Re: [PATCH v4] iomap: follow the alignment requirement for iomap_dio_hole_iter()

[Qu Wenruo <quwenruo.btrfs@gmx.com>]

在 2026/10/26 13:20, kernel test robot 写道:
> 
> Hello,
> 
> kernel test robot noticed "xfstests.generic.091.fail" on:
> 
> commit: 4025185cf953e3f52ab63cf5fec50c1485ee7bb7 ("[PATCH v4] iomap: follow the alignment requirement for iomap_dio_hole_iter()") https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master patch link: https://lore.kernel.org/all/73bd4d9b39e821453028ca8e567eaad2ebe0fc18.1785880740.git.wqu@suse.com/
> patch subject: [PATCH v4] iomap: follow the alignment requirement for iomap_dio_hole_iter()
> 
> [test failed on linux-next/master 20a33c6b6cbe7e4805dfb6cd68816937443aa148]
> 
> in testcase: xfstests
> version: xfstests-x86_64-22348afe-1_20261019
> with following parameters:
> 
> 	disk: 4HDD
> 	fs: ext4
> 	test: generic-group-09
> 
> 
> config: x86_64-rhel-9.4-func
> compiler: gcc-14
> test machine: 8 threads Intel(R) Core(TM) i7-6700 CPU @ 3.40GHz (Skylake) with 28G memory
> 
> (please refer to attached dmesg/kmsg for entire log/backtrace)
> 
> 
> If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <yi1.lai@intel.com>
> | Closes: https://lore.kernel.org/oe-lkp/202610261048.3dc64aea-lkp@intel.com
> 
> 
> 2026-10-24 23:20:06 cd /lkp/benchmarks/xfstests
> 2026-10-24 23:20:07 export TEST_DIR=/fs/sda1
> 2026-10-24 23:20:07 export TEST_DEV=/dev/sda1
> 2026-10-24 23:20:07 export FSTYP=ext4
> 2026-10-24 23:20:07 export SCRATCH_MNT=/fs/scratch
> 2026-10-24 23:20:07 mkdir /fs/scratch -p
> 2026-10-24 23:20:07 export SCRATCH_DEV=/dev/sda4
> 2026-10-24 23:20:07 sed "s:^:generic/:" /lkp/benchmarks/xfstests/tests/generic-group-09
> 2026-10-24 23:20:07 ./check -E tests/exclude/ext4 generic/091 generic/092 generic/093 generic/094 generic/095 generic/096 generic/097 generic/098
> FSTYP         -- ext4
> PLATFORM      -- Linux/x86_64 lkp-skl-d01 7.3.0-rc2+ #1 SMP PREEMPT Sat Oct 24 20:47:23 CST 2026
> MKFS_OPTIONS  -- -F /dev/sda4
> MOUNT_OPTIONS -- -o acl,user_xattr /dev/sda4 /fs/scratch
> 
> generic/091        [failed, exit status 1]- output mismatch (see /lkp/benchmarks/xfstests/results//generic/091.out.bad)
>      --- tests/generic/091.out	2026-10-19 16:30:12.000000000 +0000
>      +++ /lkp/benchmarks/xfstests/results//generic/091.out.bad	2026-10-24 23:20:20.173485972 +0000
>      @@ -1,7 +1,25 @@
>       QA output created by 091
>       fsx -N 10000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
>      -fsx -N 10000 -o 8192 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
>      -fsx -N 10000 -o 32768 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
>      -fsx -N 10000 -o 8192 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
>      -fsx -N 10000 -o 32768 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -R -W
>      -fsx -N 10000 -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -W
>      ...
>      (Run 'diff -u /lkp/benchmarks/xfstests/tests/generic/091.out /lkp/benchmarks/xfstests/results//generic/091.out.bad'  to see the entire diff)

Indeed I can reproduce the problem on both ext4 and xfs.

It looks like ext4 and non-COW xfs inodes relies on the unaligned hole 
iterating behavior.

Considering only btrfs and COW xfs inodes requests FSBLOCK_ALIGNED flag, 
for now I'll fix it by only do the revert if it's requested by the dio flag.

Thanks,
Qu

> generic/092         1s
> generic/093         1s
> generic/094         14s
> generic/095         51s
> generic/096         1s
> generic/097         2s
> generic/098         11s
> Ran: generic/091 generic/092 generic/093 generic/094 generic/095 generic/096 generic/097 generic/098
> Failures: generic/091
> Failed 1 of 8 tests
> 
> 
> The kernel config and materials to reproduce are available at:
> https://download.01.org/0day-ci/archive/20261026/202610261048.3dc64aea-lkp@intel.com
> 
> 
> --
> 0-DAY CI Kernel Test Service
> https://github.com/intel/lkp-tests/wiki