Re: [PATCH v2] f2fs/023: test sanity check condition w/ error injection

[Chao Yu <chao@kernel.org>]

On 11/15/2025 4:18 PM, Zorro Lang wrote:
> On Mon, Nov 03, 2025 at 02:21:43PM +0800, Chao Yu wrote:
>> After commit 5c1768b67250 ("f2fs: fix to do sanity check correctly on
>> i_inline_xattr_size"), f2fs should handle corrupted i_inline_xattr_size
>> correctly, let's add this regression testcase to check that.
>>
>> Cc: Jaegeuk Kim <jaegeuk@kernel.org>
>> Signed-off-by: Chao Yu <chao@kernel.org>
>> ---
>> v2:
>> - use _filter_scratch to filter scratch mountpoint
>> - check output message of inject.f2fs rather than return value
>> - add _require_attrs
>> - add f2fs/023 to attr group
>> - fail testcase if mkfs failed
>>   tests/f2fs/023     | 49 ++++++++++++++++++++++++++++++++++++++++++++++
>>   tests/f2fs/023.out |  2 ++
>>   2 files changed, 51 insertions(+)
>>   create mode 100755 tests/f2fs/023
>>   create mode 100644 tests/f2fs/023.out
>>
>> diff --git a/tests/f2fs/023 b/tests/f2fs/023
>> new file mode 100755
>> index 00000000..33108421
>> --- /dev/null
>> +++ b/tests/f2fs/023
>> @@ -0,0 +1,49 @@
>> +#! /bin/bash
>> +# SPDX-License-Identifier: GPL-2.0
>> +# Copyright (c) 2025 Chao Yu <chao@kernel.org>
>> +#
>> +# FS QA Test No. f2fs/023
>> +#
>> +# This testcase tries to inject fault into inode.i_inline_xattr_size,
>> +# and check whether sanity check of f2fs can handle fault correctly.
>> +#
>> +
>> +. ./common/preamble
>> +. ./common/attr
>> +. ./common/filter
>> +_begin_fstest auto quick rw attr
>> +
>> +_fixed_by_kernel_commit 5c1768b67250 \
>> +        "f2fs: fix to do sanity check correctly on i_inline_xattr_size"
>> +
>> +_require_attrs
>> +_require_scratch_nocheck
>> +_require_inject_f2fs_command node i_inline
>> +_require_inject_f2fs_command node i_inline_xattr_size
>> +
>> +testfile=$SCRATCH_MNT/testfile
>> +
>> +# remove all mkfs options to avoid layout change of on-disk inode
>> +export MKFS_OPTIONS=""
>> +
>> +_scratch_mkfs "-O extra_attr,flexible_inline_xattr" >> $seqres.full || _fail "mkfs failed"
>> +_scratch_mount "-o inline_xattr_size=512" >>$seqres.full 2>&1
>> +touch $testfile
>> +_scratch_unmount
>> +
>> +# inject .i_inline field: clear F2FS_EXTRA_ATTR bit
>> +output=`$F2FS_INJECT_PROG --node --mb i_inline --nid 4 --val 0x1 $SCRATCH_DEV`
>> +echo "$output" >> $seqres.full
>> +echo "$output" | grep "unknown or unsupported member" && _fail "i_inline is unsupported"
>> +
>> +# inject .i_inline_xattr_size field from 512 to 2048
>> +output=`$F2FS_INJECT_PROG --node --mb i_inline_xattr_size --nid 4 --val 2048 $SCRATCH_DEV`
>> +echo "$output" >> $seqres.full
>> +echo "$output" | grep "unknown or unsupported member" && _fail "i_inline_xattr_size is unsupported"
> 
> I think we might need a common helper to do f2fs injection and return
> error if injection fails, if F2FS_INJECT_PROG always can't return error

Yes, let me find free slot to clean up, thanks for the suggestion and
review. :)

Thanks.

> properly. Anyway, that's another story. For this patch:
> 
> Reviewed-by: Zorro Lang <zlang@redhat.com>
> 
>> +
>> +_scratch_mount
>> +
>> +$GETFATTR_PROG -n user.test "$testfile" 2>&1 | _filter_scratch
>> +
>> +status=0
>> +exit
>> diff --git a/tests/f2fs/023.out b/tests/f2fs/023.out
>> new file mode 100644
>> index 00000000..f9635d38
>> --- /dev/null
>> +++ b/tests/f2fs/023.out
>> @@ -0,0 +1,2 @@
>> +QA output created by 023
>> +getfattr: SCRATCH_MNT/testfile: Structure needs cleaning
>> -- 
>> 2.49.0
>>
>