| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
This patch modifies the attr set routines to be delay ready. This means
they no longer roll or commit transactions, but instead return -EAGAIN
to have the calling routine roll and refresh the transaction. In this
series, xfs_attr_set_args has become xfs_attr_set_iter, which uses a
state machine like switch to keep track of where it was when EAGAIN was
returned. See xfs_attr.h for a more detailed diagram of the states.
Two new helper functions have been added: xfs_attr_rmtval_find_space and
xfs_attr_rmtval_set_blk. They provide a subset of logic similar to
xfs_attr_rmtval_set, but they store the current block in the delay attr
context to allow the caller to roll the transaction between allocations.
This helps to simplify and consolidate code used by
xfs_attr_leaf_addname and xfs_attr_node_addname. xfs_attr_set_args has
now become a simple loop to refresh the transaction until the operation
is completed. Lastly, xfs_attr_rmtval_remove is no longer used, and is
removed.
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
|
|
This patch modifies the attr remove routines to be delay ready. This
means they no longer roll or commit transactions, but instead return
-EAGAIN to have the calling routine roll and refresh the transaction. In
this series, xfs_attr_remove_args is merged with
xfs_attr_node_removename become a new function, xfs_attr_remove_iter.
This new version uses a sort of state machine like switch to keep track
of where it was when EAGAIN was returned. A new version of
xfs_attr_remove_args consists of a simple loop to refresh the
transaction until the operation is completed. A new XFS_DAC_DEFER_FINISH
flag is used to finish the transaction where ever the existing code used
to.
Calls to xfs_attr_rmtval_remove are replaced with the delay ready
version __xfs_attr_rmtval_remove. We will rename
__xfs_attr_rmtval_remove back to xfs_attr_rmtval_remove when we are
done.
xfs_attr_rmtval_remove itself is still in use by the set routines (used
during a rename). For reasons of preserving existing function, we
modify xfs_attr_rmtval_remove to call xfs_defer_finish when the flag is
set. Similar to how xfs_attr_remove_args does here. Once we transition
the set routines to be delay ready, xfs_attr_rmtval_remove is no longer
used and will be removed.
This patch also adds a new struct xfs_delattr_context, which we will use
to keep track of the current state of an attribute operation. The new
xfs_delattr_state enum is used to track various operations that are in
progress so that we know not to repeat them, and resume where we left
off before EAGAIN was returned to cycle out the transaction. Other
members take the place of local variables that need to retain their
values across multiple function calls. See xfs_attr.h for a more
detailed diagram of the states.
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
|
|
This patch basically hoists the node transaction handling around the
leaf code we just hoisted. This will helps setup this area for the
state machine since the goto is easily replaced with a state since it
ends with a transaction roll.
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
|
|
This patch hoists xfs_attr_leaf_addname into the calling function. The
goal being to get all the code that will require state management into
the same scope. This isn't particularly aesthetic right away, but it is a
preliminary step to merging in the state machine code.
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
|
|
This patch hoists the later half of xfs_attr_node_addname into
the calling function. We do this because it is this area that
will need the most state management, and we want to keep such
code in the same scope as much as possible
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
|
|
This patch separates the first half of xfs_attr_node_addname into a
helper function xfs_attr_node_addname_find_attr. It also replaces the
restart goto with an EAGAIN return code driven by a loop in the calling
function. This looks odd now, but will clean up nicly once we introduce
the state machine. It will also enable hoisting the last state out of
xfs_attr_node_addname with out having to plumb in a "done" parameter to
know if we need to move to the next state or not.
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
|
|
This patch separate xfs_attr_node_addname into two functions. This will
help to make it easier to hoist parts of xfs_attr_node_addname that need
state management
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
|
|
This patch is actually the combination of patches from the previous
version (v18). Initially patch 3 hoisted xfs_attr_set_shortform, and
the next added the helper xfs_attr_set_fmt. xfs_attr_set_fmt is similar
the old xfs_attr_set_shortform. It returns 0 when the attr has been set
and no further action is needed. It returns -EAGAIN when shortform has
been transformed to leaf, and the calling function should proceed the
set the attr in leaf form.
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
|
|
This patch pulls a new helper function xfs_attr_node_remove_name out
of xfs_attr_node_remove_step. This helps to modularize
xfs_attr_node_remove_step which will help make the delayed attribute
code easier to follow
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
|
|
Originally we added this patch to help modularize the attr code in
preparation for delayed attributes and the state machine it requires.
However, later reviews found that this slightly alters the transaction
handling as the helper function is ambiguous as to whether the
transaction is diry or clean. This may cause a dirty transaction to be
included in the next roll, where previously it had not. To preserve the
existing code flow, we reverse apply this commit.
Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
|
|
large directory block size operations are assert failing because
xfs_bunmapi() is not completely removing fragmented directory blocks
like so:
XFS: Assertion failed: done, file: fs/xfs/libxfs/xfs_dir2.c, line: 677
....
Call Trace:
xfs_dir2_shrink_inode+0x1a8/0x210
xfs_dir2_block_to_sf+0x2ae/0x410
xfs_dir2_block_removename+0x21a/0x280
xfs_dir_removename+0x195/0x1d0
xfs_rename+0xb79/0xc50
? avc_has_perm+0x8d/0x1a0
? avc_has_perm_noaudit+0x9a/0x120
xfs_vn_rename+0xdb/0x150
vfs_rename+0x719/0xb50
? __lookup_hash+0x6a/0xa0
do_renameat2+0x413/0x5e0
__x64_sys_rename+0x45/0x50
do_syscall_64+0x3a/0x70
entry_SYSCALL_64_after_hwframe+0x44/0xae
We are aborting the bunmapi() pass because of this specific chunk of
code:
/*
* Make sure we don't touch multiple AGF headers out of order
* in a single transaction, as that could cause AB-BA deadlocks.
*/
if (!wasdel && !isrt) {
agno = XFS_FSB_TO_AGNO(mp, del.br_startblock);
if (prev_agno != NULLAGNUMBER && prev_agno > agno)
break;
prev_agno = agno;
}
This is designed to prevent deadlocks in AGF locking when freeing
multiple extents by ensuring that we only ever lock in increasing
AG number order. Unfortunately, this also violates the "bunmapi will
always succeed" semantic that some high level callers depend on,
such as xfs_dir2_shrink_inode(), xfs_da_shrink_inode() and
xfs_inactive_symlink_rmt().
This AG lock ordering was introduced back in 2017 to fix deadlocks
triggered by generic/299 as reported here:
https://lore.kernel.org/linux-xfs/800468eb-3ded-9166-20a4-047de8018582@gmail.com/
This codebase is old enough that it was before we were defering all
AG based extent freeing from within xfs_bunmapi(). THat is, we never
actually lock AGs in xfs_bunmapi() any more - every non-rt based
extent free is added to the defer ops list, as is all BMBT block
freeing. And RT extents are not RT based, so there's no lock
ordering issues associated with them.
Hence this AGF lock ordering code is both broken and dead. Let's
just remove it so that the large directory block code works reliably
again.
Tested against xfs/538 and generic/299 which is the original test
that exposed the deadlocks that this code fixed.
Fixes: 5b094d6dac04 ("xfs: fix multi-AG deadlock in xfs_bunmapi")
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
|
|
xfs/538 is assert failing with this trace when testing with
directory block sizes of 64kB:
XFS: Assertion failed: !xfs_need_iread_extents(ifp), file: fs/xfs/libxfs/xfs_bmap.c, line: 608
....
Call Trace:
xfs_bmap_btree_to_extents+0x2a9/0x470
? kmem_cache_alloc+0xe7/0x220
__xfs_bunmapi+0x4ca/0xdf0
xfs_bunmapi+0x1a/0x30
xfs_dir2_shrink_inode+0x71/0x210
xfs_dir2_block_to_sf+0x2ae/0x410
xfs_dir2_block_removename+0x21a/0x280
xfs_dir_removename+0x195/0x1d0
xfs_remove+0x244/0x460
xfs_vn_unlink+0x53/0xa0
? selinux_inode_unlink+0x13/0x20
vfs_unlink+0x117/0x220
do_unlinkat+0x1a2/0x2d0
__x64_sys_unlink+0x42/0x60
do_syscall_64+0x3a/0x70
entry_SYSCALL_64_after_hwframe+0x44/0xae
This is a check to ensure that the extents have been read into
memory before we are doing a ifork btree manipulation. This assert
is bogus in the above case.
We have a fragmented directory block that has more extents in it
than can fit in extent format, so the inode data fork is in btree
format. xfs_dir2_shrink_inode() asks to remove all remaining 16
filesystem blocks from the inode so it can convert to short form,
and __xfs_bunmapi() removes all the extents. We now have a data fork
in btree format but have zero extents in the fork. This incorrectly
trips the xfs_need_iread_extents() assert because it assumes that an
empty extent btree means the extent tree has not been read into
memory yet. This is clearly not the case with xfs_bunmapi(), as it
has an explicit call to xfs_iread_extents() in it to pull the
extents into memory before it starts unmapping.
Also, the assert directly after this bogus one is:
ASSERT(ifp->if_format == XFS_DINODE_FMT_BTREE);
Which covers the context in which it is legal to call
xfs_bmap_btree_to_extents just fine. Hence we should just remove the
bogus assert as it is clearly wrong and causes a regression.
The returns the test behaviour to the pre-existing assert failure in
xfs_dir2_shrink_inode() that indicates xfs_bunmapi() has failed to
remove all the extents in the range it was asked to unmap.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
|
|
Add our new OFTC channel to the MAINTAINERS list so everyone will know
where to go. Ignore the XFS wikis, we have no access to them.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
|
|
The RTINHERIT bit can be set on a directory so that newly created
regular files will have the REALTIME bit set to store their data on the
realtime volume. If an extent size hint (and EXTSZINHERIT) are set on
the directory, the hint will also be copied into the new file.
As pointed out in previous patches, for realtime files we require the
extent size hint be an integer multiple of the realtime extent, but we
don't perform the same validation on a directory with both RTINHERIT and
EXTSZINHERIT set, even though the only use-case of that combination is
to propagate extent size hints into new realtime files. This leads to
inode corruption errors when the bad values are propagated.
Because there may be existing filesystems with such a configuration, we
cannot simply amend the inode verifier to trip on these directories and
call it a day because that will cause previously "working" filesystems
to start throwing errors abruptly. Note that it's valid to have
directories with rtinherit set even if there is no realtime volume, in
which case the problem does not manifest because rtinherit is ignored if
there's no realtime device; and it's possible that someone set the flag,
crashed, repaired the filesystem (which clears the hint on the realtime
file) and continued.
Therefore, mitigate this issue in several ways: First, if we try to
write out an inode with both rtinherit/extszinherit set and an unaligned
extent size hint, turn off the hint to correct the error. Second, if
someone tries to misconfigure a directory via the fssetxattr ioctl, fail
the ioctl. Third, reverify both extent size hint values when we
propagate heritable inode attributes from parent to child, to prevent
misconfigurations from spreading.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
|
|
While chasing a bug involving invalid extent size hints being propagated
into newly created realtime files, I noticed that the xfs_ioctl_setattr
checks for the extent size hints weren't the same as the ones now
encoded in libxfs and used for validation in repair and mkfs.
Because the checks in libxfs are more stringent than the ones in the
ioctl, it's possible for a live system to set inode flags that
immediately result in corruption warnings. Specifically, it's possible
to set an extent size hint on an rtinherit directory without checking if
the hint is aligned to the realtime extent size, which makes no sense
since that combination is used only to seed new realtime files.
Replace the open-coded and inadequate checks with the libxfs verifier
versions and update the code comments a bit.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
The new online shrink code exposed a gap in the per-AG reservation
code, which is that we only return ENOSPC to callers if the entire fs
doesn't have enough free blocks. Except for debugging mode, the
reservation init code doesn't ever check that there's enough free space
in that AG to cover the reservation.
Not having enough space is not considered an immediate fatal error that
requires filesystem offlining because (a) it's shouldn't be possible to
wind up in that state through normal file operations and (b) even if
one did, freeing data blocks would recover the situation.
However, online shrink now needs to know if shrinking would not leave
enough space so that it can abort the shrink operation. Hence we need
to promote this assertion into an actual error return.
Observed by running xfs/168 with a 1k block size, though in theory this
could happen with any configuration.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Gao Xiang <hsiangkao@linux.alibaba.com>
|
|
These ioctl definitions in xfs_fs.h are part of the userspace ABI and
were mistakenly removed during the 5.13 merge window.
Fixes: 9fefd5db08ce ("xfs: convert to fileattr")
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
|
|
sc->ip is the inode that's being scrubbed, which means that it's not set
for scrub types that don't involve inodes. If one of those scrubbers
(e.g. inode btrees) returns EDEADLOCK, we'll trip over the null pointer.
Fix that by reporting either the file being examined or the file that
was used to call scrub.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
|
|
If a realtime allocation fails because we can't find a sufficiently
large free extent satisfying locality rules, relax the locality rules
and try again. This reduces the occurrence of short writes to realtime
files when the write size is large and the free space is fragmented.
This was originally discovered by running generic/186 with the realtime
reflink patchset and a 128k cow extent size hint, but the short write
symptoms can manifest with a 128k extent size hint and no reflink, so
apply the fix now.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
|
|
xfs_bmap_rtalloc doesn't handle realtime extent files with extent size
hints larger than the rt volume's extent size properly, because
xfs_bmap_extsize_align can adjust the offset/length parameters to try to
fit the extent size hint.
Under these conditions, minlen has to be large enough so that any
allocation returned by xfs_rtallocate_extent will be large enough to
cover at least one of the blocks that the caller asked for. If the
allocation is too short, bmapi_write will return no mapping for the
requested range, which causes ENOSPC errors in other parts of the
filesystem.
Therefore, adjust minlen upwards to fix this. This can be found by
running generic/263 (g/127 or g/522) with a realtime extent size hint
that's larger than the rt volume extent size.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
|
|
|
|
syzbot is reporting OOB write at vga16fb_imageblit() [1], for
resize_screen() from ioctl(VT_RESIZE) returns 0 without checking whether
requested rows/columns fit the amount of memory reserved for the graphical
screen if current mode is KD_GRAPHICS.
----------
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#include <linux/kd.h>
#include <linux/vt.h>
int main(int argc, char *argv[])
{
const int fd = open("/dev/char/4:1", O_RDWR);
struct vt_sizes vt = { 0x4100, 2 };
ioctl(fd, KDSETMODE, KD_GRAPHICS);
ioctl(fd, VT_RESIZE, &vt);
ioctl(fd, KDSETMODE, KD_TEXT);
return 0;
}
----------
Allow framebuffer drivers to return -EINVAL, by moving vc->vc_mode !=
KD_GRAPHICS check from resize_screen() to fbcon_resize().
Link: https://syzkaller.appspot.com/bug?extid=1f29e126cf461c4de3b3 [1]
Reported-by: syzbot <syzbot+1f29e126cf461c4de3b3@syzkaller.appspotmail.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Tested-by: syzbot <syzbot+1f29e126cf461c4de3b3@syzkaller.appspotmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
iomap_max_page_shift is expected to contain a page shift, so it can't be a
'bool', has to be an 'unsigned int'
And fix the default values: P4D_SHIFT is when huge iomap is allowed.
However, on some architectures (eg: powerpc book3s/64), P4D_SHIFT is not a
constant so it can't be used to initialise a static variable. So,
initialise iomap_max_page_shift with a maximum shift supported by the
architecture, it is gated by P4D_SHIFT in vmap_try_huge_p4d() anyway.
Link: https://lkml.kernel.org/r/ad2d366015794a9f21320dcbdd0a8eb98979e9df.1620898113.git.christophe.leroy@csgroup.eu
Fixes: bbc180a5adb0 ("mm: HUGE_VMAP arch support cleanup")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
When I added CONFIG_MODPROBE_PATH, I neglected to update Documentation/.
It's still true that this defaults to /sbin/modprobe, but now via a level
of indirection. So document that the kernel might have been built with
something other than /sbin/modprobe as the initial value.
Link: https://lkml.kernel.org/r/20210420125324.1246826-1-linux@rasmusvillemoes.dk
Fixes: 17652f4240f7a ("modules: add CONFIG_MODPROBE_PATH")
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jessica Yu <jeyu@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
I believe there are some issues introduced by commit 31651c607151
("hfsplus: avoid deadlock on file truncation")
HFS+ has extent records which always contains 8 extents. In case the
first extent record in catalog file gets full, new ones are allocated from
extents overflow file.
In case shrinking truncate happens to middle of an extent record which
locates in extents overflow file, the logic in hfsplus_file_truncate() was
changed so that call to hfs_brec_remove() is not guarded any more.
Right action would be just freeing the extents that exceed the new size
inside extent record by calling hfsplus_free_extents(), and then check if
the whole extent record should be removed. However since the guard
(blk_cnt > start) is now after the call to hfs_brec_remove(), this has
unfortunate effect that the last matching extent record is removed
unconditionally.
To reproduce this issue, create a file which has at least 10 extents, and
then perform shrinking truncate into middle of the last extent record, so
that the number of remaining extents is not under or divisible by 8. This
causes the last extent record (8 extents) to be removed totally instead of
truncating into middle of it. Thus this causes corruption, and lost data.
Fix for this is simply checking if the new truncated end is below the
start of this extent record, making it safe to remove the full extent
record. However call to hfs_brec_remove() can't be moved to it's previous
place since we're dropping ->tree_lock and it can cause a race condition
and the cached info being invalidated possibly corrupting the node data.
Another issue is related to this one. When entering into the block
(blk_cnt > start) we are not holding the ->tree_lock. We break out from
the loop not holding the lock, but hfs_find_exit() does unlock it. Not
sure if it's possible for someone else to take the lock under our feet,
but it can cause hard to debug errors and premature unlocking. Even if
there's no real risk of it, the locking should still always be kept in
balance. Thus taking the lock now just before the check.
Link: https://lkml.kernel.org/r/20210429165139.3082828-1-jouni.roivas@tuxera.com
Fixes: 31651c607151f ("hfsplus: avoid deadlock on file truncation")
Signed-off-by: Jouni Roivas <jouni.roivas@tuxera.com>
Reviewed-by: Anton Altaparmakov <anton@tuxera.com>
Cc: Anatoly Trosinenko <anatoly.trosinenko@gmail.com>
Cc: Viacheslav Dubeyko <slava@dubeyko.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
A readahead request will not allocate more memory than can be represented
by a size_t, even on systems that have HIGHMEM available. Change the
length functions from returning an loff_t to a size_t.
Link: https://lkml.kernel.org/r/20210510201201.1558972-1-willy@infradead.org
Fixes: 32c0a6bcaa1f57 ("btrfs: add and use readahead_batch_length")
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
These tests deliberately access these arrays out of bounds, which will
cause the dynamic local bounds checks inserted by
CONFIG_UBSAN_LOCAL_BOUNDS to fail and panic the kernel. To avoid this
problem, access the arrays via volatile pointers, which will prevent the
compiler from being able to determine the array bounds.
These accesses use volatile pointers to char (char *volatile) rather than
the more conventional pointers to volatile char (volatile char *) because
we want to prevent the compiler from making inferences about the pointer
itself (i.e. its array bounds), not the data that it refers to.
Link: https://lkml.kernel.org/r/20210507025915.1464056-1-pcc@google.com
Link: https://linux-review.googlesource.com/id/I90b1713fbfa1bf68ff895aef099ea77b98a7c3b9
Signed-off-by: Peter Collingbourne <pcc@google.com>
Tested-by: Alexander Potapenko <glider@google.com>
Reviewed-by: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Peter Collingbourne <pcc@google.com>
Cc: George Popescu <georgepope@android.com>
Cc: Elena Petrova <lenaptr@google.com>
Cc: Evgenii Stepanov <eugenis@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
32-bit architectures which expect 8-byte alignment for 8-byte integers and
need 64-bit DMA addresses (arm, mips, ppc) had their struct page
inadvertently expanded in 2019. When the dma_addr_t was added, it forced
the alignment of the union to 8 bytes, which inserted a 4 byte gap between
'flags' and the union.
Fix this by storing the dma_addr_t in one or two adjacent unsigned longs.
This restores the alignment to that of an unsigned long. We always
store the low bits in the first word to prevent the PageTail bit from
being inadvertently set on a big endian platform. If that happened,
get_user_pages_fast() racing against a page which was freed and
reallocated to the page_pool could dereference a bogus compound_head(),
which would be hard to trace back to this cause.
Link: https://lkml.kernel.org/r/20210510153211.1504886-1-willy@infradead.org
Fixes: c25fff7171be ("mm: add dma_addr_t to struct page")
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Tested-by: Matteo Croce <mcroce@linux.microsoft.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
This reverts commit 3e96b6a2e9ad929a3230a22f4d64a74671a0720b. General
Protection Fault in rmap_walk_ksm() under memory pressure:
remove_rmap_item_from_tree() needs to take page lock, of course.
Link: https://lkml.kernel.org/r/alpine.LSU.2.11.2105092253500.1127@eggly.anvils
Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Consider the following sequence of events:
1. Userspace issues a UFFD ioctl, which ends up calling into
shmem_mfill_atomic_pte(). We successfully account the blocks, we
shmem_alloc_page(), but then the copy_from_user() fails. We return
-ENOENT. We don't release the page we allocated.
2. Our caller detects this error code, tries the copy_from_user() after
dropping the mmap_lock, and retries, calling back into
shmem_mfill_atomic_pte().
3. Meanwhile, let's say another process filled up the tmpfs being used.
4. So shmem_mfill_atomic_pte() fails to account blocks this time, and
immediately returns - without releasing the page.
This triggers a BUG_ON in our caller, which asserts that the page
should always be consumed, unless -ENOENT is returned.
To fix this, detect if we have such a "dangling" page when accounting
fails, and if so, release it before returning.
Link: https://lkml.kernel.org/r/20210428230858.348400-1-axelrasmussen@google.com
Fixes: cb658a453b93 ("userfaultfd: shmem: avoid leaking blocks and used blocks in UFFDIO_COPY")
Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Reported-by: Hugh Dickins <hughd@google.com>
Acked-by: Hugh Dickins <hughd@google.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Sysbot has reported a "divide error" which has been identified as being
caused by a corrupted file_size value within the file inode. This value
has been corrupted to a much larger value than expected.
Calculate_skip() is passed i_size_read(inode) >> msblk->block_log. Due to
the file_size value corruption this overflows the int argument/variable in
that function, leading to the divide error.
This patch changes the function to use u64. This will accommodate any
unexpectedly large values due to corruption.
The value returned from calculate_skip() is clamped to be never more than
SQUASHFS_CACHED_BLKS - 1, or 7. So file_size corruption does not lead to
an unexpectedly large return result here.
Link: https://lkml.kernel.org/r/20210507152618.9447-1-phillip@squashfs.org.uk
Signed-off-by: Phillip Lougher <phillip@squashfs.org.uk>
Reported-by: <syzbot+e8f781243ce16ac2f962@syzkaller.appspotmail.com>
Reported-by: <syzbot+7b98870d4fec9447b951@syzkaller.appspotmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Splitting an earlier version of a patch that allowed calling
__request_region() while holding the resource lock into a series of
patches required changing the return code for the newly introduced
__request_region_locked().
Unfortunately this change was not carried through to a subsequent commit
56fd94919b8b ("kernel/resource: fix locking in request_free_mem_region")
in the series. This resulted in a use-after-free due to freeing the
struct resource without properly releasing it. Fix this by correcting the
return code check so that the struct is not freed if the request to add it
was successful.
Link: https://lkml.kernel.org/r/20210512073528.22334-1-apopple@nvidia.com
Fixes: 56fd94919b8b ("kernel/resource: fix locking in request_free_mem_region")
Signed-off-by: Alistair Popple <apopple@nvidia.com>
Reported-by: kernel test robot <oliver.sang@intel.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Muchun Song <smuchun@gmail.com>
Cc: Oliver Sang <oliver.sang@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Paul E. McKenney reported [1] that commit 1f0723a4c0df ("mm, slub: enable
slub_debug static key when creating cache with explicit debug flags")
results in the lockdep complaint:
======================================================
WARNING: possible circular locking dependency detected
5.12.0+ #15 Not tainted
------------------------------------------------------
rcu_torture_sta/109 is trying to acquire lock:
ffffffff96063cd0 (cpu_hotplug_lock){++++}-{0:0}, at: static_key_enable+0x9/0x20
but task is already holding lock:
ffffffff96173c28 (slab_mutex){+.+.}-{3:3}, at: kmem_cache_create_usercopy+0x2d/0x250
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 (slab_mutex){+.+.}-{3:3}:
lock_acquire+0xb9/0x3a0
__mutex_lock+0x8d/0x920
slub_cpu_dead+0x15/0xf0
cpuhp_invoke_callback+0x17a/0x7c0
cpuhp_invoke_callback_range+0x3b/0x80
_cpu_down+0xdf/0x2a0
cpu_down+0x2c/0x50
device_offline+0x82/0xb0
remove_cpu+0x1a/0x30
torture_offline+0x80/0x140
torture_onoff+0x147/0x260
kthread+0x10a/0x140
ret_from_fork+0x22/0x30
-> #0 (cpu_hotplug_lock){++++}-{0:0}:
check_prev_add+0x8f/0xbf0
__lock_acquire+0x13f0/0x1d80
lock_acquire+0xb9/0x3a0
cpus_read_lock+0x21/0xa0
static_key_enable+0x9/0x20
__kmem_cache_create+0x38d/0x430
kmem_cache_create_usercopy+0x146/0x250
kmem_cache_create+0xd/0x10
rcu_torture_stats+0x79/0x280
kthread+0x10a/0x140
ret_from_fork+0x22/0x30
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(slab_mutex);
lock(cpu_hotplug_lock);
lock(slab_mutex);
lock(cpu_hotplug_lock);
*** DEADLOCK ***
1 lock held by rcu_torture_sta/109:
#0: ffffffff96173c28 (slab_mutex){+.+.}-{3:3}, at: kmem_cache_create_usercopy+0x2d/0x250
stack backtrace:
CPU: 3 PID: 109 Comm: rcu_torture_sta Not tainted 5.12.0+ #15
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1ubuntu1.1 04/01/2014
Call Trace:
dump_stack+0x6d/0x89
check_noncircular+0xfe/0x110
? lock_is_held_type+0x98/0x110
check_prev_add+0x8f/0xbf0
__lock_acquire+0x13f0/0x1d80
lock_acquire+0xb9/0x3a0
? static_key_enable+0x9/0x20
? mark_held_locks+0x49/0x70
cpus_read_lock+0x21/0xa0
? static_key_enable+0x9/0x20
static_key_enable+0x9/0x20
__kmem_cache_create+0x38d/0x430
kmem_cache_create_usercopy+0x146/0x250
? rcu_torture_stats_print+0xd0/0xd0
kmem_cache_create+0xd/0x10
rcu_torture_stats+0x79/0x280
? rcu_torture_stats_print+0xd0/0xd0
kthread+0x10a/0x140
? kthread_park+0x80/0x80
ret_from_fork+0x22/0x30
This is because there's one order of locking from the hotplug callbacks:
lock(cpu_hotplug_lock); // from hotplug machinery itself
lock(slab_mutex); // in e.g. slab_mem_going_offline_callback()
And commit 1f0723a4c0df made the reverse sequence possible:
lock(slab_mutex); // in kmem_cache_create_usercopy()
lock(cpu_hotplug_lock); // kmem_cache_open() -> static_key_enable()
The simplest fix is to move static_key_enable() to a place before slab_mutex is
taken. That means kmem_cache_create_usercopy() in mm/slab_common.c which is not
ideal for SLUB-specific code, but the #ifdef CONFIG_SLUB_DEBUG makes it
at least self-contained and obvious.
[1] https://lore.kernel.org/lkml/20210502171827.GA3670492@paulmck-ThinkPad-P17-Gen-1/
Link: https://lkml.kernel.org/r/20210504120019.26791-1-vbabka@suse.cz
Fixes: 1f0723a4c0df ("mm, slub: enable slub_debug static key when creating cache with explicit debug flags")
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reported-by: Paul E. McKenney <paulmck@kernel.org>
Tested-by: Paul E. McKenney <paulmck@kernel.org>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
When rework early cow of pinned hugetlb pages, we moved huge_ptep_get()
upper but overlooked a side effect that the huge_ptep_get() will fetch the
pte after wr-protection. After moving it upwards, we need explicit
wr-protect of child pte or we will keep the write bit set in the child
process, which could cause data corrution where the child can write to the
original page directly.
This issue can also be exposed by "memfd_test hugetlbfs" kselftest.
Link: https://lkml.kernel.org/r/20210503234356.9097-3-peterx@redhat.com
Fixes: 4eae4efa2c299 ("hugetlb: do early cow when page pinned on src mm")
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Joel Fernandes (Google) <joel@joelfernandes.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Patch series "mm/hugetlb: Fix issues on file sealing and fork", v2.
Hugh reported issue with F_SEAL_FUTURE_WRITE not applied correctly to
hugetlbfs, which I can easily verify using the memfd_test program, which
seems that the program is hardly run with hugetlbfs pages (as by default
shmem).
Meanwhile I found another probably even more severe issue on that hugetlb
fork won't wr-protect child cow pages, so child can potentially write to
parent private pages. Patch 2 addresses that.
After this series applied, "memfd_test hugetlbfs" should start to pass.
This patch (of 2):
F_SEAL_FUTURE_WRITE is missing for hugetlb starting from the first day.
There is a test program for that and it fails constantly.
$ ./memfd_test hugetlbfs
memfd-hugetlb: CREATE
memfd-hugetlb: BASIC
memfd-hugetlb: SEAL-WRITE
memfd-hugetlb: SEAL-FUTURE-WRITE
mmap() didn't fail as expected
Aborted (core dumped)
I think it's probably because no one is really running the hugetlbfs test.
Fix it by checking FUTURE_WRITE also in hugetlbfs_file_mmap() as what we
do in shmem_mmap(). Generalize a helper for that.
Link: https://lkml.kernel.org/r/20210503234356.9097-1-peterx@redhat.com
Link: https://lkml.kernel.org/r/20210503234356.9097-2-peterx@redhat.com
Fixes: ab3948f58ff84 ("mm/memfd: add an F_SEAL_FUTURE_WRITE seal to memfd")
Signed-off-by: Peter Xu <peterx@redhat.com>
Reported-by: Hugh Dickins <hughd@google.com>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Joel Fernandes (Google) <joel@joelfernandes.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
To ensure that instructions are observable in a new mapping, the arm64
set_pte_at() implementation cleans the D-cache and invalidates the
I-cache to the PoU. As an optimisation, this is only done on executable
mappings and the PG_dcache_clean page flag is set to avoid future cache
maintenance on the same page.
When two different processes map the same page (e.g. private executable
file or shared mapping) there's a potential race on checking and setting
PG_dcache_clean via set_pte_at() -> __sync_icache_dcache(). While on the
fault paths the page is locked (PG_locked), mprotect() does not take the
page lock. The result is that one process may see the PG_dcache_clean
flag set but the I/D cache maintenance not yet performed.
Avoid test_and_set_bit(PG_dcache_clean) in favour of separate test_bit()
and set_bit(). In the rare event of a race, the cache maintenance is
done twice.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: <stable@vger.kernel.org>
Cc: Will Deacon <will@kernel.org>
Cc: Steven Price <steven.price@arm.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20210514095001.13236-1-catalin.marinas@arm.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
|
Fix the following kernel-doc warning:
block/partitions/efi.c:685: warning: wrong kernel-doc identifier on line:
* efi_partition(struct parsed_partitions *state)
Cc: Alexander Viro <viro@math.psu.edu>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://lore.kernel.org/r/20210513171708.8391-1-bvanassche@acm.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
If a tag set is shared across request queues (e.g. SCSI LUNs) then the
block layer core keeps track of the number of active request queues in
tags->active_queues. blk_mq_tag_busy() and blk_mq_tag_idle() update that
atomic counter if the hctx flag BLK_MQ_F_TAG_QUEUE_SHARED is set. Make
sure that blk_mq_exit_queue() calls blk_mq_tag_idle() before that flag is
cleared by blk_mq_del_queue_tag_set().
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Hannes Reinecke <hare@suse.com>
Fixes: 0d2602ca30e4 ("blk-mq: improve support for shared tags maps")
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20210513171529.7977-1-bvanassche@acm.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
In case of shared sbitmap, request won't be held in plug list any more
sine commit 32bc15afed04 ("blk-mq: Facilitate a shared sbitmap per
tagset"), this way makes request merge from flush plug list & batching
submission not possible, so cause performance regression.
Yanhui reports performance regression when running sequential IO
test(libaio, 16 jobs, 8 depth for each job) in VM, and the VM disk
is emulated with image stored on xfs/megaraid_sas.
Fix the issue by recovering original behavior to allow to hold request
in plug list.
Cc: Yanhui Ma <yama@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Bart Van Assche <bvanassche@acm.org>
Cc: kashyap.desai@broadcom.com
Fixes: 32bc15afed04 ("blk-mq: Facilitate a shared sbitmap per tagset")
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20210514022052.1047665-1-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
xen_swiotlb_init calls swiotlb_late_init_with_tbl, which fails with
-ENOMEM if the swiotlb has already been initialized.
Add an explicit check io_tlb_default_mem != NULL at the beginning of
xen_swiotlb_init. If the swiotlb is already initialized print a warning
and return -EEXIST.
On x86, the error propagates.
On ARM, we don't actually need a special swiotlb buffer (yet), any
buffer would do. So ignore the error and continue.
CC: boris.ostrovsky@oracle.com
CC: jgross@suse.com
Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com>
Reviewed-by: Boris Ostrovsky <boris.ostrvsky@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20210512201823.1963-3-sstabellini@kernel.org
Signed-off-by: Juergen Gross <jgross@suse.com>
|
|
Although SWIOTLB_NO_FORCE is meant to allow later calls to swiotlb_init,
today dma_direct_map_page returns error if SWIOTLB_NO_FORCE.
For now, without a larger overhaul of SWIOTLB_NO_FORCE, the best we can
do is to avoid setting SWIOTLB_NO_FORCE in mem_init when we know that it
is going to be required later (e.g. Xen requires it).
CC: boris.ostrovsky@oracle.com
CC: jgross@suse.com
CC: catalin.marinas@arm.com
CC: will@kernel.org
CC: linux-arm-kernel@lists.infradead.org
Fixes: 2726bf3ff252 ("swiotlb: Make SWIOTLB_NO_FORCE perform no allocation")
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20210512201823.1963-2-sstabellini@kernel.org
Signed-off-by: Juergen Gross <jgross@suse.com>
|
|
Move xen_swiotlb_detect to a static inline function to make it available
to !CONFIG_XEN builds.
CC: boris.ostrovsky@oracle.com
CC: jgross@suse.com
Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/20210512201823.1963-1-sstabellini@kernel.org
Signed-off-by: Juergen Gross <jgross@suse.com>
|
|
Mohammed reports (https://bugzilla.kernel.org/show_bug.cgi?id=213029)
the commit e4ab4658f1cf ("clocksource/drivers/hyper-v: Handle vDSO
differences inline") broke vDSO on x86. The problem appears to be that
VDSO_CLOCKMODE_HVCLOCK is an enum value in 'enum vdso_clock_mode' and
'#ifdef VDSO_CLOCKMODE_HVCLOCK' branch evaluates to false (it is not
a define).
Use a dedicated HAVE_VDSO_CLOCKMODE_HVCLOCK define instead.
Fixes: e4ab4658f1cf ("clocksource/drivers/hyper-v: Handle vDSO differences inline")
Reported-by: Mohammed Gamal <mgamal@redhat.com>
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Link: https://lore.kernel.org/r/20210513073246.1715070-1-vkuznets@redhat.com
|
|
Since recent changes instead of storing a large array of struct
io_mapped_ubuf, we store pointers to them, that is 4 times slimmer and
we should not to so worry about restricting max number of registererd
buffer slots, increase the limit 4 times.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d3dee1da37f46da416aa96a16bf9e5094e10584d.1620990371.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
There are three types of requests that left disabled for sqpoll, namely
epoll ctx, statx, and resources update. Since SQPOLL task is now closely
mimics a userspace thread, remove the restrictions.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/909b52d70c45636d8d7897582474ea5aab5eed34.1620990306.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Always remove linked timeout on io_link_timeout_fn() from the master
request link list, otherwise we may get use-after-free when first
io_link_timeout_fn() puts linked timeout in the fail path, and then
will be found and put on master's free.
Cc: stable@vger.kernel.org # 5.10+
Fixes: 90cd7e424969d ("io_uring: track link timeout's master explicitly")
Reported-and-tested-by: syzbot+5a864149dd970b546223@syzkaller.appspotmail.com
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/69c46bf6ce37fec4fdcd98f0882e18eb07ce693a.1620990121.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Some interrupt handlers have an "extra" that saves 1 or 2
registers (r14, r15) in the paca save area and makes them available to
use by the handler.
The change to always save nvgprs in exception handlers lead to some
interrupt handlers saving those scratch r14 / r15 registers into the
interrupt frame's GPR saves, which get restored on interrupt exit.
Fix this by always reloading those scratch registers from paca before
the EXCEPTION_COMMON that saves nvgprs.
Fixes: 4228b2c3d20e ("powerpc/64e/interrupt: always save nvgprs on interrupt")
Reported-by: Christian Zigotzky <chzigotzky@xenosoft.de>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Tested-by: Christian Zigotzky <chzigotzky@xenosoft.de>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210514044008.1955783-1-npiggin@gmail.com
|
|
scv support introduced the notion of code that implicitly soft-masks
irqs due to the instruction addresses. This is required because scv
enters the kernel with MSR[EE]=1.
If a NMI (including soft-NMI) interrupt hits when we are implicitly
soft-masked then its regs->softe does not reflect this because it is
derived from the explicit soft mask state (paca->irq_soft_mask). This
makes arch_irq_disabled_regs(regs) return false.
This can trigger a warning in the soft-NMI watchdog code (shown below).
Fix it by having NMI interrupts set regs->softe to disabled in case of
interrupting an implicit soft-masked region.
------------[ cut here ]------------
WARNING: CPU: 41 PID: 1103 at arch/powerpc/kernel/watchdog.c:259 soft_nmi_interrupt+0x3e4/0x5f0
CPU: 41 PID: 1103 Comm: (spawn) Not tainted
NIP: c000000000039534 LR: c000000000039234 CTR: c000000000009a00
REGS: c000007fffbcf940 TRAP: 0700 Not tainted
MSR: 9000000000021033 <SF,HV,ME,IR,DR,RI,LE> CR: 22042482 XER: 200400ad
CFAR: c000000000039260 IRQMASK: 3
GPR00: c000000000039204 c000007fffbcfbe0 c000000001d6c300 0000000000000003
GPR04: 00007ffffa45d078 0000000000000000 0000000000000008 0000000000000020
GPR08: 0000007ffd4e0000 0000000000000000 c000007ffffceb00 7265677368657265
GPR12: 9000000000009033 c000007ffffceb00 00000f7075bf4480 000000000000002a
GPR16: 00000f705745a528 00007ffffa45ddd8 00000f70574d0008 0000000000000000
GPR20: 00000f7075c58d70 00000f7057459c38 0000000000000001 0000000000000040
GPR24: 0000000000000000 0000000000000029 c000000001dae058 0000000000000029
GPR28: 0000000000000000 0000000000000800 0000000000000009 c000007fffbcfd60
NIP [c000000000039534] soft_nmi_interrupt+0x3e4/0x5f0
LR [c000000000039234] soft_nmi_interrupt+0xe4/0x5f0
Call Trace:
[c000007fffbcfbe0] [c000000000039204] soft_nmi_interrupt+0xb4/0x5f0 (unreliable)
[c000007fffbcfcf0] [c00000000000c0e8] soft_nmi_common+0x138/0x1c4
--- interrupt: 900 at end_real_trampolines+0x0/0x1000
NIP: c000000000003000 LR: 00007ca426adb03c CTR: 900000000280f033
REGS: c000007fffbcfd60 TRAP: 0900
MSR: 9000000000009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 44042482 XER: 200400ad
CFAR: 00007ca426946020 IRQMASK: 0
GPR00: 00000000000000ad 00007ffffa45d050 00007ca426b07f00 0000000000000035
GPR04: 00007ffffa45d078 0000000000000000 0000000000000008 0000000000000020
GPR08: 0000000000000000 0000000000100000 0000000010000000 00007ffffa45d110
GPR12: 0000000000000001 00007ca426d4e680 00000f7075bf4480 000000000000002a
GPR16: 00000f705745a528 00007ffffa45ddd8 00000f70574d0008 0000000000000000
GPR20: 00000f7075c58d70 00000f7057459c38 0000000000000001 0000000000000040
GPR24: 0000000000000000 00000f7057473f68 0000000000000003 000000000000041b
GPR28: 00007ffffa45d4c4 0000000000000035 0000000000000000 00000f7057473f68
NIP [c000000000003000] end_real_trampolines+0x0/0x1000
LR [00007ca426adb03c] 0x7ca426adb03c
--- interrupt: 900
Instruction dump:
60000000 60000000 60420000 38600001 482b3ae5 60000000 e93f0138 a36d0008
7daa6b78 71290001 7f7907b4 4082fd34 <0fe00000> 4bfffd2c 60420000 ea6100a8
---[ end trace dc75f67d819779da ]---
Fixes: 118178e62e2e ("powerpc: move NMI entry/exit code into wrapper")
Reported-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210503111708.758261-1-npiggin@gmail.com
|
|
The stf entry barrier fallback is unsafe to execute in a semi-patched
state, which can happen when enabling/disabling the mitigation with
strict kernel RWX enabled and using the hash MMU.
See the previous commit for more details.
Fix it by changing the order in which we patch the instructions.
Note the stf barrier fallback is only used on Power6 or earlier.
Fixes: bd573a81312f ("powerpc/mm/64s: Allow STRICT_KERNEL_RWX again")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210513140800.1391706-2-mpe@ellerman.id.au
|
|
The entry flush mitigation can be enabled/disabled at runtime. When this
happens it results in the kernel patching its own instructions to
enable/disable the mitigation sequence.
With strict kernel RWX enabled instruction patching happens via a
secondary mapping of the kernel text, so that we don't have to make the
primary mapping writable. With the hash MMU this leads to a hash fault,
which causes us to execute the exception entry which contains the entry
flush mitigation.
This means we end up executing the entry flush in a semi-patched state,
ie. after we have patched the first instruction but before we patch the
second or third instruction of the sequence.
On machines with updated firmware the entry flush is a series of special
nops, and it's safe to to execute in a semi-patched state.
However when using the fallback flush the sequence is mflr/branch/mtlr,
and so it's not safe to execute if we have patched out the mflr but not
the other two instructions. Doing so leads to us corrputing LR, leading
to an oops, for example:
# echo 0 > /sys/kernel/debug/powerpc/entry_flush
kernel tried to execute exec-protected page (c000000002971000) - exploit attempt? (uid: 0)
BUG: Unable to handle kernel instruction fetch
Faulting instruction address: 0xc000000002971000
Oops: Kernel access of bad area, sig: 11 [#1]
LE PAGE_SIZE=64K MMU=Hash SMP NR_CPUS=2048 NUMA pSeries
CPU: 0 PID: 2215 Comm: bash Not tainted 5.13.0-rc1-00010-gda3bb206c9ce #1
NIP: c000000002971000 LR: c000000002971000 CTR: c000000000120c40
REGS: c000000013243840 TRAP: 0400 Not tainted (5.13.0-rc1-00010-gda3bb206c9ce)
MSR: 8000000010009033 <SF,EE,ME,IR,DR,RI,LE> CR: 48428482 XER: 00000000
...
NIP 0xc000000002971000
LR 0xc000000002971000
Call Trace:
do_patch_instruction+0xc4/0x340 (unreliable)
do_entry_flush_fixups+0x100/0x3b0
entry_flush_set+0x50/0xe0
simple_attr_write+0x160/0x1a0
full_proxy_write+0x8c/0x110
vfs_write+0xf0/0x340
ksys_write+0x84/0x140
system_call_exception+0x164/0x2d0
system_call_common+0xec/0x278
The simplest fix is to change the order in which we patch the
instructions, so that the sequence is always safe to execute. For the
non-fallback flushes it doesn't matter what order we patch in.
Fixes: bd573a81312f ("powerpc/mm/64s: Allow STRICT_KERNEL_RWX again")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210513140800.1391706-1-mpe@ellerman.id.au
|
Allison Henderson
Dave Chinner
Darrick J. Wong
Linus Torvalds
Tetsuo Handa
Christophe Leroy
Rasmus Villemoes
Jouni Roivas
Matthew Wilcox (Oracle)
Peter Collingbourne
Hugh Dickins
Axel Rasmussen
Phillip Lougher
Alistair Popple
Vlastimil Babka
Peter Xu
Catalin Marinas
Bart Van Assche
Ming Lei
Stefano Stabellini
Christoph Hellwig
Vitaly Kuznetsov
Pavel Begunkov
Nicholas Piggin
Michael Ellerman