Re: [syzbot] [mm?] KMSAN: uninit-value in zswap_store

Re: [syzbot] [mm?] KMSAN: uninit-value in zswap_store

[Hugh Dickins]

[-- Attachment #1: Type: text/plain, Size: 8876 bytes --]

On Mon, 10 Jun 2024, Yosry Ahmed wrote:
> On Sat, Jun 8, 2024 at 2:09 PM syzbot
> <syzbot+9c1fe13fcb51574b249b@syzkaller.appspotmail.com> wrote:
> >
> > Hello,
> >
> > syzbot found the following issue on:
> >
> > HEAD commit:    614da38e2f7a Merge tag 'hid-for-linus-2024051401' of git:/..
> > git tree:       upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=113f3fd6980000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=f5d2cbf33633f507
> > dashboard link: https://syzkaller.appspot.com/bug?extid=9c1fe13fcb51574b249b
> > compiler:       Debian clang version 15.0.6, GNU ld (GNU Binutils for Debian) 2.40
> >
> > Unfortunately, I don't have any reproducer for this issue yet.
> >
> > Downloadable assets:
> > disk image: https://storage.googleapis.com/syzbot-assets/89eafb874b71/disk-614da38e.raw.xz
> > vmlinux: https://storage.googleapis.com/syzbot-assets/356000512ad9/vmlinux-614da38e.xz
> > kernel image: https://storage.googleapis.com/syzbot-assets/839c73939115/bzImage-614da38e.xz
> >
> > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > Reported-by: syzbot+9c1fe13fcb51574b249b@syzkaller.appspotmail.com
> 
> This doesn't look like a zswap bug, and perhaps not a bug at all but I
> am not sure. Zswap is reading the data in a folio to check if it is
> filled with a repeated pattern, and it is encountering uninitialized
> data. I am not sure if this is expected behavior or not.

I agree, it doesn't look like a zswap bug: zswap is just following orders.

> 
> Could it just be that the size of the data written by syzbot is not
> divisible by PAGE_SIZE, so part of the final page is left
> uninitialized? Do we keep the unwritten parts of a shmem page
> uninitialized by any chance?

Shmem can keep uninitialized data pages around, but should be zeroing
what's still uninitialized before it can reach the outside world -
in this case being written to swap.  I don't know of a shmem bug here.

> 
> Hugh, do you mind taking a quick look at whether this is a real bug?
> 
> If this is expected behavior, perhaps there is some annotation we can
> use in zswap that it is fine to encounter uninitialized data when
> reading the folio.

I've not been faced with a KMSAN report before, so I might well be
misunderstanding its language: but this looks like an ext4 "bug" to me.

I think the story that the three KMSAN stacks are telling is this:
syzbot has an ext4 filesystem on a loop device on a tmpfs file (I do
exactly that too); ext4 is doing some ext4_xattr_inode_write() business,
in the course of which it writes back a not-fully-initialized block to
the loop device; shmem faithfully copies that data into its pagecache,
then later when under memory pressure that page gets "written" out to
zswap: where zswap_is_page_same_filled() takes an interest in the data,
and KMSAN objects because some of it was not originally initialized.

If that's a correct interpretation, then it's probably not a big deal:
it's probably the uninitialized end of a buffer that's written out,
not any part of the "disk" which ext4 would ever show to a user; but
I do agree with KMSAN that ext4 would do better to clear that area,
rather than accidentally storing someone else's super-secret info.

Hugh

> 
> >
> > =====================================================
> > BUG: KMSAN: uninit-value in zswap_is_page_same_filled mm/zswap.c:1481 [inline]
> > BUG: KMSAN: uninit-value in zswap_store+0x1008/0x2ca0 mm/zswap.c:1553
> >  zswap_is_page_same_filled mm/zswap.c:1481 [inline]
> >  zswap_store+0x1008/0x2ca0 mm/zswap.c:1553
> >  swap_writepage+0x126/0x4c0 mm/page_io.c:198
> >  shmem_writepage+0x1826/0x1f70 mm/shmem.c:1518
> >  pageout mm/vmscan.c:660 [inline]
> >  shrink_folio_list+0x4a55/0x7910 mm/vmscan.c:1323
> >  evict_folios+0x9d7f/0xcc20 mm/vmscan.c:4537
> >  try_to_shrink_lruvec+0x160e/0x1a50 mm/vmscan.c:4733
> >  shrink_one+0x66f/0xd40 mm/vmscan.c:4772
> >  shrink_many mm/vmscan.c:4835 [inline]
> >  lru_gen_shrink_node mm/vmscan.c:4935 [inline]
> >  shrink_node+0x4856/0x55f0 mm/vmscan.c:5894
> >  kswapd_shrink_node mm/vmscan.c:6704 [inline]
> >  balance_pgdat mm/vmscan.c:6895 [inline]
> >  kswapd+0x1eba/0x4460 mm/vmscan.c:7164
> >  kthread+0x3e2/0x540 kernel/kthread.c:389
> >  ret_from_fork+0x6d/0x90 arch/x86/kernel/process.c:147
> >  ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
> >
> > Uninit was stored to memory at:
> >  memcpy_from_iter lib/iov_iter.c:73 [inline]
> >  iterate_bvec include/linux/iov_iter.h:122 [inline]
> >  iterate_and_advance2 include/linux/iov_iter.h:249 [inline]
> >  iterate_and_advance include/linux/iov_iter.h:271 [inline]
> >  __copy_from_iter lib/iov_iter.c:249 [inline]
> >  copy_page_from_iter_atomic+0x12b7/0x2ae0 lib/iov_iter.c:481
> >  generic_perform_write+0x4c1/0xc60 mm/filemap.c:3982
> >  shmem_file_write_iter+0x2bd/0x2f0 mm/shmem.c:2920
> >  do_iter_readv_writev+0x7e6/0x960
> >  vfs_iter_write+0x459/0xd00 fs/read_write.c:895
> >  lo_write_bvec drivers/block/loop.c:246 [inline]
> >  lo_write_simple drivers/block/loop.c:267 [inline]
> >  do_req_filebacked drivers/block/loop.c:491 [inline]
> >  loop_handle_cmd drivers/block/loop.c:1907 [inline]
> >  loop_process_work+0x1502/0x3440 drivers/block/loop.c:1942
> >  loop_rootcg_workfn+0x2b/0x40 drivers/block/loop.c:1973
> >  process_one_work kernel/workqueue.c:3267 [inline]
> >  process_scheduled_works+0xa81/0x1bd0 kernel/workqueue.c:3348
> >  worker_thread+0xea5/0x1560 kernel/workqueue.c:3429
> >  kthread+0x3e2/0x540 kernel/kthread.c:389
> >  ret_from_fork+0x6d/0x90 arch/x86/kernel/process.c:147
> >  ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
> >
> > Uninit was created at:
> >  __alloc_pages+0x9d6/0xe70 mm/page_alloc.c:4598
> >  alloc_pages_mpol+0x299/0x990 mm/mempolicy.c:2264
> >  alloc_pages mm/mempolicy.c:2335 [inline]
> >  folio_alloc+0x1d0/0x230 mm/mempolicy.c:2342
> >  filemap_alloc_folio+0xa6/0x440 mm/filemap.c:984
> >  __filemap_get_folio+0xa10/0x14b0 mm/filemap.c:1926
> >  grow_dev_folio fs/buffer.c:1042 [inline]
> >  grow_buffers fs/buffer.c:1108 [inline]
> >  __getblk_slow fs/buffer.c:1134 [inline]
> >  bdev_getblk+0x39b/0xc80 fs/buffer.c:1429
> >  __getblk include/linux/buffer_head.h:355 [inline]
> >  sb_getblk include/linux/buffer_head.h:361 [inline]
> >  ext4_getblk+0x3da/0xe00 fs/ext4/inode.c:843
> >  ext4_xattr_inode_write fs/ext4/xattr.c:1421 [inline]
> >  ext4_xattr_inode_lookup_create fs/ext4/xattr.c:1596 [inline]
> >  ext4_xattr_set_entry+0x574d/0x6880 fs/ext4/xattr.c:1718
> >  ext4_xattr_block_set+0xb94/0x4fb0 fs/ext4/xattr.c:2037
> >  ext4_xattr_move_to_block fs/ext4/xattr.c:2654 [inline]
> >  ext4_xattr_make_inode_space fs/ext4/xattr.c:2729 [inline]
> >  ext4_expand_extra_isize_ea+0x20bd/0x3560 fs/ext4/xattr.c:2821
> >  __ext4_expand_extra_isize+0x5dc/0x680 fs/ext4/inode.c:5789
> >  ext4_try_to_expand_extra_isize fs/ext4/inode.c:5832 [inline]
> >  __ext4_mark_inode_dirty+0x70d/0xa10 fs/ext4/inode.c:5910
> >  ext4_delete_inline_entry+0x650/0x7d0 fs/ext4/inline.c:1753
> >  ext4_delete_entry+0x13f/0x7d0 fs/ext4/namei.c:2719
> >  __ext4_unlink+0x9b8/0x11b0 fs/ext4/namei.c:3273
> >  ext4_unlink+0x226/0x630 fs/ext4/namei.c:3321
> >  vfs_unlink+0x676/0xa30 fs/namei.c:4343
> >  do_unlinkat+0x823/0xe10 fs/namei.c:4407
> >  __do_sys_unlinkat fs/namei.c:4450 [inline]
> >  __se_sys_unlinkat fs/namei.c:4443 [inline]
> >  __x64_sys_unlinkat+0x17c/0x230 fs/namei.c:4443
> >  x64_sys_call+0x846/0x3b50 arch/x86/include/generated/asm/syscalls_64.h:264
> >  do_syscall_x64 arch/x86/entry/common.c:52 [inline]
> >  do_syscall_64+0xcf/0x1e0 arch/x86/entry/common.c:83
> >  entry_SYSCALL_64_after_hwframe+0x77/0x7f
> >
> > CPU: 1 PID: 88 Comm: kswapd1 Not tainted 6.9.0-syzkaller-02707-g614da38e2f7a #0
> > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 04/02/2024
> > =====================================================
> >
> >
> > ---
> > This report is generated by a bot. It may contain errors.
> > See https://goo.gl/tpsmEJ for more information about syzbot.
> > syzbot engineers can be reached at syzkaller@googlegroups.com.
> >
> > syzbot will keep track of this issue. See:
> > https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
> >
> > If the report is already addressed, let syzbot know by replying with:
> > #syz fix: exact-commit-title
> >
> > If you want to overwrite report's subsystems, reply with:
> > #syz set subsystems: new-subsystem
> > (See the list of subsystem names on the web dashboard)
> >
> > If the report is a duplicate of another one, reply with:
> > #syz dup: exact-subject-of-another-report
> >
> > If you want to undo deduplication, reply with:
> > #syz undup
> 

Re: [syzbot] [mm?] KMSAN: uninit-value in zswap_store

[Yosry Ahmed]

On Tue, Jun 11, 2024 at 2:14 AM Hugh Dickins <hughd@google.com> wrote:
>
> On Mon, 10 Jun 2024, Yosry Ahmed wrote:
> > On Sat, Jun 8, 2024 at 2:09 PM syzbot
> > <syzbot+9c1fe13fcb51574b249b@syzkaller.appspotmail.com> wrote:
> > >
> > > Hello,
> > >
> > > syzbot found the following issue on:
> > >
> > > HEAD commit:    614da38e2f7a Merge tag 'hid-for-linus-2024051401' of git:/..
> > > git tree:       upstream
> > > console output: https://syzkaller.appspot.com/x/log.txt?x=113f3fd6980000
> > > kernel config:  https://syzkaller.appspot.com/x/.config?x=f5d2cbf33633f507
> > > dashboard link: https://syzkaller.appspot.com/bug?extid=9c1fe13fcb51574b249b
> > > compiler:       Debian clang version 15.0.6, GNU ld (GNU Binutils for Debian) 2.40
> > >
> > > Unfortunately, I don't have any reproducer for this issue yet.
> > >
> > > Downloadable assets:
> > > disk image: https://storage.googleapis.com/syzbot-assets/89eafb874b71/disk-614da38e.raw.xz
> > > vmlinux: https://storage.googleapis.com/syzbot-assets/356000512ad9/vmlinux-614da38e.xz
> > > kernel image: https://storage.googleapis.com/syzbot-assets/839c73939115/bzImage-614da38e.xz
> > >
> > > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > > Reported-by: syzbot+9c1fe13fcb51574b249b@syzkaller.appspotmail.com
> >
> > This doesn't look like a zswap bug, and perhaps not a bug at all but I
> > am not sure. Zswap is reading the data in a folio to check if it is
> > filled with a repeated pattern, and it is encountering uninitialized
> > data. I am not sure if this is expected behavior or not.
>
> I agree, it doesn't look like a zswap bug: zswap is just following orders.
>
> >
> > Could it just be that the size of the data written by syzbot is not
> > divisible by PAGE_SIZE, so part of the final page is left
> > uninitialized? Do we keep the unwritten parts of a shmem page
> > uninitialized by any chance?
>
> Shmem can keep uninitialized data pages around, but should be zeroing
> what's still uninitialized before it can reach the outside world -
> in this case being written to swap.  I don't know of a shmem bug here.

Thanks for taking a look at this and stating that shmem will not pass
uninitialized memory to swap/zswap.

>
> >
> > Hugh, do you mind taking a quick look at whether this is a real bug?
> >
> > If this is expected behavior, perhaps there is some annotation we can
> > use in zswap that it is fine to encounter uninitialized data when
> > reading the folio.
>
> I've not been faced with a KMSAN report before, so I might well be
> misunderstanding its language: but this looks like an ext4 "bug" to me.
>
> I think the story that the three KMSAN stacks are telling is this:
> syzbot has an ext4 filesystem on a loop device on a tmpfs file (I do
> exactly that too); ext4 is doing some ext4_xattr_inode_write() business,
> in the course of which it writes back a not-fully-initialized block to
> the loop device; shmem faithfully copies that data into its pagecache,
> then later when under memory pressure that page gets "written" out to
> zswap: where zswap_is_page_same_filled() takes an interest in the data,
> and KMSAN objects because some of it was not originally initialized.
>
> If that's a correct interpretation, then it's probably not a big deal:
> it's probably the uninitialized end of a buffer that's written out,
> not any part of the "disk" which ext4 would ever show to a user; but
> I do agree with KMSAN that ext4 would do better to clear that area,
> rather than accidentally storing someone else's super-secret info.

Interesting, I never saw an ext4 filesystem on a loop device on a
tmpfs file :) My understanding of the KMSAN stacks seems to match
yours. Thanks for looping in ext4 folks, hopefully they will shed more
light on this from their end.

>
> Hugh
>
> >
> > >
> > > =====================================================
> > > BUG: KMSAN: uninit-value in zswap_is_page_same_filled mm/zswap.c:1481 [inline]
> > > BUG: KMSAN: uninit-value in zswap_store+0x1008/0x2ca0 mm/zswap.c:1553
> > >  zswap_is_page_same_filled mm/zswap.c:1481 [inline]
> > >  zswap_store+0x1008/0x2ca0 mm/zswap.c:1553
> > >  swap_writepage+0x126/0x4c0 mm/page_io.c:198
> > >  shmem_writepage+0x1826/0x1f70 mm/shmem.c:1518
> > >  pageout mm/vmscan.c:660 [inline]
> > >  shrink_folio_list+0x4a55/0x7910 mm/vmscan.c:1323
> > >  evict_folios+0x9d7f/0xcc20 mm/vmscan.c:4537
> > >  try_to_shrink_lruvec+0x160e/0x1a50 mm/vmscan.c:4733
> > >  shrink_one+0x66f/0xd40 mm/vmscan.c:4772
> > >  shrink_many mm/vmscan.c:4835 [inline]
> > >  lru_gen_shrink_node mm/vmscan.c:4935 [inline]
> > >  shrink_node+0x4856/0x55f0 mm/vmscan.c:5894
> > >  kswapd_shrink_node mm/vmscan.c:6704 [inline]
> > >  balance_pgdat mm/vmscan.c:6895 [inline]
> > >  kswapd+0x1eba/0x4460 mm/vmscan.c:7164
> > >  kthread+0x3e2/0x540 kernel/kthread.c:389
> > >  ret_from_fork+0x6d/0x90 arch/x86/kernel/process.c:147
> > >  ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
> > >
> > > Uninit was stored to memory at:
> > >  memcpy_from_iter lib/iov_iter.c:73 [inline]
> > >  iterate_bvec include/linux/iov_iter.h:122 [inline]
> > >  iterate_and_advance2 include/linux/iov_iter.h:249 [inline]
> > >  iterate_and_advance include/linux/iov_iter.h:271 [inline]
> > >  __copy_from_iter lib/iov_iter.c:249 [inline]
> > >  copy_page_from_iter_atomic+0x12b7/0x2ae0 lib/iov_iter.c:481
> > >  generic_perform_write+0x4c1/0xc60 mm/filemap.c:3982
> > >  shmem_file_write_iter+0x2bd/0x2f0 mm/shmem.c:2920
> > >  do_iter_readv_writev+0x7e6/0x960
> > >  vfs_iter_write+0x459/0xd00 fs/read_write.c:895
> > >  lo_write_bvec drivers/block/loop.c:246 [inline]
> > >  lo_write_simple drivers/block/loop.c:267 [inline]
> > >  do_req_filebacked drivers/block/loop.c:491 [inline]
> > >  loop_handle_cmd drivers/block/loop.c:1907 [inline]
> > >  loop_process_work+0x1502/0x3440 drivers/block/loop.c:1942
> > >  loop_rootcg_workfn+0x2b/0x40 drivers/block/loop.c:1973
> > >  process_one_work kernel/workqueue.c:3267 [inline]
> > >  process_scheduled_works+0xa81/0x1bd0 kernel/workqueue.c:3348
> > >  worker_thread+0xea5/0x1560 kernel/workqueue.c:3429
> > >  kthread+0x3e2/0x540 kernel/kthread.c:389
> > >  ret_from_fork+0x6d/0x90 arch/x86/kernel/process.c:147
> > >  ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
> > >
> > > Uninit was created at:
> > >  __alloc_pages+0x9d6/0xe70 mm/page_alloc.c:4598
> > >  alloc_pages_mpol+0x299/0x990 mm/mempolicy.c:2264
> > >  alloc_pages mm/mempolicy.c:2335 [inline]
> > >  folio_alloc+0x1d0/0x230 mm/mempolicy.c:2342
> > >  filemap_alloc_folio+0xa6/0x440 mm/filemap.c:984
> > >  __filemap_get_folio+0xa10/0x14b0 mm/filemap.c:1926
> > >  grow_dev_folio fs/buffer.c:1042 [inline]
> > >  grow_buffers fs/buffer.c:1108 [inline]
> > >  __getblk_slow fs/buffer.c:1134 [inline]
> > >  bdev_getblk+0x39b/0xc80 fs/buffer.c:1429
> > >  __getblk include/linux/buffer_head.h:355 [inline]
> > >  sb_getblk include/linux/buffer_head.h:361 [inline]
> > >  ext4_getblk+0x3da/0xe00 fs/ext4/inode.c:843
> > >  ext4_xattr_inode_write fs/ext4/xattr.c:1421 [inline]
> > >  ext4_xattr_inode_lookup_create fs/ext4/xattr.c:1596 [inline]
> > >  ext4_xattr_set_entry+0x574d/0x6880 fs/ext4/xattr.c:1718
> > >  ext4_xattr_block_set+0xb94/0x4fb0 fs/ext4/xattr.c:2037
> > >  ext4_xattr_move_to_block fs/ext4/xattr.c:2654 [inline]
> > >  ext4_xattr_make_inode_space fs/ext4/xattr.c:2729 [inline]
> > >  ext4_expand_extra_isize_ea+0x20bd/0x3560 fs/ext4/xattr.c:2821
> > >  __ext4_expand_extra_isize+0x5dc/0x680 fs/ext4/inode.c:5789
> > >  ext4_try_to_expand_extra_isize fs/ext4/inode.c:5832 [inline]
> > >  __ext4_mark_inode_dirty+0x70d/0xa10 fs/ext4/inode.c:5910
> > >  ext4_delete_inline_entry+0x650/0x7d0 fs/ext4/inline.c:1753
> > >  ext4_delete_entry+0x13f/0x7d0 fs/ext4/namei.c:2719
> > >  __ext4_unlink+0x9b8/0x11b0 fs/ext4/namei.c:3273
> > >  ext4_unlink+0x226/0x630 fs/ext4/namei.c:3321
> > >  vfs_unlink+0x676/0xa30 fs/namei.c:4343
> > >  do_unlinkat+0x823/0xe10 fs/namei.c:4407
> > >  __do_sys_unlinkat fs/namei.c:4450 [inline]
> > >  __se_sys_unlinkat fs/namei.c:4443 [inline]
> > >  __x64_sys_unlinkat+0x17c/0x230 fs/namei.c:4443
> > >  x64_sys_call+0x846/0x3b50 arch/x86/include/generated/asm/syscalls_64.h:264
> > >  do_syscall_x64 arch/x86/entry/common.c:52 [inline]
> > >  do_syscall_64+0xcf/0x1e0 arch/x86/entry/common.c:83
> > >  entry_SYSCALL_64_after_hwframe+0x77/0x7f
> > >
> > > CPU: 1 PID: 88 Comm: kswapd1 Not tainted 6.9.0-syzkaller-02707-g614da38e2f7a #0
> > > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 04/02/2024
> > > =====================================================
> > >
> > >
> > > ---
> > > This report is generated by a bot. It may contain errors.
> > > See https://goo.gl/tpsmEJ for more information about syzbot.
> > > syzbot engineers can be reached at syzkaller@googlegroups.com.
> > >
> > > syzbot will keep track of this issue. See:
> > > https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
> > >
> > > If the report is already addressed, let syzbot know by replying with:
> > > #syz fix: exact-commit-title
> > >
> > > If you want to overwrite report's subsystems, reply with:
> > > #syz set subsystems: new-subsystem
> > > (See the list of subsystem names on the web dashboard)
> > >
> > > If the report is a duplicate of another one, reply with:
> > > #syz dup: exact-subject-of-another-report
> > >
> > > If you want to undo deduplication, reply with:
> > > #syz undup
> >

Re: [syzbot] [mm?] KMSAN: uninit-value in zswap_store

[Jan Kara]

On Tue 11-06-24 02:13:59, Hugh Dickins wrote:
> On Mon, 10 Jun 2024, Yosry Ahmed wrote:
> > Hugh, do you mind taking a quick look at whether this is a real bug?
> > 
> > If this is expected behavior, perhaps there is some annotation we can
> > use in zswap that it is fine to encounter uninitialized data when
> > reading the folio.
> 
> I've not been faced with a KMSAN report before, so I might well be
> misunderstanding its language: but this looks like an ext4 "bug" to me.
> 
> I think the story that the three KMSAN stacks are telling is this:
> syzbot has an ext4 filesystem on a loop device on a tmpfs file (I do
> exactly that too); ext4 is doing some ext4_xattr_inode_write() business,
> in the course of which it writes back a not-fully-initialized block to
> the loop device; shmem faithfully copies that data into its pagecache,
> then later when under memory pressure that page gets "written" out to
> zswap: where zswap_is_page_same_filled() takes an interest in the data,
> and KMSAN objects because some of it was not originally initialized.
> 
> If that's a correct interpretation, then it's probably not a big deal:
> it's probably the uninitialized end of a buffer that's written out,
> not any part of the "disk" which ext4 would ever show to a user; but
> I do agree with KMSAN that ext4 would do better to clear that area,
> rather than accidentally storing someone else's super-secret info.

Yes, that seems to be accurate.  ext4_xattr_inode_write() stores large
extended attribute in the inode and we don't bother to zero out the tail of
the block we use since we never access bytes beyond xattr size. Frankly I
don't consider this a bug since the uninitialized bytes are never exposed
to (unpriviledged) userspace. But I agree that out of pure precaution and
because it doesn't cost much in terms of performance we could zero out the
block tail.

								Honza
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [syzbot] [mm?] KMSAN: uninit-value in zswap_store

[Eric Biggers]

On Wed, Jun 12, 2024 at 09:50:27AM +0200, Jan Kara wrote:
> On Tue 11-06-24 02:13:59, Hugh Dickins wrote:
> > On Mon, 10 Jun 2024, Yosry Ahmed wrote:
> > > Hugh, do you mind taking a quick look at whether this is a real bug?
> > > 
> > > If this is expected behavior, perhaps there is some annotation we can
> > > use in zswap that it is fine to encounter uninitialized data when
> > > reading the folio.
> > 
> > I've not been faced with a KMSAN report before, so I might well be
> > misunderstanding its language: but this looks like an ext4 "bug" to me.
> > 
> > I think the story that the three KMSAN stacks are telling is this:
> > syzbot has an ext4 filesystem on a loop device on a tmpfs file (I do
> > exactly that too); ext4 is doing some ext4_xattr_inode_write() business,
> > in the course of which it writes back a not-fully-initialized block to
> > the loop device; shmem faithfully copies that data into its pagecache,
> > then later when under memory pressure that page gets "written" out to
> > zswap: where zswap_is_page_same_filled() takes an interest in the data,
> > and KMSAN objects because some of it was not originally initialized.
> > 
> > If that's a correct interpretation, then it's probably not a big deal:
> > it's probably the uninitialized end of a buffer that's written out,
> > not any part of the "disk" which ext4 would ever show to a user; but
> > I do agree with KMSAN that ext4 would do better to clear that area,
> > rather than accidentally storing someone else's super-secret info.
> 
> Yes, that seems to be accurate.  ext4_xattr_inode_write() stores large
> extended attribute in the inode and we don't bother to zero out the tail of
> the block we use since we never access bytes beyond xattr size. Frankly I
> don't consider this a bug since the uninitialized bytes are never exposed
> to (unpriviledged) userspace. But I agree that out of pure precaution and
> because it doesn't cost much in terms of performance we could zero out the
> block tail.

Writing uninitialized memory to disk is definitely a bug.

- Eric