On Wed, Dec 18, 2024 at 11:32:51AM +0100, Mikhail Zaslonko wrote:
> Since the input data length passed to zlib_compress_folios() can be
> arbitrary, always setting strm.avail_in to a multiple of PAGE_SIZE may
> cause read-in bytes to exceed the input range. Currently this triggers
> an assert in btrfs_compress_folios() on the debug kernel (see below).
> Fix strm.avail_in calculation for S390 hardware acceleration path.
>
> assertion failed: *total_in <= orig_len, in fs/btrfs/compression.c:1041
> ------------[ cut here ]------------
> kernel BUG at fs/btrfs/compression.c:1041!
> monitor event: 0040 ilc:2 [#1] PREEMPT SMP
> CPU: 16 UID: 0 PID: 325 Comm: kworker/u273:3 Not tainted 6.13.0-20241204.rc1.git6.fae3b21430ca.300.fc41.s390x+debug #1
> Hardware name: IBM 3931 A01 703 (z/VM 7.4.0)
> Workqueue: btrfs-delalloc btrfs_work_helper
> Krnl PSW : 0704d00180000000 0000021761df6538 (btrfs_compress_folios+0x198/0x1a0)
> R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:0 AS:3 CC:1 PM:0 RI:0 EA:3
> Krnl GPRS: 0000000080000000 0000000000000001 0000000000000047 0000000000000000
> 0000000000000006 ffffff01757bb000 000001976232fcc0 000000000000130c
> 000001976232fcd0 000001976232fcc8 00000118ff4a0e30 0000000000000001
> 00000111821ab400 0000011100000000 0000021761df6534 000001976232fb58
> Krnl Code: 0000021761df6528: c020006f5ef4 larl %r2,0000021762be2310
> 0000021761df652e: c0e5ffbd09d5 brasl %r14,00000217615978d8
> #0000021761df6534: af000000 mc 0,0
> >0000021761df6538: 0707 bcr 0,%r7
> 0000021761df653a: 0707 bcr 0,%r7
> 0000021761df653c: 0707 bcr 0,%r7
> 0000021761df653e: 0707 bcr 0,%r7
> 0000021761df6540: c004004bb7ec brcl 0,000002176276d518
> Call Trace:
> [<0000021761df6538>] btrfs_compress_folios+0x198/0x1a0
> ([<0000021761df6534>] btrfs_compress_folios+0x194/0x1a0)
> [<0000021761d97788>] compress_file_range+0x3b8/0x6d0
> [<0000021761dcee7c>] btrfs_work_helper+0x10c/0x160
> [<0000021761645760>] process_one_work+0x2b0/0x5d0
> [<000002176164637e>] worker_thread+0x20e/0x3e0
> [<000002176165221a>] kthread+0x15a/0x170
> [<00000217615b859c>] __ret_from_fork+0x3c/0x60
> [<00000217626e72d2>] ret_from_fork+0xa/0x38
> INFO: lockdep is turned off.
> Last Breaking-Event-Address:
> [<0000021761597924>] _printk+0x4c/0x58
> Kernel panic - not syncing: Fatal exception: panic_on_oops
>
> Fixes: fd1e75d0105d ("btrfs: make compression path to be subpage compatible")
> Signed-off-by: Mikhail Zaslonko <zaslonko@xxxxxxxxxxxxx>
> Acked-by: Ilya Leoshkevich <iii@xxxxxxxxxxxxx>
> Reviewed-by: Qu Wenruo <wqu@xxxxxxxx>
Added to for-next, thanks.
