diff options
Diffstat (limited to 'net/sunrpc/cache.c')
-rw-r--r-- | net/sunrpc/cache.c | 239 |
1 files changed, 129 insertions, 110 deletions
diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index bd843a81afa0..f075a9fb5ccc 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -32,13 +32,15 @@ #include <linux/sunrpc/cache.h> #include <linux/sunrpc/stats.h> #include <linux/sunrpc/rpc_pipe_fs.h> +#include <trace/events/sunrpc.h> + #include "netns.h" +#include "fail.h" #define RPCDBG_FACILITY RPCDBG_CACHE static bool cache_defer_req(struct cache_req *req, struct cache_head *item); static void cache_revisit_request(struct cache_head *item); -static bool cache_listeners_exist(struct cache_detail *detail); static void cache_init(struct cache_head *h, struct cache_detail *detail) { @@ -65,13 +67,14 @@ static struct cache_head *sunrpc_cache_find_rcu(struct cache_detail *detail, rcu_read_lock(); hlist_for_each_entry_rcu(tmp, head, cache_list) { - if (detail->match(tmp, key)) { - if (cache_is_expired(detail, tmp)) - continue; - tmp = cache_get_rcu(tmp); - rcu_read_unlock(); - return tmp; - } + if (!detail->match(tmp, key)) + continue; + if (test_bit(CACHE_VALID, &tmp->flags) && + cache_is_expired(detail, tmp)) + continue; + tmp = cache_get_rcu(tmp); + rcu_read_unlock(); + return tmp; } rcu_read_unlock(); return NULL; @@ -113,18 +116,21 @@ static struct cache_head *sunrpc_cache_add_entry(struct cache_detail *detail, spin_lock(&detail->hash_lock); /* check if entry appeared while we slept */ - hlist_for_each_entry_rcu(tmp, head, cache_list) { - if (detail->match(tmp, key)) { - if (cache_is_expired(detail, tmp)) { - sunrpc_begin_cache_remove_entry(tmp, detail); - freeme = tmp; - break; - } - cache_get(tmp); - spin_unlock(&detail->hash_lock); - cache_put(new, detail); - return tmp; + hlist_for_each_entry_rcu(tmp, head, cache_list, + lockdep_is_held(&detail->hash_lock)) { + if (!detail->match(tmp, key)) + continue; + if (test_bit(CACHE_VALID, &tmp->flags) && + cache_is_expired(detail, tmp)) { + sunrpc_begin_cache_remove_entry(tmp, detail); + trace_cache_entry_expired(detail, tmp); + freeme = tmp; + break; } + cache_get(tmp); + spin_unlock(&detail->hash_lock); + cache_put(new, detail); + return tmp; } hlist_add_head_rcu(&new->cache_list, head); @@ -174,6 +180,25 @@ static void cache_fresh_unlocked(struct cache_head *head, } } +static void cache_make_negative(struct cache_detail *detail, + struct cache_head *h) +{ + set_bit(CACHE_NEGATIVE, &h->flags); + trace_cache_entry_make_negative(detail, h); +} + +static void cache_entry_update(struct cache_detail *detail, + struct cache_head *h, + struct cache_head *new) +{ + if (!test_bit(CACHE_NEGATIVE, &new->flags)) { + detail->update(h, new); + trace_cache_entry_update(detail, h); + } else { + cache_make_negative(detail, h); + } +} + struct cache_head *sunrpc_cache_update(struct cache_detail *detail, struct cache_head *new, struct cache_head *old, int hash) { @@ -186,10 +211,7 @@ struct cache_head *sunrpc_cache_update(struct cache_detail *detail, if (!test_bit(CACHE_VALID, &old->flags)) { spin_lock(&detail->hash_lock); if (!test_bit(CACHE_VALID, &old->flags)) { - if (test_bit(CACHE_NEGATIVE, &new->flags)) - set_bit(CACHE_NEGATIVE, &old->flags); - else - detail->update(old, new); + cache_entry_update(detail, old, new); cache_fresh_locked(old, new->expiry_time, detail); spin_unlock(&detail->hash_lock); cache_fresh_unlocked(old, detail); @@ -207,10 +229,7 @@ struct cache_head *sunrpc_cache_update(struct cache_detail *detail, detail->init(tmp, old); spin_lock(&detail->hash_lock); - if (test_bit(CACHE_NEGATIVE, &new->flags)) - set_bit(CACHE_NEGATIVE, &tmp->flags); - else - detail->update(tmp, new); + cache_entry_update(detail, tmp, new); hlist_add_head(&tmp->cache_list, &detail->hash_table[hash]); detail->entries++; cache_get(tmp); @@ -224,13 +243,6 @@ struct cache_head *sunrpc_cache_update(struct cache_detail *detail, } EXPORT_SYMBOL_GPL(sunrpc_cache_update); -static int cache_make_upcall(struct cache_detail *cd, struct cache_head *h) -{ - if (cd->cache_upcall) - return cd->cache_upcall(cd, h); - return sunrpc_cache_pipe_upcall(cd, h); -} - static inline int cache_is_valid(struct cache_head *h) { if (!test_bit(CACHE_VALID, &h->flags)) @@ -259,7 +271,7 @@ static int try_to_negate_entry(struct cache_detail *detail, struct cache_head *h spin_lock(&detail->hash_lock); rv = cache_is_valid(h); if (rv == -EAGAIN) { - set_bit(CACHE_NEGATIVE, &h->flags); + cache_make_negative(detail, h); cache_fresh_locked(h, seconds_since_boot()+CACHE_NEW_EXPIRY, detail); rv = -ENOENT; @@ -303,17 +315,14 @@ int cache_check(struct cache_detail *detail, (h->expiry_time != 0 && age > refresh_age/2)) { dprintk("RPC: Want update, refage=%lld, age=%lld\n", refresh_age, age); - if (!test_and_set_bit(CACHE_PENDING, &h->flags)) { - switch (cache_make_upcall(detail, h)) { - case -EINVAL: - rv = try_to_negate_entry(detail, h); - break; - case -EAGAIN: - cache_fresh_unlocked(h, detail); - break; - } - } else if (!cache_listeners_exist(detail)) + switch (detail->cache_upcall(detail, h)) { + case -EINVAL: rv = try_to_negate_entry(detail, h); + break; + case -EAGAIN: + cache_fresh_unlocked(h, detail); + break; + } } if (rv == -EAGAIN) { @@ -468,6 +477,7 @@ static int cache_clean(void) continue; sunrpc_begin_cache_remove_entry(ch, current_detail); + trace_cache_entry_expired(current_detail, ch); rv = 1; break; } @@ -490,16 +500,17 @@ static int cache_clean(void) */ static void do_cache_clean(struct work_struct *work) { - int delay = 5; - if (cache_clean() == -1) - delay = round_jiffies_relative(30*HZ); + int delay; if (list_empty(&cache_list)) - delay = 0; + return; - if (delay) - queue_delayed_work(system_power_efficient_wq, - &cache_cleaner, delay); + if (cache_clean() == -1) + delay = round_jiffies_relative(30*HZ); + else + delay = 5; + + queue_delayed_work(system_power_efficient_wq, &cache_cleaner, delay); } @@ -521,7 +532,6 @@ void cache_purge(struct cache_detail *detail) { struct cache_head *ch = NULL; struct hlist_head *head = NULL; - struct hlist_node *tmp = NULL; int i = 0; spin_lock(&detail->hash_lock); @@ -533,7 +543,9 @@ void cache_purge(struct cache_detail *detail) dprintk("RPC: %d entries in %s cache\n", detail->entries, detail->name); for (i = 0; i < detail->hash_size; i++) { head = &detail->hash_table[i]; - hlist_for_each_entry_safe(ch, tmp, head, cache_list) { + while (!hlist_empty(head)) { + ch = hlist_entry(head->first, struct cache_head, + cache_list); sunrpc_begin_cache_remove_entry(ch, detail); spin_unlock(&detail->hash_lock); sunrpc_end_cache_remove_entry(ch, detail); @@ -665,7 +677,7 @@ static void cache_limit_defers(void) /* Consider removing either the first or the last */ if (cache_defer_cnt > DFR_MAX) { - if (prandom_u32() & 1) + if (prandom_u32_max(2)) discard = list_entry(cache_defer_list.next, struct cache_deferred_req, recent); else @@ -678,16 +690,30 @@ static void cache_limit_defers(void) discard->revisit(discard, 1); } +#if IS_ENABLED(CONFIG_FAIL_SUNRPC) +static inline bool cache_defer_immediately(void) +{ + return !fail_sunrpc.ignore_cache_wait && + should_fail(&fail_sunrpc.attr, 1); +} +#else +static inline bool cache_defer_immediately(void) +{ + return false; +} +#endif + /* Return true if and only if a deferred request is queued. */ static bool cache_defer_req(struct cache_req *req, struct cache_head *item) { struct cache_deferred_req *dreq; - if (req->thread_wait) { + if (!cache_defer_immediately()) { cache_wait_req(req, item); if (!test_bit(CACHE_PENDING, &item->flags)) return false; } + dreq = req->defer(req); if (dreq == NULL) return false; @@ -768,7 +794,6 @@ void cache_clean_deferred(void *owner) */ static DEFINE_SPINLOCK(queue_lock); -static DEFINE_MUTEX(queue_io_mutex); struct cache_queue { struct list_head list; @@ -794,7 +819,7 @@ static int cache_request(struct cache_detail *detail, detail->cache_request(detail, crq->item, &bp, &len); if (len < 0) - return -EAGAIN; + return -E2BIG; return PAGE_SIZE - len; } @@ -896,44 +921,26 @@ static ssize_t cache_do_downcall(char *kaddr, const char __user *buf, return ret; } -static ssize_t cache_slow_downcall(const char __user *buf, - size_t count, struct cache_detail *cd) -{ - static char write_buf[8192]; /* protected by queue_io_mutex */ - ssize_t ret = -EINVAL; - - if (count >= sizeof(write_buf)) - goto out; - mutex_lock(&queue_io_mutex); - ret = cache_do_downcall(write_buf, buf, count, cd); - mutex_unlock(&queue_io_mutex); -out: - return ret; -} - static ssize_t cache_downcall(struct address_space *mapping, const char __user *buf, size_t count, struct cache_detail *cd) { - struct page *page; - char *kaddr; + char *write_buf; ssize_t ret = -ENOMEM; - if (count >= PAGE_SIZE) - goto out_slow; + if (count >= 32768) { /* 32k is max userland buffer, lets check anyway */ + ret = -EINVAL; + goto out; + } - page = find_or_create_page(mapping, 0, GFP_KERNEL); - if (!page) - goto out_slow; + write_buf = kvmalloc(count + 1, GFP_KERNEL); + if (!write_buf) + goto out; - kaddr = kmap(page); - ret = cache_do_downcall(kaddr, buf, count, cd); - kunmap(page); - unlock_page(page); - put_page(page); + ret = cache_do_downcall(write_buf, buf, count, cd); + kvfree(write_buf); +out: return ret; -out_slow: - return cache_slow_downcall(buf, count, cd); } static ssize_t cache_write(struct file *filp, const char __user *buf, @@ -1195,20 +1202,12 @@ static bool cache_listeners_exist(struct cache_detail *detail) * * Each request is at most one page long. */ -int sunrpc_cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h) +static int cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h) { - char *buf; struct cache_request *crq; int ret = 0; - if (!detail->cache_request) - return -EINVAL; - - if (!cache_listeners_exist(detail)) { - warn_no_listener(detail); - return -EINVAL; - } if (test_bit(CACHE_CLEANED, &h->flags)) /* Too late to make an upcall */ return -EAGAIN; @@ -1231,6 +1230,7 @@ int sunrpc_cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h) if (test_bit(CACHE_PENDING, &h->flags)) { crq->item = cache_get(h); list_add_tail(&crq->q.list, &detail->queue); + trace_cache_entry_upcall(detail, h); } else /* Lost a race, no longer PENDING, so don't enqueue */ ret = -EAGAIN; @@ -1242,8 +1242,27 @@ int sunrpc_cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h) } return ret; } + +int sunrpc_cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h) +{ + if (test_and_set_bit(CACHE_PENDING, &h->flags)) + return 0; + return cache_pipe_upcall(detail, h); +} EXPORT_SYMBOL_GPL(sunrpc_cache_pipe_upcall); +int sunrpc_cache_pipe_upcall_timeout(struct cache_detail *detail, + struct cache_head *h) +{ + if (!cache_listeners_exist(detail)) { + warn_no_listener(detail); + trace_cache_entry_no_listener(detail, h); + return -EINVAL; + } + return sunrpc_cache_pipe_upcall(detail, h); +} +EXPORT_SYMBOL_GPL(sunrpc_cache_pipe_upcall_timeout); + /* * parse a message from user-space and pass it * to an appropriate cache @@ -1415,10 +1434,10 @@ static int c_show(struct seq_file *m, void *p) cache_get(cp); if (cache_check(cd, cp, NULL)) /* cache_check does a cache_put on failure */ - seq_printf(m, "# "); + seq_puts(m, "# "); else { if (cache_is_expired(cd, cp)) - seq_printf(m, "# "); + seq_puts(m, "# "); cache_put(cp, cd); } @@ -1533,7 +1552,7 @@ static ssize_t write_flush(struct file *file, const char __user *buf, static ssize_t cache_read_procfs(struct file *filp, char __user *buf, size_t count, loff_t *ppos) { - struct cache_detail *cd = PDE_DATA(file_inode(filp)); + struct cache_detail *cd = pde_data(file_inode(filp)); return cache_read(filp, buf, count, ppos, cd); } @@ -1541,14 +1560,14 @@ static ssize_t cache_read_procfs(struct file *filp, char __user *buf, static ssize_t cache_write_procfs(struct file *filp, const char __user *buf, size_t count, loff_t *ppos) { - struct cache_detail *cd = PDE_DATA(file_inode(filp)); + struct cache_detail *cd = pde_data(file_inode(filp)); return cache_write(filp, buf, count, ppos, cd); } static __poll_t cache_poll_procfs(struct file *filp, poll_table *wait) { - struct cache_detail *cd = PDE_DATA(file_inode(filp)); + struct cache_detail *cd = pde_data(file_inode(filp)); return cache_poll(filp, wait, cd); } @@ -1557,21 +1576,21 @@ static long cache_ioctl_procfs(struct file *filp, unsigned int cmd, unsigned long arg) { struct inode *inode = file_inode(filp); - struct cache_detail *cd = PDE_DATA(inode); + struct cache_detail *cd = pde_data(inode); return cache_ioctl(inode, filp, cmd, arg, cd); } static int cache_open_procfs(struct inode *inode, struct file *filp) { - struct cache_detail *cd = PDE_DATA(inode); + struct cache_detail *cd = pde_data(inode); return cache_open(inode, filp, cd); } static int cache_release_procfs(struct inode *inode, struct file *filp) { - struct cache_detail *cd = PDE_DATA(inode); + struct cache_detail *cd = pde_data(inode); return cache_release(inode, filp, cd); } @@ -1588,14 +1607,14 @@ static const struct proc_ops cache_channel_proc_ops = { static int content_open_procfs(struct inode *inode, struct file *filp) { - struct cache_detail *cd = PDE_DATA(inode); + struct cache_detail *cd = pde_data(inode); return content_open(inode, filp, cd); } static int content_release_procfs(struct inode *inode, struct file *filp) { - struct cache_detail *cd = PDE_DATA(inode); + struct cache_detail *cd = pde_data(inode); return content_release(inode, filp, cd); } @@ -1609,14 +1628,14 @@ static const struct proc_ops content_proc_ops = { static int open_flush_procfs(struct inode *inode, struct file *filp) { - struct cache_detail *cd = PDE_DATA(inode); + struct cache_detail *cd = pde_data(inode); return open_flush(inode, filp, cd); } static int release_flush_procfs(struct inode *inode, struct file *filp) { - struct cache_detail *cd = PDE_DATA(inode); + struct cache_detail *cd = pde_data(inode); return release_flush(inode, filp, cd); } @@ -1624,7 +1643,7 @@ static int release_flush_procfs(struct inode *inode, struct file *filp) static ssize_t read_flush_procfs(struct file *filp, char __user *buf, size_t count, loff_t *ppos) { - struct cache_detail *cd = PDE_DATA(file_inode(filp)); + struct cache_detail *cd = pde_data(file_inode(filp)); return read_flush(filp, buf, count, ppos, cd); } @@ -1633,7 +1652,7 @@ static ssize_t write_flush_procfs(struct file *filp, const char __user *buf, size_t count, loff_t *ppos) { - struct cache_detail *cd = PDE_DATA(file_inode(filp)); + struct cache_detail *cd = pde_data(file_inode(filp)); return write_flush(filp, buf, count, ppos, cd); } |