mmap dio and DAX
cover mmap dio and DAX
common/rc : requires SCRATCH_DEV support DAXsrc/t_mmap_dio.c : intro mmap and O_DIRECT rw through files tests/generic/405 : IO between DAX/non-DAX mountpoints
tests/xfs/138 : IO between DAX/non-DAX xfs files(per-inode flag)
Xiong Zhou (4):
common/rc: add _require_scratch_dax
src/t_mmap_dio: add mmap dio test
xfs: test per-inode DAX flag by IO
generic: test mmap dio through DAX and non-DAX
.gitignore | 1 + common/rc | 10 +++++ src/Makefile | 2 +- src/t_mmap_dio.c | 81 ++++++++++++++++++++++++++++++++++++++ tests/generic/405 | 100 +++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + tests/xfs/138 | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 10 files changed, 304 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out
Signature · 1 line
1.8.3.1
PATCH 1/4 common/rc: add _require_scratch_dax
To make sure we can test DAX feature on SCRATCH_DEV.
--- common/rc | 10 ++++++++++ 1 file changed, 10 insertions(+)
common/rc +10 -0
@@ -2632,6 +2632,16 @@ _require_scratch_shutdown() _scratch_unmount } +# Does dax mount option work on this dev/fs? +_require_scratch_dax() +{ + _scratch_unmount + _scratch_mkfs > /dev/null 2>&1 + _scratch_mount -o dax || \ + _notrun "$SCRATCH_DEV $FSTYP does not support -o dax" + _scratch_unmount +} + # Does norecovery support by this fs? _require_norecovery() {
Signature · 1 line
1.8.3.1
Re: [PATCH 1/4] common/rc: add _require_scratch_dax
Quoted 12 lines · Xiong Zhou
To make sure we can test DAX feature on
SCRATCH_DEV.Signed-off-by Xiong Zhou <[email protected]>--- common/rc | 10 ++++++++++ 1 file changed, 10 insertions(+)common/rc +0
-0
common/rc _require_scratch_shutdown() · line 2638
Quoted 7 lines
common/rc +4
-0@@ -2632,6 +2632,16 @@ _require_scratch_shutdown() _scratch_unmount } +# Does dax mount option work on this dev/fs? +_require_scratch_dax() +{ + _scratch_unmount
Use _require_scratch instead of _scratch_unmount here. _require_scratch ensures we have SCRATCH_DEV defined and unmounts it if it's mounted.
Thanks,
Eryu
Quoted 16 lines
common/rc +6
-0+ _scratch_mkfs > /dev/null 2>&1 + _scratch_mount -o dax || \ + _notrun "$SCRATCH_DEV $FSTYP does not support -o dax" + _scratch_unmount +} + # Does norecovery support by this fs? _require_norecovery() {Signature · 6 lines
1.8.3.1 -- To unsubscribe from this list: send the line "unsubscribe fstests" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
PATCH 2/4 src/t_mmap_dio: add mmap dio test
In reply to Xiong ZhouTo test read/write between a mmap area of one file and another file directly, with different size.
--- .gitignore | 1 + src/Makefile | 2 +- src/t_mmap_dio.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 83 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c
.gitignore +1 -0
@@ -129,6 +129,7 @@ /src/cloner /src/renameat2 /src/t_rename_overwrite +/src/t_mmap_dio # dmapi/ binaries /dmapi/src/common/cmd/read_invis
src/Makefile +1 -1
@@ -12,7 +12,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \ godown resvtest writemod makeextents itrash rename \ multi_open_unlink dmiperf unwritten_sync genhashnames t_holes \ t_mmap_writev t_truncate_cmtime dirhash_collide t_rename_overwrite \ - holetest t_truncate_self + holetest t_truncate_self t_mmap_dio LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \ preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
src/t_mmap_dio.c new file +81 -0
81 lines not on this page · raw message
Signature · 1 line
1.8.3.1
PATCH 3/4 xfs: test per-inode DAX flag by IO
In reply to Xiong ZhouWe do mmap and O_DIRECT read/write, through files with and without DAX flag.
--- tests/xfs/138 | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 ++ tests/xfs/group | 1 + 3 files changed, 108 insertions(+) create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out
tests/xfs/138 new file +105 -0
@@ -0,0 +1,105 @@ +#! /bin/bash +# FS QA Test 138 +# +# Test per-inode DAX flag by mmap direct IO. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017 Red Hat 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 +#----------------------------------------------------------------------- +# + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +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 + +_cleanup() +{ + cd / + rm -f $tmp.* +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# remove previous $seqres.full before test +rm -f $seqres.full + +_supported_fs xfs +_supported_os Linux +_require_scratch_dax +_require_test_program "feature" +_require_xfs_io_command "chattr" "+/-x" + +# $1 mmap read/write size +_dax_flag_mmap_dio() +{ + # both dax + $XFS_IO_PROG -c "chattr +x" $SCRATCH_MNT/tf_s + $XFS_IO_PROG -c "chattr +x" $SCRATCH_MNT/tf_d + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 + + # from non dax to dax + $XFS_IO_PROG -c "chattr -x" $SCRATCH_MNT/tf_s + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 + + # from dax to non dax + $XFS_IO_PROG -c "chattr +x" $SCRATCH_MNT/tf_s + $XFS_IO_PROG -c "chattr -x" $SCRATCH_MNT/tf_d + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 + + # both non dax + $XFS_IO_PROG -c "chattr -x" $SCRATCH_MNT/tf_s + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 +} + +_do_tests() +{ + # less than page size + _dax_flag_mmap_dio 1024 + # page size + _dax_flag_mmap_dio `src/feature -s` + # bigger sizes, for PMD faults + _dax_flag_mmap_dio 16777216 + _dax_flag_mmap_dio 67108864 +} + +_scratch_mkfs >> $seqres.full 2>&1 || _fail "mkfs failed" + +# mount with dax option +_scratch_mount "-o dax" >> $seqres.full 2>&1 || \ + _fail "mount -o dax failed" + +dd if=/dev/zero of=$SCRATCH_MNT/tf_s bs=4k count=256k > /dev/null 2>&1 +dd if=/dev/zero of=$SCRATCH_MNT/tf_d bs=4k count=256k > /dev/null 2>&1 + +_do_tests +_scratch_unmount >> $seqres.full 2>&1 || _fail "umount failed" + +# mount again wo/ dax option +_scratch_mount >> $seqres.full 2>&1 || _fail "mount failed" +_do_tests +_scratch_unmount >> $seqres.full 2>&1 + +# success, all done +echo "Silence is golden" +status=0 +exit
tests/xfs/138.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 138 +Silence is golden
tests/xfs/group +1 -0
@@ -135,6 +135,7 @@ 135 auto logprint quick v2log 136 attr2 137 auto metadata v2log +138 auto attr quick 139 auto quick clone 140 auto clone 141 auto log metadata
Signature · 1 line
1.8.3.1
Re: [PATCH 3/4] xfs: test per-inode DAX flag by IO
Quoted 4 lines · Xiong Zhou
We do mmap and
O_DIRECTread/write, through files
with and without DAX flag.Signed-off-by Xiong Zhou <[email protected]>
I think patch 1/4 and 2/4 could be folded into this patch, where the new helper and test program are used for the first time.
Quoted 53 lines
--- tests/xfs/138 | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 ++ tests/xfs/group | 1 + 3 files changed, 108 insertions(+) create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.outtests/xfs/138 new file +39
-0@@ -0,0 +1,105 @@ +#! /bin/bash +# FS QA Test 138 +# +# Test per-inode DAX flag by mmap direct IO. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017 Red Hat 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 +#----------------------------------------------------------------------- +# + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +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 + +_cleanup() +{ + cd / + rm -f $tmp.* +} + +# get standard environment, filters and checks
tests/xfs/138 line 49
Quoted 10 lines
tests/xfs/138 new file +10
-0+. ./common/rc +. ./common/filter + +# remove previous $seqres.full before test +rm -f $seqres.full + +_supported_fs xfs +_supported_os Linux +_require_scratch_dax +_require_test_program "feature"
Need _require_test_program "t_mmap_dio"
tests/xfs/138 line 53
Quoted 4 lines
tests/xfs/138 new file +4
-0+_require_xfs_io_command "chattr" "+/-x" + +# $1 mmap read/write size +_dax_flag_mmap_dio()
Name local functions without leading underscore, it's reserved for common helpers.
Quoted 26 lines
tests/xfs/138 new file +26
-0+{ + # both dax + $XFS_IO_PROG -c "chattr +x" $SCRATCH_MNT/tf_s + $XFS_IO_PROG -c "chattr +x" $SCRATCH_MNT/tf_d + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 + + # from non dax to dax + $XFS_IO_PROG -c "chattr -x" $SCRATCH_MNT/tf_s + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 + + # from dax to non dax + $XFS_IO_PROG -c "chattr +x" $SCRATCH_MNT/tf_s + $XFS_IO_PROG -c "chattr -x" $SCRATCH_MNT/tf_d + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 + + # both non dax + $XFS_IO_PROG -c "chattr -x" $SCRATCH_MNT/tf_s + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 +} + +_do_tests() +{ + # less than page size + _dax_flag_mmap_dio 1024 + # page size + _dax_flag_mmap_dio `src/feature -s`
tests/xfs/138 line 89
Quoted 10 lines
tests/xfs/138 new file +10
-0+ # bigger sizes, for PMD faults + _dax_flag_mmap_dio 16777216 + _dax_flag_mmap_dio 67108864 +} + +_scratch_mkfs >> $seqres.full 2>&1 || _fail "mkfs failed" + +# mount with dax option +_scratch_mount "-o dax" >> $seqres.full 2>&1 || \ + _fail "mount -o dax failed"
No need to _fail, test would fail if mkfs & mount failed.
tests/xfs/138 line 100
Quoted 11 lines
tests/xfs/138 new file +11
-0+ +dd if=/dev/zero of=$SCRATCH_MNT/tf_s bs=4k count=256k > /dev/null 2>&1 +dd if=/dev/zero of=$SCRATCH_MNT/tf_d bs=4k count=256k > /dev/null 2>&1 + +_do_tests +_scratch_unmount >> $seqres.full 2>&1 || _fail "umount failed" + +# mount again wo/ dax option +_scratch_mount >> $seqres.full 2>&1 || _fail "mount failed" +_do_tests +_scratch_unmount >> $seqres.full 2>&1
No need to unmount, test harness does this for you after test :)
Thanks,
Eryu
Quoted 32 lines
tests/xfs/138 new file +5
-0+ +# success, all done +echo "Silence is golden" +status=0 +exittests/xfs/138.out new file +2
-0@@ -0,0 +1,2 @@ +QA output created by 138 +Silence is goldentests/xfs/group +1
-0@@ -135,6 +135,7 @@ 135 auto logprint quick v2log 136 attr2 137 auto metadata v2log +138 auto attr quick 139 auto quick clone 140 auto clone 141 auto log metadataSignature · 6 lines
1.8.3.1 -- To unsubscribe from this list: send the line "unsubscribe fstests" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
PATCH 4/4 generic: test mmap dio through DAX and non-DAX
In reply to Xiong Zhou--- tests/generic/405 | 100 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + 3 files changed, 103 insertions(+) create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out
tests/generic/405 new file +100 -0
@@ -0,0 +1,100 @@ +#! /bin/bash +# FS QA Test 405 +# +# mmap dio between DAX and non-DAX mountpoints. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017 Red Hat 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 +#----------------------------------------------------------------------- +# + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +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 + +_cleanup() +{ + cd / + rm -f $tmp.* +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# remove previous $seqres.full before test +rm -f $seqres.full + +# real QA test starts here + +# Modify as appropriate. +_supported_fs xfs ext4 +_supported_os Linux +_require_test +_require_scratch_dax +_require_test_program "feature" + +# $1 mmap read/write size +_dax_flag_mmap_dio() +{ + # both dax + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 + + # from non dax to dax + src/t_mmap_dio $TEST_DIR/tf_s $SCRATCH_MNT/tf_d $1 + + # from dax to non dax + src/t_mmap_dio $SCRATCH_MNT/tf_s $TEST_DIR/tf_d $1 + + # both non dax + src/t_mmap_dio $TEST_DIR/tf_s $TEST_DIR/tf_d $1 +} + +_do_tests() +{ + # less than page size + _dax_flag_mmap_dio 1024 + # page size + _dax_flag_mmap_dio `src/feature -s` + # bigger sizes, for PMD faults + _dax_flag_mmap_dio 16777216 + _dax_flag_mmap_dio 67108864 +} + +_scratch_mkfs >> $seqres.full 2>&1 || _fail "mkfs failed" + +# SCRATCH_DEV mount with dax option +_scratch_mount "-o dax" >> $seqres.full 2>&1 || \ + _fail "mount -o dax failed" + +dd if=/dev/zero of=$SCRATCH_MNT/tf_s bs=4k count=256k > /dev/null 2>&1 +dd if=/dev/zero of=$SCRATCH_MNT/tf_d bs=4k count=256k > /dev/null 2>&1 +dd if=/dev/zero of=$TEST_DIR/tf_s bs=4k count=256k > /dev/null 2>&1 +dd if=/dev/zero of=$TEST_DIR/tf_d bs=4k count=256k > /dev/null 2>&1 + +_do_tests +_test_unmount +_scratch_unmount + +# success, all done +echo "Silence is golden" +status=0 +exit
tests/generic/405.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 405 +Silence is golden
tests/generic/group +1 -0
@@ -407,3 +407,4 @@ 402 auto quick rw 403 auto quick attr 404 auto quick insert +405 auto quick
Signature · 1 line
1.8.3.1
Re: [PATCH 4/4] generic: test mmap dio through DAX and non-DAX
Quoted 1 line · Xiong Zhou
Signed-off-by Xiong Zhou <[email protected]>
Better to have detailed commit log.
Quoted 53 lines
--- tests/generic/405 | 100 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + 3 files changed, 103 insertions(+) create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.outtests/generic/405 new file +39
-0@@ -0,0 +1,100 @@ +#! /bin/bash +# FS QA Test 405 +# +# mmap dio between DAX and non-DAX mountpoints. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017 Red Hat 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 +#----------------------------------------------------------------------- +# + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +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 + +_cleanup() +{ + cd / + rm -f $tmp.* +} + +# get standard environment, filters and checks
tests/generic/405 line 49
Quoted 10 lines
tests/generic/405 new file +10
-0+. ./common/rc +. ./common/filter + +# remove previous $seqres.full before test +rm -f $seqres.full + +# real QA test starts here + +# Modify as appropriate. +_supported_fs xfs ext4
_supported_fs generic
Other _require rules make sure we have everything we need, so no need to restrict supported fs to xfs and ext4. (And if we really want to do so, the test should be a "shared" test, not generic.)
tests/generic/405 line 56
Quoted 7 lines
tests/generic/405 new file +7
-0+_supported_os Linux +_require_test +_require_scratch_dax +_require_test_program "feature" + +# $1 mmap read/write size +_dax_flag_mmap_dio()
Not testing per-file flag now, the name seems a bit misleading.
tests/generic/405 line 62
Quoted 6 lines
tests/generic/405 new file +6
-0+{ + # both dax + src/t_mmap_dio $SCRATCH_MNT/tf_s $SCRATCH_MNT/tf_d $1 + + # from non dax to dax + src/t_mmap_dio $TEST_DIR/tf_s $SCRATCH_MNT/tf_d $1
Assuming TEST_DIR isn't a DAX mount, which seems not always the case.
Perhaps you can check mount options of TEST_DEV using _fs_options() to see if it's mounted with dax, and _notrun if it is.
Thanks,
Eryu
Quoted 62 lines
tests/generic/405 new file +38
-0+ + # from dax to non dax + src/t_mmap_dio $SCRATCH_MNT/tf_s $TEST_DIR/tf_d $1 + + # both non dax + src/t_mmap_dio $TEST_DIR/tf_s $TEST_DIR/tf_d $1 +} + +_do_tests() +{ + # less than page size + _dax_flag_mmap_dio 1024 + # page size + _dax_flag_mmap_dio `src/feature -s` + # bigger sizes, for PMD faults + _dax_flag_mmap_dio 16777216 + _dax_flag_mmap_dio 67108864 +} + +_scratch_mkfs >> $seqres.full 2>&1 || _fail "mkfs failed" + +# SCRATCH_DEV mount with dax option +_scratch_mount "-o dax" >> $seqres.full 2>&1 || \ + _fail "mount -o dax failed" + +dd if=/dev/zero of=$SCRATCH_MNT/tf_s bs=4k count=256k > /dev/null 2>&1 +dd if=/dev/zero of=$SCRATCH_MNT/tf_d bs=4k count=256k > /dev/null 2>&1 +dd if=/dev/zero of=$TEST_DIR/tf_s bs=4k count=256k > /dev/null 2>&1 +dd if=/dev/zero of=$TEST_DIR/tf_d bs=4k count=256k > /dev/null 2>&1 + +_do_tests +_test_unmount +_scratch_unmount + +# success, all done +echo "Silence is golden" +status=0 +exittests/generic/405.out new file +2
-0@@ -0,0 +1,2 @@ +QA output created by 405 +Silence is goldentests/generic/group +1
-0@@ -407,3 +407,4 @@ 402 auto quick rw 403 auto quick attr 404 auto quick insert +405 auto quickSignature · 6 lines
1.8.3.1 -- To unsubscribe from this list: send the line "unsubscribe fstests" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
v2 cover mmap dio and DAX
v2 :
Merge helper function changes into the first patch; Rewrite _require_dax, check options for sure; Print msg in t_mmap_dio.c to show which test going wrong; Empty mount options and check after mount to ensure we
wont mount with wrong option;
Remove unnecessary leading underscore and _fail; Use xfs_io instead of dd;
Other minor fixes.
Test status:
Both cases not run on normal block device; Both cases PASS on ramdisk based pmem devices; Both cases FAIL on nvdimm devices with:
+write(Bad address) len 1024 dax to nondax +write(Bad address) len 4096 dax to nondax +write(Bad address) len 16777216 dax to nondax +write(Bad address) len 67108864 dax to nondax
Xiong Zhou (2):
xfs: test per-inode DAX flag by IO
generic: test mmap dio through DAX and non-DAX
.gitignore | 1 + common/rc | 14 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 81 +++++++++++++++++++++++++++++++++++ tests/generic/405 | 116 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + tests/xfs/138 | 110 +++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 10 files changed, 329 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out
Signature · 1 line
1.8.3.1
PATCH v2 1/2 xfs: test per-inode DAX flag by IO
In a DAX mountpoint, do IO betwen files with and without DAX per-inode flag. We do mmap and O_DIRECT read/write IO in this case. Then test again in the same device without dax mountoption.
Add help _require_scratch_dax to make sure we can test DAX feature on SCRATCH_DEV.
Add mmap dio test programme to test read/write between a mmap area of one file and another file directly, with different size.
--- .gitignore | 1 + common/rc | 14 +++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 81 ++++++++++++++++++++++++++++++++++++++++ tests/xfs/138 | 110 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 7 files changed, 210 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out
.gitignore +1 -0
@@ -129,6 +129,7 @@ /src/cloner /src/renameat2 /src/t_rename_overwrite +/src/t_mmap_dio # dmapi/ binaries /dmapi/src/common/cmd/read_invis
common/rc +14 -0
@@ -2632,6 +2632,20 @@ _require_scratch_shutdown() _scratch_unmount } +# Does dax mount option work on this dev/fs? +_require_scratch_dax() +{ + _require_scratch + _scratch_mkfs > /dev/null 2>&1 + _scratch_mount -o dax + # Check options to be sure. XFS ignores dax option + # and goes on if dev underneath does not support dax. + _fs_options $SCRATCH_DEV | grep -w "dax" > /dev/null 2>&1 + [ $? -ne 0 ] && \ + _notrun "$SCRATCH_DEV $FSTYP does not support -o dax" + _scratch_unmount +} + # Does norecovery support by this fs? _require_norecovery() {
src/Makefile +1 -1
@@ -12,7 +12,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \ godown resvtest writemod makeextents itrash rename \ multi_open_unlink dmiperf unwritten_sync genhashnames t_holes \ t_mmap_writev t_truncate_cmtime dirhash_collide t_rename_overwrite \ - holetest t_truncate_self + holetest t_truncate_self t_mmap_dio LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \ preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
src/t_mmap_dio.c new file +81 -0
81 lines not on this page · raw message
tests/xfs/138 new file +110 -0
110 lines not on this page · raw message
tests/xfs/138.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 138 +Silence is golden
tests/xfs/group +1 -0
@@ -135,6 +135,7 @@ 135 auto logprint quick v2log 136 attr2 137 auto metadata v2log +138 auto attr quick 139 auto quick clone 140 auto clone 141 auto log metadata
Signature · 1 line
1.8.3.1
Re: [PATCH v2 1/2] xfs: test per-inode DAX flag by IO
Quoted 150 lines · Xiong Zhou
In a DAX mountpoint, do IO betwen files with and without DAX per-inode flag. We do mmap and
O_DIRECTread/write IO in this case. Then test again in the same device without dax mountoption.Add help
_require_scratch_daxto make sure we can test DAX feature onSCRATCH_DEV.Add mmap dio test programme to test read/write between a mmap area of one file and another file directly, with different size.
Signed-off-by Xiong Zhou <[email protected]>--- .gitignore | 1 + common/rc | 14 +++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 81 ++++++++++++++++++++++++++++++++++++++++ tests/xfs/138 | 110 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 7 files changed, 210 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out.gitignore +1
-0@@ -129,6 +129,7 @@ /src/cloner /src/renameat2 /src/t_rename_overwrite +/src/t_mmap_dio # dmapi/ binaries /dmapi/src/common/cmd/read_inviscommon/rc +14
-0@@ -2632,6 +2632,20 @@ _require_scratch_shutdown() _scratch_unmount } +# Does dax mount option work on this dev/fs? +_require_scratch_dax() +{ + _require_scratch + _scratch_mkfs > /dev/null 2>&1 + _scratch_mount -o dax + # Check options to be sure. XFS ignores dax option + # and goes on if dev underneath does not support dax. + _fs_options $SCRATCH_DEV | grep -w "dax" > /dev/null 2>&1 + [ $? -ne 0 ] && \ + _notrun "$SCRATCH_DEV $FSTYP does not support -o dax" + _scratch_unmount +} + # Does norecovery support by this fs? _require_norecovery() {src/Makefile +1
-1@@ -12,7 +12,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \ godown resvtest writemod makeextents itrash rename \ multi_open_unlink dmiperf unwritten_sync genhashnames t_holes \ t_mmap_writev t_truncate_cmtime dirhash_collide t_rename_overwrite \ - holetest t_truncate_self + holetest t_truncate_self t_mmap_dio LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \ preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \src/t_mmap_dio.c new file +68
-0@@ -0,0 +1,81 @@ +/* +* This programme was originally written by +* Jeff Moyer <[email protected]> +*/ +#define _GNU_SOURCE 1 +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <unistd.h> +#include <fcntl.h> +#include <sys/mman.h> +#include <libaio.h> +#include <errno.h> +#include <sys/time.h> + +void usage (char *prog) +{ +fprintf (stderr, +"usage: %s <src file> <dest file> <size> <msg>\n" , + prog); +exit (1 ); +} + +void err_exit (char *op,unsigned long len,char *s) +{ +fprintf (stderr,"%s(%s) len %lu %s\n" , + op,strerror (errno), len, s); +exit (1 ); +} + +int main (int argc,char **argv) +{ +int fd, fd2, ret; +char *map; +unsigned long len; + +if (argc <3 ) +usage (basename (argv[0 ])); + + len =strtoul (argv[3 ],NULL ,10 ); +if (errno ==ERANGE ) +err_exit ("strtoul" ,0 , argv[4 ]); + +/* Open source file and mmap*/ + fd =open (argv[1 ],O_RDWR ,0644 ); +if (fd <0 ) +err_exit ("open s" , len, argv[4 ]); + + map = (char *)mmap (NULL , len, +PROT_READ |PROT_WRITE ,MAP_SHARED , fd,0 ); +if (map ==MAP_FAILED ) +err_exit ("mmap" , len, argv[4 ]); + +/* Open dest file with O_DIRECT */ + fd2 =open (argv[2 ],O_RDWR |O_DIRECT ,0644 ); +if (fd2 <0 ) +err_exit ("open d" , len, argv[4 ]); + +/* First, test storing to dest file from source mapping */ + ret =write (fd2, map, len); +if (ret != len) +err_exit ("write" , len, argv[4 ]); + + ret = (int )lseek (fd2,0 ,SEEK_SET ); +if (ret ==-1 ) +err_exit ("lseek" , len, argv[4 ]); + +/* Next, test reading from dest file into source mapping */
src/t_mmap_dio.c line 78
Quoted 10 lines
src/t_mmap_dio.c new file +10
-0+ ret =read (fd2, map, len); +if (ret != len) +err_exit ("read" , len, argv[4 ]); + ret =msync (map, len,MS_SYNC ); +if (ret <0 ) +err_exit ("msync" , len, argv[4 ]); + + ret =munmap (map, len); +if (ret <0 ) +err_exit ("munmap" , len, argv[4 ]);
Small nit: you munmap() to clean up the mmap(), but you don't close fd or fd2.
Re: [PATCH v2 1/2] xfs: test per-inode DAX flag by IO
Quoted 1 line · Ross Zwisler
On Thu, Jan 19, 2017 at 06:13:57PM +0800, Xiong Zhou wrote:
...snip..
Quoted 5 lines
src/t_mmap_dio.c line 78
Quoted 3 lines
src/t_mmap_dio.c new file +3
-0+ ret =munmap (map, len); +if (ret <0 ) +err_exit ("munmap" , len, argv[4 ]);Small nit: you
munmap()to clean up themmap(), but you don't close fd or fd2.
Thanks Ross and Eryu reviewing!
I'll send v3 of this patch shortly.
--
Xiong
PATCH v3 1/1 xfs: test per-inode DAX flag by IO
In reply to Ross ZwislerIn a DAX mountpoint, do IO betwen files with and without DAX per-inode flag. We do mmap and O_DIRECT read/write IO in this case. Then test again in the same device without dax mountoption.
Add help _require_scratch_dax to make sure we can test DAX feature on SCRATCH_DEV.
Add mmap dio test programme to test read/write between a mmap area of one file and another file directly, with different size.
---
v3:
close fds in C test programme for clean up.
.gitignore | 1 + common/rc | 14 +++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 89 +++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138 | 110 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 7 files changed, 218 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out
.gitignore +1 -0
@@ -129,6 +129,7 @@ /src/cloner /src/renameat2 /src/t_rename_overwrite +/src/t_mmap_dio # dmapi/ binaries /dmapi/src/common/cmd/read_invis
common/rc +14 -0
@@ -2632,6 +2632,20 @@ _require_scratch_shutdown() _scratch_unmount } +# Does dax mount option work on this dev/fs? +_require_scratch_dax() +{ + _require_scratch + _scratch_mkfs > /dev/null 2>&1 + _scratch_mount -o dax + # Check options to be sure. XFS ignores dax option + # and goes on if dev underneath does not support dax. + _fs_options $SCRATCH_DEV | grep -w "dax" > /dev/null 2>&1 + [ $? -ne 0 ] && \ + _notrun "$SCRATCH_DEV $FSTYP does not support -o dax" + _scratch_unmount +} + # Does norecovery support by this fs? _require_norecovery() {
src/Makefile +1 -1
@@ -12,7 +12,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \ godown resvtest writemod makeextents itrash rename \ multi_open_unlink dmiperf unwritten_sync genhashnames t_holes \ t_mmap_writev t_truncate_cmtime dirhash_collide t_rename_overwrite \ - holetest t_truncate_self + holetest t_truncate_self t_mmap_dio LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \ preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
src/t_mmap_dio.c new file +89 -0
89 lines not on this page · raw message
tests/xfs/138 new file +110 -0
110 lines not on this page · raw message
tests/xfs/138.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 138 +Silence is golden
tests/xfs/group +1 -0
@@ -135,6 +135,7 @@ 135 auto logprint quick v2log 136 attr2 137 auto metadata v2log +138 auto attr quick 139 auto quick clone 140 auto clone 141 auto log metadata
Signature · 1 line
1.8.3.1
v4 cover mmap dio and DAX
common/rc : requires SCRATCH_DEV support DAXsrc/t_mmap_dio.c : intro mmap and O_DIRECT rw through files tests/generic/405 : IO between DAX/non-DAX mountpoints
tests/xfs/138 : IO between DAX/non-DAX xfs files(per-inode flag)
v2 :
Merge helper function changes into the first patch; Rewrite _require_dax, check options for sure; Print msg in t_mmap_dio.c to show which test going wrong; Empty mount options and check after mount to ensure we
wont mount with wrong option;
Remove unnecessary leading underscore and _fail; Use xfs_io instead of dd;
Other minor fixes.
v3:
close fds in C test programme for clean up.
v4:
Test both buffered and O_DIRECT IO;
Fix arg numbers in C test programme;
Fix fs options check after mount.
Cc Jeff Moyer since this test is based on his code. (Sorry for the late cc!)
Test status:
Both cases not run on normal block device; Both cases PASS on ramdisk based pmem devices; DIO in both cases FAIL on brd based ramdisk with:
DIO in both cases FAIL on nvdimm devices with:
+write(Bad address) len 1024 dio dax to nondax +write(Bad address) len 4096 dio dax to nondax +write(Bad address) len 16777216 dio dax to nondax +write(Bad address) len 67108864 dio dax to nondax
I've reported this to nvdimm list.
https://lists.01.org/pipermail/linux-nvdimm/2017-January/008600.html
Xiong Zhou (2):
xfs: test per-inode DAX flag by IO
generic: test mmap io through DAX and non-DAX
.gitignore | 1 + common/rc | 13 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 105 ++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405 | 119 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + tests/xfs/138 | 116 ++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 10 files changed, 361 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out
Signature · 1 line
1.8.3.1
PATCH v4 1/2 xfs: test per-inode DAX flag by IO
In a DAX mountpoint, do IO betwen files with and without DAX per-inode flag. We do mmap, both O_DIRECT and buffered read/write IO in this case. Then test again in the same device without dax mountoption.
Add help _require_scratch_dax to make sure we can test DAX feature on SCRATCH_DEV.
Add mmap dio test programme to test read/write between a mmap area of one file and another file directly or buffered, with different size.
--- .gitignore | 1 + common/rc | 13 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 105 ++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138 | 116 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 7 files changed, 239 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out
.gitignore +1 -0
@@ -129,6 +129,7 @@ /src/cloner /src/renameat2 /src/t_rename_overwrite +/src/t_mmap_dio # dmapi/ binaries /dmapi/src/common/cmd/read_invis
common/rc +13 -0
@@ -2632,6 +2632,19 @@ _require_scratch_shutdown() _scratch_unmount } +# Does dax mount option work on this dev/fs? +_require_scratch_dax() +{ + _require_scratch + _scratch_mkfs > /dev/null 2>&1 + _scratch_mount -o dax + # Check options to be sure. XFS ignores dax option + # and goes on if dev underneath does not support dax. + _fs_options $SCRATCH_DEV | grep -qw "dax" || \ + _notrun "$SCRATCH_DEV $FSTYP does not support -o dax" + _scratch_unmount +} + # Does norecovery support by this fs? _require_norecovery() {
src/Makefile +1 -1
@@ -12,7 +12,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \ godown resvtest writemod makeextents itrash rename \ multi_open_unlink dmiperf unwritten_sync genhashnames t_holes \ t_mmap_writev t_truncate_cmtime dirhash_collide t_rename_overwrite \ - holetest t_truncate_self + holetest t_truncate_self t_mmap_dio LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \ preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
src/t_mmap_dio.c new file +105 -0
105 lines not on this page · raw message
tests/xfs/138 new file +116 -0
116 lines not on this page · raw message
tests/xfs/138.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 138 +Silence is golden
tests/xfs/group +1 -0
@@ -135,6 +135,7 @@ 135 auto logprint quick v2log 136 attr2 137 auto metadata v2log +138 auto attr quick 139 auto quick clone 140 auto clone 141 auto log metadata
Signature · 1 line
1.8.3.1
PATCH 1/1 fixup! xfs: test per-inode DAX flag by IO
This makes a few changes to allow xfs/138 to use PMD faults, and has a few other changes which I think are improvements.
For PMD faults:
- We need to use MKFS_OPTIONS="-d su=2m,sw=1" to ensure that the block allocations we get from XFS are 2MiB sized and aligned. Without this we will fall back to 4k DAX pages for all the tests.
- We need to preallocate the file size with fallocate, rather than using "xfs_io -c pwrite". This lets XFS increase the size of the test files before the test starts, giving us 2MiB sized and aligned blocks. This also has the added bonus that we don't start out the mmap portion of the test with all the areas of the files written to and allocated. This means that on the first read pass through a portion of a file we exercise the zero page DAX path for holes, then on subsequent passes we'll fill the holes and exercise the normal DAX path, giving us better overall coverage of the DAX code. It also makes the test run slightly faster.
- For some reason I was having trouble getting XFS to give me 2MiB aligned and sized block allocations with file size 1GiB? The test only needs file sizes of 64 MiB, and dropping the files used to that size ensured that I always got 2MiB allocations, but truthfully I don't understand this behavior from XFS.
And one cleanup thing:
- Changed the sizes above 4k to use bash math instead of having hard coded values. I personally don't have 67108864 memorized to be 64 MiB (maybe everyone else does), but the meaning of $((64 * 1024 * 1024)) is more obvious to the reader.
I'm working on similar changes for your generic/405, but am still having trouble getting PMD faults from ext4, even with the mkfs options I had previously been using for other testing.
--- tests/xfs/138 | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-)
tests/xfs/138 +6 -7
@@ -79,22 +79,21 @@ do_tests() # page size t_dax_flag_mmap_dio `src/feature -s` # bigger sizes, for PMD faults - t_dax_flag_mmap_dio 16777216 - t_dax_flag_mmap_dio 67108864 + t_dax_flag_mmap_dio $((16 * 1024 * 1024)) + t_dax_flag_mmap_dio $((64 * 1024 * 1024)) } +export MKFS_OPTIONS="-d su=2m,sw=1" _scratch_mkfs > /dev/null 2>&1 # mount with dax option _scratch_mount "-o dax" psize=`src/feature -s` -tsize=$((1024 * 1024 * 1024)) +tsize=$((64 * 1024 * 1024)) -$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ - $SCRATCH_MNT/tf_s >> $seqres.full 2>&1 -$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ - $SCRATCH_MNT/tf_d >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "falloc 0 $tsize" $SCRATCH_MNT/tf_s >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "falloc 0 $tsize" $SCRATCH_MNT/tf_d >> $seqres.full 2>&1 do_tests _scratch_unmount
Signature · 1 line
2.7.4
v5 cover mmap dio and DAX
Hi Ross,
Thanks for the fixup!
Here are v5 for testing and reviewing. And I need your permission to add you in the signed-off list :)
---
common/rc : requires SCRATCH_DEV support DAXsrc/t_mmap_dio.c : intro mmap and O_DIRECT rw through files tests/generic/405 : IO between DAX/non-DAX mountpoints
tests/xfs/138 : IO between DAX/non-DAX xfs files(per-inode flag)
v2 :
Merge helper function changes into the first patch; Rewrite _require_dax, check options for sure; Print msg in t_mmap_dio.c to show which test going wrong; Empty mount options and check after mount to ensure we
wont mount with wrong option;
Remove unnecessary leading underscore and _fail; Use xfs_io instead of dd;
Other minor fixes.
v3:
close fds in C test programme for clean up.
v4:
Test both buffered and O_DIRECT IO;
Fix arg numbers in C test programme;
Fix fs options check after mount.
Cc Jeff Moyer since this test is based on his code. (Sorry for the late cc!)
v5:
Mainly from Ross Zwisler <ross.zwisler-VuQAYsv1563Yd54FQh9/[email protected]> Add mkfs options to make fs 2M aligned, to hit PMD faults;
Use falloc instead of pwrite to init test file aviding 4k
alignment;
64M test file instead of 1G;
Bumping test seq numners;
Other minor fixes.
Test status:
Both cases not run on normal block device;
Both cases PASS on ramdisk based pmem devices;
Both cases PASS on nvdimm(memory namespace) devices; DIO in both cases FAIL on brd based ramdisk with:
DIO in both cases FAIL on nvdimm(raw namespace) devices with:
+write(Bad address) len 1024 dio dax to nondax +write(Bad address) len 4096 dio dax to nondax +write(Bad address) len 16777216 dio dax to nondax +write(Bad address) len 67108864 dio dax to nondax
(expected behavior)
https://lists.01.org/pipermail/linux-nvdimm/2017-January/008600.html
Xiong Zhou (2):
xfs: test per-inode DAX flag by IO
generic: test mmap io through DAX and non-DAX
.gitignore | 1 + common/rc | 13 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 107 ++++++++++++++++++++++++++++++++++++++++++ tests/generic/407 | 125 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/407.out | 2 + tests/generic/group | 1 + tests/xfs/196 | 124 +++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/196.out | 2 + tests/xfs/group | 1 + 10 files changed, 377 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/407 create mode 100644 tests/generic/407.out create mode 100755 tests/xfs/196 create mode 100644 tests/xfs/196.out
Signature · 1 line
1.8.3.1
PATCH v5 1/2 xfs: test per-inode DAX flag by IO
In a DAX mountpoint, do IO betwen files with and without DAX per-inode flag. We do mmap, both O_DIRECT and buffered read/write IO in this case. Then test again in the same device without dax mountoption.
Add help _require_scratch_dax to make sure we can test DAX feature on SCRATCH_DEV.
Add mmap dio test programme to test read/write between a mmap area of one file and another file directly or buffered, with different size.
--- .gitignore | 1 + common/rc | 13 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 107 ++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/196 | 124 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/196.out | 2 + tests/xfs/group | 1 + 7 files changed, 249 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/xfs/196 create mode 100644 tests/xfs/196.out
.gitignore +1 -0
@@ -129,6 +129,7 @@ /src/cloner /src/renameat2 /src/t_rename_overwrite +/src/t_mmap_dio # dmapi/ binaries /dmapi/src/common/cmd/read_invis
common/rc +13 -0
@@ -2633,6 +2633,19 @@ _require_scratch_shutdown() _scratch_unmount } +# Does dax mount option work on this dev/fs? +_require_scratch_dax() +{ + _require_scratch + _scratch_mkfs > /dev/null 2>&1 + _scratch_mount -o dax + # Check options to be sure. XFS ignores dax option + # and goes on if dev underneath does not support dax. + _fs_options $SCRATCH_DEV | grep -qw "dax" || \ + _notrun "$SCRATCH_DEV $FSTYP does not support -o dax" + _scratch_unmount +} + # Does norecovery support by this fs? _require_norecovery() {
src/Makefile +1 -1
@@ -12,7 +12,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \ godown resvtest writemod makeextents itrash rename \ multi_open_unlink dmiperf unwritten_sync genhashnames t_holes \ t_mmap_writev t_truncate_cmtime dirhash_collide t_rename_overwrite \ - holetest t_truncate_self + holetest t_truncate_self t_mmap_dio LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \ preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
src/t_mmap_dio.c new file +107 -0
107 lines not on this page · raw message
tests/xfs/196 new file +124 -0
124 lines not on this page · raw message
tests/xfs/196.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 196 +Silence is golden
tests/xfs/group +1 -0
@@ -193,6 +193,7 @@ 193 auto quick clone 194 rw auto 195 ioctl dump auto quick +196 auto attr quick 197 dir auto quick 198 auto quick clone 199 mount auto quick
Signature · 1 line
1.8.3.1
Re: [PATCH v5 1/2] xfs: test per-inode DAX flag by IO
Quoted 6 lines · Xiong Zhou
In a DAX mountpoint, do IO betwen files with and without DAX per-inode flag. We do mmap, both
O_DIRECTand buffered read/write IO in this case. Then test again in the same device without dax mountoption.
Quoted 8 lines
Add help
_require_scratch_daxto make sure we can test DAX feature onSCRATCH_DEV.Add mmap dio test programme to test read/write between a mmap area of one file and another file directly or buffered, with different size.
Yep, this fine. :)
Quoted 1 line
Signed-off-by Xiong Zhou <[email protected]>
These changes all look good.
PATCH v5 2/2 generic: test mmap io through DAX and non-DAX
In reply to Xiong ZhouMount TEST_DEV as non-DAX, SCRATCH_DEV as DAX, then do some IO between them. In this case we use mmap and dio/buffered IO read/write test programme.
--- tests/generic/407 | 125 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/407.out | 2 + tests/generic/group | 1 + 3 files changed, 128 insertions(+) create mode 100755 tests/generic/407 create mode 100644 tests/generic/407.out
tests/generic/407 new file +125 -0
@@ -0,0 +1,125 @@ +#! /bin/bash +# FS QA Test 407 +# +# mmap direct/buffered io between DAX and non-DAX mountpoints. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017 Red Hat 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 +#----------------------------------------------------------------------- +# + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +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 + +_cleanup() +{ + cd / + rm -f $tmp.* +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# remove previous $seqres.full before test +rm -f $seqres.full + +_supported_fs generic +_supported_os Linux +_require_test +_require_scratch_dax +_require_test_program "feature" +_require_test_program "t_mmap_dio" +_require_xfs_io_command "falloc" + +# $1 mmap read/write size +t_mmap_dio_dax() +{ + # both dax + # with O_DIRECT first + src/t_mmap_dio $SCRATCH_MNT/tf_{s,d} $1 "dio both dax" + # again with buffered IO + src/t_mmap_dio -b $SCRATCH_MNT/tf_{s,d} \ + $1 "buffered both dax" + + # from non dax to dax + src/t_mmap_dio $TEST_DIR/tf_s \ + $SCRATCH_MNT/tf_d $1 "dio nondax to dax" + src/t_mmap_dio -b $TEST_DIR/tf_s \ + $SCRATCH_MNT/tf_d $1 "buffered nondax to dax" + + # from dax to non dax + src/t_mmap_dio $SCRATCH_MNT/tf_s \ + $TEST_DIR/tf_d $1 "dio dax to nondax" + src/t_mmap_dio -b $SCRATCH_MNT/tf_s \ + $TEST_DIR/tf_d $1 "buffered dax to nondax" + + # both non dax + src/t_mmap_dio $TEST_DIR/tf_{s,d} $1 "dio both nondax" + src/t_mmap_dio -b $TEST_DIR/tf_{s,d} \ + $1 "buffered both nondax" +} + +do_tests() +{ + # less than page size + t_mmap_dio_dax 1024 + # page size + t_mmap_dio_dax `src/feature -s` + # bigger sizes, for PMD faults + t_mmap_dio_dax $((16 * 1024 * 1024)) + t_mmap_dio_dax $((64 * 1024 * 1024)) +} + +# make fs 2Mb aligned for PMD fault testing +if [[ $FSTYP =~ ext ]] ; then + _scratch_mkfs "-E stride=512,stripe_width=1" \ + > /dev/null 2>&1 +elif [[ $FSTYP =~ xfs ]] ; then + _scratch_mkfs "-d su=2m,sw=1" > /dev/null 2>&1 +fi + +# mount SCRATCH_DEV with dax option, TEST_DEV not +export MOUNT_OPTIONS="" +export TEST_FS_MOUNT_OPTS="" +_test_cycle_mount +_fs_options $TEST_DEV | grep -qw "dax" && \ + _notrun "we need $TEST_DEV mount without dax" +_scratch_mount "-o dax" + +tsize=$((64 * 1024 * 1024)) + +$XFS_IO_PROG -f -c "falloc 0 $tsize" \ + $SCRATCH_MNT/tf_s >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "falloc 0 $tsize" \ + $SCRATCH_MNT/tf_d >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "falloc 0 $tsize" \ + $TEST_DIR/tf_s >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "falloc 0 $tsize" \ + $TEST_DIR/tf_d >> $seqres.full 2>&1 + +do_tests + +# success, all done +echo "Silence is golden" +status=0 +exit
tests/generic/407.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 407 +Silence is golden
tests/generic/group +1 -0
@@ -409,3 +409,4 @@ 404 auto quick insert 405 auto mkfs 406 auto quick dangerous +407 auto quick
Signature · 1 line
1.8.3.1
Re: [PATCH v5 2/2] generic: test mmap io through DAX and non-DAX
Quoted 5 lines · Xiong Zhou
Mount
TEST_DEVas non-DAX,SCRATCH_DEVas DAX, then do some IO between them. In this case we use mmap and dio/buffered IO read/write test programme.Signed-off-by Ross Zwisler <[email protected]>
Yep, this is fine.
Quoted 1 line
Signed-off-by Xiong Zhou <[email protected]>
<>
tests/generic/407
Quoted 1 line
tests/generic/407 +0
-0
<>
Quoted 18 lines
tests/generic/407 +18
-0+do_tests() +{ + # less than page size + t_mmap_dio_dax 1024 + # page size + t_mmap_dio_dax `src/feature -s` + # bigger sizes, for PMD faults + t_mmap_dio_dax $((16 * 1024 * 1024)) + t_mmap_dio_dax $((64 * 1024 * 1024)) +} + +# make fs 2Mb aligned for PMD fault testing +if [[ $FSTYP =~ ext ]] ; then + _scratch_mkfs "-E stride=512,stripe_width=1" \ + > /dev/null 2>&1 +elif [[ $FSTYP =~ xfs ]] ; then + _scratch_mkfs "-d su=2m,sw=1" > /dev/null 2>&1 +fi
tests/generic/407 line 109
Quoted 10 lines
tests/generic/407 +10
-0+ +# mount SCRATCH_DEV with dax option, TEST_DEV not +export MOUNT_OPTIONS="" +export TEST_FS_MOUNT_OPTS="" +_test_cycle_mount +_fs_options $TEST_DEV | grep -qw "dax" && \ + _notrun "we need $TEST_DEV mount without dax" +_scratch_mount "-o dax" + +tsize=$((64 * 1024 * 1024))
Ugh - for some reason 64 MiB files for me don't get PMDs on ext4 because the ext4 block allocator gives us misaligned blocks. I'll follow up with them in a new thread on the ext4 list and I'll cc you, Xiong.
If we up this to 128 MiB, so
tsize=$((128 * 1024 * 1024))
I get PMDs for both ext4 and XFS. Can we make this one change before applying?
The rest looks good to me. Thanks for the tests!
v6 cover mmap dio and DAX
Hi Ross and all,
After adding some initialization of test files, I've got different test results that the write call succeeds while the following read call fails.
Please help test and review, Thanks!
-----
common/rc : requires SCRATCH_DEV support DAXsrc/t_mmap_dio.c : intro mmap and O_DIRECT rw through files tests/generic/405 : IO between DAX/non-DAX mountpoints
tests/xfs/138 : IO between DAX/non-DAX xfs files(per-inode flag)
v2 :
Merge helper function changes into the first patch; Rewrite _require_dax, check options for sure; Print msg in t_mmap_dio.c to show which test going wrong; Empty mount options and check after mount to ensure we
wont mount with wrong option;
Remove unnecessary leading underscore and _fail; Use xfs_io instead of dd;
Other minor fixes.
v3:
close fds in C test programme for clean up.
v4:
Test both buffered and O_DIRECT IO;
Fix arg numbers in C test programme;
Fix fs options check after mount.
Cc Jeff Moyer since this test is based on his code. (Sorry for the late cc!)
v5:
Mainly from Ross Zwisler <[email protected]> Add mkfs options to make fs 2M aligned, to hit PMD faults; Use falloc instead of pwrite to init test file aviding 4k
alignment;
64M test file instead of 1G;
Bumping test seq numners;
Other minor fixes.
v6:
Test file size increase to 128M;
Add fsync after write in t_mmap_dio.c;
Split sub tests in both cases, re-initialize test files
before every sub test. This change reveals the new read failure issue.
Test status:
Both cases not run on normal block device; Both cases not run on ext2/3 fs(no falloc); Both cases PASS on memory-mode nvdimm devices; DIO in both cases FAIL on brd based ramdisk with:
DIO in both cases FAIL on raw-mode nvdimm devices with:
+read(Bad address) len 1024 dio dax to nondax +read(Bad address) len 4096 dio dax to nondax +read(Bad address) len 16777216 dio dax to nondax +read(Bad address) len 67108864 dio dax to nondax
I've reported this as an issue:
https://lists.01.org/pipermail/linux-nvdimm/2017-February/008781.html
Xiong Zhou (2):
xfs: test per-inode DAX flag by IO
generic: test mmap io through DAX and non-DAX
.gitignore | 1 + common/rc | 13 +++++ src/Makefile | 2 +- src/t_mmap_dio.c | 111 +++++++++++++++++++++++++++++++++++ tests/generic/407 | 152 ++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/407.out | 2 + tests/generic/group | 1 + tests/xfs/196 | 158 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/196.out | 2 + tests/xfs/group | 1 + 10 files changed, 442 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/407 create mode 100644 tests/generic/407.out create mode 100755 tests/xfs/196 create mode 100644 tests/xfs/196.out
Signature · 1 line
1.8.3.1
PATCH v6 1/2 xfs: test per-inode DAX flag by IO
In a DAX mountpoint, do IO betwen files with and without DAX per-inode flag. We do mmap, both O_DIRECT and buffered read/write IO in this case. Then test again in the same device without dax mountoption.
Add help _require_scratch_dax to make sure we can test DAX feature on SCRATCH_DEV.
Add mmap dio test programme to test read/write between a mmap area of one file and another file directly or buffered, with different size.
--- .gitignore | 1 + common/rc | 13 +++++ src/Makefile | 2 +- src/t_mmap_dio.c | 111 ++++++++++++++++++++++++++++++++++++++ tests/xfs/196 | 158 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/196.out | 2 + tests/xfs/group | 1 + 7 files changed, 287 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/xfs/196 create mode 100644 tests/xfs/196.out
.gitignore +1 -0
@@ -129,6 +129,7 @@ /src/cloner /src/renameat2 /src/t_rename_overwrite +/src/t_mmap_dio # dmapi/ binaries /dmapi/src/common/cmd/read_invis
common/rc +13 -0
@@ -2633,6 +2633,19 @@ _require_scratch_shutdown() _scratch_unmount } +# Does dax mount option work on this dev/fs? +_require_scratch_dax() +{ + _require_scratch + _scratch_mkfs > /dev/null 2>&1 + _scratch_mount -o dax + # Check options to be sure. XFS ignores dax option + # and goes on if dev underneath does not support dax. + _fs_options $SCRATCH_DEV | grep -qw "dax" || \ + _notrun "$SCRATCH_DEV $FSTYP does not support -o dax" + _scratch_unmount +} + # Does norecovery support by this fs? _require_norecovery() {
src/Makefile +1 -1
@@ -12,7 +12,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \ godown resvtest writemod makeextents itrash rename \ multi_open_unlink dmiperf unwritten_sync genhashnames t_holes \ t_mmap_writev t_truncate_cmtime dirhash_collide t_rename_overwrite \ - holetest t_truncate_self + holetest t_truncate_self t_mmap_dio LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \ preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
src/t_mmap_dio.c new file +111 -0
111 lines not on this page · raw message
tests/xfs/196 new file +158 -0
158 lines not on this page · raw message
tests/xfs/196.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 196 +Silence is golden
tests/xfs/group +1 -0
@@ -193,6 +193,7 @@ 193 auto quick clone 194 rw auto 195 ioctl dump auto quick +196 auto attr quick 197 dir auto quick 198 auto quick clone 199 mount auto quick
Signature · 1 line
1.8.3.1
PATCH v6 2/2 generic: test mmap io through DAX and non-DAX
In reply to Xiong ZhouMount TEST_DEV as non-DAX, SCRATCH_DEV as DAX, then do some IO between them. In this case we use mmap and dio/buffered IO read/write test programme.
--- tests/generic/407 | 152 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/407.out | 2 + tests/generic/group | 1 + 3 files changed, 155 insertions(+) create mode 100755 tests/generic/407 create mode 100644 tests/generic/407.out
tests/generic/407 new file +152 -0
@@ -0,0 +1,152 @@ +#! /bin/bash +# FS QA Test 407 +# +# mmap direct/buffered io between DAX and non-DAX mountpoints. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017 Red Hat 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 +#----------------------------------------------------------------------- +# + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +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 + +_cleanup() +{ + cd / + rm -f $tmp.* +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# remove previous $seqres.full before test +rm -f $seqres.full + +_supported_fs generic +_supported_os Linux +_require_test +_require_scratch_dax +_require_test_program "feature" +_require_test_program "t_mmap_dio" +_require_xfs_io_command "falloc" + +prep_files() +{ + rm -f $SCRATCH_MNT/tf_{s,d} + rm -f $TEST_DIR/tf_{s,d} + + $XFS_IO_PROG -f -c "falloc 0 $tsize" \ + $SCRATCH_MNT/tf_{s,d} >> $seqres.full 2>&1 + $XFS_IO_PROG -f -c "falloc 0 $tsize" \ + $TEST_DIR/tf_{s,d} >> $seqres.full 2>&1 +} + +t_both_dax() +{ + prep_files + # with O_DIRECT first + src/t_mmap_dio $SCRATCH_MNT/tf_{s,d} $1 "dio both dax" + + prep_files + # again with buffered IO + src/t_mmap_dio -b $SCRATCH_MNT/tf_{s,d} \ + $1 "buffered both dax" +} + +t_nondax_to_dax() +{ + prep_files + src/t_mmap_dio $TEST_DIR/tf_s \ + $SCRATCH_MNT/tf_d $1 "dio nondax to dax" + + prep_files + src/t_mmap_dio -b $TEST_DIR/tf_s \ + $SCRATCH_MNT/tf_d $1 "buffered nondax to dax" +} + +t_dax_to_nondax() +{ + prep_files + src/t_mmap_dio $SCRATCH_MNT/tf_s \ + $TEST_DIR/tf_d $1 "dio dax to nondax" + + prep_files + src/t_mmap_dio -b $SCRATCH_MNT/tf_s \ + $TEST_DIR/tf_d $1 "buffered dax to nondax" +} + +t_both_nondax() +{ + prep_files + src/t_mmap_dio $TEST_DIR/tf_{s,d} $1 "dio both nondax" + + prep_files + src/t_mmap_dio -b $TEST_DIR/tf_{s,d} \ + $1 "buffered both nondax" +} + +# $1 mmap read/write size +t_mmap_dio_dax() +{ + t_both_dax $1 + t_dax_to_nondax $1 + t_nondax_to_dax $1 + t_both_nondax $1 +} + +do_tests() +{ + # less than page size + t_mmap_dio_dax 1024 + # page size + t_mmap_dio_dax `src/feature -s` + # bigger sizes, for PMD faults + t_mmap_dio_dax $((16 * 1024 * 1024)) + t_mmap_dio_dax $((64 * 1024 * 1024)) +} + +# make fs 2Mb aligned for PMD fault testing +if [[ $FSTYP =~ ext ]] ; then + _scratch_mkfs "-E stride=512,stripe_width=1" \ + > /dev/null 2>&1 +elif [[ $FSTYP =~ xfs ]] ; then + _scratch_mkfs "-d su=2m,sw=1" > /dev/null 2>&1 +fi + +# mount SCRATCH_DEV with dax option, TEST_DEV not +export MOUNT_OPTIONS="" +export TEST_FS_MOUNT_OPTS="" +_test_cycle_mount +_fs_options $TEST_DEV | grep -qw "dax" && \ + _notrun "we need $TEST_DEV mount without dax" +_scratch_mount "-o dax" + +tsize=$((128 * 1024 * 1024)) + +do_tests + +# success, all done +echo "Silence is golden" +status=0 +exit
tests/generic/407.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 407 +Silence is golden
tests/generic/group +1 -0
@@ -409,3 +409,4 @@ 404 auto quick insert 405 auto mkfs 406 auto quick dangerous +407 auto quick
Signature · 1 line
1.8.3.1
Re: [PATCH v6 2/2] generic: test mmap io through DAX and non-DAX
Quoted 24 lines · Xiong Zhou
Mount
TEST_DEVas non-DAX,SCRATCH_DEVas DAX, then do some IO between them. In this case we use mmap and dio/buffered IO read/write test programme.Signed-off-by Ross Zwisler <[email protected]>Signed-off-by Xiong Zhou <[email protected]>--- tests/generic/407 | 152 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/407.out | 2 + tests/generic/group | 1 + 3 files changed, 155 insertions(+) create mode 100755 tests/generic/407 create mode 100644 tests/generic/407.outtests/generic/407 new file +4
-0@@ -0,0 +1,152 @@ +#! /bin/bash +# FS QA Test 407 +# +# mmap direct/buffered io between DAX and non-DAX mountpoints.
[snip]
tests/generic/407 line 135
Quoted 8 lines
tests/generic/407 new file +8
-0+ +# make fs 2Mb aligned for PMD fault testing +if [[ $FSTYP =~ ext ]] ; then + _scratch_mkfs "-E stride=512,stripe_width=1" \ + > /dev/null 2>&1 +elif [[ $FSTYP =~ xfs ]] ; then + _scratch_mkfs "-d su=2m,sw=1" > /dev/null 2>&1 +fi
If FSTYP isn't extN nor xfs, then scratch device is not mkfs'ed, I think you're missing a default "else" here, which calls _scratch_mkfs without any options.
If you're OK with this update, I can fix it at commit time.
Thanks,
Eryu
Re: [PATCH v6 2/2] generic: test mmap io through DAX and non-DAX
Quoted 39 lines · Eryu Guan
Quoted 24 lines · Xiong Zhou
Mount
TEST_DEVas non-DAX,SCRATCH_DEVas DAX, then do some IO between them. In this case we use mmap and dio/buffered IO read/write test programme.Signed-off-by Xiong Zhou <[email protected]>--- tests/generic/407 | 152 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/407.out | 2 + tests/generic/group | 1 + 3 files changed, 155 insertions(+) create mode 100755 tests/generic/407 create mode 100644 tests/generic/407.outtests/generic/407 new file +4
-0@@ -0,0 +1,152 @@ +#! /bin/bash +# FS QA Test 407 +# +# mmap direct/buffered io between DAX and non-DAX mountpoints.[snip]
tests/generic/407 line 135
Quoted 8 lines
tests/generic/407 new file +8
-0+ +# make fs 2Mb aligned for PMD fault testing +if [[ $FSTYP =~ ext ]] ; then + _scratch_mkfs "-E stride=512,stripe_width=1" \ + > /dev/null 2>&1 +elif [[ $FSTYP =~ xfs ]] ; then + _scratch_mkfs "-d su=2m,sw=1" > /dev/null 2>&1 +fiIf FSTYP isn't extN nor xfs, then scratch device is not mkfs'ed, I think you're missing a default "else" here, which calls
_scratch_mkfswithout any options.
Quoted 1 line
If you're OK with this update, I can fix it at commit time.
Okay, Thank you!
Quoted 3 lines
Thanks,
Eryu
PATCH v4 2/2 generic: test mmap io through DAX and non-DAX
In reply to Xiong ZhouMount TEST_DEV as non-DAX, SCRATCH_DEV as DAX, then do some IO between them. In this case we use mmap and dio/buffered IO read/write test programme.
--- tests/generic/405 | 119 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + 3 files changed, 122 insertions(+) create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out
tests/generic/405 new file +119 -0
@@ -0,0 +1,119 @@ +#! /bin/bash +# FS QA Test 405 +# +# mmap direct/buffered io between DAX and non-DAX mountpoints. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017 Red Hat 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 +#----------------------------------------------------------------------- +# + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +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 + +_cleanup() +{ + cd / + rm -f $tmp.* +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# remove previous $seqres.full before test +rm -f $seqres.full + +_supported_fs generic +_supported_os Linux +_require_test +_require_scratch_dax +_require_test_program "feature" +_require_test_program "t_mmap_dio" + +# $1 mmap read/write size +t_mmap_dio_dax() +{ + # both dax + # with O_DIRECT first + src/t_mmap_dio $SCRATCH_MNT/tf_{s,d} $1 "dio both dax" + # again with buffered IO + src/t_mmap_dio -b $SCRATCH_MNT/tf_{s,d} \ + $1 "buffered both dax" + + # from non dax to dax + src/t_mmap_dio $TEST_DIR/tf_s \ + $SCRATCH_MNT/tf_d $1 "dio nondax to dax" + src/t_mmap_dio -b $TEST_DIR/tf_s \ + $SCRATCH_MNT/tf_d $1 "buffered nondax to dax" + + # from dax to non dax + src/t_mmap_dio $SCRATCH_MNT/tf_s \ + $TEST_DIR/tf_d $1 "dio dax to nondax" + src/t_mmap_dio -b $SCRATCH_MNT/tf_s \ + $TEST_DIR/tf_d $1 "buffered dax to nondax" + + # both non dax + src/t_mmap_dio $TEST_DIR/tf_{s,d} $1 "dio both nondax" + src/t_mmap_dio -b $TEST_DIR/tf_{s,d} \ + $1 "buffered both nondax" +} + +do_tests() +{ + # less than page size + t_mmap_dio_dax 1024 + # page size + t_mmap_dio_dax `src/feature -s` + # bigger sizes, for PMD faults + t_mmap_dio_dax 16777216 + t_mmap_dio_dax 67108864 +} + +_scratch_mkfs > /dev/null 2>&1 + +# mount SCRATCH_DEV with dax option, TEST_DEV not +export MOUNT_OPTIONS="" +export TEST_FS_MOUNT_OPTS="" +_test_cycle_mount +_fs_options $TEST_DEV | grep -qw "dax" && \ + _notrun "we need $TEST_DEV mount without dax" +_scratch_mount "-o dax" + +psize=`src/feature -s` +tsize=$((1024 * 1024 * 1024)) + +$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ + $SCRATCH_MNT/tf_s >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ + $SCRATCH_MNT/tf_d >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ + $TEST_DIR/tf_s >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ + $TEST_DIR/tf_d >> $seqres.full 2>&1 + +do_tests + +# success, all done +echo "Silence is golden" +status=0 +exit
tests/generic/405.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 405 +Silence is golden
tests/generic/group +1 -0
@@ -407,3 +407,4 @@ 402 auto quick rw 403 auto quick attr 404 auto quick insert +405 auto quick
Signature · 1 line
1.8.3.1
Quoted 42 lines · Xiong Zhou
common/rc : requires
SCRATCH_DEVsupport DAXsrc/t_mmap_dio.c: intro mmap andO_DIRECTrw through files tests/generic/405 : IO between DAX/non-DAX mountpoints
tests/xfs/138 : IO between DAX/non-DAX xfs files(per-inode flag)v2 :
Merge helper function changes into the first patch; Rewrite_require_dax, check options for sure; Print msg int_mmap_dio.cto show which test going wrong; Empty mount options and check after mount to ensure we
wont mount with wrong option;
Remove unnecessary leading underscore and_fail; Usexfs_ioinstead of dd;
Other minor fixes.v3:
close fds in C test programme for clean up.v4:
Test both buffered andO_DIRECTIO;
Fix arg numbers in C test programme;
Fix fs options check after mount.
Cc Jeff Moyer since this test is based on his code. (Sorry for the late cc!)Test status:
Both cases not run on normal block device; Both cases PASS on ramdisk based pmem devices; DIO in both cases FAIL on brd based ramdisk with:
DIO in both cases FAIL on nvdimm devices with:
+write(Bad address) len 1024 dio dax to nondax +write(Bad address) len 4096 dio dax to nondax +write(Bad address) len 16777216 dio dax to nondax +write(Bad address) len 67108864 dio dax to nondaxI've reported this to nvdimm list.
https://lists.01.org/pipermail/linux-nvdimm/2017-January/008600.htmlXiong Zhou (2):
xfs: test per-inode DAX flag by IO
generic: test mmap io through DAX and non-DAX
Quoted 19 lines
.gitignore | 1 + common/rc | 13 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 105 ++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405 | 119 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + tests/xfs/138 | 116 ++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 10 files changed, 361 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.outSignature · 1 line
1.8.3.1
I just wanted to let you know that I'm testing with these new xfstests right now, and so far I've been unable to successfully get any PMD faults. I'm looking into why that is right now, and should hopefully have some changes so we can do both PTE and PMD testing with this set.
Also, it looks like the test number "generic/405" was already used in xfstests/master by this commit:
66768bc generic/405: test mkfs against thin provision device
So this may need to be generic/406. :)
Quoted 60 lines · Ross Zwisler
Quoted 59 lines · Xiong Zhou
common/rc : requires
SCRATCH_DEVsupport DAXsrc/t_mmap_dio.c: intro mmap andO_DIRECTrw through files tests/generic/405 : IO between DAX/non-DAX mountpoints
tests/xfs/138 : IO between DAX/non-DAX xfs files(per-inode flag)v2 :
Merge helper function changes into the first patch; Rewrite_require_dax, check options for sure; Print msg int_mmap_dio.cto show which test going wrong; Empty mount options and check after mount to ensure we
wont mount with wrong option;
Remove unnecessary leading underscore and_fail; Usexfs_ioinstead of dd;
Other minor fixes.v3:
close fds in C test programme for clean up.v4:
Test both buffered andO_DIRECTIO;
Fix arg numbers in C test programme;
Fix fs options check after mount.
Cc Jeff Moyer since this test is based on his code. (Sorry for the late cc!)Test status:
Both cases not run on normal block device; Both cases PASS on ramdisk based pmem devices; DIO in both cases FAIL on brd based ramdisk with:
DIO in both cases FAIL on nvdimm devices with:
+write(Bad address) len 1024 dio dax to nondax +write(Bad address) len 4096 dio dax to nondax +write(Bad address) len 16777216 dio dax to nondax +write(Bad address) len 67108864 dio dax to nondaxI've reported this to nvdimm list.
https://lists.01.org/pipermail/linux-nvdimm/2017-January/008600.htmlXiong Zhou (2):
xfs: test per-inode DAX flag by IO
generic: test mmap io through DAX and non-DAX.gitignore | 1 + common/rc | 13 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 105 ++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405 | 119 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + tests/xfs/138 | 116 ++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 10 files changed, 361 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.out
Quoted 7 lines
Quoted 2 lines
Signature · 1 line
1.8.3.1I just wanted to let you know that I'm testing with these new xfstests right now, and so far I've been unable to successfully get any PMD faults. I'm looking into why that is right now, and should hopefully have some changes so we can do both PTE and PMD testing with this set.
Thank you very much for looking into this!
Adding a printk msg in dax_iomap_pmd_fault in fs/dax.c shows that these 2 cases called this function, so do __radix_tree_insert in lib/radix-tree.c with order > 0. I must have missed something..
Quoted 7 lines
Also, it looks like the test number "generic/405" was already used in xfstests/master by this commit:
66768bc generic/405: test mkfs against thin provision device
So this may need to be generic/406. :)
Ya, i can update this or Eryu can handle it while applying.
Thanks,
Xiong
Quoted 12 lines · Xiong Zhou
Quoted 4 lines
I just wanted to let you know that I'm testing with these new xfstests right now, and so far I've been unable to successfully get any PMD faults. I'm looking into why that is right now, and should hopefully have some changes so we can do both PTE and PMD testing with this set.
Thank you very much for looking into this!
Adding a printk msg in
dax_iomap_pmd_faultinfs/dax.cshows that these 2 cases called this function, so do__radix_tree_insertinlib/radix-tree.cwith order > 0. I must have missed something..Quoted 1 line
Quoted 8 lines
Quoted 6 lines
Also, it looks like the test number "generic/405" was already used in xfstests/master by this commit:
66768bc generic/405: test mkfs against thin provision device
So this may need to be generic/406. :)
Ya, i can update this or Eryu can handle it while applying.
I'll do renumber after patch being reviewed, no need to update patch only to change the seq number. (Better to use latest tree when writing new case though.)
Thanks,
Eryu
Quoted 69 lines · Xiong Zhou
Quoted 60 lines · Ross Zwisler
Quoted 59 lines · Xiong Zhou
common/rc : requires
SCRATCH_DEVsupport DAXsrc/t_mmap_dio.c: intro mmap andO_DIRECTrw through files tests/generic/405 : IO between DAX/non-DAX mountpoints
tests/xfs/138 : IO between DAX/non-DAX xfs files(per-inode flag)v2 :
Merge helper function changes into the first patch; Rewrite_require_dax, check options for sure; Print msg int_mmap_dio.cto show which test going wrong; Empty mount options and check after mount to ensure we
wont mount with wrong option;
Remove unnecessary leading underscore and_fail; Usexfs_ioinstead of dd;
Other minor fixes.v3:
close fds in C test programme for clean up.v4:
Test both buffered andO_DIRECTIO;
Fix arg numbers in C test programme;
Fix fs options check after mount.
Cc Jeff Moyer since this test is based on his code. (Sorry for the late cc!)Test status:
Both cases not run on normal block device; Both cases PASS on ramdisk based pmem devices; DIO in both cases FAIL on brd based ramdisk with:
DIO in both cases FAIL on nvdimm devices with:
+write(Bad address) len 1024 dio dax to nondax +write(Bad address) len 4096 dio dax to nondax +write(Bad address) len 16777216 dio dax to nondax +write(Bad address) len 67108864 dio dax to nondaxI've reported this to nvdimm list.
https://lists.01.org/pipermail/linux-nvdimm/2017-January/008600.htmlXiong Zhou (2):
xfs: test per-inode DAX flag by IO
generic: test mmap io through DAX and non-DAX.gitignore | 1 + common/rc | 13 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 105 ++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405 | 119 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + tests/xfs/138 | 116 ++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 10 files changed, 361 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.outQuoted 7 lines
Quoted 2 lines
Signature · 1 line
1.8.3.1I just wanted to let you know that I'm testing with these new xfstests right now, and so far I've been unable to successfully get any PMD faults. I'm looking into why that is right now, and should hopefully have some changes so we can do both PTE and PMD testing with this set.
Quoted 5 lines
Thank you very much for looking into this!
Adding a printk msg in
dax_iomap_pmd_faultinfs/dax.cshows that these 2 cases called this function, so do__radix_tree_insertinlib/radix-tree.cwith order > 0. I must have missed something..
Ah, yea, the flow is a little confusing. When we first try to do a PMD fault we insert a 2MiB "empty" entry into the radix tree that basically just allows us to lock an entire 2MiB range. This happens in dax_iomap_pmd_fault() via grab_mapping_entry(). This is most likely what you're seeing with your debug.
Then, with that empty 2MiB entry in place we try and actually service the fault and insert a real mapping to a 2MiB huge page. There are still cases when this can fall back to 4k pages, and one of them is if the block allocation we are given by the filesystem isn't 2MiB aligned. That is the alignment check against PG_PMD_COLOUR in dax_pmd_insert_mapping(), and that's what we were hitting. The way to get around this is to tell XFS that we would like 2MiB aligned and sized block allocations via the following mkfs options:
export MKFS_OPTIONS="-d su=2m,sw=1"
We also need to fallocate our storage space so that we get 2 MiB allocations instead of 4k allocations.
I've been working on patches that do all of this - I'll try and send them out today.
This has taken a little longer than I would have liked because when debugging this issue I found an issue with DAX + DIO in the kernel. So, your test has already found an important bug in the kernel before it was even committed to xfstests! :)
BTW, if we fallocate our files, is there additional value in writing data into the files before we start testing as you do via these lines?
$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \
$SCRATCH_MNT/tf_s >> $seqres.full 2>&1
$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \
$SCRATCH_MNT/tf_d >> $seqres.full 2>&1This puts a known pattern into the files and means that reads are handled from media instead of from hole pages, but we never verify the data pattern and it slows down the test quite a bit. What do you think?
Quoted 89 lines · Ross Zwisler
Quoted 69 lines · Xiong Zhou
Quoted 60 lines · Ross Zwisler
Quoted 59 lines · Xiong Zhou
common/rc : requires
SCRATCH_DEVsupport DAXsrc/t_mmap_dio.c: intro mmap andO_DIRECTrw through files tests/generic/405 : IO between DAX/non-DAX mountpoints
tests/xfs/138 : IO between DAX/non-DAX xfs files(per-inode flag)v2 :
Merge helper function changes into the first patch; Rewrite_require_dax, check options for sure; Print msg int_mmap_dio.cto show which test going wrong; Empty mount options and check after mount to ensure we
wont mount with wrong option;
Remove unnecessary leading underscore and_fail; Usexfs_ioinstead of dd;
Other minor fixes.v3:
close fds in C test programme for clean up.v4:
Test both buffered andO_DIRECTIO;
Fix arg numbers in C test programme;
Fix fs options check after mount.
Cc Jeff Moyer since this test is based on his code. (Sorry for the late cc!)Test status:
Both cases not run on normal block device; Both cases PASS on ramdisk based pmem devices; DIO in both cases FAIL on brd based ramdisk with:
DIO in both cases FAIL on nvdimm devices with:
+write(Bad address) len 1024 dio dax to nondax +write(Bad address) len 4096 dio dax to nondax +write(Bad address) len 16777216 dio dax to nondax +write(Bad address) len 67108864 dio dax to nondaxI've reported this to nvdimm list.
https://lists.01.org/pipermail/linux-nvdimm/2017-January/008600.htmlXiong Zhou (2):
xfs: test per-inode DAX flag by IO
generic: test mmap io through DAX and non-DAX.gitignore | 1 + common/rc | 13 ++++++ src/Makefile | 2 +- src/t_mmap_dio.c | 105 ++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405 | 119 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + tests/xfs/138 | 116 ++++++++++++++++++++++++++++++++++++++++++++++++ tests/xfs/138.out | 2 + tests/xfs/group | 1 + 10 files changed, 361 insertions(+), 1 deletion(-) create mode 100644 src/t_mmap_dio.c create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out create mode 100755 tests/xfs/138 create mode 100644 tests/xfs/138.outQuoted 7 lines
Quoted 2 lines
Signature · 1 line
1.8.3.1I just wanted to let you know that I'm testing with these new xfstests right now, and so far I've been unable to successfully get any PMD faults. I'm looking into why that is right now, and should hopefully have some changes so we can do both PTE and PMD testing with this set.
Quoted 5 lines
Thank you very much for looking into this!
Adding a printk msg in
dax_iomap_pmd_faultinfs/dax.cshows that these 2 cases called this function, so do__radix_tree_insertinlib/radix-tree.cwith order > 0. I must have missed something..Ah, yea, the flow is a little confusing. When we first try to do a PMD fault we insert a 2MiB "empty" entry into the radix tree that basically just allows us to lock an entire 2MiB range. This happens in
dax_iomap_pmd_fault()viagrab_mapping_entry(). This is most likely what you're seeing with your debug.Then, with that empty 2MiB entry in place we try and actually service the fault and insert a real mapping to a 2MiB huge page. There are still cases when this can fall back to 4k pages, and one of them is if the block allocation we are given by the filesystem isn't 2MiB aligned. That is the alignment check against
PG_PMD_COLOURindax_pmd_insert_mapping(), and that's what we were hitting. The way to get around this is to tell XFS that we would like 2MiB aligned and sized block allocations via the following mkfs options:
Quoted 4 lines
export MKFS_OPTIONS="-d su=2m,sw=1"
We also need to fallocate our storage space so that we get 2 MiB allocations instead of 4k allocations.
Aha, I forgot to checking return status of fault handler. Thanks very much for the detailed explanation and instructions. :)
Quoted 8 lines
I've been working on patches that do all of this - I'll try and send them out today.
This has taken a little longer than I would have liked because when debugging this issue I found an issue with DAX + DIO in the kernel. So, your test has already found an important bug in the kernel before it was even committed to xfstests! :)
Good to know. :)
Quoted 4 lines
BTW, if we fallocate our files, is there additional value in writing data into the files before we start testing as you do via these lines?
Quoted 8 lines
$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ $SCRATCH_MNT/tf_s >> $seqres.full 2>&1 $XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ $SCRATCH_MNT/tf_d >> $seqres.full 2>&1This puts a known pattern into the files and means that reads are handled from media instead of from hole pages, but we never verify the data pattern and it slows down the test quite a bit. What do you think?
falloc is better for this job. I'll send next version after more tests.
Thanks for reviewing!
--
Xiong
PATCH v2 2/2 generic: test mmap dio through DAX and non-DAX
In reply to Xiong ZhouMount TEST_DEV as non-DAX, SCRATCH_DEV as DAX, then do some IO between them. In this case we use mmap and dio read/write test programme.
--- tests/generic/405 | 116 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/405.out | 2 + tests/generic/group | 1 + 3 files changed, 119 insertions(+) create mode 100755 tests/generic/405 create mode 100644 tests/generic/405.out
tests/generic/405 new file +116 -0
@@ -0,0 +1,116 @@ +#! /bin/bash +# FS QA Test 405 +# +# mmap dio between DAX and non-DAX mountpoints. +# +#----------------------------------------------------------------------- +# Copyright (c) 2017 Red Hat 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 +#----------------------------------------------------------------------- +# + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +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 + +_cleanup() +{ + cd / + rm -f $tmp.* +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# remove previous $seqres.full before test +rm -f $seqres.full + +_supported_fs generic +_supported_os Linux +_require_test +_require_scratch_dax +_require_test_program "feature" +_require_test_program "t_mmap_dio" + +# $1 mmap read/write size +t_mmap_dio_dax() +{ + # both dax + src/t_mmap_dio $SCRATCH_MNT/tf_{s,d} $1 "both dax" + + # from non dax to dax + src/t_mmap_dio $TEST_DIR/tf_s \ + $SCRATCH_MNT/tf_d $1 "nondax to dax" + + # from dax to non dax + src/t_mmap_dio $SCRATCH_MNT/tf_s \ + $TEST_DIR/tf_d $1 "dax to nondax" + + # both non dax + src/t_mmap_dio $TEST_DIR/tf_{s,d} $1 "both nondax" +} + +do_tests() +{ + # less than page size + t_mmap_dio_dax 1024 + # page size + t_mmap_dio_dax `src/feature -s` + # bigger sizes, for PMD faults + t_mmap_dio_dax 16777216 + t_mmap_dio_dax 67108864 +} + +_scratch_mkfs > /dev/null 2>&1 + +# mount SCRATCH_DEV with dax option, TEST_DEV not +export MOUNT_OPTIONS="" +export TEST_FS_MOUNT_OPTS="" +_test_unmount +_test_mount +_fs_options $TEST_DEV | grep -w "dax" > /dev/null 2>&1 +if [ $? -eq 0 ] ; then + _notrun "we need $TEST_DEV mount without dax" +fi +_scratch_mount "-o dax" +_fs_options $SCRATCH_DEV | grep -w "dax" > /dev/null 2>&1 +if [ $? -ne 0 ] ; then + _notrun "we need $SCRATCH_DEV mount with dax" +fi + +psize=`src/feature -s` +tsize=$((1024 * 1024 * 1024)) + +$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ + $SCRATCH_MNT/tf_s >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ + $SCRATCH_MNT/tf_d >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ + $TEST_DIR/tf_s >> $seqres.full 2>&1 +$XFS_IO_PROG -f -c "pwrite -W -b $psize 0 $tsize" \ + $TEST_DIR/tf_d >> $seqres.full 2>&1 + +do_tests + +# success, all done +echo "Silence is golden" +status=0 +exit
tests/generic/405.out new file +2 -0
@@ -0,0 +1,2 @@ +QA output created by 405 +Silence is golden
tests/generic/group +1 -0
@@ -407,3 +407,4 @@ 402 auto quick rw 403 auto quick attr 404 auto quick insert +405 auto quick
Signature · 1 line
1.8.3.1