xfstests: fix 122 for recent xfsprogs
PATCH 1/3 xfstests: fix 122 for recent xfsprogs
- filter out xfs_alloctype_t, this was an internal emum that got removed
- filter out xfs_bmbt_rec_32_t, this is a variant of the xfs_bmbt_rec_t that had almost no users and was removed
- filter out xfs_dinode_core_t, the separate dinode core is gone, and just checking the size of the full dinode is enough
- accept xfs_bmbt_rec_t as the new canonical name for xfs_bmbt_rec_64_t, and replace the old name with the new one in the output stream.
122 +10 -1
@@ -57,6 +57,12 @@ s/sizeof\( xfs_agf_t \) = 64/sizeof( xfs fi } +# filter out known changes to xfs type names +_type_name_filter() +{ + sed -e 's/xfs_bmbt_rec_64_t/xfs_bmbt_rec_t/' +} + _attribute_filter() { sed -e 's/__attribute__ *[(][(]packed[)][)]//' \
@@ -73,7 +79,9 @@ cat >$cprog <<EOF EOF cat >$tmp.ignore <<EOF +xfs_alloctype_t xfs_buf_cancel_t +xfs_bmbt_rec_32_t xfs_swapext_t xfs_dirops_t xfs_efi_log_item_t
@@ -160,6 +168,7 @@ xfs_log_item_chunk_t xfs_trans_t xfs_dirent_t xfs_fsop_getparents_handlereq_t +xfs_dinode_core_t EOF echo 'int main(int argc, char *argv[]) {' >>$cprog
@@ -193,7 +202,7 @@ echo 'return 0; }' >>$cprog # create and run program cc -o $oprog $cprog >> $seq.full 2>&1 || \ _notrun "Could not compile test program (see end of $seq.full)" -$oprog | _type_size_filter > $progout +$oprog | _type_size_filter | _type_name_filter > $progout # # add addition sizes and xfs_sb_t fields that don't exist in the version
122.out +1 -4
@@ -50,7 +50,6 @@ sizeof( xfs_agfl_t ) = 4 sizeof( xfs_agi_t ) = 296 sizeof( xfs_alloc_rec_incore_t ) = 8 sizeof( xfs_alloc_rec_t ) = 8 -sizeof( xfs_alloctype_t ) = 4 sizeof( xfs_attr_inactive_list_t ) = 8 sizeof( xfs_attr_leaf_entry_t ) = 8 sizeof( xfs_attr_leaf_hdr_t ) = 32
@@ -61,14 +60,12 @@ sizeof( xfs_attr_leafblock_t ) = 56 sizeof( xfs_attr_shortform_t ) = 8 sizeof( xfs_attrlist_cursor_t ) = 16 sizeof( xfs_bmbt_key_t ) = 8 -sizeof( xfs_bmbt_rec_32_t ) = 16 -sizeof( xfs_bmbt_rec_64_t ) = 16 +sizeof( xfs_bmbt_rec_t ) = 16 sizeof( xfs_bmdr_block_t ) = 4 sizeof( xfs_btnum_t ) = 4 sizeof( xfs_buf_log_format_t ) = 24 sizeof( xfs_da_blkinfo_t ) = 12 sizeof( xfs_da_intnode_t ) = 24 -sizeof( xfs_dinode_core_t ) = 96 sizeof( xfs_dinode_fmt_t ) = 4 sizeof( xfs_dir2_data_free_t ) = 4 sizeof( xfs_dir2_data_hdr_t ) = 16
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
PATCH 2/2 xfstests: split mapped writes vs direct I/O tests from 091
This effectively reverts
xfstests: add mapped write fsx operations to 091
and adds a new test case for it. It tests something slightly different, and regressions in existing tests due to new features are pretty nasty in a test suite.
263 new file +64 -0
@@ -0,0 +1,64 @@ +#! /bin/bash +# FS QA Test No. 263 +# +# fsx exercising direct IO vs sub-block buffered I/O +# +#----------------------------------------------------------------------- +# Copyright (c) 2000-2004 Silicon Graphics, 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 [email protected] + +seq=`basename $0` +echo "QA output created by $seq" + +here=`pwd` +tmp=/tmp/$$ +status=1 # failure is the default! +trap "rm -f $tmp.*; exit \$status" 0 1 2 3 15 + +# get standard environment, filters and checks +. ./common.rc +. ./common.filter + +# real QA test starts here +_supported_fs generic +_supported_os Linux + +rm -f $seq.full + +run_fsx() +{ + echo fsx $@ | tee -a $seq.full + args=`echo $@ | sed -e "s/ BSIZE / $bsize /g" -e "s/ PSIZE / $psize /g"` + rm -f $TEST_DIR/junk + $here/ltp/fsx $args $TEST_DIR/junk >>$seq.full 2>&1 + if [ $? -ne 0 ]; then + cat $seq.full + exit 1 + fi +} + +psize=`$here/src/feature -s` +bsize=`blockdev --getss $TEST_DEV` + +run_fsx -N 10000 -o 8192 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z +run_fsx -N 10000 -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z + +status=0 +exit
group +1 -0
@@ -376,3 +376,4 @@ deprecated 260 auto quick trim 261 auto quick quota 262 auto quick quota +263 rw auto quick
263.out new file +3 -0
@@ -0,0 +1,3 @@ +QA output created by 263 +fsx -N 10000 -o 8192 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z +fsx -N 10000 -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z
091 +0 -6
@@ -88,11 +88,5 @@ kernel=`uname -r | sed -e 's/\(2\..\).* #run_fsx -N 10000 -o 128000 -l 500000 -r PSIZE -t PSIZE -w PSIZE -Z -W run_fsx -N 10000 -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -W - run_fsx -N 10000 -o 8192 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z - run_fsx -N 10000 -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z - -# Commented out calls above are less likely to pick up issues, so -# save time by commenting them out (leave 'em for manual testing). - status=0 exit
091.out +0 -2
@@ -5,5 +5,3 @@ fsx -N 10000 -o 32768 -l 500000 -r PSIZE 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 -fsx -N 10000 -o 8192 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z -fsx -N 10000 -o 128000 -l 500000 -r PSIZE -t BSIZE -w BSIZE -Z
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
PATCH 3/3 xfstests: allow 091 for non-xfs filesystems
The tests in 091 are entirely generic and pass e.g. on ext4 and jfs. btrfs fails it, but that looks a like a btrfs-specific issue to me.
091 +1 -1
@@ -37,7 +37,7 @@ trap "rm -f $tmp.*; exit \$status" 0 1 2 . ./common.filter # real QA test starts here -_supported_fs xfs +_supported_fs generic _supported_os Linux IRIX [ $HOSTOS == IRIX ] && _notrun "Not working on IRIX yet"
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Re: [PATCH 3/3] xfstests: allow 091 for non-xfs filesystems
Quoted 10 lines · Christoph Hellwig
The tests in 091 are entirely generic and pass e.g. on ext4 and jfs. btrfs fails it, but that looks a like a btrfs-specific issue to me.
Signed-off-by Christoph Hellwig <[email protected]>Index: xfstests-dev/091
===================================================================
091 trap "rm -f $tmp.*; exit \$status" 0 1 2 · line 43
Quoted 8 lines
091 +1
-1@@ -37,7 +37,7 @@ trap "rm -f $tmp.*; exit \$status" 0 1 2 . ./common.filter # real QA test starts here -_supported_fs xfs +_supported_fs generic _supported_os Linux IRIX [ $HOSTOS == IRIX ] && _notrun "Not working on IRIX yet"
Given that the test is not run on irix, I'd change the _supported_os line to just Linux as well, and kill the separate check for Irix while you are there.
Cheers,
Dave.
Signature · 7 lines
Dave Chinner [email protected] _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Re: [PATCH 3/3] xfstests: allow 091 for non-xfs filesystems
Quoted 3 lines
Given that the test is not run on irix, I'd change the
_supported_osline to just Linux as well, and kill the separate check for Irix while you are there.
Ok.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
PATCH v2 3/3 xfstests: allow 091 for non-xfs filesystems
In reply to Christoph HellwigThe tests in 091 are entirely generic and pass e.g. on ext4 and jfs. btrfs fails it, but that looks a like a btrfs-specific issue to me.
Also use _supported_os properly instead of erroring out manually on IRIX.
091 +2 -4
@@ -37,10 +37,8 @@ trap "rm -f $tmp.*; exit \$status" 0 1 2 . ./common.filter # real QA test starts here -_supported_fs xfs -_supported_os Linux IRIX - -[ $HOSTOS == IRIX ] && _notrun "Not working on IRIX yet" +_supported_fs generic +_supported_os Linux rm -f $seq.full
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Re: [PATCH 3/3 v2] xfstests: allow 091 for non-xfs filesystems
Quoted 7 lines · Christoph Hellwig
The tests in 091 are entirely generic and pass e.g. on ext4 and jfs. btrfs fails it, but that looks a like a btrfs-specific issue to me.
Also use
_supported_osproperly instead of erroring out manually on IRIX.Signed-off-by Christoph Hellwig <[email protected]>
Signature · 7 lines
Dave Chinner [email protected] _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Re: [PATCH 3/3 v2] xfstests: allow 091 for non-xfs filesystems
In reply to Christoph HellwigQuoted 7 lines · Christoph Hellwig
The tests in 091 are entirely generic and pass e.g. on ext4 and jfs. btrfs fails it, but that looks a like a btrfs-specific issue to me.
Also use
_supported_osproperly instead of erroring out manually on IRIX.Signed-off-by Christoph Hellwig <[email protected]>
Looks OK to me.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Re: [PATCH 2/2] xfstests: split mapped writes vs direct I/O tests from 091
In reply to Christoph HellwigQuoted 7 lines · Christoph Hellwig
This effectively reverts
xfstests: add mapped write fsx operations to 091
and adds a new test case for it. It tests something slightly different, and regressions in existing tests due to new features are pretty nasty in a test suite.
Makes sense.
Quoted 22 lines
Signed-off-by Christoph Hellwig <[email protected]>Index: xfstests-dev/263
===================================================================263 new file +15
-0@@ -0,0 +1,64 @@ +#! /bin/bash +# FS QA Test No. 263 +# +# fsx exercising direct IO vs sub-block buffered I/O +# +#----------------------------------------------------------------------- +# Copyright (c) 2000-2004 Silicon Graphics, 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
263 line 25
Quoted 10 lines
263 new file +10
-0+# 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 [email protected]
No point keeping an invalid email address as the owner of a new test.
FWIW, I'm considering removing this from the tests as the information is in the commit history, and the use of the field (emailing the owner when the automated test infrastructure SGI used 10 years ago failed) is no longer used...
Otherwise looks OK.
Signature · 7 lines
Dave Chinner [email protected] _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Re: [PATCH 2/2] xfstests: split mapped writes vs direct I/O tests from 091
Quoted 10 lines · Dave Chinner
263 line 25
Quoted 2 lines
263 new file +2
-0+# creator [email protected]No point keeping an invalid email address as the owner of a new test.
FWIW, I'm considering removing this from the tests as the information is in the commit history, and the use of the field (emailing the owner when the automated test infrastructure SGI used 10 years ago failed) is no longer used...
We should really just kill it. If anyone runs the test automated again he should get the mails and/or cc the xfs lists.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Re: [PATCH 2/2] xfstests: split mapped writes vs direct I/O tests from 091
In reply to Dave ChinnerQuoted 38 lines · Dave Chinner
Quoted 7 lines · Christoph Hellwig
This effectively reverts
xfstests: add mapped write fsx operations to 091
and adds a new test case for it. It tests something slightly different, and regressions in existing tests due to new features are pretty nasty in a test suite.
Makes sense.
Quoted 27 lines
Signed-off-by Christoph Hellwig <[email protected]>Index: xfstests-dev/263
===================================================================263 new file +20
-0@@ -0,0 +1,64 @@ +#! /bin/bash +# FS QA Test No. 263 +# +# fsx exercising direct IO vs sub-block buffered I/O +# +#----------------------------------------------------------------------- +# Copyright (c) 2000-2004 Silicon Graphics, 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
Quoted 8 lines
263 line 25
Quoted 5 lines
263 new file +5
-0+# +#----------------------------------------------------------------------- +# +# creator [email protected]No point keeping an invalid email address as the owner of a new test.
My only comment was that there is some small value in knowing who the original author of the test.
Quoted 4 lines
FWIW, I'm considering removing this from the tests as the information is in the commit history, and the use of the field (emailing the owner when the automated test infrastructure SGI used 10 years ago failed) is no longer used...
...but you're right, it would be in the commit history so that's probably sufficient. I have no objection to deleting it from all the tests (and the template).
-Alex
Quoted 3 lines
Otherwise looks OK.
Reviewed-by Dave Chinner <[email protected]>
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Re: [PATCH 2/2] xfstests: split mapped writes vs direct I/O tests from 091
In reply to Christoph HellwigQuoted 9 lines · Christoph Hellwig
This effectively reverts
xfstests: add mapped write fsx operations to 091
and adds a new test case for it. It tests something slightly different, and regressions in existing tests due to new features are pretty nasty in a test suite.
Signed-off-by Christoph Hellwig <[email protected]>
Looks good. I see no need to re-post just to delete the "author" line. Do that as a separate commit that updates all the tests at once.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Quoted 9 lines · Christoph Hellwig
- filter out
xfs_alloctype_t, this was an internal emum that got removed
- filter outxfs_bmbt_rec_32_t, this is a variant of thexfs_bmbt_rec_tthat had almost no users and was removed
- filter outxfs_dinode_core_t, the separate dinode core is gone, and just checking the size of the full dinode is enough
- acceptxfs_bmbt_rec_tas the new canonical name forxfs_bmbt_rec_64_t, and replace the old name with the new one in the output stream.Signed-off-by Christoph Hellwig <[email protected]>
Signature · 7 lines
Dave Chinner [email protected] _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Quoted 9 lines · Christoph Hellwig
- filter out
xfs_alloctype_t, this was an internal emum that got removed
- filter outxfs_bmbt_rec_32_t, this is a variant of thexfs_bmbt_rec_tthat had almost no users and was removed
- filter outxfs_dinode_core_t, the separate dinode core is gone, and just checking the size of the full dinode is enough
- acceptxfs_bmbt_rec_tas the new canonical name forxfs_bmbt_rec_64_t, and replace the old name with the new one in the output stream.Signed-off-by Christoph Hellwig <[email protected]>
Looks good.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs