Re: [PATCH] generic: test i_blocks for large files

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]



On Fri, Aug 12, 2022 at 04:50:50PM +0200, Pavel Reichl wrote:
> This is a regression test for an incorrect
> computation of i_blocks for files larger than
> 4 GiB. Bug was filed for exFAT.
> 
> Signed-off-by: Pavel Reichl <preichl@xxxxxxxxxx>
> ---
>  tests/generic/693-inodes-for-large-files     | 49 ++++++++++++++++++++
>  tests/generic/693-inodes-for-large-files.out |  1 +
>  2 files changed, 50 insertions(+)
>  create mode 100755 tests/generic/693-inodes-for-large-files
>  create mode 100644 tests/generic/693-inodes-for-large-files.out
> 
> diff --git a/tests/generic/693-inodes-for-large-files b/tests/generic/693-inodes-for-large-files

I think case *number* is enough.

> new file mode 100755
> index 00000000..797d6a21
> --- /dev/null
> +++ b/tests/generic/693-inodes-for-large-files
> @@ -0,0 +1,49 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2022  Red Hat Inc. All Rights Reserved.
> +#
> +# FS QA Test 693
> +#
> +# Verify that i_blocks for files larger than 4 GiB have correct
> +# values.
> +#
> +# This test verifies the problem fixed in kernel with commit
> +# 0c336d6e33f4 exfat: fix incorrect loading of i_blocks for large files
> +#
> +. ./common/preamble
> +_begin_fstest auto
> +
> +# Override the default cleanup function.
> +_cleanup()
> +{
> +	cd /
> +	rm -r -f $tmp.* $junk_dir
> +}
> +
> +_supported_fs generic
> +_require_test

If this test need 4G free space, I'm wondering if we need to make sure
there's enough space in $TEST_DIR, as we do in _require_scratch_size.

> +
> +junk_dir=$TEST_DIR/$seq
> +junk_file=$junk_dir/junk
> +mkdir -p $junk_dir
> +
> +$XFS_IO_PROG -f -c "truncate 4G" $junk_file

It helps to allocate 4g real space on exfat, due to exfat doesn't
support sparse file, but on other filesystems (e.g. xfs, ext4, btrfs)
it will get a file which i_blocks=0.

As a generic test case, if you really want to allocate 4g real space,
I think you have to real write data:

  $XFS_IO_PROG -f -c "pwrite -W 0 4g" $junk_file

If anyone knows any better idea, feel free to reply :)

> +
> +iblocks=`$XFS_IO_PROG -c "stat" $junk_file | grep 'stat.blocks'`

stat -c %b $junk_file ??

> +test -z "$iblocks" && echo 'Failed to parse the blocks'
> +
> +_test_cycle_mount
> +
> +iblocks_after_remount=`$XFS_IO_PROG -c "stat" $junk_file | grep 'stat.blocks'`

stat -c %b $junk_file ??

> +
> +test -z "$iblocks_after_remount" && echo 'Failed to parse the blocks'

When will we get an empty "$iblocks_after_remount" ?

> +
> +if [ "$iblocks" != "$iblocks_after_remount" ]; then
> +	echo "Number of blocks needs to be same: $iblocks, $iblocks_after_remount"
> +	status=1

If you "exit" directly, the status is 1 by default. But I think you don't need
to exit at here, due to above output will break golden image and fail this test.

> +	exit
> +fi
> +
> +status=0
> +
> +exit
> diff --git a/tests/generic/693-inodes-for-large-files.out b/tests/generic/693-inodes-for-large-files.out
> new file mode 100644
> index 00000000..f39aa77c
> --- /dev/null
> +++ b/tests/generic/693-inodes-for-large-files.out
> @@ -0,0 +1 @@
> +QA output created by 693-inodes-for-large-files

Silence is golden ??

> -- 
> 2.37.1
> 




[Index of Archives]     [Linux Filesystems Development]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux