xfs
[Top] [All Lists]

Re: [PATCH] Test to ensure that the EOFBLOCK_FL gets set/unset correctly

To: Eric Sandeen <sandeen@xxxxxxxxxxx>
Subject: Re: [PATCH] Test to ensure that the EOFBLOCK_FL gets set/unset correctly.
From: Akshay Lal <akshaylal@xxxxxxxxxx>
Date: Wed, 8 Sep 2010 11:52:43 -0700
Cc: xfs@xxxxxxxxxxx
Dkim-signature: v=1; a=rsa-sha1; c=relaxed/relaxed; d=google.com; s=beta; t=1283971996; bh=lg0C2/f74pc7rbW84NAQVurwEKM=; h=MIME-Version:Reply-To:In-Reply-To:References:From:Date:Message-ID: Subject:To:Cc:Content-Type:Content-Transfer-Encoding; b=WXcB7I5PKxllN2SlhmK2IZLEzQtLwuA6qkuZdxZyyaHx9lbIvyjQEQdFC/TE/2n7q VuKdbUlWHTigwAr7/AxKA==
Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=beta; h=domainkey-signature:received:mime-version:received:reply-to :in-reply-to:references:from:date:message-id:subject:to:cc :content-type:content-transfer-encoding; bh=SWjBDZpExlpWnYaE/44HzNg9Y+sW172fuWZbNWtRoNw=; b=t/C0UF7pRZkYk83ZsvjnRhGvHszC6NGVBNdohv43UCO38+M6LpIklhJEJdeQl5VaNt kRRWZ3Z8fnVx3rHvoLMw==
Domainkey-signature: a=rsa-sha1; c=nofws; d=google.com; s=beta; h=mime-version:reply-to:in-reply-to:references:from:date:message-id :subject:to:cc:content-type:content-transfer-encoding; b=pZw/dGfa9na8kxLsDqKzOzEczhSTKxGVSiBasvx09t331pw1QFKvS46wVEQV/yXsLU QGpkqADtDPQY54VM9Vxg==
In-reply-to: <4C87DB3D.4030709@xxxxxxxxxxx>
References: <1282941224-5805-1-git-send-email-alal@xxxxxxxxxx> <20100827233216.GJ705@dastard> <4C785254.2020708@xxxxxxxxxxx> <20100828001703.GK705@dastard> <4C785710.9070405@xxxxxxxxxxx> <AANLkTi=6gw_H0GkimuLHexsfkp-pe-9k8WVXYCT9SYZ+@xxxxxxxxxxxxxx> <4C868314.9090503@xxxxxxxxxxx> <AANLkTi=1kv=ihFOO4pBoNkVqJedZ-yc-K1Tp4D7hYM4j@xxxxxxxxxxxxxx> <4C87DB3D.4030709@xxxxxxxxxxx>
Reply-to: alal@xxxxxxxxxx
Thanks Eric, as we move fwd, these patch will be better.


---
Cheers!
Akshay Lal



On Wed, Sep 8, 2010 at 11:51 AM, Eric Sandeen <sandeen@xxxxxxxxxxx> wrote:
> Akshay Lal wrote:
>> There was a missing fi at the the end of the _check_ext4_eof_flag method.
>>
>> Also corrected some indentation.
>
> Thanks, I've committed this version.
>
> In the future, keeping patch resubmissions tidy is helpful, i.e.
>
> Subject: [PATCH Vx] subsystem: description
>
> patch description
>
> Signed-off-by:
> ---
> other information that won't go into the commit
>
> patch itself
>
> ... without the other cruft, forwarded snippets, etc :)
>
> see also Documentation/SubmittingPatches
> in the kernel tree.
>
> Thanks,
>
> -Eric
>
>> Signed-off-by: Akshay Lal <alal@xxxxxxxxxx>
>> ---------------------------------------------------------------------------------------
>> Updated patch:
>> ---------------------------------------------------------------------------------------
>> From e6906071ab6c0ad38d3ee0a463b5c7944e71fd00 Mon Sep 17 00:00:00 2001
>> From: Akshay Lal <alal@xxxxxxxxxx>
>> Date: Tue, 7 Sep 2010 12:54:33 -0700
>> Subject: [PATCH] Test to ensure that the EOFBLOCK_FL gets set/unset 
>> correctly.
>>
>> As found by Theodore Ts'o:
>> If a 128K file is falloc'ed using the KEEP_SIZE flag, and then
>> write exactly 128K, the EOFBLOCK_FL doesn't get cleared correctly.
>> This is bad since it forces e2fsck to complain about that inode.
>> If you have a large number of inodes that are written with fallocate
>> using KEEP_SIZE, and then fill them up to their expected size,
>> e2fsck will potentially complain about a _huge_ number of inodes.
>> This would also cause a huge increase in the time taken by e2fsck
>> to complete its check.
>>
>> Test scenarios covered:
>> 1. Fallocating X bytes and writing Y (Y<X) (buffered and direct io)
>> 2. Fallocating X bytes and writing Y (Y=X) (buffered and direct io)
>> 3. Fallocating X bytes and writing Y (Y>X) (buffered and direct io)
>>
>> These test cases exercise the normal and edge case conditions using
>> falloc (and KEEP_SIZE).
>>
>> Ref: http://thread.gmane.org/gmane.comp.file-systems.ext4/20682
>> ---
>>  243     |  178 
>> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>  243.out |   13 +++++
>>  group   |    1 +
>>  3 files changed, 192 insertions(+), 0 deletions(-)
>>  create mode 100644 243
>>  create mode 100644 243.out
>>
>> diff --git a/243 b/243
>> new file mode 100644
>> index 0000000..8b2e647
>> --- /dev/null
>> +++ b/243
>> @@ -0,0 +1,178 @@
>> +#! /bin/bash
>> +# FS QA Test No. 243
>> +#
>> +# Test to ensure that the EOFBLOCK_FL gets set/unset correctly.
>> +#
>> +# As found by Theodore Ts'o:
>> +# If a 128K file is falloc'ed using the KEEP_SIZE flag, and then
>> +# write exactly 128K, the EOFBLOCK_FL doesn't get cleared correctly.
>> +# This is bad since it forces e2fsck to complain about that inode.
>> +# If you have a large number of inodes that are written with fallocate
>> +# using KEEP_SIZE, and then fill them up to their expected size,
>> +# e2fsck will potentially complain about a _huge_ number of inodes.
>> +# This would also cause a huge increase in the time taken by e2fsck
>> +# to complete its check.
>> +#
>> +# Test scenarios covered:
>> +# 1. Fallocating X bytes and writing Y (Y<X) (buffered and direct io)
>> +# 2. Fallocating X bytes and writing Y (Y=X) (buffered and direct io)
>> +# 3. Fallocating X bytes and writing Y (Y>X) (buffered and direct io)
>> +#
>> +# These test cases exercise the normal and edge case conditions using
>> +# falloc (and KEEP_SIZE).
>> +#
>> +# Ref: http://thread.gmane.org/gmane.comp.file-systems.ext4/20682
>> +#
>> +#-----------------------------------------------------------------------
>> +# Copyright (c) 2010 Google, Inc.  All Rights Reserved.
>> +#
>> +# This program is free software; you can redistribute it and/or
>> +# modify it under the terms of the GNU General Public License as
>> +# published by the Free Software Foundation.
>> +#
>> +# This program is distributed in the hope that it would be useful,
>> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +# GNU General Public License for more details.
>> +#
>> +# You should have received a copy of the GNU General Public License
>> +# along with this program; if not, write the Free Software Foundation,
>> +# Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
>> +#-----------------------------------------------------------------------
>> +#
>> +# creator
>> +owner=alal@xxxxxxxxxx
>> +
>> +seq=`basename $0`
>> +echo "QA output created by $seq"
>> +
>> +here=`pwd`
>> +tmp=/tmp/$$
>> +status=1        # failure is the default!
>> +trap "_cleanup; exit \$status" 0 1 2 3 15
>> +
>> +# Test specific macros.
>> +BIT_NOT_SET=0   # inode flag - 0x400000 bit is not set.
>> +BIT_SET=1       # inode flag - 0x400000 bit is set.
>> +
>> +# Generic test cleanup function.
>> +_cleanup()
>> +{
>> +  cd /
>> +  rm -f $tmp.*
>> +}
>> +
>> +# Ext4 uses the EOFBLOCKS_FL bit when fallocating blocks with KEEP_SIZE
>> +# enabled. The only time this bit should be set is when extending the 
>> allocated
>> +# blocks further than what the i_size represents. In the situations wherein 
>> the
>> +# i_size covers all allocated blocks, this bit should be cleared.
>> +
>> +# Checks the state of the sample file in the filesystem and returns whether
>> +# the inode flag 0x400000 is set or not.
>> +_check_ext4_eof_flag()
>> +{
>> +  # Check whether EOFBLOCK_FL is set.
>> +  # For ext4 filesystems: use debugfs to check if EOFBLOCKS_FL is set.
>> +  # Other filesystems: do nothing. The default fsck at the end of the test
>> +  # should catch any potential errors.
>> +  if [ "${FSTYP}" == "ext4" ]; then
>> +    bit_set=1
>> +
>> +    # Unmount the ${TEST_DEV}
>> +    umount ${TEST_DEV}
>> +
>> +    # Run debugfs to gather file_parameters - specifically iflags.
>> +    file_params=`debugfs ${TEST_DEV} -R "stat ${1}" 2>&1 | grep -e Flags:`
>> +    iflags=${file_params#*Flags: }
>> +
>> +    # Ensure that the iflags value was parsed correctly.
>> +    if [ -z ${iflags} ]; then
>> +      echo "iFlags value was not parsed successfully." >> $seq.full
>> +      status=1
>> +      exit ${status}
>> +    fi
>> +
>> +    # Check if EOFBLOCKS_FL is set.
>> +    if ((${iflags} & 0x400000)); then
>> +      echo "EOFBLOCK_FL bit is set." >> $seq.full
>> +      bit_set=1
>> +    else
>> +      echo "EOFBLOCK_FL bit is not set." >> $seq.full
>> +      bit_set=0
>> +    fi
>> +
>> +    # Check current bit state to expected value.
>> +    if [ ${bit_set} -ne ${2} ]; then
>> +      echo "Error: Current bit state incorrect." >> $seq.full
>> +      status=1
>> +      exit ${status}
>> +    fi
>> +
>> +    # Mount the ${TEST_DEV}
>> +    mount ${TEST_DEV} -t ${FSTYP} ${TEST_DIR}
>> +  fi
>> +}
>> +
>> +# Get standard environment, filters and checks.
>> +. ./common.rc
>> +. ./common.filter
>> +
>> +# Prerequisites for the test run.
>> +_supported_fs ext4 xfs btrfs gfs2
>> +_supported_os Linux
>> +_require_xfs_io_falloc
>> +
>> +# Real QA test starts here.
>> +rm -f $seq.full
>> +
>> +# Begin test cases.
>> +echo "Test 1: Fallocate 40960 bytes and write 4096 bytes (buffered io)." \
>> +    >> $seq.full
>> +${XFS_IO_PROG} -F -f                    \
>> +    -c 'falloc -k 0 40960'              \
>> +    -c 'pwrite 0 4096'                  \
>> +    ${TEST_DIR}/test_1 | _filter_xfs_io_unique
>> +_check_ext4_eof_flag test_1 ${BIT_SET}
>> +
>> +echo "Test 2: Fallocate 40960 bytes and write 4096 bytes (direct io)." \
>> +    >> $seq.full
>> +${XFS_IO_PROG} -F -f -d                 \
>> +    -c 'falloc -k 0 40960'              \
>> +    -c 'pwrite 0 4096'                  \
>> +    ${TEST_DIR}/test_2 | _filter_xfs_io_unique
>> +_check_ext4_eof_flag test_2 ${BIT_SET}
>> +
>> +echo "Test 3: Fallocate 40960 bytes and write 40960 bytes (buffered io)." \
>> +    >> $seq.full
>> +${XFS_IO_PROG} -F -f                    \
>> +    -c 'falloc -k 0 40960'              \
>> +    -c 'pwrite 0 40960'                 \
>> +    ${TEST_DIR}/test_3 | _filter_xfs_io_unique
>> +_check_ext4_eof_flag test_3 ${BIT_NOT_SET}
>> +
>> +echo "Test 4: Fallocate 40960 bytes and write 40960 bytes (direct io)." \
>> +    >> $seq.full
>> +${XFS_IO_PROG} -F -f -d                 \
>> +    -c 'falloc -k 0 40960'              \
>> +    -c 'pwrite 0 40960'                 \
>> +    ${TEST_DIR}/test_4 | _filter_xfs_io_unique
>> +_check_ext4_eof_flag test_4 ${BIT_NOT_SET}
>> +
>> +echo "Test 5: Fallocate 128k, seek 256k and write 4k block (buffered io)." \
>> +    >> $seq.full
>> +${XFS_IO_PROG} -F -f                    \
>> +    -c 'falloc -k 0 128k'               \
>> +    -c 'pwrite 256k 4k'                 \
>> +    ${TEST_DIR}/test_5 | _filter_xfs_io_unique
>> +_check_ext4_eof_flag test_5 ${BIT_NOT_SET}
>> +
>> +echo "Test 6: Fallocate 128k, seek to 256k and write a 4k block (direct 
>> io)." \
>> +    >> $seq.full
>> +${XFS_IO_PROG} -F -f -d                 \
>> +    -c 'falloc -k 0 128k'               \
>> +    -c 'pwrite 256k 4k'                 \
>> +    ${TEST_DIR}/test_6 | _filter_xfs_io_unique
>> +_check_ext4_eof_flag test_6 ${BIT_NOT_SET}
>> +
>> +status=0
>> +exit ${status}
>> diff --git a/243.out b/243.out
>> new file mode 100644
>> index 0000000..290a005
>> --- /dev/null
>> +++ b/243.out
>> @@ -0,0 +1,13 @@
>> +QA output created by 243
>> +wrote 4096/4096 bytes at offset 0
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +wrote 4096/4096 bytes at offset 0
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +wrote 40960/40960 bytes at offset 0
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +wrote 40960/40960 bytes at offset 0
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +wrote 4096/4096 bytes at offset 262144
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +wrote 4096/4096 bytes at offset 262144
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> diff --git a/group b/group
>> index ff16bb3..e6dab13 100644
>> --- a/group
>> +++ b/group
>> @@ -356,3 +356,4 @@ deprecated
>>  240 auto aio quick rw
>>  241 auto
>>  242 auto quick prealloc
>> +243 auto quick prealloc
>
>

<Prev in Thread] Current Thread [Next in Thread>