fix kernels without v5 support
Hi all,
this series ensures tests pass on kernels without v5 support. As a side effect it also removes support for historic kernels and xfsprogs without any v5 support, and without mkfs input validation.
PATCH 1/6 xfs: remove support for tools and kernels with v5 support
v5 file systems have been the default for more than 10 years. Drop support for non-v5 enabled kernels and xfsprogs.
--- common/config | 13 ------------- common/xfs | 26 -------------------------- tests/xfs/002 | 3 --- tests/xfs/005 | 1 - tests/xfs/077 | 1 - tests/xfs/083 | 2 -- tests/xfs/132 | 1 - tests/xfs/148 | 1 - tests/xfs/263 | 5 ----- tests/xfs/299 | 2 -- tests/xfs/304 | 2 -- tests/xfs/305 | 2 -- 12 files changed, 59 deletions(-)
common/config +0 -13
@@ -329,19 +329,6 @@ if [ -x /usr/sbin/selinuxenabled ] && /usr/sbin/selinuxenabled; then export SELINUX_MOUNT_OPTIONS fi -# check if mkfs.xfs supports v5 xfs -if [ "$FSTYP" == "xfs" ]; then - XFS_MKFS_HAS_NO_META_SUPPORT="" - touch /tmp/crc_check.img - $MKFS_XFS_PROG -N -d file,name=/tmp/crc_check.img,size=32m -m crc=0 \ - >/dev/null 2>&1; - if [ $? -ne 0 ]; then - XFS_MKFS_HAS_NO_META_SUPPORT=true - fi - rm -f /tmp/crc_check.img - export XFS_MKFS_HAS_NO_META_SUPPORT -fi - _common_mount_opts() { case $FSTYP in
common/xfs +0 -26
@@ -59,11 +59,6 @@ _scratch_mkfs_xfs_opts() { mkfs_opts=$* - # remove metadata related mkfs options if mkfs.xfs doesn't them - if [ -n "$XFS_MKFS_HAS_NO_META_SUPPORT" ]; then - mkfs_opts=`echo $mkfs_opts | sed "s/-m\s\+\S\+//g"` - fi - _scratch_options mkfs echo "$MKFS_XFS_PROG $SCRATCH_OPTIONS $mkfs_opts"
@@ -439,24 +434,6 @@ _require_projid16bit() || _notrun "16 bit project IDs not supported on $SCRATCH_DEV" } -# this test requires the crc feature to be available in mkfs.xfs -# -_require_xfs_mkfs_crc() -{ - _scratch_mkfs_xfs_supported -m crc=1 >/dev/null 2>&1 \ - || _notrun "mkfs.xfs doesn't have crc feature" -} - -# this test requires the xfs kernel support crc feature -# -_require_xfs_crc() -{ - _scratch_mkfs_xfs -m crc=1 >/dev/null 2>&1 - _try_scratch_mount >/dev/null 2>&1 \ - || _notrun "Kernel doesn't support crc feature" - _scratch_unmount -} - # If the xfs_info output for the given XFS filesystem mount mentions the given # feature. If so, return 0 for success. If not, return 1 for failure. If the # third option is -v, echo 1 for success and 0 for not.
@@ -1268,9 +1245,6 @@ _require_scratch_xfs_shrink() # _require_meta_uuid() { - # This will create a crc fs on $SCRATCH_DEV - _require_xfs_crc - _scratch_xfs_db -x -c "uuid restore" 2>&1 \ | grep -q "invalid UUID\|supported on V5 fs" \ && _notrun "Userspace doesn't support meta_uuid feature"
tests/xfs/002 +0 -3
@@ -24,9 +24,6 @@ _supported_fs xfs _require_scratch_nocheck _require_no_large_scratch_dev -# So we can explicitly turn it _off_: -_require_xfs_mkfs_crc - _scratch_mkfs_xfs -m crc=0 -d size=128m >> $seqres.full 2>&1 || _fail "mkfs failed" # Scribble past a couple V4 secondary superblocks to populate sb_crc
tests/xfs/005 +0 -1
@@ -20,7 +20,6 @@ _begin_fstest auto quick _supported_fs xfs _require_scratch_nocheck -_require_xfs_mkfs_crc _scratch_mkfs_xfs -m crc=1 >> $seqres.full 2>&1 || _fail "mkfs failed"
tests/xfs/077 +0 -1
@@ -24,7 +24,6 @@ _supported_fs xfs _require_xfs_copy _require_scratch _require_no_large_scratch_dev -_require_xfs_crc _require_meta_uuid # Takes 2 args, 2nd optional:
tests/xfs/083 +0 -2
@@ -28,8 +28,6 @@ _cleanup() _supported_fs xfs _require_scratch -#_require_xfs_crc # checksum not required, but you probably want it anyway... -#_require_xfs_mkfs_crc _require_attrs _require_populate_commands
tests/xfs/132 +0 -1
@@ -24,7 +24,6 @@ _require_scratch_nocheck # due to transaction cancellation. Hence we don't want to check dmesg here. _disable_dmesg_check -_require_xfs_mkfs_crc _scratch_mkfs -m crc=0 > $seqres.full 2>&1 # The files that EIO in the golden output changes if we have quotas enabled
tests/xfs/148 +0 -1
@@ -27,7 +27,6 @@ _cleanup() _supported_fs xfs _require_test _require_attrs -_require_xfs_mkfs_crc _disable_dmesg_check imgfile=$TEST_DIR/img-$seq
tests/xfs/263 +0 -5
@@ -21,11 +21,6 @@ _supported_fs xfs _require_scratch _require_xfs_quota -# We could test older, non-project capable kernels but keep it simpler; -# Only test crc and beyond (but we will test with and without the feature) -_require_xfs_mkfs_crc -_require_xfs_crc - function option_string() { VAL=$1
tests/xfs/299 +0 -2
@@ -31,8 +31,6 @@ chmod a+rwx $seqres.full # arbitrary users will write here _require_scratch _require_xfs_quota -_require_xfs_mkfs_crc -_require_xfs_crc # The actual point at which limit enforcement takes place for the # hard block limit is variable depending on filesystem blocksize,
tests/xfs/304 +0 -2
@@ -19,8 +19,6 @@ _supported_fs xfs _require_scratch _require_xfs_quota -_require_xfs_mkfs_crc -_require_xfs_crc _scratch_mkfs_xfs -m crc=1 >/dev/null 2>&1
tests/xfs/305 +0 -2
@@ -19,8 +19,6 @@ _supported_fs xfs _require_scratch _require_xfs_quota -_require_xfs_mkfs_crc -_require_xfs_crc _require_command "$KILLALL_PROG" killall _scratch_mkfs_xfs -m crc=1 >/dev/null 2>&1
Signature · 1 line
2.39.2
Re: [PATCH 1/6] xfs: remove support for tools and kernels with v5 support
Quoted 1 line · Christoph Hellwig
xfs: remove support for tools and kernels with v5 support
I think you mean tools and kernels *without* v5 support.
- Eric
Re: [PATCH 1/6] xfs: remove support for tools and kernels with v5 support
Quoted 4 lines · Eric Biggers
Quoted 1 line · Christoph Hellwig
xfs: remove support for tools and kernels with v5 support
I think you mean tools and kernels *without* v5 support.
Yes.
Re: [PATCH 1/6] xfs: remove support for tools and kernels with v5 support
In reply to Christoph HellwigQuoted 4 lines · Christoph Hellwig
v5 file systems have been the default for more than 10 years. Drop support for non-v5 enabled kernels and xfsprogs.
Signed-off-by Christoph Hellwig <[email protected]>
Seems fine to me.
--D
Quoted 224 lines
--- common/config | 13 ------------- common/xfs | 26 -------------------------- tests/xfs/002 | 3 --- tests/xfs/005 | 1 - tests/xfs/077 | 1 - tests/xfs/083 | 2 -- tests/xfs/132 | 1 - tests/xfs/148 | 1 - tests/xfs/263 | 5 ----- tests/xfs/299 | 2 -- tests/xfs/304 | 2 -- tests/xfs/305 | 2 -- 12 files changed, 59 deletions(-)common/config +0
-13@@ -329,19 +329,6 @@ if [ -x /usr/sbin/selinuxenabled ] && /usr/sbin/selinuxenabled; then export SELINUX_MOUNT_OPTIONS fi -# check if mkfs.xfs supports v5 xfs -if [ "$FSTYP" == "xfs" ]; then - XFS_MKFS_HAS_NO_META_SUPPORT="" - touch /tmp/crc_check.img - $MKFS_XFS_PROG -N -d file,name=/tmp/crc_check.img,size=32m -m crc=0 \ - >/dev/null 2>&1; - if [ $? -ne 0 ]; then - XFS_MKFS_HAS_NO_META_SUPPORT=true - fi - rm -f /tmp/crc_check.img - export XFS_MKFS_HAS_NO_META_SUPPORT -fi - _common_mount_opts() { case $FSTYP incommon/xfs +0
-26@@ -59,11 +59,6 @@ _scratch_mkfs_xfs_opts() { mkfs_opts=$* - # remove metadata related mkfs options if mkfs.xfs doesn't them - if [ -n "$XFS_MKFS_HAS_NO_META_SUPPORT" ]; then - mkfs_opts=`echo $mkfs_opts | sed "s/-m\s\+\S\+//g"` - fi - _scratch_options mkfs echo "$MKFS_XFS_PROG $SCRATCH_OPTIONS $mkfs_opts"@@ -439,24 +434,6 @@ _require_projid16bit() || _notrun "16 bit project IDs not supported on $SCRATCH_DEV" } -# this test requires the crc feature to be available in mkfs.xfs -# -_require_xfs_mkfs_crc() -{ - _scratch_mkfs_xfs_supported -m crc=1 >/dev/null 2>&1 \ - || _notrun "mkfs.xfs doesn't have crc feature" -} - -# this test requires the xfs kernel support crc feature -# -_require_xfs_crc() -{ - _scratch_mkfs_xfs -m crc=1 >/dev/null 2>&1 - _try_scratch_mount >/dev/null 2>&1 \ - || _notrun "Kernel doesn't support crc feature" - _scratch_unmount -} - # If the xfs_info output for the given XFS filesystem mount mentions the given # feature. If so, return 0 for success. If not, return 1 for failure. If the # third option is -v, echo 1 for success and 0 for not.@@ -1268,9 +1245,6 @@ _require_scratch_xfs_shrink() # _require_meta_uuid() { - # This will create a crc fs on $SCRATCH_DEV - _require_xfs_crc - _scratch_xfs_db -x -c "uuid restore" 2>&1 \ | grep -q "invalid UUID\|supported on V5 fs" \ && _notrun "Userspace doesn't support meta_uuid feature"tests/xfs/002 +0
-3@@ -24,9 +24,6 @@ _supported_fs xfs _require_scratch_nocheck _require_no_large_scratch_dev -# So we can explicitly turn it _off_: -_require_xfs_mkfs_crc - _scratch_mkfs_xfs -m crc=0 -d size=128m >> $seqres.full 2>&1 || _fail "mkfs failed" # Scribble past a couple V4 secondary superblocks to populate sb_crctests/xfs/005 +0
-1@@ -20,7 +20,6 @@ _begin_fstest auto quick _supported_fs xfs _require_scratch_nocheck -_require_xfs_mkfs_crc _scratch_mkfs_xfs -m crc=1 >> $seqres.full 2>&1 || _fail "mkfs failed"tests/xfs/077 +0
-1@@ -24,7 +24,6 @@ _supported_fs xfs _require_xfs_copy _require_scratch _require_no_large_scratch_dev -_require_xfs_crc _require_meta_uuid # Takes 2 args, 2nd optional:tests/xfs/083 +0
-2@@ -28,8 +28,6 @@ _cleanup() _supported_fs xfs _require_scratch -#_require_xfs_crc # checksum not required, but you probably want it anyway... -#_require_xfs_mkfs_crc _require_attrs _require_populate_commandstests/xfs/132 +0
-1@@ -24,7 +24,6 @@ _require_scratch_nocheck # due to transaction cancellation. Hence we don't want to check dmesg here. _disable_dmesg_check -_require_xfs_mkfs_crc _scratch_mkfs -m crc=0 > $seqres.full 2>&1 # The files that EIO in the golden output changes if we have quotas enabledtests/xfs/148 +0
-1@@ -27,7 +27,6 @@ _cleanup() _supported_fs xfs _require_test _require_attrs -_require_xfs_mkfs_crc _disable_dmesg_check imgfile=$TEST_DIR/img-$seqtests/xfs/263 +0
-5@@ -21,11 +21,6 @@ _supported_fs xfs _require_scratch _require_xfs_quota -# We could test older, non-project capable kernels but keep it simpler; -# Only test crc and beyond (but we will test with and without the feature) -_require_xfs_mkfs_crc -_require_xfs_crc - function option_string() { VAL=$1tests/xfs/299 +0
-2@@ -31,8 +31,6 @@ chmod a+rwx $seqres.full # arbitrary users will write here _require_scratch _require_xfs_quota -_require_xfs_mkfs_crc -_require_xfs_crc # The actual point at which limit enforcement takes place for the # hard block limit is variable depending on filesystem blocksize,tests/xfs/304 +0
-2@@ -19,8 +19,6 @@ _supported_fs xfs _require_scratch _require_xfs_quota -_require_xfs_mkfs_crc -_require_xfs_crc _scratch_mkfs_xfs -m crc=1 >/dev/null 2>&1tests/xfs/305 +0
-2@@ -19,8 +19,6 @@ _supported_fs xfs _require_scratch _require_xfs_quota -_require_xfs_mkfs_crc -_require_xfs_crc _require_command "$KILLALL_PROG" killall _scratch_mkfs_xfs -m crc=1 >/dev/null 2>&1Signature · 1 line
2.39.2
Re: [PATCH 1/6] xfs: remove support for tools and kernels with v5 support
In reply to Christoph HellwigQuoted 3 lines · Christoph Hellwig
v5 file systems have been the default for more than 10 years. Drop support for non-v5 enabled kernels and xfsprogs.
Hi,
This patch is causing xfs/077 to fail as shown below,
# ./check xfs/077 FSTYP -- xfs (debug) PLATFORM -- Linux/x86_64 xfs-crc 6.9.0-rc4+ #2 SMP PREEMPT_DYNAMIC Mon Apr 29 08:08:05 GMT 2024 MKFS_OPTIONS -- -f -f -m crc=1,reflink=0,rmapbt=0, -i sparse=0 /dev/loop6 MOUNT_OPTIONS -- -o usrquota,grpquota,prjquota -o context=system_u:object_r:root_t:s0 /dev/loop6 /media/scratch
xfs/077 9s ... [not run] Kernel doesn't support meta_uuid feature
Ran: xfs/077
Not run: xfs/077
Passed all 1 tests
The corresponding configuration file used had the following,
export TEST_DEV=/dev/loop5
export TEST_DIR=/media/test
export SCRATCH_DEV=/dev/loop6
export SCRATCH_MNT=/media/scratch
MKFS_OPTIONS='-f -m crc=1,reflink=0,rmapbt=0, -i sparse=0' MOUNT_OPTIONS='-o usrquota,grpquota,prjquota'
Signature · 1 line
Chandan
Re: [PATCH 1/6] xfs: remove support for tools and kernels with v5 support
Quoted 8 lines · Chandan Babu R
Quoted 3 lines · Christoph Hellwig
v5 file systems have been the default for more than 10 years. Drop support for non-v5 enabled kernels and xfsprogs.
Hi,
This patch is causing xfs/077 to fail as shown below,
That's not a failure, but it isn't run. AFAICS it is because _require_meta_uuid expects to work on a scratch fs, but none is generated beforehand as the _require_xfs_mkfs_crc that did that under the hood now went away.
The fix is to probably just nuke _require_meta_uuid and explicitly create a CRC-enabled fs. I'll look into that.
Re: [PATCH 1/6] xfs: remove support for tools and kernels with v5 support
In reply to Chandan Babu RQuoted 22 lines · Chandan Babu R
Quoted 3 lines · Christoph Hellwig
v5 file systems have been the default for more than 10 years. Drop support for non-v5 enabled kernels and xfsprogs.
Hi,
This patch is causing xfs/077 to fail as shown below,
# ./check xfs/077 FSTYP -- xfs (debug) PLATFORM -- Linux/x86_64 xfs-crc 6.9.0-rc4+ #2 SMP PREEMPT_DYNAMIC Mon Apr 29 08:08:05 GMT 2024 MKFS_OPTIONS -- -f -f -m crc=1,reflink=0,rmapbt=0, -i sparse=0 /dev/loop6 MOUNT_OPTIONS -- -o usrquota,grpquota,prjquota -o context=system_u:object_r:root_t:s0 /dev/loop6 /media/scratchxfs/077 9s ... [not run] Kernel doesn't support
meta_uuidfeature
Ran: xfs/077
Not run: xfs/077
Passed all 1 testsThe corresponding configuration file used had the following,
Quoted 7 lines
export TEST_DEV=/dev/loop5
export TEST_DIR=/media/test
export SCRATCH_DEV=/dev/loop6
export SCRATCH_MNT=/media/scratchMKFS_OPTIONS='-f -m crc=1,reflink=0,rmapbt=0, -i sparse=0' MOUNT_OPTIONS='-o usrquota,grpquota,prjquota'
Thanks for finding this issue, the _require_meta_uuid helper or the calling of _require_meta_uuid looks not correct. The xfs/077 calls _require_meta_uuid as:
_supported_fs xfs
_require_xfs_copy
_require_scratch
_require_no_large_scratch_dev
_require_meta_uuid <=====
...
The x/077 hasn't mkfs on SCRATCH_DEV, but the _require_meta_uuid does _scratch_xfs_db directly. That looks not right.
I think we should do "_scratch_mkfs ..." before doing _require_meta_uuid. E.g. mkfs at the beginning of _require_meta_uuid.
Any thoughts?
Thanks,
Zorro
Quoted 4 lines
Signature · 2 lines
-- Chandan
PATCH 2/6 remove xfs/096
In reply to Christoph HellwigThis test exercises mkfs error handling before strict validation was added and thus is useless for xfsprogs > 4.5.
--- .gitignore | 1 - common/xfs | 9 --- tests/xfs/096 | 130 ------------------------------------- tests/xfs/096.out.external | 50 -------------- tests/xfs/096.out.internal | 51 --------------- 5 files changed, 241 deletions(-) delete mode 100755 tests/xfs/096 delete mode 100644 tests/xfs/096.out.external delete mode 100644 tests/xfs/096.out.internal
.gitignore +0 -1
@@ -211,7 +211,6 @@ tags /tests/generic/050.out /tests/xfs/033.out /tests/xfs/071.out -/tests/xfs/096.out /tests/xfs/216.out # cscope files
common/xfs +0 -9
@@ -1211,15 +1211,6 @@ _require_xfs_mkfs_validation() fi } -# The opposite of _require_xfs_mkfs_validation. -_require_xfs_mkfs_without_validation() -{ - _xfs_mkfs_validation_check - if [ "$?" -ne 0 ]; then - _notrun "Requires older mkfs without strict input checks: the last supported version of xfsprogs is 4.5." - fi -} - _require_scratch_xfs_shrink() { _require_scratch
tests/xfs/096 deleted +0 -130
130 lines not on this page · raw message
tests/xfs/096.out.external deleted +0 -50
50 lines not on this page · raw message
tests/xfs/096.out.internal deleted +0 -51
51 lines not on this page · raw message
Signature · 1 line
2.39.2
Re: [PATCH 2/6] remove xfs/096
Quoted 4 lines · Christoph Hellwig
This test exercises mkfs error handling before strict validation was added and thus is useless for xfsprogs > 4.5.
Signed-off-by Christoph Hellwig <[email protected]>
Yesssssssss decruftifying. :)
--D
Quoted 296 lines
--- .gitignore | 1 - common/xfs | 9 --- tests/xfs/096 | 130 ------------------------------------- tests/xfs/096.out.external | 50 -------------- tests/xfs/096.out.internal | 51 --------------- 5 files changed, 241 deletions(-) delete mode 100755 tests/xfs/096 delete mode 100644 tests/xfs/096.out.external delete mode 100644 tests/xfs/096.out.internal.gitignore +0
-1@@ -211,7 +211,6 @@ tags /tests/generic/050.out /tests/xfs/033.out /tests/xfs/071.out -/tests/xfs/096.out /tests/xfs/216.out # cscope filescommon/xfs +0
-9@@ -1211,15 +1211,6 @@ _require_xfs_mkfs_validation() fi } -# The opposite of _require_xfs_mkfs_validation. -_require_xfs_mkfs_without_validation() -{ - _xfs_mkfs_validation_check - if [ "$?" -ne 0 ]; then - _notrun "Requires older mkfs without strict input checks: the last supported version of xfsprogs is 4.5." - fi -} - _require_scratch_xfs_shrink() { _require_scratchtests/xfs/096 deleted +0
-130@@ -1,130 +0,0 @@ -#! /bin/bash -# SPDX-License-Identifier: GPL-2.0 -# Copyright (c) 2000-2004 Silicon Graphics, Inc. All Rights Reserved. -# -# FS QA Test No. 096 -# -# test out mkfs_xfs output on IRIX/Linux and some of its error handling -# ensure pv#920679 is addressed -# -seqfull=$0 -. ./common/preamble -_begin_fstest mkfs v2log auto quick - -# Import common functions. -. ./common/filter -. ./common/log - -# -# filter out counts which will vary -# - extsz, blocks, agsize, agcount, device name, rtextents -# - log version varies for crc enabled fs -# - lsunit varies for 512/4k sector devices -# filter out differences between linux and irix: -# - sectsz on Linux -# - mmr, mixed-case on IRIX -# - lazy-count on IRIX -# - inode-paths on IRIX -# - trailing spaces on Linux but not on IRIX -# -# Example output: -# meta-data=DEV isize=256 agcount=N, agsize=N blks -# data = bsize=4096 blocks=N, imaxpct=25 -# = sunit=65 swidth=65 blks, unwritten=1 -# naming =version 2 bsize=4096 -# log =internal log bsize=4096 blocks=N, version=1 -# = sunit=0 blks -# realtime =none extsz=65536 blocks=N, rtextents=N -# -mkfs_filter() -{ - tee -a $seqres.full | \ - sed \ - -e 's/extsz=[0-9][0-9]*[ ]*/extsz=N, /' \ - -e 's/blocks=[0-9][0-9]*/blocks=N/' \ - -e 's/imaxpct=[0-9][0-9]*/imaxpct=N/' \ - -e 's/agsize=[0-9][0-9]*/agsize=N/' \ - -e 's/agcount=[0-9][0-9]*/agcount=N/' \ - -e 's/swidth=[0-9][0-9]* blks$/&, unwritten=1/' \ - -e 's/rtextents=[0-9][0-9]*/rtextents=N/' \ - -e 's/meta-data=[^ ]*/meta-data=DEV/' \ - -e 's/ *isize=[0-9]* / isize=N /' \ - -e '/ *= *sectsz=[0-9][0-9]* *attr=[0-9][0-9]*.*$/d' \ - -e '/ *= *mmr=[0-9][0-9]* *$/d' \ - -e 's/ *mixed-case=[YN]//' \ - -e 's/ *ascii-ci=[01]//' \ - -e 's/\(version=\)\([12]\)/\1N/' \ - -e 's/\(sunit=\)\([018] blks\)/\1N blks/' \ - -e 's/sectsz=[0-9][0-9]* *//' \ - -e 's/, lazy-count.*//' \ - -e '/inode-paths/d' \ - -e 's/\(log[ ]*=\).*bsize/\1LOG bsize/' \ - -e 's/\(realtime[ ]*=\).*extsz/\1REALTIME extsz/' \ - -e '/.*crc=/d' \ - -e 's/ *$//' \ - -e 's/ ftype=[01]//' \ - -e '/^log stripe unit.*too large/d' \ - -e '/^log stripe unit adjusted/d' \ - -e '/Discarding/d' \ - | grep -v parent -} - -# real QA test starts here - -# Modify as appropriate. -_supported_fs xfs -_require_scratch -_require_v2log -_require_xfs_mkfs_without_validation - -# choose .out file based on internal/external log -rm -f $seqfull.out -if [ "$USE_EXTERNAL" = yes ]; then - ln -s $seq.out.external $seqfull.out -else - ln -s $seq.out.internal $seqfull.out -fi - -# maximum log record size -max_lr_size=`expr 256 \* 1024` - -big_su=`expr $max_lr_size + 4096` - -# -# Test out various mkfs param combinations -# -cat >$tmp.seq.params <<EOF -# su too big but must be a multiple of fs block size too - -l version=2,su=`expr $max_lr_size + 512` -# test log stripe greater than LR size - -l version=2,su=$big_su -# same test but get log stripe from data stripe - -l version=2 -d su=$big_su,sw=1 -# test out data stripe - -m crc=0 -l version=1 -d su=$big_su,sw=1 -# test out data stripe the same but using sunit & swidth - -m crc=0 -l version=1 -d sunit=`expr $big_su / 512`,swidth=`expr $big_su / 512` -EOF - -# -# call mkfs in a loop for various params -# -echo "" -cat $tmp.seq.params \ -| while read mkfs -do - if echo $mkfs | grep -q '^#'; then - # print out header & ignore comment - echo $mkfs - continue - fi - echo "--- mkfs=$mkfs ---" - export MKFS_OPTIONS="$mkfs" - _scratch_mkfs_xfs 2>&1 | mkfs_filter - echo "" - echo "" -done - -# success, all done -status=0 -exittests/xfs/096.out.external deleted +0
-50@@ -1,50 +0,0 @@ -QA output created by 096 - -# su too big but must be a multiple of fs block size too ---- mkfs=-l version=2,su=262656 --- -log stripe unit (262656) must be a multiple of the block size (4096) - - -# test log stripe greater than LR size ---- mkfs=-l version=2,su=266240 --- -meta-data=DEV isize=N agcount=N, agsize=N blks -data = bsize=4096 blocks=N, imaxpct=N - = sunit=0 swidth=0 blks, unwritten=1 -naming =version 2 bsize=4096 -log =LOG bsize=4096 blocks=N, version=N -realtime =REALTIME extsz=N, blocks=N, rtextents=N - - -# same test but get log stripe from data stripe ---- mkfs=-l version=2 -d su=266240,sw=1 --- -meta-data=DEV isize=N agcount=N, agsize=N blks -data = bsize=4096 blocks=N, imaxpct=N - = sunit=65 swidth=65 blks, unwritten=1 -naming =version 2 bsize=4096 -log =LOG bsize=4096 blocks=N, version=N - = sunit=N blks -realtime =REALTIME extsz=N, blocks=N, rtextents=N - - -# test out data stripe ---- mkfs=-m crc=0 -l version=1 -d su=266240,sw=1 --- -meta-data=DEV isize=N agcount=N, agsize=N blks -data = bsize=4096 blocks=N, imaxpct=N - = sunit=65 swidth=65 blks, unwritten=1 -naming =version 2 bsize=4096 -log =LOG bsize=4096 blocks=N, version=N - = sunit=N blks -realtime =REALTIME extsz=N, blocks=N, rtextents=N - - -# test out data stripe the same but using sunit & swidth ---- mkfs=-m crc=0 -l version=1 -d sunit=520,swidth=520 --- -meta-data=DEV isize=N agcount=N, agsize=N blks -data = bsize=4096 blocks=N, imaxpct=N - = sunit=65 swidth=65 blks, unwritten=1 -naming =version 2 bsize=4096 -log =LOG bsize=4096 blocks=N, version=N - = sunit=N blks -realtime =REALTIME extsz=N, blocks=N, rtextents=N - -tests/xfs/096.out.internal deleted +0
-51@@ -1,51 +0,0 @@ -QA output created by 096 - -# su too big but must be a multiple of fs block size too ---- mkfs=-l version=2,su=262656 --- -log stripe unit (262656) must be a multiple of the block size (4096) - - -# test log stripe greater than LR size ---- mkfs=-l version=2,su=266240 --- -meta-data=DEV isize=N agcount=N, agsize=N blks -data = bsize=4096 blocks=N, imaxpct=N - = sunit=0 swidth=0 blks, unwritten=1 -naming =version 2 bsize=4096 -log =LOG bsize=4096 blocks=N, version=N - = sunit=N blks -realtime =REALTIME extsz=N, blocks=N, rtextents=N - - -# same test but get log stripe from data stripe ---- mkfs=-l version=2 -d su=266240,sw=1 --- -meta-data=DEV isize=N agcount=N, agsize=N blks -data = bsize=4096 blocks=N, imaxpct=N - = sunit=65 swidth=65 blks, unwritten=1 -naming =version 2 bsize=4096 -log =LOG bsize=4096 blocks=N, version=N - = sunit=N blks -realtime =REALTIME extsz=N, blocks=N, rtextents=N - - -# test out data stripe ---- mkfs=-m crc=0 -l version=1 -d su=266240,sw=1 --- -meta-data=DEV isize=N agcount=N, agsize=N blks -data = bsize=4096 blocks=N, imaxpct=N - = sunit=65 swidth=65 blks, unwritten=1 -naming =version 2 bsize=4096 -log =LOG bsize=4096 blocks=N, version=N - = sunit=N blks -realtime =REALTIME extsz=N, blocks=N, rtextents=N - - -# test out data stripe the same but using sunit & swidth ---- mkfs=-m crc=0 -l version=1 -d sunit=520,swidth=520 --- -meta-data=DEV isize=N agcount=N, agsize=N blks -data = bsize=4096 blocks=N, imaxpct=N - = sunit=65 swidth=65 blks, unwritten=1 -naming =version 2 bsize=4096 -log =LOG bsize=4096 blocks=N, version=N - = sunit=N blks -realtime =REALTIME extsz=N, blocks=N, rtextents=N - -Signature · 1 line
2.39.2
PATCH 3/6 xfs/078: remove the 512 byte block size sub-case
In reply to Christoph Hellwig512 byte block sizes are only supported for v4 file systems, and xfs/078 crudely forces use of v4 file systems for it. This doesn't work if the kernel is built without v4 support. Given that v4 support is slowly being phased out and 512 byte block sizes have never been common, drop this part of the test.
--- tests/xfs/078 | 9 ++------- tests/xfs/078.out | 18 ------------------ 2 files changed, 2 insertions(+), 25 deletions(-)
tests/xfs/078 +2 -7
@@ -69,12 +69,8 @@ _grow_loop() echo echo "*** mkfs loop file (size=$original)" - mkfs_crc_opts="" - if [ $bsize -lt 1024 -a -z "$XFS_MKFS_HAS_NO_META_SUPPORT" ]; then - mkfs_crc_opts="-m crc=0" - fi - $MKFS_XFS_PROG $mkfs_crc_opts -b size=$bsize $dparam $LOOP_DEV \ - | _filter_mkfs 2>/dev/null + $MKFS_XFS_PROG -b size=$bsize $dparam $LOOP_DEV | \ + _filter_mkfs 2>/dev/null echo "*** extend loop file" _destroy_loop_device $LOOP_DEV
@@ -104,7 +100,6 @@ _grow_loop $((168024*4096)) 1376452608 4096 1 # Some other blocksize cases... _grow_loop $((168024*2048)) 1376452608 2048 1 -_grow_loop $((168024*512)) 1376452608 512 1 16m _grow_loop $((168024*1024)) 688230400 1024 1 # Other corner cases suggested by dgc
tests/xfs/078.out +0 -18
@@ -37,24 +37,6 @@ data blocks changed from 168024 to 672096 *** unmount *** check -=== GROWFS (from 86028288 to 1376452608, 512 blocksize) - -*** mkfs loop file (size=86028288) -meta-data=DDEV isize=XXX agcount=N, agsize=XXX blks -data = bsize=XXX blocks=XXX, imaxpct=PCT - = sunit=XXX swidth=XXX, unwritten=X -naming =VERN bsize=XXX -log =LDEV bsize=XXX blocks=XXX -realtime =RDEV extsz=XXX blocks=XXX, rtextents=XXX -*** extend loop file -wrote 512/512 bytes at offset 1376452608 -*** mount loop filesystem -*** grow loop filesystem -xfs_growfs --BlockSize=512 --Blocks=163840 -data blocks changed from 163840 to 2688384 -*** unmount -*** check - === GROWFS (from 172056576 to 688230400, 1024 blocksize) *** mkfs loop file (size=172056576)
Signature · 1 line
2.39.2
Re: [PATCH 3/6] xfs/078: remove the 512 byte block size sub-case
Quoted 5 lines · Christoph Hellwig
512 byte block sizes are only supported for v4 file systems, and xfs/078 crudely forces use of v4 file systems for it. This doesn't work if the kernel is built without v4 support. Given that v4 support is slowly being phased out and 512 byte block sizes have never been common, drop this part of the test.
I've long wondered just how many people actually used that blocksize...
Quoted 11 lines
Signed-off-by Christoph Hellwig <[email protected]>--- tests/xfs/078 | 9 ++------- tests/xfs/078.out | 18 ------------------ 2 files changed, 2 insertions(+), 25 deletions(-)tests/xfs/078 +0
-0
tests/xfs/078 _grow_loop() · line 75
Quoted 7 lines
tests/xfs/078 +0
-4@@ -69,12 +69,8 @@ _grow_loop() echo echo "*** mkfs loop file (size=$original)" - mkfs_crc_opts="" - if [ $bsize -lt 1024 -a -z "$XFS_MKFS_HAS_NO_META_SUPPORT" ]; then - mkfs_crc_opts="-m crc=0" - fi
...because this was particularly nasty. Why wouldn't this test have skipped this iteration if the fs config doesn't support 512 blocksizes?
Who cares. Anyway,
--D
Quoted 48 lines
tests/xfs/078 +2
-3- $MKFS_XFS_PROG $mkfs_crc_opts -b size=$bsize $dparam $LOOP_DEV \ - | _filter_mkfs 2>/dev/null + $MKFS_XFS_PROG -b size=$bsize $dparam $LOOP_DEV | \ + _filter_mkfs 2>/dev/null echo "*** extend loop file" _destroy_loop_device $LOOP_DEV@@ -104,7 +100,6 @@ _grow_loop $((168024*4096)) 1376452608 4096 1 # Some other blocksize cases... _grow_loop $((168024*2048)) 1376452608 2048 1 -_grow_loop $((168024*512)) 1376452608 512 1 16m _grow_loop $((168024*1024)) 688230400 1024 1 # Other corner cases suggested by dgctests/xfs/078.out +0
-18@@ -37,24 +37,6 @@ data blocks changed from 168024 to 672096 *** unmount *** check -=== GROWFS (from 86028288 to 1376452608, 512 blocksize) - -*** mkfs loop file (size=86028288) -meta-data=DDEV isize=XXX agcount=N, agsize=XXX blks -data = bsize=XXX blocks=XXX, imaxpct=PCT - = sunit=XXX swidth=XXX, unwritten=X -naming =VERN bsize=XXX -log =LDEV bsize=XXX blocks=XXX -realtime =RDEV extsz=XXX blocks=XXX, rtextents=XXX -*** extend loop file -wrote 512/512 bytes at offset 1376452608 -*** mount loop filesystem -*** grow loop filesystem -xfs_growfs --BlockSize=512 --Blocks=163840 -data blocks changed from 163840 to 2688384 -*** unmount -*** check - === GROWFS (from 172056576 to 688230400, 1024 blocksize) *** mkfs loop file (size=172056576)Signature · 1 line
2.39.2
PATCH 4/6 xfs/263: remove the nocrc sub-test
In reply to Christoph HellwigRemove the test on nocrc file systems as v5 has been the default for 10 years and the kernel has made v4 support optional, which would fail this sub-case.
--- tests/xfs/263 | 5 --- tests/xfs/263.out | 85 ----------------------------------------------- 2 files changed, 90 deletions(-)
tests/xfs/263 +0 -5
@@ -66,11 +66,6 @@ function test_all_state() done } -echo "==== NO CRC ====" -_scratch_mkfs_xfs "-m crc=0 -n ftype=0" >> $seqres.full -test_all_state - -echo "==== CRC ====" _scratch_mkfs_xfs "-m crc=1" >>$seqres.full test_all_state
tests/xfs/263.out +0 -85
89 lines not on this page · raw message
Signature · 1 line
2.39.2
Re: [PATCH 4/6] xfs/263: remove the nocrc sub-test
Quoted 15 lines · Christoph Hellwig
Remove the test on nocrc file systems as v5 has been the default for 10 years and the kernel has made v4 support optional, which would fail this sub-case.
Signed-off-by Christoph Hellwig <[email protected]>--- tests/xfs/263 | 5 --- tests/xfs/263.out | 85 ----------------------------------------------- 2 files changed, 90 deletions(-)tests/xfs/263 +0
-0
tests/xfs/263 function test_all_state() · line 71
Quoted 6 lines
tests/xfs/263 +0
-3@@ -66,11 +66,6 @@ function test_all_state() done } -echo "==== NO CRC ====" -_scratch_mkfs_xfs "-m crc=0 -n ftype=0" >> $seqres.full -test_all_state
I think we should continue testing V4 quota options all the way to the end of support (~2030) by splitting these into two tests, one of which can use the _require_xfs_nocrc predicate introduced in the next patch. Thoughts?
--D
Quoted 103 lines
tests/xfs/263 +0
-2- -echo "==== CRC ====" _scratch_mkfs_xfs "-m crc=1" >>$seqres.full test_all_statetests/xfs/263.out +0
-85@@ -1,89 +1,4 @@ QA output created by 263 -==== NO CRC ==== -== Options: rw == -== Options: usrquota,rw == -User quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: ON - Enforcement: ON - Inode #XXX (1 blocks, 1 extents) -Group quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: OFF - Enforcement: OFF - Inode: N/A -Project quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: OFF - Enforcement: OFF - Inode: N/A -Blocks grace time: [7 days] -Inodes grace time: [7 days] -Realtime Blocks grace time: [7 days] -== Options: grpquota,rw == -User quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: OFF - Enforcement: OFF - Inode #XXX (1 blocks, 1 extents) -Group quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: ON - Enforcement: ON - Inode #XXX (1 blocks, 1 extents) -Project quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: OFF - Enforcement: OFF - Inode: N/A -Blocks grace time: [7 days] -Inodes grace time: [7 days] -Realtime Blocks grace time: [7 days] -== Options: usrquota,grpquota,rw == -User quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: ON - Enforcement: ON - Inode #XXX (1 blocks, 1 extents) -Group quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: ON - Enforcement: ON - Inode #XXX (1 blocks, 1 extents) -Project quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: OFF - Enforcement: OFF - Inode: N/A -Blocks grace time: [7 days] -Inodes grace time: [7 days] -Realtime Blocks grace time: [7 days] -== Options: prjquota,rw == -User quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: OFF - Enforcement: OFF - Inode #XXX (1 blocks, 1 extents) -Group quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: OFF - Enforcement: OFF - Inode: N/A -Project quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: ON - Enforcement: ON - Inode #XXX (1 blocks, 1 extents) -Blocks grace time: [7 days] -Inodes grace time: [7 days] -Realtime Blocks grace time: [7 days] -== Options: usrquota,prjquota,rw == -User quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: ON - Enforcement: ON - Inode #XXX (1 blocks, 1 extents) -Group quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: OFF - Enforcement: OFF - Inode: N/A -Project quota state on SCRATCH_MNT (SCRATCH_DEV) - Accounting: ON - Enforcement: ON - Inode #XXX (1 blocks, 1 extents) -Blocks grace time: [7 days] -Inodes grace time: [7 days] -Realtime Blocks grace time: [7 days] -== Options: grpquota,prjquota,rw == -== Options: usrquota,grpquota,prjquota,rw == -==== CRC ==== == Options: rw == == Options: usrquota,rw == User quota state on SCRATCH_MNT (SCRATCH_DEV)Signature · 1 line
2.39.2
Re: [PATCH 4/6] xfs/263: remove the nocrc sub-test
Quoted 4 lines · Darrick J. Wong
I think we should continue testing V4 quota options all the way to the end of support (~2030) by splitting these into two tests, one of which can use the
_require_xfs_nocrcpredicate introduced in the next patch. Thoughts?
Sure.
PATCH 5/6 xfs/522: use reflink instead of crc as test feature
In reply to Christoph HellwigReplace crc as the main test feature with reflink so that this test do not require v4 file system support.
--- tests/xfs/522 | 58 +++++++++++++++++++++++++-------------------------- 1 file changed, 29 insertions(+), 29 deletions(-)
tests/xfs/522 +29 -29
@@ -46,58 +46,58 @@ test_mkfs_config() { echo Simplest config file cat > $def_cfgfile << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config $def_cfgfile echo Piped-in config file test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL [metadata] -crc = 1 +reflink = 1 ENDL echo Full line comment test_mkfs_config << ENDL # This is a full line comment. [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL # This is a full line comment. [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL #This is a full line comment. [metadata] -crc = 0 +reflink = 0 ENDL echo End of line comment test_mkfs_config << ENDL [metadata] -crc = 0 ; This is an eol comment. +reflink = 0 ; This is an eol comment. ENDL test_mkfs_config << ENDL [metadata] -crc = 0 ;This is an eol comment. +reflink = 0 ;This is an eol comment. ENDL echo Multiple directives test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 finobt = 0 ENDL echo Multiple sections test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 [inode] sparse = 0
@@ -111,92 +111,92 @@ ENDL echo Space around the section name test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL echo Single space around the key/value directive test_mkfs_config << ENDL [metadata] - crc=0 + reflink=0 ENDL test_mkfs_config << ENDL [metadata] -crc =0 +reflink =0 ENDL test_mkfs_config << ENDL [metadata] -crc= 0 +reflink= 0 ENDL test_mkfs_config << ENDL [metadata] -crc=0 +reflink=0 ENDL echo Two spaces around the key/value directive test_mkfs_config << ENDL [metadata] - crc =0 + reflink =0 ENDL test_mkfs_config << ENDL [metadata] - crc= 0 + reflink= 0 ENDL test_mkfs_config << ENDL [metadata] - crc=0 + reflink=0 ENDL test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL [metadata] -crc =0 +reflink =0 ENDL test_mkfs_config << ENDL [metadata] -crc= 0 +reflink= 0 ENDL echo Three spaces around the key/value directive test_mkfs_config << ENDL [metadata] - crc = 0 + reflink = 0 ENDL test_mkfs_config << ENDL [metadata] - crc= 0 + reflink= 0 ENDL test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL echo Four spaces around the key/value directive test_mkfs_config << ENDL [metadata] - crc = 0 + reflink = 0 ENDL echo Arbitrary spaces and tabs test_mkfs_config << ENDL [metadata] - crc = 0 + reflink = 0 ENDL echo ambiguous comment/section names test_mkfs_config << ENDL [metadata] #[data] -crc = 0 +reflink = 0 ENDL echo ambiguous comment/variable names
Signature · 1 line
2.39.2
Re: [PATCH 5/6] xfs/522: use reflink instead of crc as test feature
Quoted 4 lines · Christoph Hellwig
Replace crc as the main test feature with reflink so that this test do not require v4 file system support.
Signed-off-by Christoph Hellwig <[email protected]>
Pretty straight conversion, so
--D
Quoted 193 lines
--- tests/xfs/522 | 58 +++++++++++++++++++++++++-------------------------- 1 file changed, 29 insertions(+), 29 deletions(-)tests/xfs/522 +29
-29@@ -46,58 +46,58 @@ test_mkfs_config() { echo Simplest config file cat > $def_cfgfile << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config $def_cfgfile echo Piped-in config file test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL [metadata] -crc = 1 +reflink = 1 ENDL echo Full line comment test_mkfs_config << ENDL # This is a full line comment. [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL # This is a full line comment. [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL #This is a full line comment. [metadata] -crc = 0 +reflink = 0 ENDL echo End of line comment test_mkfs_config << ENDL [metadata] -crc = 0 ; This is an eol comment. +reflink = 0 ; This is an eol comment. ENDL test_mkfs_config << ENDL [metadata] -crc = 0 ;This is an eol comment. +reflink = 0 ;This is an eol comment. ENDL echo Multiple directives test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 finobt = 0 ENDL echo Multiple sections test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 [inode] sparse = 0@@ -111,92 +111,92 @@ ENDL echo Space around the section name test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL echo Single space around the key/value directive test_mkfs_config << ENDL [metadata] - crc=0 + reflink=0 ENDL test_mkfs_config << ENDL [metadata] -crc =0 +reflink =0 ENDL test_mkfs_config << ENDL [metadata] -crc= 0 +reflink= 0 ENDL test_mkfs_config << ENDL [metadata] -crc=0 +reflink=0 ENDL echo Two spaces around the key/value directive test_mkfs_config << ENDL [metadata] - crc =0 + reflink =0 ENDL test_mkfs_config << ENDL [metadata] - crc= 0 + reflink= 0 ENDL test_mkfs_config << ENDL [metadata] - crc=0 + reflink=0 ENDL test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL test_mkfs_config << ENDL [metadata] -crc =0 +reflink =0 ENDL test_mkfs_config << ENDL [metadata] -crc= 0 +reflink= 0 ENDL echo Three spaces around the key/value directive test_mkfs_config << ENDL [metadata] - crc = 0 + reflink = 0 ENDL test_mkfs_config << ENDL [metadata] - crc= 0 + reflink= 0 ENDL test_mkfs_config << ENDL [metadata] -crc = 0 +reflink = 0 ENDL echo Four spaces around the key/value directive test_mkfs_config << ENDL [metadata] - crc = 0 + reflink = 0 ENDL echo Arbitrary spaces and tabs test_mkfs_config << ENDL [metadata] - crc = 0 + reflink = 0 ENDL echo ambiguous comment/section names test_mkfs_config << ENDL [metadata] #[data] -crc = 0 +reflink = 0 ENDL echo ambiguous comment/variable namesSignature · 1 line
2.39.2
PATCH 6/6 xfs: don't run tests that require v4 file systems when not supported
In reply to Christoph HellwigAdd a _require_xfs_nocrc helper that checks that we can mkfs and mount a crc=0 file systems before running tests that rely on it to avoid failures on kernels with CONFIG_XFS_SUPPORT_V4 disabled.
--- common/xfs | 10 ++++++++++ tests/xfs/002 | 1 + tests/xfs/045 | 1 + tests/xfs/095 | 1 + tests/xfs/132 | 1 + tests/xfs/148 | 2 ++ tests/xfs/158 | 1 + tests/xfs/160 | 1 + tests/xfs/194 | 2 ++ tests/xfs/199 | 1 + tests/xfs/300 | 1 + tests/xfs/513 | 1 + tests/xfs/526 | 1 + 13 files changed, 24 insertions(+)
common/xfs +10 -0
@@ -1852,3 +1852,13 @@ _xfs_discard_max_offset_kb() $XFS_IO_PROG -c 'statfs' "$1" | \ awk '{g[$1] = $3} END {print (g["geom.bsize"] * g["geom.datablocks"] / 1024)}' } + +# check if mkfs and the kernel support nocrc (v4) file systems +_require_xfs_nocrc() +{ + _scratch_mkfs_xfs -m crc=0 > /dev/null 2>&1 || \ + _notrun "v4 file systems not supported" + _try_scratch_mount > /dev/null 2>&1 || \ + _notrun "v4 file systems not supported" + _scratch_unmount +}
tests/xfs/002 +1 -0
@@ -23,6 +23,7 @@ _begin_fstest auto quick growfs _supported_fs xfs _require_scratch_nocheck _require_no_large_scratch_dev +_require_xfs_nocrc _scratch_mkfs_xfs -m crc=0 -d size=128m >> $seqres.full 2>&1 || _fail "mkfs failed"
tests/xfs/045 +1 -0
@@ -22,6 +22,7 @@ _supported_fs xfs _require_test _require_scratch_nocheck +_require_xfs_nocrc echo "*** get uuid" uuid=`_get_existing_uuid`
tests/xfs/095 +1 -0
@@ -19,6 +19,7 @@ _begin_fstest log v2log auto _supported_fs xfs _require_scratch _require_v2log +_require_xfs_nocrc if [ "$(blockdev --getss $SCRATCH_DEV)" != "512" ]; then _notrun "need 512b sector size"
tests/xfs/132 +1 -0
@@ -19,6 +19,7 @@ _supported_fs xfs # we intentionally corrupt the filesystem, so don't check it after the test _require_scratch_nocheck +_require_xfs_nocrc # on success, we'll get a shutdown filesystem with a really noisy log message # due to transaction cancellation. Hence we don't want to check dmesg here.
tests/xfs/148 +2 -0
@@ -27,6 +27,8 @@ _cleanup() _supported_fs xfs _require_test _require_attrs +_require_xfs_nocrc + _disable_dmesg_check imgfile=$TEST_DIR/img-$seq
tests/xfs/158 +1 -0
@@ -18,6 +18,7 @@ _supported_fs xfs _require_scratch_xfs_inobtcount _require_command "$XFS_ADMIN_PROG" "xfs_admin" _require_xfs_repair_upgrade inobtcount +_require_xfs_nocrc # Make sure we can't format a filesystem with inobtcount and not finobt. _scratch_mkfs -m crc=1,inobtcount=1,finobt=0 &> $seqres.full && \
tests/xfs/160 +1 -0
@@ -18,6 +18,7 @@ _supported_fs xfs _require_command "$XFS_ADMIN_PROG" "xfs_admin" _require_scratch_xfs_bigtime _require_xfs_repair_upgrade bigtime +_require_xfs_nocrc date --date='Jan 1 00:00:00 UTC 2040' > /dev/null 2>&1 || \ _notrun "Userspace does not support dates past 2038."
tests/xfs/194 +2 -0
@@ -30,6 +30,8 @@ _supported_fs xfs # real QA test starts here _require_scratch +_require_xfs_nocrc + _scratch_mkfs_xfs >/dev/null 2>&1 # For this test we use block size = 1/8 page size
tests/xfs/199 +1 -0
@@ -26,6 +26,7 @@ _cleanup() _supported_fs xfs _require_scratch +_require_xfs_nocrc # clear any mkfs options so that we can directly specify the options we need to # be able to test the features bitmask behaviour correctly.
tests/xfs/300 +1 -0
@@ -13,6 +13,7 @@ _begin_fstest auto fsr . ./common/filter _require_scratch +_require_xfs_nocrc # real QA test starts here
tests/xfs/513 +1 -0
@@ -37,6 +37,7 @@ _fixed_by_kernel_commit 237d7887ae72 \ _require_test _require_loop _require_xfs_io_command "falloc" +_require_xfs_nocrc LOOP_IMG=$TEST_DIR/$seq.dev LOOP_SPARE_IMG=$TEST_DIR/$seq.logdev
tests/xfs/526 +1 -0
@@ -26,6 +26,7 @@ _supported_fs xfs _require_test _require_scratch_nocheck _require_xfs_mkfs_cfgfile +_require_xfs_nocrc cfgfile=$TEST_DIR/a rm -rf $cfgfile
Signature · 1 line
2.39.2
Re: [PATCH 6/6] xfs: don't run tests that require v4 file systems when not supported
Quoted 8 lines · Christoph Hellwig
Add a
_require_xfs_nocrchelper that checks that we can mkfs and mount a crc=0 file systems before running tests that rely on it to avoid failures on kernels withCONFIG_XFS_SUPPORT_V4disabled.Signed-off-by Christoph Hellwig <[email protected]>--- common/xfs | 10 ++++++++++ tests/xfs/002 | 1 +
Looks fine to me.
Quoted 1 line
tests/xfs/045 | 1 +
xfs_db can change uuids on v5 filesystems now, so we don't nee the -mcrc=0 in this test.
Quoted 2 lines
tests/xfs/095 | 1 + tests/xfs/132 | 1 +
Looks fine to me.
Quoted 1 line
tests/xfs/148 | 2 ++
I wonder if we could rewrite this test to use either the xfs_db write -d command on dirents or attrs directly; or the link/attrset commands, since AFAICT the dir/attr code doesn't itself run namecheck when creating entries/attrs.
Quoted 2 lines
tests/xfs/158 | 1 + tests/xfs/160 | 1 +
inobtcount and bigtime are new features, maybe these two tests should lose the clause that checks that we can't upgrade a V4 filesystem?
Quoted 1 line
tests/xfs/194 | 2 ++
Not sure why this one is fixated on $pagesize/8. Was that a requirement to induce an error? Or would this work just as well on a 1k fsblock fs?
(Eric?)
Quoted 2 lines
tests/xfs/199 | 1 + tests/xfs/300 | 1 +
Looks fine to me.
Quoted 1 line
tests/xfs/513 | 1 +
I think we should split this into separate tests for V4/V5 options and only _require_xfs_nocrc the one with V4 options, because I wouldn't want to stop testing V5 codepaths simply because someone turned off V4 support in the kernle.
Quoted 1 line
tests/xfs/526 | 1 +
I'm at a loss on this one -- what it does is useful, but there aren't any V5 mkfs options that conflict as nicely as crc=0 does.
Quoted 170 lines
13 files changed, 24 insertions(+)common/xfs +10
-0@@ -1852,3 +1852,13 @@ _xfs_discard_max_offset_kb() $XFS_IO_PROG -c 'statfs' "$1" | \ awk '{g[$1] = $3} END {print (g["geom.bsize"] * g["geom.datablocks"] / 1024)}' } + +# check if mkfs and the kernel support nocrc (v4) file systems +_require_xfs_nocrc() +{ + _scratch_mkfs_xfs -m crc=0 > /dev/null 2>&1 || \ + _notrun "v4 file systems not supported" + _try_scratch_mount > /dev/null 2>&1 || \ + _notrun "v4 file systems not supported" + _scratch_unmount +}tests/xfs/002 +1
-0@@ -23,6 +23,7 @@ _begin_fstest auto quick growfs _supported_fs xfs _require_scratch_nocheck _require_no_large_scratch_dev +_require_xfs_nocrc _scratch_mkfs_xfs -m crc=0 -d size=128m >> $seqres.full 2>&1 || _fail "mkfs failed"tests/xfs/045 +1
-0@@ -22,6 +22,7 @@ _supported_fs xfs _require_test _require_scratch_nocheck +_require_xfs_nocrc echo "*** get uuid" uuid=`_get_existing_uuid`tests/xfs/095 +1
-0@@ -19,6 +19,7 @@ _begin_fstest log v2log auto _supported_fs xfs _require_scratch _require_v2log +_require_xfs_nocrc if [ "$(blockdev --getss $SCRATCH_DEV)" != "512" ]; then _notrun "need 512b sector size"tests/xfs/132 +1
-0@@ -19,6 +19,7 @@ _supported_fs xfs # we intentionally corrupt the filesystem, so don't check it after the test _require_scratch_nocheck +_require_xfs_nocrc # on success, we'll get a shutdown filesystem with a really noisy log message # due to transaction cancellation. Hence we don't want to check dmesg here.tests/xfs/148 +2
-0@@ -27,6 +27,8 @@ _cleanup() _supported_fs xfs _require_test _require_attrs +_require_xfs_nocrc + _disable_dmesg_check imgfile=$TEST_DIR/img-$seqtests/xfs/158 +1
-0@@ -18,6 +18,7 @@ _supported_fs xfs _require_scratch_xfs_inobtcount _require_command "$XFS_ADMIN_PROG" "xfs_admin" _require_xfs_repair_upgrade inobtcount +_require_xfs_nocrc # Make sure we can't format a filesystem with inobtcount and not finobt. _scratch_mkfs -m crc=1,inobtcount=1,finobt=0 &> $seqres.full && \tests/xfs/160 +1
-0@@ -18,6 +18,7 @@ _supported_fs xfs _require_command "$XFS_ADMIN_PROG" "xfs_admin" _require_scratch_xfs_bigtime _require_xfs_repair_upgrade bigtime +_require_xfs_nocrc date --date='Jan 1 00:00:00 UTC 2040' > /dev/null 2>&1 || \ _notrun "Userspace does not support dates past 2038."tests/xfs/194 +2
-0@@ -30,6 +30,8 @@ _supported_fs xfs # real QA test starts here _require_scratch +_require_xfs_nocrc + _scratch_mkfs_xfs >/dev/null 2>&1 # For this test we use block size = 1/8 page sizetests/xfs/199 +1
-0@@ -26,6 +26,7 @@ _cleanup() _supported_fs xfs _require_scratch +_require_xfs_nocrc # clear any mkfs options so that we can directly specify the options we need to # be able to test the features bitmask behaviour correctly.tests/xfs/300 +1
-0@@ -13,6 +13,7 @@ _begin_fstest auto fsr . ./common/filter _require_scratch +_require_xfs_nocrc # real QA test starts heretests/xfs/513 +1
-0@@ -37,6 +37,7 @@ _fixed_by_kernel_commit 237d7887ae72 \ _require_test _require_loop _require_xfs_io_command "falloc" +_require_xfs_nocrc LOOP_IMG=$TEST_DIR/$seq.dev LOOP_SPARE_IMG=$TEST_DIR/$seq.logdevtests/xfs/526 +1
-0@@ -26,6 +26,7 @@ _supported_fs xfs _require_test _require_scratch_nocheck _require_xfs_mkfs_cfgfile +_require_xfs_nocrc cfgfile=$TEST_DIR/a rm -rf $cfgfileSignature · 1 line
2.39.2
Re: [PATCH 6/6] xfs: don't run tests that require v4 file systems when not supported
Quoted 11 lines · Darrick J. Wong
Quoted 4 lines
Signed-off-by Christoph Hellwig <[email protected]>--- common/xfs | 10 ++++++++++ tests/xfs/002 | 1 +Looks fine to me.
Quoted 1 line
tests/xfs/045 | 1 +
xfs_dbcan change uuids on v5 filesystems now, so we don't nee the -mcrc=0 in this test.
Ok, I'll look into that.
Quoted 8 lines
Looks fine to me.
Quoted 1 line
tests/xfs/148 | 2 ++I wonder if we could rewrite this test to use either the
xfs_dbwrite -d command on dirents or attrs directly; or the link/attrset commands, since AFAICT the dir/attr code doesn't itself run namecheck when creating entries/attrs.
Can I leave that to you? :)
Quoted 5 lines
Quoted 2 lines
tests/xfs/158 | 1 + tests/xfs/160 | 1 +inobtcount and bigtime are new features, maybe these two tests should lose the clause that checks that we can't upgrade a V4 filesystem?
I'll take a look.
Quoted 6 lines
Quoted 1 line
tests/xfs/194 | 2 ++Not sure why this one is fixated on $pagesize/8. Was that a requirement to induce an error? Or would this work just as well on a 1k fsblock fs?
(Eric?)
I can check if it could be made to work on $pagesize/4, but I'll need to defer to Eric if that even makes sense.
Quoted 6 lines
Quoted 1 line
tests/xfs/513 | 1 +I think we should split this into separate tests for V4/V5 options and only
_require_xfs_nocrcthe one with V4 options, because I wouldn't want to stop testing V5 codepaths simply because someone turned off V4 support in the kernle.
Ok.
Quoted 4 lines
Quoted 1 line
tests/xfs/526 | 1 +I'm at a loss on this one -- what it does is useful, but there aren't any V5 mkfs options that conflict as nicely as crc=0 does.
Yes, I tried to look for conflicting options, but I couldn't find anything. Maybe we'll grow some before the v4 support is retired for real :)
Re: [PATCH 6/6] xfs: don't run tests that require v4 file systems when not supported
Quoted 17 lines · Christoph Hellwig
Quoted 11 lines · Darrick J. Wong
Quoted 4 lines
Signed-off-by Christoph Hellwig <[email protected]>--- common/xfs | 10 ++++++++++ tests/xfs/002 | 1 +Looks fine to me.
Quoted 1 line
tests/xfs/045 | 1 +
xfs_dbcan change uuids on v5 filesystems now, so we don't nee the -mcrc=0 in this test.Ok, I'll look into that.
Quoted 2 lines
Looks fine to me.
Quoted 8 lines
Quoted 6 lines
Quoted 1 line
tests/xfs/148 | 2 ++I wonder if we could rewrite this test to use either the
xfs_dbwrite -d command on dirents or attrs directly; or the link/attrset commands, since AFAICT the dir/attr code doesn't itself run namecheck when creating entries/attrs.Can I leave that to you? :)
Yes.
Quoted 27 lines
Quoted 5 lines
Quoted 2 lines
tests/xfs/158 | 1 + tests/xfs/160 | 1 +inobtcount and bigtime are new features, maybe these two tests should lose the clause that checks that we can't upgrade a V4 filesystem?
I'll take a look.
Quoted 6 lines
Quoted 1 line
tests/xfs/194 | 2 ++Not sure why this one is fixated on $pagesize/8. Was that a requirement to induce an error? Or would this work just as well on a 1k fsblock fs?
(Eric?)
I can check if it could be made to work on $pagesize/4, but I'll need to defer to Eric if that even makes sense.
Quoted 6 lines
Quoted 1 line
tests/xfs/513 | 1 +I think we should split this into separate tests for V4/V5 options and only
_require_xfs_nocrcthe one with V4 options, because I wouldn't want to stop testing V5 codepaths simply because someone turned off V4 support in the kernle.Ok.
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
tests/xfs/526 | 1 +I'm at a loss on this one -- what it does is useful, but there aren't any V5 mkfs options that conflict as nicely as crc=0 does.
Yes, I tried to look for conflicting options, but I couldn't find anything. Maybe we'll grow some before the v4 support is retired for real :)
Well hilariously just yesterday djwong-wtf just grew one now that you can't format rtgroups=1 without exchange=1 so I guess there's some hope.
--D
--D
Quoted 5 lines · Christoph Hellwig
Hi all,
this series ensures tests pass on kernels without v5 support. As a side effect it also removes support for historic kernels and xfsprogs without any v5 support, and without mkfs input validation.
Thanks for doing this! I'm wondering if fstests should do this "removing" earlier than xfs? Hope to hear more opinions from xfs list and other fstests users (especially from some LTS distro) :)
Thanks,
Zorro
Quoted 1 line
Quoted 7 lines · Zorro Lang
Quoted 3 lines
this series ensures tests pass on kernels without v5 support. As a side effect it also removes support for historic kernels and xfsprogs without any v5 support, and without mkfs input validation.
Thanks for doing this! I'm wondering if fstests should do this "removing" earlier than xfs? Hope to hear more opinions from xfs list and other fstests users (especially from some LTS distro) :)
What is being removed is support for kernels and xfsprogs that do not support v5 file systems at all, not testing on v4 file system for the test device and the large majority of tests using the scratch device without specifying an explicit version.
The exception from the above are two sub-cases for v4 that are removed in the this series - if we really care about them I could move them into separate tests, but I doubt it's worth it.
Quoted 5 lines · Christoph Hellwig
Quoted 4 lines · Zorro Lang
Quoted 3 lines
this series ensures tests pass on kernels without v5 support. As a side effect it also removes support for historic kernels and xfsprogs without any v5 support, and without mkfs input validation.
Quoted 8 lines
Quoted 3 lines
Thanks for doing this! I'm wondering if fstests should do this "removing" earlier than xfs? Hope to hear more opinions from xfs list and other fstests users (especially from some LTS distro) :)
What is being removed is support for kernels and xfsprogs that do not support v5 file systems at all, not testing on v4 file system for the test device and the large majority of tests using the scratch device without specifying an explicit version.
Sure, I think most of systems testing doesn't need this, except some old LTS distros.
Quoted 4 lines
The exception from the above are two sub-cases for v4 that are removed in the this series - if we really care about them I could move them into separate tests, but I doubt it's worth it.
Let me check and test this patchset more, before acking it. And give some time to get more review. Thanks for this patchset!
Thanks,
Zorro
Quoted 1 line
Quoted 19 lines · Zorro Lang
Quoted 5 lines · Christoph Hellwig
Quoted 4 lines · Zorro Lang
Quoted 3 lines
this series ensures tests pass on kernels without v5 support. As a side effect it also removes support for historic kernels and xfsprogs without any v5 support, and without mkfs input validation.
Quoted 8 lines
Quoted 3 lines
Thanks for doing this! I'm wondering if fstests should do this "removing" earlier than xfs? Hope to hear more opinions from xfs list and other fstests users (especially from some LTS distro) :)
What is being removed is support for kernels and xfsprogs that do not support v5 file systems at all, not testing on v4 file system for the test device and the large majority of tests using the scratch device without specifying an explicit version.
Sure, I think most of systems testing doesn't need this, except some old LTS distros.
Quoted 1 line
Quoted 6 lines
Quoted 3 lines
The exception from the above are two sub-cases for v4 that are removed in the this series - if we really care about them I could move them into separate tests, but I doubt it's worth it.
Let me check and test this patchset more, before acking it. And give some time to get more review. Thanks for this patchset!
My testing done on some old and new distro/kernels with v4 and v5 xfs. It looks good, doesn't bring in more issues. So I'd like to have this change.
But as the review points from XFS list, looks like the patch 4/6 is better to not be merged, and the patch 6/6 need further changes? So I'll look forward the V2, to take more actions.
Thanks,
Zorro
Quoted 5 lines
Thanks,
ZorroQuoted 1 line
Quoted 3 lines · Zorro Lang
But as the review points from XFS list, looks like the patch 4/6 is better to not be merged, and the patch 6/6 need further changes? So I'll look forward the V2, to take more actions.
Maybe you can take 1-3 and 5 for now and I'll rework the others?
Quoted 6 lines · Christoph Hellwig
Quoted 3 lines · Zorro Lang
But as the review points from XFS list, looks like the patch 4/6 is better to not be merged, and the patch 6/6 need further changes? So I'll look forward the V2, to take more actions.
Maybe you can take 1-3 and 5 for now and I'll rework the others?
I second that. I'll deal with xfs/158, 160, and 526, but the other patches that were ok could go in.
--D
Quoted 2 lines · Darrick J. Wong
I second that. I'll deal with xfs/158, 160, and 526, but the other patches that were ok could go in.
Should I resend with the other review comments fixed but the _require_xfs_nocrc left in those for now?
Quoted 6 lines · Christoph Hellwig
Quoted 2 lines · Darrick J. Wong
I second that. I'll deal with xfs/158, 160, and 526, but the other patches that were ok could go in.
Should I resend with the other review comments fixed but the
_require_xfs_nocrcleft in those for now?
Yes. You could even do:
_require_xfs_nocrc # FIXME: djwong said he'd fix this for real
If you like. ;)
--D
Quoted 6 lines · Christoph Hellwig
Quoted 3 lines · Zorro Lang
But as the review points from XFS list, looks like the patch 4/6 is better to not be merged, and the patch 6/6 need further changes? So I'll look forward the V2, to take more actions.
Maybe you can take 1-3 and 5 for now and I'll rework the others?
Sure, you can find these 4 patches in "patches-in-queue" branch. Feel free to tell me if you hope to change something.
Thanks,
Zorro