linux-stable.git/fs/xfs, branch linux-6.3.y

xfs: verify buffer contents when we skip log replay

2023-06-09T08:48:25+00:00

commit 22ed903eee23a5b174e240f1cdfa9acf393a5210 upstream.

syzbot detected a crash during log recovery:

XFS (loop0): Mounting V5 Filesystem bfdc47fc-10d8-4eed-a562-11a831b3f791
XFS (loop0): Torn write (CRC failure) detected at log block 0x180. Truncating head block from 0x200.
XFS (loop0): Starting recovery (logdev: internal)
==================================================================
BUG: KASAN: slab-out-of-bounds in xfs_btree_lookup_get_block+0x15c/0x6d0 fs/xfs/libxfs/xfs_btree.c:1813
Read of size 8 at addr ffff88807e89f258 by task syz-executor132/5074

CPU: 0 PID: 5074 Comm: syz-executor132 Not tainted 6.2.0-rc1-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/26/2022
Call Trace:
 
 __dump_stack lib/dump_stack.c:88 [inline]
 dump_stack_lvl+0x1b1/0x290 lib/dump_stack.c:106
 print_address_description+0x74/0x340 mm/kasan/report.c:306
 print_report+0x107/0x1f0 mm/kasan/report.c:417
 kasan_report+0xcd/0x100 mm/kasan/report.c:517
 xfs_btree_lookup_get_block+0x15c/0x6d0 fs/xfs/libxfs/xfs_btree.c:1813
 xfs_btree_lookup+0x346/0x12c0 fs/xfs/libxfs/xfs_btree.c:1913
 xfs_btree_simple_query_range+0xde/0x6a0 fs/xfs/libxfs/xfs_btree.c:4713
 xfs_btree_query_range+0x2db/0x380 fs/xfs/libxfs/xfs_btree.c:4953
 xfs_refcount_recover_cow_leftovers+0x2d1/0xa60 fs/xfs/libxfs/xfs_refcount.c:1946
 xfs_reflink_recover_cow+0xab/0x1b0 fs/xfs/xfs_reflink.c:930
 xlog_recover_finish+0x824/0x920 fs/xfs/xfs_log_recover.c:3493
 xfs_log_mount_finish+0x1ec/0x3d0 fs/xfs/xfs_log.c:829
 xfs_mountfs+0x146a/0x1ef0 fs/xfs/xfs_mount.c:933
 xfs_fs_fill_super+0xf95/0x11f0 fs/xfs/xfs_super.c:1666
 get_tree_bdev+0x400/0x620 fs/super.c:1282
 vfs_get_tree+0x88/0x270 fs/super.c:1489
 do_new_mount+0x289/0xad0 fs/namespace.c:3145
 do_mount fs/namespace.c:3488 [inline]
 __do_sys_mount fs/namespace.c:3697 [inline]
 __se_sys_mount+0x2d3/0x3c0 fs/namespace.c:3674
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7f89fa3f4aca
Code: 83 c4 08 5b 5d c3 66 2e 0f 1f 84 00 00 00 00 00 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007fffd5fb5ef8 EFLAGS: 00000206 ORIG_RAX: 00000000000000a5
RAX: ffffffffffffffda RBX: 00646975756f6e2c RCX: 00007f89fa3f4aca
RDX: 0000000020000100 RSI: 0000000020009640 RDI: 00007fffd5fb5f10
RBP: 00007fffd5fb5f10 R08: 00007fffd5fb5f50 R09: 000000000000970d
R10: 0000000000200800 R11: 0000000000000206 R12: 0000000000000004
R13: 0000555556c6b2c0 R14: 0000000000200800 R15: 00007fffd5fb5f50
 

The fuzzed image contains an AGF with an obviously garbage
agf_refcount_level value of 32, and a dirty log with a buffer log item
for that AGF.  The ondisk AGF has a higher LSN than the recovered log
item.  xlog_recover_buf_commit_pass2 reads the buffer, compares the
LSNs, and decides to skip replay because the ondisk buffer appears to be
newer.

Unfortunately, the ondisk buffer is corrupt, but recovery just read the
buffer with no buffer ops specified:

	error = xfs_buf_read(mp->m_ddev_targp, buf_f->blf_blkno,
			buf_f->blf_len, buf_flags, &bp, NULL);

Skipping the buffer leaves its contents in memory unverified.  This sets
us up for a kernel crash because xfs_refcount_recover_cow_leftovers
reads the buffer (which is still around in XBF_DONE state, so no read
verification) and creates a refcountbt cursor of height 32.  This is
impossible so we run off the end of the cursor object and crash.

Fix this by invoking the verifier on all skipped buffers and aborting
log recovery if the ondisk buffer is corrupt.  It might be smarter to
force replay the log item atop the buffer and then see if it'll pass the
write verifier (like ext4 does) but for now let's go with the
conservative option where we stop immediately.

Link: https://syzkaller.appspot.com/bug?extid=7e9494b8b399902e994e
Signed-off-by: Darrick J. Wong 
Reviewed-by: Dave Chinner 
Signed-off-by: Dave Chinner 
Signed-off-by: Greg Kroah-Hartman

xfs: fix livelock in delayed allocation at ENOSPC

2023-05-30T13:17:22+00:00

commit 9419092fb2630c30e4ffeb9ef61007ef0c61827a upstream.

On a filesystem with a non-zero stripe unit and a large sequential
write, delayed allocation will set a minimum allocation length of
the stripe unit. If allocation fails because there are no extents
long enough for an aligned minlen allocation, it is supposed to
fall back to unaligned allocation which allows single block extents
to be allocated.

When the allocator code was rewritting in the 6.3 cycle, this
fallback was broken - the old code used args->fsbno as the both the
allocation target and the allocation result, the new code passes the
target as a separate parameter. The conversion didn't handle the
aligned->unaligned fallback path correctly - it reset args->fsbno to
the target fsbno on failure which broke allocation failure detection
in the high level code and so it never fell back to unaligned
allocations.

This resulted in a loop in writeback trying to allocate an aligned
block, getting a false positive success, trying to insert the result
in the BMBT. This did nothing because the extent already was in the
BMBT (merge results in an unchanged extent) and so it returned the
prior extent to the conversion code as the current iomap.

Because the iomap returned didn't cover the offset we tried to map,
xfs_convert_blocks() then retries the allocation, which fails in the
same way and now we have a livelock.

Reported-and-tested-by: Brian Foster 
Fixes: 85843327094f ("xfs: factor xfs_bmap_btalloc()")
Signed-off-by: Dave Chinner 
Reviewed-by: Darrick J. Wong 
Signed-off-by: Greg Kroah-Hartman

xfs: don't consider future format versions valid

2023-05-11T14:16:53+00:00

commit aa88019851a85df80cb77f143758b13aee09e3d9 upstream.

In commit fe08cc504448 we reworked the valid superblock version
checks. If it is a V5 filesystem, it is always valid, then we
checked if the version was less than V4 (reject) and then checked
feature fields in the V4 flags to determine if it was valid.

What we missed was that if the version is not V4 at this point,
we shoudl reject the fs. i.e. the check current treats V6+
filesystems as if it was a v4 filesystem. Fix this.

cc: stable@vger.kernel.org
Fixes: fe08cc504448 ("xfs: open code sb verifier feature checks")
Signed-off-by: Dave Chinner 
Reviewed-by: Darrick J. Wong 
Signed-off-by: Dave Chinner 
Signed-off-by: Greg Kroah-Hartman

xfs: fix mismerged tracepoints

2023-03-24T20:16:01+00:00

At some point in between sending this patch to the list and merging it
into for-next, the tracepoints got all mixed up because I've
over-reliant on automated tools not sucking.  The end result is that the
tracepoints are all wrong, so fix them.

Signed-off-by: Darrick J. Wong

xfs: clear incore AGFL_RESET state if it's not needed

2023-03-24T15:40:02+00:00

Prior to commit 7ac2ff8bb371, when we loaded the incore perag structure
with information from the AGF header, we would set or clear the
pagf_agfl_reset field based on whether or not the AGFL list was
misaligned within the block.  IOWs, it's an incore state bit that's
supposed to cache something in the ondisk metadata.  Therefore, the code
still needs to support clearing the incore bit if (somehow) the AGFL
were to correct itself.

It turns out that xfs_repair does exactly this -- phase 4 loads the AGF
to scan the rmapbt for corrupt records, which can set NEEDS_AGFL_RESET.
The scan unsets AGF_INIT but doesn't unset NEEDS_AGFL_RESET.  Phase 5
totally rewrites the AGFL and fixes the alignment problem, didn't clear
NEEDS_AGFL_RESET historically, and reloads the perag state to fix the
freelist.  This results in the AGFL being reset based on stale data,
which then causes the new AGFL blocks to be leaked.  A subsequent
xfs_repair -n then complains about the leaks.

One could argue that phase 5 ought to clear this bit directly when it
reloads the perag AGF data after rewriting the AGFL, but libxfs used to
handle this for us, so it should go back to doing that.

Found by fuzzing flfirst = ones in xfs/352.

Fixes: 7ac2ff8bb371 ("xfs: perags need atomic operational state")
Signed-off-by: Darrick J. Wong

xfs: pass the correct cursor to xfs_iomap_prealloc_size

2023-03-24T15:40:01+00:00

In xfs_buffered_write_iomap_begin, @icur is the iext cursor for the data
fork and @ccur is the cursor for the cow fork.  Pass in whichever cursor
corresponds to allocfork, because otherwise the xfs_iext_prev_extent
call can use the data fork cursor to walk off the end of the cow fork
structure.  Best case it returns the wrong results, worst case it does
this:

stack segment: 0000 [#1] PREEMPT SMP
CPU: 2 PID: 3141909 Comm: fsstress Tainted: G        W          6.3.0-rc2-xfsx #6.3.0-rc2 7bf5cc2e98997627cae5c930d890aba3aeec65dd
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20171121_152543-x86-ol7-builder-01.us.oracle.com-4.el7.1 04/01/2014
RIP: 0010:xfs_iext_prev+0x71/0x150 [xfs]
RSP: 0018:ffffc90002233aa8 EFLAGS: 00010297
RAX: 000000000000000f RBX: 000000000000000e RCX: 000000000000000c
RDX: 0000000000000002 RSI: 000000000000000e RDI: ffff8883d0019ba0
RBP: 989642409af8a7a7 R08: ffffea0000000001 R09: 0000000000000002
R10: 0000000000000000 R11: 000000000000000c R12: ffffc90002233b00
R13: ffff8883d0019ba0 R14: 989642409af8a6bf R15: 000ffffffffe0000
FS:  00007fdf8115f740(0000) GS:ffff88843fd00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007fdf8115e000 CR3: 0000000357256000 CR4: 00000000003506e0
Call Trace:
 
 xfs_iomap_prealloc_size.constprop.0.isra.0+0x1a6/0x410 [xfs 619a268fb2406d68bd34e007a816b27e70abc22c]
 xfs_buffered_write_iomap_begin+0xa87/0xc60 [xfs 619a268fb2406d68bd34e007a816b27e70abc22c]
 iomap_iter+0x132/0x2f0
 iomap_file_buffered_write+0x92/0x330
 xfs_file_buffered_write+0xb1/0x330 [xfs 619a268fb2406d68bd34e007a816b27e70abc22c]
 vfs_write+0x2eb/0x410
 ksys_write+0x65/0xe0
 do_syscall_64+0x2b/0x80
 entry_SYSCALL_64_after_hwframe+0x46/0xb0

Found by xfs/538 in alwayscow mode, but this doesn't seem particular to
that test.

Fixes: 590b16516ef3 ("xfs: refactor xfs_iomap_prealloc_size")
Actually-Fixes: 66ae56a53f0e ("xfs: introduce an always_cow mode")
Signed-off-by: Darrick J. Wong

xfs: test dir/attr hash when loading module

2023-03-19T16:55:49+00:00

Back in the 6.2-rc1 days, Eric Whitney reported a fstests regression in
ext4 against generic/454.  The cause of this test failure was the
unfortunate combination of setting an xattr name containing UTF8 encoded
emoji, an xattr hash function that accepted a char pointer with no
explicit signedness, signed type extension of those chars to an int, and
the 6.2 build tools maintainers deciding to mandate -funsigned-char
across the board.  As a result, the ondisk extended attribute structure
written out by 6.1 and 6.2 were not the same.

This discrepancy, in fact, had been noticeable if a filesystem with such
an xattr were moved between any two architectures that don't employ the
same signedness of a raw "char" declaration.  The only reason anyone
noticed is that x86 gcc defaults to signed, and no such -funsigned-char
update was made to e2fsprogs, so e2fsck immediately started reporting
data corruption.

After a day and a half of discussing how to handle this use case (xattrs
with bit 7 set anywhere in the name) without breaking existing users,
Linus merged his own patch and didn't tell the maintainer.  None of the
ext4 developers realized this until AUTOSEL announced that the commit
had been backported to stable.

In the end, this problem could have been detected much earlier if there
had been any useful tests of hash function(s) in use inside ext4 to make
sure that they always produce the same outputs given the same inputs.

The XFS dirent/xattr name hash takes a uint8_t*, so I don't think it's
vulnerable to this problem.  However, let's avoid all this drama by
adding our own self test to check that the da hash produces the same
outputs for a static pile of inputs on various platforms.  This enables
us to fix any breakage that may result in a controlled fashion.  The
buffer and test data are identical to the patches submitted to xfsprogs.

Link: https://lore.kernel.org/linux-ext4/Y8bpkm3jA3bDm3eL@debian-BULLSEYE-live-builder-AMD64/
Link: https://lore.kernel.org/linux-xfs/ZBUKCRR7xvIqPrpX@destitution/T/#md38272cc684e2c0d61494435ccbb91f022e8dee4
Signed-off-by: Darrick J. Wong 
Reviewed-by: Dave Chinner

xfs: add tracepoints for each of the externally visible allocators

2023-03-19T16:55:49+00:00

There are now five separate space allocator interfaces exposed to the
rest of XFS for five different strategies to find space.  Add
tracepoints for each of them so that I can tell from a trace dump
exactly which ones got called and what happened underneath them.  Add a
sixth so it's more obvious if an allocation actually happened.

Signed-off-by: Darrick J. Wong 
Reviewed-by: Dave Chinner

xfs: walk all AGs if TRYLOCK passed to xfs_alloc_vextent_iterate_ags

2023-03-19T16:55:48+00:00

Callers of xfs_alloc_vextent_iterate_ags that pass in the TRYLOCK flag
want us to perform a non-blocking scan of the AGs for free space.  There
are no ordering constraints for non-blocking AGF lock acquisition, so
the scan can freely start over at AG 0 even when minimum_agno > 0.

This manifests fairly reliably on xfs/294 on 6.3-rc2 with the parent
pointer patchset applied and the realtime volume enabled.  I observed
the following sequence as part of an xfs_dir_createname call:

0. Fragment the free space, then allocate nearly all the free space in
   all AGs except AG 0.

1. Create a directory in AG 2 and let it grow for a while.

2. Try to allocate 2 blocks to expand the dirent part of a directory.
   The space will be allocated out of AG 0, but the allocation will not
   be contiguous.  This (I think) activates the LOWMODE allocator.

3. The bmapi call decides to convert from extents to bmbt format and
   tries to allocate 1 block.  This allocation request calls
   xfs_alloc_vextent_start_ag with the inode number, which starts the
   scan at AG 2.  We ignore AG 0 (with all its free space) and instead
   scrape AG 2 and 3 for more space.  We find one block, but this now
   kicks t_highest_agno to 3.

4. The createname call decides it needs to split the dabtree.  It tries
   to allocate even more space with xfs_alloc_vextent_start_ag, but now
   we're constrained to AG 3, and we don't find the space.  The
   createname returns ENOSPC and the filesystem shuts down.

This change fixes the problem by making the trylock scan wrap around to
AG 0 if it doesn't like the AGs that it finds.  Since the current
transaction itself holds AGF 0, the trylock of AGF 0 will succeed, and
we take space from the AG that has plenty.

Signed-off-by: Darrick J. Wong 
Reviewed-by: Dave Chinner

xfs: try to idiot-proof the allocators

2023-03-16T16:32:18+00:00

In porting his development branch to 6.3-rc1, yours truly has
repeatedly screwed up the args->pag being fed to the xfs_alloc_vextent*
functions.  Add some debugging assertions to test the preconditions
required of the callers.

Signed-off-by: Darrick J. Wong 
Reviewed-by: Dave Chinner