs.h index e9ec915f7553..1a2d5e3c7f4e 100644 --- a/fs/affs/affs.h +++ b/fs/affs/affs.h @@ -199,8 +199,7 @@ extern const struct address_space_operations affs_symlink_aops; extern const struct address_space_operations affs_aops; extern const struct address_space_operations affs_aops_ofs; -extern struct dentry_operations affs_dentry_operations; -extern struct dentry_operations affs_dentry_operations_intl; +extern const struct dentry_operations affs_dentry_operations; static inline void affs_set_blocksize(struct super_block *sb, int size) diff --git a/fs/affs/namei.c b/fs/affs/namei.c index cfcf1b6cf82b..960d336ec694 100644 --- a/fs/affs/namei.c +++ b/fs/affs/namei.c @@ -19,12 +19,12 @@ static int affs_intl_toupper(int ch); static int affs_intl_hash_dentry(struct dentry *, struct qstr *); static int affs_intl_compare_dentry(struct dentry *, struct qstr *, struct qstr *); -struct dentry_operations affs_dentry_operations = { +const struct dentry_operations affs_dentry_operations = { .d_hash = affs_hash_dentry, .d_compare = affs_compare_dentry, }; -static struct dentry_operations affs_intl_dentry_operations = { +static const struct dentry_operations affs_intl_dentry_operations = { .d_hash = affs_intl_hash_dentry, .d_compare = affs_intl_compare_dentry, }; diff --git a/fs/coda/dir.c b/fs/coda/dir.c index 75b1fa90b2cb..4bb9d0a5decc 100644 --- a/fs/coda/dir.c +++ b/fs/coda/dir.c @@ -59,7 +59,7 @@ static int coda_return_EIO(void) } #define CODA_EIO_ERROR ((void *) (coda_return_EIO)) -static struct dentry_operations coda_dentry_operations = +static const struct dentry_operations coda_dentry_operations = { .d_revalidate = coda_dentry_revalidate, .d_delete = coda_dentry_delete, diff --git a/fs/hfs/hfs_fs.h b/fs/hfs/hfs_fs.h index 9955232fdf8c..052387e11671 100644 --- a/fs/hfs/hfs_fs.h +++ b/fs/hfs/hfs_fs.h @@ -213,7 +213,7 @@ extern void hfs_mdb_put(struct super_block *); extern int hfs_part_find(struct super_block *, sector_t *, sector_t *); /* string.c */ -extern struct dentry_operations hfs_dentry_operations; +extern const struct dentry_operations hfs_dentry_operations; extern int hfs_hash_dentry(struct dentry *, struct qstr *); extern int hfs_strcmp(const unsigned char *, unsigned int, diff --git a/fs/hfs/sysdep.c b/fs/hfs/sysdep.c index 5bf89ec01cd4..7478f5c219aa 100644 --- a/fs/hfs/sysdep.c +++ b/fs/hfs/sysdep.c @@ -31,7 +31,7 @@ static int hfs_revalidate_dentry(struct dentry *dentry, struct nameidata *nd) return 1; } -struct dentry_operations hfs_dentry_operations = +const struct dentry_operations hfs_dentry_operations = { .d_revalidate = hfs_revalidate_dentry, .d_hash = hfs_hash_dentry, diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index f027a905225f..5c10d803d9df 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -327,7 +327,7 @@ void hfsplus_file_truncate(struct inode *); /* inode.c */ extern const struct address_space_operations hfsplus_aops; extern const struct address_space_operations hfsplus_btree_aops; -extern struct dentry_operations hfsplus_dentry_operations; +extern const struct dentry_operations hfsplus_dentry_operations; void hfsplus_inode_read_fork(struct inode *, struct hfsplus_fork_raw *); void hfsplus_inode_write_fork(struct inode *, struct hfsplus_fork_raw *); diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c index f105ee9e1cc4..1bcf597c0562 100644 --- a/fs/hfsplus/inode.c +++ b/fs/hfsplus/inode.c @@ -137,7 +137,7 @@ const struct address_space_operations hfsplus_aops = { .writepages = hfsplus_writepages, }; -struct dentry_operations hfsplus_dentry_operations = { +const struct dentry_operations hfsplus_dentry_operations = { .d_hash = hfsplus_hash_dentry, .d_compare = hfsplus_compare_dentry, }; diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index 5c538e0ec14b..fe02ad4740e7 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -31,12 +31,12 @@ static inline struct hostfs_inode_info *HOSTFS_I(struct inode *inode) #define FILE_HOSTFS_I(file) HOSTFS_I((file)->f_path.dentry->d_inode) -int hostfs_d_delete(struct dentry *dentry) +static int hostfs_d_delete(struct dentry *dentry) { return 1; } -struct dentry_operations hostfs_dentry_ops = { +static const struct dentry_operations hostfs_dentry_ops = { .d_delete = hostfs_d_delete, }; diff --git a/fs/hpfs/dentry.c b/fs/hpfs/dentry.c index 08319126b2af..940d6d150bee 100644 --- a/fs/hpfs/dentry.c +++ b/fs/hpfs/dentry.c @@ -49,7 +49,7 @@ static int hpfs_compare_dentry(struct dentry *dentry, struct qstr *a, struct qst return 0; } -static struct dentry_operations hpfs_dentry_operations = { +static const struct dentry_operations hpfs_dentry_operations = { .d_hash = hpfs_hash_dentry, .d_compare = hpfs_compare_dentry, }; diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c index 6147ec3643a0..13d2eddd0692 100644 --- a/fs/isofs/inode.c +++ b/fs/isofs/inode.c @@ -114,7 +114,7 @@ static const struct super_operations isofs_sops = { }; -static struct dentry_operations isofs_dentry_ops[] = { +static const struct dentry_operations isofs_dentry_ops[] = { { .d_hash = isofs_hash, .d_compare = isofs_dentry_cmp, diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c index 07e9715b8658..9c590722d87e 100644 --- a/fs/ncpfs/dir.c +++ b/fs/ncpfs/dir.c @@ -79,7 +79,7 @@ static int ncp_hash_dentry(struct dentry *, struct qstr *); static int ncp_compare_dentry (struct dentry *, struct qstr *, struct qstr *); static int ncp_delete_dentry(struct dentry *); -static struct dentry_operations ncp_dentry_operations = +static const struct dentry_operations ncp_dentry_operations = { .d_revalidate = ncp_lookup_validate, .d_hash = ncp_hash_dentry, @@ -87,7 +87,7 @@ static struct dentry_operations ncp_dentry_operations = .d_delete = ncp_delete_dentry, }; -struct dentry_operations ncp_root_dentry_operations = +const struct dentry_operations ncp_root_dentry_operations = { .d_hash = ncp_hash_dentry, .d_compare = ncp_compare_dentry, diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index ad92461cbfc3..ae881ccd2f03 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -1136,7 +1136,7 @@ xattr_lookup_poison(struct dentry *dentry, struct qstr *q1, struct qstr *name) return 1; } -static struct dentry_operations xattr_lookup_poison_ops = { +static const struct dentry_operations xattr_lookup_poison_ops = { .d_compare = xattr_lookup_poison, }; diff --git a/fs/smbfs/dir.c b/fs/smbfs/dir.c index e7ddd0328ddc..3e4803b4427e 100644 --- a/fs/smbfs/dir.c +++ b/fs/smbfs/dir.c @@ -277,7 +277,7 @@ static int smb_hash_dentry(struct dentry *, struct qstr *); static int smb_compare_dentry(struct dentry *, struct qstr *, struct qstr *); static int smb_delete_dentry(struct dentry *); -static struct dentry_operations smbfs_dentry_operations = +static const struct dentry_operations smbfs_dentry_operations = { .d_revalidate = smb_lookup_validate, .d_hash = smb_hash_dentry, @@ -285,7 +285,7 @@ static struct dentry_operations smbfs_dentry_operations = .d_delete = smb_delete_dentry, }; -static struct dentry_operations smbfs_dentry_operations_case = +static const struct dentry_operations smbfs_dentry_operations_case = { .d_revalidate = smb_lookup_validate, .d_delete = smb_delete_dentry, diff --git a/fs/sysv/namei.c b/fs/sysv/namei.c index a1f1ef33e81c..33e047b59b8d 100644 --- a/fs/sysv/namei.c +++ b/fs/sysv/namei.c @@ -38,7 +38,7 @@ static int sysv_hash(struct dentry *dentry, struct qstr *qstr) return 0; } -struct dentry_operations sysv_dentry_operations = { +const struct dentry_operations sysv_dentry_operations = { .d_hash = sysv_hash, }; diff --git a/fs/sysv/sysv.h b/fs/sysv/sysv.h index 38ebe3f85b3d..5784a318c883 100644 --- a/fs/sysv/sysv.h +++ b/fs/sysv/sysv.h @@ -170,7 +170,7 @@ extern const struct file_operations sysv_file_operations; extern const struct file_operations sysv_dir_operations; extern const struct address_space_operations sysv_aops; extern const struct super_operations sysv_sops; -extern struct dentry_operations sysv_dentry_operations; +extern const struct dentry_operations sysv_dentry_operations; enum { diff --git a/include/linux/ncp_fs.h b/include/linux/ncp_fs.h index f69e66d151cc..30b06c893944 100644 --- a/include/linux/ncp_fs.h +++ b/include/linux/ncp_fs.h @@ -204,7 +204,7 @@ void ncp_update_inode2(struct inode *, struct ncp_entry_info *); /* linux/fs/ncpfs/dir.c */ extern const struct inode_operations ncp_dir_inode_operations; extern const struct file_operations ncp_dir_operations; -extern struct dentry_operations ncp_root_dentry_operations; +extern const struct dentry_operations ncp_root_dentry_operations; int ncp_conn_logged_in(struct super_block *); int ncp_date_dos2unix(__le16 time, __le16 date); void ncp_date_unix2dos(int unix_date, __le16 * time, __le16 * date); -- cgit v1.2.3-59-g8ed1b From 585d3bc06f4ca57f975a5a1f698f65a45ea66225 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Wed, 25 Feb 2009 10:44:19 +0100 Subject: fs: move bdev code out of buffer.c Move some block device related code out from buffer.c and put it in block_dev.c. I'm trying to move non-buffer_head code out of buffer.c Signed-off-by: Al Viro --- fs/block_dev.c | 146 ++++++++++++++++++++++++++++++++++++++++++++ fs/buffer.c | 145 ------------------------------------------- include/linux/buffer_head.h | 7 --- include/linux/fs.h | 7 +++ 4 files changed, 153 insertions(+), 152 deletions(-) (limited to 'include/linux') diff --git a/fs/block_dev.c b/fs/block_dev.c index b3c1efff5e1d..8c3c6899ccf3 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -18,6 +18,7 @@ #include #include #include +#include #include #include #include @@ -174,6 +175,151 @@ blkdev_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov, iov, offset, nr_segs, blkdev_get_blocks, NULL); } +/* + * Write out and wait upon all the dirty data associated with a block + * device via its mapping. Does not take the superblock lock. + */ +int sync_blockdev(struct block_device *bdev) +{ + int ret = 0; + + if (bdev) + ret = filemap_write_and_wait(bdev->bd_inode->i_mapping); + return ret; +} +EXPORT_SYMBOL(sync_blockdev); + +/* + * Write out and wait upon all dirty data associated with this + * device. Filesystem data as well as the underlying block + * device. Takes the superblock lock. + */ +int fsync_bdev(struct block_device *bdev) +{ + struct super_block *sb = get_super(bdev); + if (sb) { + int res = fsync_super(sb); + drop_super(sb); + return res; + } + return sync_blockdev(bdev); +} + +/** + * freeze_bdev -- lock a filesystem and force it into a consistent state + * @bdev: blockdevice to lock + * + * This takes the block device bd_mount_sem to make sure no new mounts + * happen on bdev until thaw_bdev() is called. + * If a superblock is found on this device, we take the s_umount semaphore + * on it to make sure nobody unmounts until the snapshot creation is done. + * The reference counter (bd_fsfreeze_count) guarantees that only the last + * unfreeze process can unfreeze the frozen filesystem actually when multiple + * freeze requests arrive simultaneously. It counts up in freeze_bdev() and + * count down in thaw_bdev(). When it becomes 0, thaw_bdev() will unfreeze + * actually. + */ +struct super_block *freeze_bdev(struct block_device *bdev) +{ + struct super_block *sb; + int error = 0; + + mutex_lock(&bdev->bd_fsfreeze_mutex); + if (bdev->bd_fsfreeze_count > 0) { + bdev->bd_fsfreeze_count++; + sb = get_super(bdev); + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return sb; + } + bdev->bd_fsfreeze_count++; + + down(&bdev->bd_mount_sem); + sb = get_super(bdev); + if (sb && !(sb->s_flags & MS_RDONLY)) { + sb->s_frozen = SB_FREEZE_WRITE; + smp_wmb(); + + __fsync_super(sb); + + sb->s_frozen = SB_FREEZE_TRANS; + smp_wmb(); + + sync_blockdev(sb->s_bdev); + + if (sb->s_op->freeze_fs) { + error = sb->s_op->freeze_fs(sb); + if (error) { + printk(KERN_ERR + "VFS:Filesystem freeze failed\n"); + sb->s_frozen = SB_UNFROZEN; + drop_super(sb); + up(&bdev->bd_mount_sem); + bdev->bd_fsfreeze_count--; + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return ERR_PTR(error); + } + } + } + + sync_blockdev(bdev); + mutex_unlock(&bdev->bd_fsfreeze_mutex); + + return sb; /* thaw_bdev releases s->s_umount and bd_mount_sem */ +} +EXPORT_SYMBOL(freeze_bdev); + +/** + * thaw_bdev -- unlock filesystem + * @bdev: blockdevice to unlock + * @sb: associated superblock + * + * Unlocks the filesystem and marks it writeable again after freeze_bdev(). + */ +int thaw_bdev(struct block_device *bdev, struct super_block *sb) +{ + int error = 0; + + mutex_lock(&bdev->bd_fsfreeze_mutex); + if (!bdev->bd_fsfreeze_count) { + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return -EINVAL; + } + + bdev->bd_fsfreeze_count--; + if (bdev->bd_fsfreeze_count > 0) { + if (sb) + drop_super(sb); + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return 0; + } + + if (sb) { + BUG_ON(sb->s_bdev != bdev); + if (!(sb->s_flags & MS_RDONLY)) { + if (sb->s_op->unfreeze_fs) { + error = sb->s_op->unfreeze_fs(sb); + if (error) { + printk(KERN_ERR + "VFS:Filesystem thaw failed\n"); + sb->s_frozen = SB_FREEZE_TRANS; + bdev->bd_fsfreeze_count++; + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return error; + } + } + sb->s_frozen = SB_UNFROZEN; + smp_wmb(); + wake_up(&sb->s_wait_unfrozen); + } + drop_super(sb); + } + + up(&bdev->bd_mount_sem); + mutex_unlock(&bdev->bd_fsfreeze_mutex); + return 0; +} +EXPORT_SYMBOL(thaw_bdev); + static int blkdev_writepage(struct page *page, struct writeback_control *wbc) { return block_write_full_page(page, blkdev_get_block, wbc); diff --git a/fs/buffer.c b/fs/buffer.c index 891e1c78e4f1..a2fd743d97cb 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -165,151 +165,6 @@ void end_buffer_write_sync(struct buffer_head *bh, int uptodate) put_bh(bh); } -/* - * Write out and wait upon all the dirty data associated with a block - * device via its mapping. Does not take the superblock lock. - */ -int sync_blockdev(struct block_device *bdev) -{ - int ret = 0; - - if (bdev) - ret = filemap_write_and_wait(bdev->bd_inode->i_mapping); - return ret; -} -EXPORT_SYMBOL(sync_blockdev); - -/* - * Write out and wait upon all dirty data associated with this - * device. Filesystem data as well as the underlying block - * device. Takes the superblock lock. - */ -int fsync_bdev(struct block_device *bdev) -{ - struct super_block *sb = get_super(bdev); - if (sb) { - int res = fsync_super(sb); - drop_super(sb); - return res; - } - return sync_blockdev(bdev); -} - -/** - * freeze_bdev -- lock a filesystem and force it into a consistent state - * @bdev: blockdevice to lock - * - * This takes the block device bd_mount_sem to make sure no new mounts - * happen on bdev until thaw_bdev() is called. - * If a superblock is found on this device, we take the s_umount semaphore - * on it to make sure nobody unmounts until the snapshot creation is done. - * The reference counter (bd_fsfreeze_count) guarantees that only the last - * unfreeze process can unfreeze the frozen filesystem actually when multiple - * freeze requests arrive simultaneously. It counts up in freeze_bdev() and - * count down in thaw_bdev(). When it becomes 0, thaw_bdev() will unfreeze - * actually. - */ -struct super_block *freeze_bdev(struct block_device *bdev) -{ - struct super_block *sb; - int error = 0; - - mutex_lock(&bdev->bd_fsfreeze_mutex); - if (bdev->bd_fsfreeze_count > 0) { - bdev->bd_fsfreeze_count++; - sb = get_super(bdev); - mutex_unlock(&bdev->bd_fsfreeze_mutex); - return sb; - } - bdev->bd_fsfreeze_count++; - - down(&bdev->bd_mount_sem); - sb = get_super(bdev); - if (sb && !(sb->s_flags & MS_RDONLY)) { - sb->s_frozen = SB_FREEZE_WRITE; - smp_wmb(); - - __fsync_super(sb); - - sb->s_frozen = SB_FREEZE_TRANS; - smp_wmb(); - - sync_blockdev(sb->s_bdev); - - if (sb->s_op->freeze_fs) { - error = sb->s_op->freeze_fs(sb); - if (error) { - printk(KERN_ERR - "VFS:Filesystem freeze failed\n"); - sb->s_frozen = SB_UNFROZEN; - drop_super(sb); - up(&bdev->bd_mount_sem); - bdev->bd_fsfreeze_count--; - mutex_unlock(&bdev->bd_fsfreeze_mutex); - return ERR_PTR(error); - } - } - } - - sync_blockdev(bdev); - mutex_unlock(&bdev->bd_fsfreeze_mutex); - - return sb; /* thaw_bdev releases s->s_umount and bd_mount_sem */ -} -EXPORT_SYMBOL(freeze_bdev); - -/** - * thaw_bdev -- unlock filesystem - * @bdev: blockdevice to unlock - * @sb: associated superblock - * - * Unlocks the filesystem and marks it writeable again after freeze_bdev(). - */ -int thaw_bdev(struct block_device *bdev, struct super_block *sb) -{ - int error = 0; - - mutex_lock(&bdev->bd_fsfreeze_mutex); - if (!bdev->bd_fsfreeze_count) { - mutex_unlock(&bdev->bd_fsfreeze_mutex); - return -EINVAL; - } - - bdev->bd_fsfreeze_count--; - if (bdev->bd_fsfreeze_count > 0) { - if (sb) - drop_super(sb); - mutex_unlock(&bdev->bd_fsfreeze_mutex); - return 0; - } - - if (sb) { - BUG_ON(sb->s_bdev != bdev); - if (!(sb->s_flags & MS_RDONLY)) { - if (sb->s_op->unfreeze_fs) { - error = sb->s_op->unfreeze_fs(sb); - if (error) { - printk(KERN_ERR - "VFS:Filesystem thaw failed\n"); - sb->s_frozen = SB_FREEZE_TRANS; - bdev->bd_fsfreeze_count++; - mutex_unlock(&bdev->bd_fsfreeze_mutex); - return error; - } - } - sb->s_frozen = SB_UNFROZEN; - smp_wmb(); - wake_up(&sb->s_wait_unfrozen); - } - drop_super(sb); - } - - up(&bdev->bd_mount_sem); - mutex_unlock(&bdev->bd_fsfreeze_mutex); - return 0; -} -EXPORT_SYMBOL(thaw_bdev); - /* * Various filesystems appear to want __find_get_block to be non-blocking. * But it's the page lock which protects the buffers. To get around this, diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h index bd7ac793be19..f19fd9045ea0 100644 --- a/include/linux/buffer_head.h +++ b/include/linux/buffer_head.h @@ -165,15 +165,8 @@ int sync_mapping_buffers(struct address_space *mapping); void unmap_underlying_metadata(struct block_device *bdev, sector_t block); void mark_buffer_async_write(struct buffer_head *bh); -void invalidate_bdev(struct block_device *); -int sync_blockdev(struct block_device *bdev); void __wait_on_buffer(struct buffer_head *); wait_queue_head_t *bh_waitq_head(struct buffer_head *bh); -int fsync_bdev(struct block_device *); -struct super_block *freeze_bdev(struct block_device *); -int thaw_bdev(struct block_device *, struct super_block *); -int fsync_super(struct super_block *); -int fsync_no_super(struct block_device *); struct buffer_head *__find_get_block(struct block_device *bdev, sector_t block, unsigned size); struct buffer_head *__getblk(struct block_device *bdev, sector_t block, diff --git a/include/linux/fs.h b/include/linux/fs.h index 5f74d616cd7d..c2c4454a268a 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1874,6 +1874,13 @@ extern void bd_set_size(struct block_device *, loff_t size); extern void bd_forget(struct inode *inode); extern void bdput(struct block_device *); extern struct block_device *open_by_devnum(dev_t, fmode_t); +extern void invalidate_bdev(struct block_device *); +extern int sync_blockdev(struct block_device *bdev); +extern struct super_block *freeze_bdev(struct block_device *); +extern int thaw_bdev(struct block_device *bdev, struct super_block *sb); +extern int fsync_bdev(struct block_device *); +extern int fsync_super(struct super_block *); +extern int fsync_no_super(struct block_device *); #else static inline void bd_forget(struct inode *inode) {} #endif -- cgit v1.2.3-59-g8ed1b From a3ec947c85ec339884b30ef6a08133e9311fdae1 Mon Sep 17 00:00:00 2001 From: Sukadev Bhattiprolu Date: Wed, 4 Mar 2009 12:06:34 -0800 Subject: vfs: simple_set_mnt() should return void simple_set_mnt() is defined as returning 'int' but always returns 0. Callers assume simple_set_mnt() never fails and don't properly cleanup if it were to _ever_ fail. For instance, get_sb_single() and get_sb_nodev() should: up_write(sb->s_unmount); deactivate_super(sb); if simple_set_mnt() fails. Since simple_set_mnt() never fails, would be cleaner if it did not return anything. [akpm@linux-foundation.org: fix build] Signed-off-by: Sukadev Bhattiprolu Acked-by: Serge Hallyn Cc: Al Viro Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- drivers/mtd/mtdsuper.c | 7 +++++-- fs/9p/vfs_super.c | 5 +++-- fs/cifs/cifsfs.c | 3 ++- fs/devpts/inode.c | 3 ++- fs/libfs.c | 3 ++- fs/namespace.c | 3 +-- fs/proc/root.c | 3 ++- fs/super.c | 9 ++++++--- fs/ubifs/super.c | 3 ++- include/linux/fs.h | 2 +- kernel/cgroup.c | 3 ++- 11 files changed, 28 insertions(+), 16 deletions(-) (limited to 'include/linux') diff --git a/drivers/mtd/mtdsuper.c b/drivers/mtd/mtdsuper.c index 00d46e137b2a..92285d0089c2 100644 --- a/drivers/mtd/mtdsuper.c +++ b/drivers/mtd/mtdsuper.c @@ -81,13 +81,16 @@ static int get_sb_mtd_aux(struct file_system_type *fs_type, int flags, /* go */ sb->s_flags |= MS_ACTIVE; - return simple_set_mnt(mnt, sb); + simple_set_mnt(mnt, sb); + + return 0; /* new mountpoint for an already mounted superblock */ already_mounted: DEBUG(1, "MTDSB: Device %d (\"%s\") is already mounted\n", mtd->index, mtd->name); - ret = simple_set_mnt(mnt, sb); + simple_set_mnt(mnt, sb); + ret = 0; goto out_put; out_error: diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c index 93212e40221a..5f8ab8adb5f5 100644 --- a/fs/9p/vfs_super.c +++ b/fs/9p/vfs_super.c @@ -168,8 +168,9 @@ static int v9fs_get_sb(struct file_system_type *fs_type, int flags, p9stat_free(st); kfree(st); -P9_DPRINTK(P9_DEBUG_VFS, " return simple set mount\n"); - return simple_set_mnt(mnt, sb); +P9_DPRINTK(P9_DEBUG_VFS, " simple set mount, return 0\n"); + simple_set_mnt(mnt, sb); + return 0; release_sb: if (sb) { diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 13ea53251dcf..38491fd3871d 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -606,7 +606,8 @@ cifs_get_sb(struct file_system_type *fs_type, return rc; } sb->s_flags |= MS_ACTIVE; - return simple_set_mnt(mnt, sb); + simple_set_mnt(mnt, sb); + return 0; } static ssize_t cifs_file_aio_write(struct kiocb *iocb, const struct iovec *iov, diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c index 140b43144cd8..b0a76340a4cd 100644 --- a/fs/devpts/inode.c +++ b/fs/devpts/inode.c @@ -454,7 +454,8 @@ static int get_init_pts_sb(struct file_system_type *fs_type, int flags, s->s_flags |= MS_ACTIVE; } do_remount_sb(s, flags, data, 0); - return simple_set_mnt(mnt, s); + simple_set_mnt(mnt, s); + return 0; } /* diff --git a/fs/libfs.c b/fs/libfs.c index ec600bd33e75..4910a36f516e 100644 --- a/fs/libfs.c +++ b/fs/libfs.c @@ -242,7 +242,8 @@ int get_sb_pseudo(struct file_system_type *fs_type, char *name, d_instantiate(dentry, root); s->s_root = dentry; s->s_flags |= MS_ACTIVE; - return simple_set_mnt(mnt, s); + simple_set_mnt(mnt, s); + return 0; Enomem: up_write(&s->s_umount); diff --git a/fs/namespace.c b/fs/namespace.c index 06f8e63f6cb1..2432ca6bb223 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -397,11 +397,10 @@ static void __mnt_unmake_readonly(struct vfsmount *mnt) spin_unlock(&vfsmount_lock); } -int simple_set_mnt(struct vfsmount *mnt, struct super_block *sb) +void simple_set_mnt(struct vfsmount *mnt, struct super_block *sb) { mnt->mnt_sb = sb; mnt->mnt_root = dget(sb->s_root); - return 0; } EXPORT_SYMBOL(simple_set_mnt); diff --git a/fs/proc/root.c b/fs/proc/root.c index f6299a25594e..1e15a2b176e8 100644 --- a/fs/proc/root.c +++ b/fs/proc/root.c @@ -83,7 +83,8 @@ static int proc_get_sb(struct file_system_type *fs_type, ns->proc_mnt = mnt; } - return simple_set_mnt(mnt, sb); + simple_set_mnt(mnt, sb); + return 0; } static void proc_kill_sb(struct super_block *sb) diff --git a/fs/super.c b/fs/super.c index 6ce501447ada..e512fab64c93 100644 --- a/fs/super.c +++ b/fs/super.c @@ -831,7 +831,8 @@ int get_sb_bdev(struct file_system_type *fs_type, bdev->bd_super = s; } - return simple_set_mnt(mnt, s); + simple_set_mnt(mnt, s); + return 0; error_s: error = PTR_ERR(s); @@ -877,7 +878,8 @@ int get_sb_nodev(struct file_system_type *fs_type, return error; } s->s_flags |= MS_ACTIVE; - return simple_set_mnt(mnt, s); + simple_set_mnt(mnt, s); + return 0; } EXPORT_SYMBOL(get_sb_nodev); @@ -909,7 +911,8 @@ int get_sb_single(struct file_system_type *fs_type, s->s_flags |= MS_ACTIVE; } do_remount_sb(s, flags, data, 0); - return simple_set_mnt(mnt, s); + simple_set_mnt(mnt, s); + return 0; } EXPORT_SYMBOL(get_sb_single); diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 1182b66a5491..c5c98355459a 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -2034,7 +2034,8 @@ static int ubifs_get_sb(struct file_system_type *fs_type, int flags, /* 'fill_super()' opens ubi again so we must close it here */ ubi_close_volume(ubi); - return simple_set_mnt(mnt, sb); + simple_set_mnt(mnt, sb); + return 0; out_deact: up_write(&sb->s_umount); diff --git a/include/linux/fs.h b/include/linux/fs.h index c2c4454a268a..a7d73914a9f7 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1719,7 +1719,7 @@ struct super_block *sget(struct file_system_type *type, extern int get_sb_pseudo(struct file_system_type *, char *, const struct super_operations *ops, unsigned long, struct vfsmount *mnt); -extern int simple_set_mnt(struct vfsmount *mnt, struct super_block *sb); +extern void simple_set_mnt(struct vfsmount *mnt, struct super_block *sb); int __put_super_and_need_restart(struct super_block *sb); /* Alas, no aliases. Too much hassle with bringing module.h everywhere */ diff --git a/kernel/cgroup.c b/kernel/cgroup.c index b01100ebd074..c500ca7239b2 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1071,7 +1071,8 @@ static int cgroup_get_sb(struct file_system_type *fs_type, mutex_unlock(&cgroup_mutex); } - return simple_set_mnt(mnt, sb); + simple_set_mnt(mnt, sb); + return 0; free_cg_links: free_cg_links(&tmp_cg_links); -- cgit v1.2.3-59-g8ed1b From d2dd14ac1847082d4bb955619e86ed315c0ecd20 Mon Sep 17 00:00:00 2001 From: Jean Delvare Date: Sat, 28 Mar 2009 21:34:41 +0100 Subject: i2c-nforce2: Add support for MCP67, MCP73, MCP78S and MCP79 The MCP78S and MCP79 appear to be compatible with the previous nForce chips as far as the SMBus controller is concerned. The MCP67 and MCP73 were not tested yet but I'd be very surprised if they weren't compatible too. Signed-off-by: Jean Delvare Cc: Oleg Ryjkov Cc: Malcolm Lalkaka Cc: Zbigniew Luszpinski --- Documentation/i2c/busses/i2c-nforce2 | 12 ++++++++---- drivers/i2c/busses/i2c-nforce2.c | 12 ++++++++++-- include/linux/pci_ids.h | 4 ++++ 3 files changed, 22 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/Documentation/i2c/busses/i2c-nforce2 b/Documentation/i2c/busses/i2c-nforce2 index fae3495bcbaf..9698c396b830 100644 --- a/Documentation/i2c/busses/i2c-nforce2 +++ b/Documentation/i2c/busses/i2c-nforce2 @@ -7,10 +7,14 @@ Supported adapters: * nForce3 250Gb MCP 10de:00E4 * nForce4 MCP 10de:0052 * nForce4 MCP-04 10de:0034 - * nForce4 MCP51 10de:0264 - * nForce4 MCP55 10de:0368 - * nForce4 MCP61 10de:03EB - * nForce4 MCP65 10de:0446 + * nForce MCP51 10de:0264 + * nForce MCP55 10de:0368 + * nForce MCP61 10de:03EB + * nForce MCP65 10de:0446 + * nForce MCP67 10de:0542 + * nForce MCP73 10de:07D8 + * nForce MCP78S 10de:0752 + * nForce MCP79 10de:0AA2 Datasheet: not publicly available, but seems to be similar to the AMD-8111 SMBus 2.0 adapter. diff --git a/drivers/i2c/busses/i2c-nforce2.c b/drivers/i2c/busses/i2c-nforce2.c index 05af6cd7f270..2ff4683703a8 100644 --- a/drivers/i2c/busses/i2c-nforce2.c +++ b/drivers/i2c/busses/i2c-nforce2.c @@ -31,10 +31,14 @@ nForce3 250Gb MCP 00E4 nForce4 MCP 0052 nForce4 MCP-04 0034 - nForce4 MCP51 0264 - nForce4 MCP55 0368 + nForce MCP51 0264 + nForce MCP55 0368 nForce MCP61 03EB nForce MCP65 0446 + nForce MCP67 0542 + nForce MCP73 07D8 + nForce MCP78S 0752 + nForce MCP79 0AA2 This driver supports the 2 SMBuses that are included in the MCP of the nForce2/3/4/5xx chipsets. @@ -315,6 +319,10 @@ static struct pci_device_id nforce2_ids[] = { { PCI_DEVICE(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP55_SMBUS) }, { PCI_DEVICE(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP61_SMBUS) }, { PCI_DEVICE(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP65_SMBUS) }, + { PCI_DEVICE(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP67_SMBUS) }, + { PCI_DEVICE(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP73_SMBUS) }, + { PCI_DEVICE(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP78S_SMBUS) }, + { PCI_DEVICE(PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_NFORCE_MCP79_SMBUS) }, { 0 } }; diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index 05dfa7c4fb64..5109fecde284 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -1237,6 +1237,7 @@ #define PCI_DEVICE_ID_NVIDIA_NVENET_21 0x0451 #define PCI_DEVICE_ID_NVIDIA_NVENET_22 0x0452 #define PCI_DEVICE_ID_NVIDIA_NVENET_23 0x0453 +#define PCI_DEVICE_ID_NVIDIA_NFORCE_MCP67_SMBUS 0x0542 #define PCI_DEVICE_ID_NVIDIA_NVENET_24 0x054C #define PCI_DEVICE_ID_NVIDIA_NVENET_25 0x054D #define PCI_DEVICE_ID_NVIDIA_NVENET_26 0x054E @@ -1247,11 +1248,14 @@ #define PCI_DEVICE_ID_NVIDIA_NVENET_31 0x07DF #define PCI_DEVICE_ID_NVIDIA_NFORCE_MCP67_IDE 0x0560 #define PCI_DEVICE_ID_NVIDIA_NFORCE_MCP73_IDE 0x056C +#define PCI_DEVICE_ID_NVIDIA_NFORCE_MCP78S_SMBUS 0x0752 #define PCI_DEVICE_ID_NVIDIA_NFORCE_MCP77_IDE 0x0759 #define PCI_DEVICE_ID_NVIDIA_NVENET_32 0x0760 #define PCI_DEVICE_ID_NVIDIA_NVENET_33 0x0761 #define PCI_DEVICE_ID_NVIDIA_NVENET_34 0x0762 #define PCI_DEVICE_ID_NVIDIA_NVENET_35 0x0763 +#define PCI_DEVICE_ID_NVIDIA_NFORCE_MCP73_SMBUS 0x07D8 +#define PCI_DEVICE_ID_NVIDIA_NFORCE_MCP79_SMBUS 0x0AA2 #define PCI_DEVICE_ID_NVIDIA_NVENET_36 0x0AB0 #define PCI_DEVICE_ID_NVIDIA_NVENET_37 0x0AB1 #define PCI_DEVICE_ID_NVIDIA_NVENET_38 0x0AB2 -- cgit v1.2.3-59-g8ed1b From eff9ec95efaaf6b12d230f0ea7d3c295d3bc9d57 Mon Sep 17 00:00:00 2001 From: Marco Aurelio da Costa Date: Sat, 28 Mar 2009 21:34:44 +0100 Subject: i2c-algo-pca: Add PCA9665 support Add support for the PCA9665 I2C controller. Signed-off-by: Wolfram Sang Signed-off-by: Jean Delvare --- drivers/i2c/algos/i2c-algo-pca.c | 180 +++++++++++++++++++++++++++++++--- drivers/i2c/busses/Kconfig | 8 +- drivers/i2c/busses/i2c-pca-isa.c | 14 ++- drivers/i2c/busses/i2c-pca-platform.c | 9 +- include/linux/i2c-algo-pca.h | 33 ++++++- 5 files changed, 216 insertions(+), 28 deletions(-) (limited to 'include/linux') diff --git a/drivers/i2c/algos/i2c-algo-pca.c b/drivers/i2c/algos/i2c-algo-pca.c index 943d70ee5d59..a8e51bd1a4f5 100644 --- a/drivers/i2c/algos/i2c-algo-pca.c +++ b/drivers/i2c/algos/i2c-algo-pca.c @@ -46,6 +46,14 @@ static int i2c_debug; #define pca_wait(adap) adap->wait_for_completion(adap->data) #define pca_reset(adap) adap->reset_chip(adap->data) +static void pca9665_reset(void *pd) +{ + struct i2c_algo_pca_data *adap = pd; + pca_outw(adap, I2C_PCA_INDPTR, I2C_PCA_IPRESET); + pca_outw(adap, I2C_PCA_IND, 0xA5); + pca_outw(adap, I2C_PCA_IND, 0x5A); +} + /* * Generate a start condition on the i2c bus. * @@ -333,27 +341,171 @@ static const struct i2c_algorithm pca_algo = { .functionality = pca_func, }; -static int pca_init(struct i2c_adapter *adap) +static unsigned int pca_probe_chip(struct i2c_adapter *adap) { - static int freqs[] = {330,288,217,146,88,59,44,36}; - int clock; struct i2c_algo_pca_data *pca_data = adap->algo_data; - - if (pca_data->i2c_clock > 7) { - printk(KERN_WARNING "%s: Invalid I2C clock speed selected. Trying default.\n", - adap->name); - pca_data->i2c_clock = I2C_PCA_CON_59kHz; + /* The trick here is to check if there is an indirect register + * available. If there is one, we will read the value we first + * wrote on I2C_PCA_IADR. Otherwise, we will read the last value + * we wrote on I2C_PCA_ADR + */ + pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_IADR); + pca_outw(pca_data, I2C_PCA_IND, 0xAA); + pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_ITO); + pca_outw(pca_data, I2C_PCA_IND, 0x00); + pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_IADR); + if (pca_inw(pca_data, I2C_PCA_IND) == 0xAA) { + printk(KERN_INFO "%s: PCA9665 detected.\n", adap->name); + return I2C_PCA_CHIP_9665; + } else { + printk(KERN_INFO "%s: PCA9564 detected.\n", adap->name); + return I2C_PCA_CHIP_9564; } +} + +static int pca_init(struct i2c_adapter *adap) +{ + struct i2c_algo_pca_data *pca_data = adap->algo_data; adap->algo = &pca_algo; - pca_reset(pca_data); + if (pca_probe_chip(adap) == I2C_PCA_CHIP_9564) { + static int freqs[] = {330, 288, 217, 146, 88, 59, 44, 36}; + int clock; + + if (pca_data->i2c_clock > 7) { + switch (pca_data->i2c_clock) { + case 330000: + pca_data->i2c_clock = I2C_PCA_CON_330kHz; + break; + case 288000: + pca_data->i2c_clock = I2C_PCA_CON_288kHz; + break; + case 217000: + pca_data->i2c_clock = I2C_PCA_CON_217kHz; + break; + case 146000: + pca_data->i2c_clock = I2C_PCA_CON_146kHz; + break; + case 88000: + pca_data->i2c_clock = I2C_PCA_CON_88kHz; + break; + case 59000: + pca_data->i2c_clock = I2C_PCA_CON_59kHz; + break; + case 44000: + pca_data->i2c_clock = I2C_PCA_CON_44kHz; + break; + case 36000: + pca_data->i2c_clock = I2C_PCA_CON_36kHz; + break; + default: + printk(KERN_WARNING + "%s: Invalid I2C clock speed selected." + " Using default 59kHz.\n", adap->name); + pca_data->i2c_clock = I2C_PCA_CON_59kHz; + } + } else { + printk(KERN_WARNING "%s: " + "Choosing the clock frequency based on " + "index is deprecated." + " Use the nominal frequency.\n", adap->name); + } + + pca_reset(pca_data); + + clock = pca_clock(pca_data); + printk(KERN_INFO "%s: Clock frequency is %dkHz\n", + adap->name, freqs[clock]); + + pca_set_con(pca_data, I2C_PCA_CON_ENSIO | clock); + } else { + int clock; + int mode; + int tlow, thi; + /* Values can be found on PCA9665 datasheet section 7.3.2.6 */ + int min_tlow, min_thi; + /* These values are the maximum raise and fall values allowed + * by the I2C operation mode (Standard, Fast or Fast+) + * They are used (added) below to calculate the clock dividers + * of PCA9665. Note that they are slightly different of the + * real maximum, to allow the change on mode exactly on the + * maximum clock rate for each mode + */ + int raise_fall_time; + + struct i2c_algo_pca_data *pca_data = adap->algo_data; + + /* Ignore the reset function from the module, + * we can use the parallel bus reset + */ + pca_data->reset_chip = pca9665_reset; + + if (pca_data->i2c_clock > 1265800) { + printk(KERN_WARNING "%s: I2C clock speed too high." + " Using 1265.8kHz.\n", adap->name); + pca_data->i2c_clock = 1265800; + } + + if (pca_data->i2c_clock < 60300) { + printk(KERN_WARNING "%s: I2C clock speed too low." + " Using 60.3kHz.\n", adap->name); + pca_data->i2c_clock = 60300; + } + + /* To avoid integer overflow, use clock/100 for calculations */ + clock = pca_clock(pca_data) / 100; + + if (pca_data->i2c_clock > 10000) { + mode = I2C_PCA_MODE_TURBO; + min_tlow = 14; + min_thi = 5; + raise_fall_time = 22; /* Raise 11e-8s, Fall 11e-8s */ + } else if (pca_data->i2c_clock > 4000) { + mode = I2C_PCA_MODE_FASTP; + min_tlow = 17; + min_thi = 9; + raise_fall_time = 22; /* Raise 11e-8s, Fall 11e-8s */ + } else if (pca_data->i2c_clock > 1000) { + mode = I2C_PCA_MODE_FAST; + min_tlow = 44; + min_thi = 20; + raise_fall_time = 58; /* Raise 29e-8s, Fall 29e-8s */ + } else { + mode = I2C_PCA_MODE_STD; + min_tlow = 157; + min_thi = 134; + raise_fall_time = 127; /* Raise 29e-8s, Fall 98e-8s */ + } + + /* The minimum clock that respects the thi/tlow = 134/157 is + * 64800 Hz. Below that, we have to fix the tlow to 255 and + * calculate the thi factor. + */ + if (clock < 648) { + tlow = 255; + thi = 1000000 - clock * raise_fall_time; + thi /= (I2C_PCA_OSC_PER * clock) - tlow; + } else { + tlow = (1000000 - clock * raise_fall_time) * min_tlow; + tlow /= I2C_PCA_OSC_PER * clock * (min_thi + min_tlow); + thi = tlow * min_thi / min_tlow; + } + + pca_reset(pca_data); - clock = pca_clock(pca_data); - printk(KERN_INFO "%s: Clock frequency is %dkHz\n", adap->name, - freqs[clock]); + printk(KERN_INFO + "%s: Clock frequency is %dHz\n", adap->name, clock * 100); - pca_set_con(pca_data, I2C_PCA_CON_ENSIO | clock); + pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_IMODE); + pca_outw(pca_data, I2C_PCA_IND, mode); + pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_ISCLL); + pca_outw(pca_data, I2C_PCA_IND, tlow); + pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_ISCLH); + pca_outw(pca_data, I2C_PCA_IND, thi); + + pca_set_con(pca_data, I2C_PCA_CON_ENSIO); + } udelay(500); /* 500 us for oscilator to stabilise */ return 0; @@ -388,7 +540,7 @@ EXPORT_SYMBOL(i2c_pca_add_numbered_bus); MODULE_AUTHOR("Ian Campbell , " "Wolfram Sang "); -MODULE_DESCRIPTION("I2C-Bus PCA9564 algorithm"); +MODULE_DESCRIPTION("I2C-Bus PCA9564/PCA9665 algorithm"); MODULE_LICENSE("GPL"); module_param(i2c_debug, int, 0); diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig index 7f95905bbb9d..68650643d116 100644 --- a/drivers/i2c/busses/Kconfig +++ b/drivers/i2c/busses/Kconfig @@ -617,12 +617,12 @@ config I2C_ELEKTOR will be called i2c-elektor. config I2C_PCA_ISA - tristate "PCA9564 on an ISA bus" + tristate "PCA9564/PCA9665 on an ISA bus" depends on ISA select I2C_ALGOPCA default n help - This driver supports ISA boards using the Philips PCA9564 + This driver supports ISA boards using the Philips PCA9564/PCA9665 parallel bus to I2C bus controller. This driver can also be built as a module. If so, the module @@ -634,11 +634,11 @@ config I2C_PCA_ISA time). If unsure, say N. config I2C_PCA_PLATFORM - tristate "PCA9564 as platform device" + tristate "PCA9564/PCA9665 as platform device" select I2C_ALGOPCA default n help - This driver supports a memory mapped Philips PCA9564 + This driver supports a memory mapped Philips PCA9564/PCA9665 parallel bus to I2C bus controller. This driver can also be built as a module. If so, the module diff --git a/drivers/i2c/busses/i2c-pca-isa.c b/drivers/i2c/busses/i2c-pca-isa.c index c420a7c0f3e4..0cc8017b3f64 100644 --- a/drivers/i2c/busses/i2c-pca-isa.c +++ b/drivers/i2c/busses/i2c-pca-isa.c @@ -41,7 +41,7 @@ static int irq = -1; /* Data sheet recommends 59kHz for 100kHz operation due to variation * in the actual clock rate */ -static int clock = I2C_PCA_CON_59kHz; +static int clock = 59000; static wait_queue_head_t pca_wait; @@ -103,7 +103,7 @@ static struct i2c_algo_pca_data pca_isa_data = { static struct i2c_adapter pca_isa_ops = { .owner = THIS_MODULE, .algo_data = &pca_isa_data, - .name = "PCA9564 ISA Adapter", + .name = "PCA9564/PCA9665 ISA Adapter", .timeout = 100, }; @@ -196,7 +196,7 @@ static void __exit pca_isa_exit(void) } MODULE_AUTHOR("Ian Campbell "); -MODULE_DESCRIPTION("ISA base PCA9564 driver"); +MODULE_DESCRIPTION("ISA base PCA9564/PCA9665 driver"); MODULE_LICENSE("GPL"); module_param(base, ulong, 0); @@ -205,7 +205,13 @@ MODULE_PARM_DESC(base, "I/O base address"); module_param(irq, int, 0); MODULE_PARM_DESC(irq, "IRQ"); module_param(clock, int, 0); -MODULE_PARM_DESC(clock, "Clock rate as described in table 1 of PCA9564 datasheet"); +MODULE_PARM_DESC(clock, "Clock rate in hertz.\n\t\t" + "For PCA9564: 330000,288000,217000,146000," + "88000,59000,44000,36000\n" + "\t\tFor PCA9665:\tStandard: 60300 - 100099\n" + "\t\t\t\tFast: 100100 - 400099\n" + "\t\t\t\tFast+: 400100 - 10000099\n" + "\t\t\t\tTurbo: Up to 1265800"); module_init(pca_isa_init); module_exit(pca_isa_exit); diff --git a/drivers/i2c/busses/i2c-pca-platform.c b/drivers/i2c/busses/i2c-pca-platform.c index 6bb15ad0a6b6..51d179bbddf9 100644 --- a/drivers/i2c/busses/i2c-pca-platform.c +++ b/drivers/i2c/busses/i2c-pca-platform.c @@ -172,8 +172,9 @@ static int __devinit i2c_pca_pf_probe(struct platform_device *pdev) i2c->adap.nr = pdev->id >= 0 ? pdev->id : 0; i2c->adap.owner = THIS_MODULE; - snprintf(i2c->adap.name, sizeof(i2c->adap.name), "PCA9564 at 0x%08lx", - (unsigned long) res->start); + snprintf(i2c->adap.name, sizeof(i2c->adap.name), + "PCA9564/PCA9665 at 0x%08lx", + (unsigned long) res->start); i2c->adap.algo_data = &i2c->algo_data; i2c->adap.dev.parent = &pdev->dev; i2c->adap.timeout = platform_data->timeout; @@ -246,7 +247,7 @@ e_remap: e_alloc: release_mem_region(res->start, res_len(res)); e_print: - printk(KERN_ERR "Registering PCA9564 FAILED! (%d)\n", ret); + printk(KERN_ERR "Registering PCA9564/PCA9665 FAILED! (%d)\n", ret); return ret; } @@ -290,7 +291,7 @@ static void __exit i2c_pca_pf_exit(void) } MODULE_AUTHOR("Wolfram Sang "); -MODULE_DESCRIPTION("I2C-PCA9564 platform driver"); +MODULE_DESCRIPTION("I2C-PCA9564/PCA9665 platform driver"); MODULE_LICENSE("GPL"); module_init(i2c_pca_pf_init); diff --git a/include/linux/i2c-algo-pca.h b/include/linux/i2c-algo-pca.h index adcb3dc7ac26..1364d62e2fbe 100644 --- a/include/linux/i2c-algo-pca.h +++ b/include/linux/i2c-algo-pca.h @@ -1,7 +1,14 @@ #ifndef _LINUX_I2C_ALGO_PCA_H #define _LINUX_I2C_ALGO_PCA_H -/* Clock speeds for the bus */ +/* Chips known to the pca algo */ +#define I2C_PCA_CHIP_9564 0x00 +#define I2C_PCA_CHIP_9665 0x01 + +/* Internal period for PCA9665 oscilator */ +#define I2C_PCA_OSC_PER 3 /* e10-8s */ + +/* Clock speeds for the bus for PCA9564*/ #define I2C_PCA_CON_330kHz 0x00 #define I2C_PCA_CON_288kHz 0x01 #define I2C_PCA_CON_217kHz 0x02 @@ -18,6 +25,26 @@ #define I2C_PCA_ADR 0x02 /* OWN ADR Read/Write */ #define I2C_PCA_CON 0x03 /* CONTROL Read/Write */ +/* PCA9665 registers */ +#define I2C_PCA_INDPTR 0x00 /* INDIRECT Pointer Write Only */ +#define I2C_PCA_IND 0x02 /* INDIRECT Read/Write */ + +/* PCA9665 indirect registers */ +#define I2C_PCA_ICOUNT 0x00 /* Byte Count for buffered mode */ +#define I2C_PCA_IADR 0x01 /* OWN ADR */ +#define I2C_PCA_ISCLL 0x02 /* SCL LOW period */ +#define I2C_PCA_ISCLH 0x03 /* SCL HIGH period */ +#define I2C_PCA_ITO 0x04 /* TIMEOUT */ +#define I2C_PCA_IPRESET 0x05 /* Parallel bus reset */ +#define I2C_PCA_IMODE 0x06 /* I2C Bus mode */ + +/* PCA9665 I2C bus mode */ +#define I2C_PCA_MODE_STD 0x00 /* Standard mode */ +#define I2C_PCA_MODE_FAST 0x01 /* Fast mode */ +#define I2C_PCA_MODE_FASTP 0x02 /* Fast Plus mode */ +#define I2C_PCA_MODE_TURBO 0x03 /* Turbo mode */ + + #define I2C_PCA_CON_AA 0x80 /* Assert Acknowledge */ #define I2C_PCA_CON_ENSIO 0x40 /* Enable */ #define I2C_PCA_CON_STA 0x20 /* Start */ @@ -31,7 +58,9 @@ struct i2c_algo_pca_data { int (*read_byte) (void *data, int reg); int (*wait_for_completion) (void *data); void (*reset_chip) (void *data); - /* i2c_clock values are defined in linux/i2c-algo-pca.h */ + /* For PCA9564, use one of the predefined frequencies: + * 330000, 288000, 217000, 146000, 88000, 59000, 44000, 36000 + * For PCA9665, use the frequency you want here. */ unsigned int i2c_clock; }; -- cgit v1.2.3-59-g8ed1b From 8e99ada8deaa9033600cd2c7d0a9366b0e99ab68 Mon Sep 17 00:00:00 2001 From: Wolfram Sang Date: Sat, 28 Mar 2009 21:34:45 +0100 Subject: i2c-algo-pca: Rework waiting for a free bus Waiting for a free bus now accepts the timeout value in jiffies and does proper checking using time_before. Signed-off-by: Wolfram Sang Signed-off-by: Jean Delvare --- arch/sh/boards/board-sh7785lcr.c | 2 +- drivers/i2c/algos/i2c-algo-pca.c | 17 ++++++++++------- drivers/i2c/busses/i2c-pca-isa.c | 2 +- include/linux/i2c-pca-platform.h | 2 +- 4 files changed, 13 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/arch/sh/boards/board-sh7785lcr.c b/arch/sh/boards/board-sh7785lcr.c index 94c0296bc35d..6f94f17adc46 100644 --- a/arch/sh/boards/board-sh7785lcr.c +++ b/arch/sh/boards/board-sh7785lcr.c @@ -229,7 +229,7 @@ static struct resource i2c_resources[] = { static struct i2c_pca9564_pf_platform_data i2c_platform_data = { .gpio = 0, .i2c_clock_speed = I2C_PCA_CON_330kHz, - .timeout = 100, + .timeout = HZ, }; static struct platform_device i2c_device = { diff --git a/drivers/i2c/algos/i2c-algo-pca.c b/drivers/i2c/algos/i2c-algo-pca.c index a8e51bd1a4f5..9e134fad7bda 100644 --- a/drivers/i2c/algos/i2c-algo-pca.c +++ b/drivers/i2c/algos/i2c-algo-pca.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include @@ -186,14 +187,16 @@ static int pca_xfer(struct i2c_adapter *i2c_adap, int numbytes = 0; int state; int ret; - int timeout = i2c_adap->timeout; + unsigned long timeout = jiffies + i2c_adap->timeout; - while ((state = pca_status(adap)) != 0xf8 && timeout--) { - msleep(10); - } - if (state != 0xf8) { - dev_dbg(&i2c_adap->dev, "bus is not idle. status is %#04x\n", state); - return -EAGAIN; + while (pca_status(adap) != 0xf8) { + if (time_before(jiffies, timeout)) { + msleep(10); + } else { + dev_dbg(&i2c_adap->dev, "bus is not idle. status is " + "%#04x\n", state); + return -EAGAIN; + } } DEB1("{{{ XFER %d messages\n", num); diff --git a/drivers/i2c/busses/i2c-pca-isa.c b/drivers/i2c/busses/i2c-pca-isa.c index 0cc8017b3f64..b9403fdfb6d8 100644 --- a/drivers/i2c/busses/i2c-pca-isa.c +++ b/drivers/i2c/busses/i2c-pca-isa.c @@ -104,7 +104,7 @@ static struct i2c_adapter pca_isa_ops = { .owner = THIS_MODULE, .algo_data = &pca_isa_data, .name = "PCA9564/PCA9665 ISA Adapter", - .timeout = 100, + .timeout = HZ, }; static int __devinit pca_isa_match(struct device *dev, unsigned int id) diff --git a/include/linux/i2c-pca-platform.h b/include/linux/i2c-pca-platform.h index 3d191873f2d1..aba33759dec4 100644 --- a/include/linux/i2c-pca-platform.h +++ b/include/linux/i2c-pca-platform.h @@ -6,7 +6,7 @@ struct i2c_pca9564_pf_platform_data { * not supplied (negative value), but it * cannot exit some error conditions then */ int i2c_clock_speed; /* values are defined in linux/i2c-algo-pca.h */ - int timeout; /* timeout = this value * 10us */ + int timeout; /* timeout in jiffies */ }; #endif /* I2C_PCA9564_PLATFORM_H */ -- cgit v1.2.3-59-g8ed1b From 506a8b6c27cb08998dc13069fbdf6eb7ec748b99 Mon Sep 17 00:00:00 2001 From: Flavio Leitner Date: Sat, 28 Mar 2009 21:34:46 +0100 Subject: i2c-piix4: Add support for the Broadcom HT1100 chipset Add support for the Broadcom HT1100 LD chipset (SMBus function.) Signed-off-by: Flavio Leitner Signed-off-by: Jean Delvare --- Documentation/i2c/busses/i2c-piix4 | 2 +- drivers/i2c/busses/Kconfig | 1 + drivers/i2c/busses/i2c-piix4.c | 4 +++- include/linux/pci_ids.h | 1 + 4 files changed, 6 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/Documentation/i2c/busses/i2c-piix4 b/Documentation/i2c/busses/i2c-piix4 index ef1efa79b1df..f889481762b5 100644 --- a/Documentation/i2c/busses/i2c-piix4 +++ b/Documentation/i2c/busses/i2c-piix4 @@ -4,7 +4,7 @@ Supported adapters: * Intel 82371AB PIIX4 and PIIX4E * Intel 82443MX (440MX) Datasheet: Publicly available at the Intel website - * ServerWorks OSB4, CSB5, CSB6 and HT-1000 southbridges + * ServerWorks OSB4, CSB5, CSB6, HT-1000 and HT-1100 southbridges Datasheet: Only available via NDA from ServerWorks * ATI IXP200, IXP300, IXP400, SB600, SB700 and SB800 southbridges Datasheet: Not publicly available diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig index 68650643d116..da809ad0996a 100644 --- a/drivers/i2c/busses/Kconfig +++ b/drivers/i2c/busses/Kconfig @@ -132,6 +132,7 @@ config I2C_PIIX4 Serverworks CSB5 Serverworks CSB6 Serverworks HT-1000 + Serverworks HT-1100 SMSC Victory66 This driver can also be built as a module. If so, the module diff --git a/drivers/i2c/busses/i2c-piix4.c b/drivers/i2c/busses/i2c-piix4.c index 63d5e5978046..0249a7d762b9 100644 --- a/drivers/i2c/busses/i2c-piix4.c +++ b/drivers/i2c/busses/i2c-piix4.c @@ -20,7 +20,7 @@ /* Supports: Intel PIIX4, 440MX - Serverworks OSB4, CSB5, CSB6, HT-1000 + Serverworks OSB4, CSB5, CSB6, HT-1000, HT-1100 ATI IXP200, IXP300, IXP400, SB600, SB700, SB800 SMSC Victory66 @@ -487,6 +487,8 @@ static struct pci_device_id piix4_ids[] = { PCI_DEVICE_ID_SERVERWORKS_CSB6) }, { PCI_DEVICE(PCI_VENDOR_ID_SERVERWORKS, PCI_DEVICE_ID_SERVERWORKS_HT1000SB) }, + { PCI_DEVICE(PCI_VENDOR_ID_SERVERWORKS, + PCI_DEVICE_ID_SERVERWORKS_HT1100LD) }, { 0, } }; diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index 5109fecde284..2c9e8080da5e 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -1479,6 +1479,7 @@ #define PCI_DEVICE_ID_SERVERWORKS_HT1000IDE 0x0214 #define PCI_DEVICE_ID_SERVERWORKS_CSB6IDE2 0x0217 #define PCI_DEVICE_ID_SERVERWORKS_CSB6LPC 0x0227 +#define PCI_DEVICE_ID_SERVERWORKS_HT1100LD 0x0408 #define PCI_VENDOR_ID_SBE 0x1176 #define PCI_DEVICE_ID_SBE_WANXL100 0x0301 -- cgit v1.2.3-59-g8ed1b