| Message ID | 1458029534-17578-1-git-send-email-eguan@redhat.com |
|---|---|
| State | Not Applicable |
| Headers | show |
On Tue, Mar 15, 2016 at 04:12:14PM +0800, Eryu Guan wrote: > This is a test that performs simple I/O on dm error device, which > returns EIO on all I/O request. > > This is motivated by an ext4 bug that crashes kernel on error path when > trying to update atime. Following kernel patch should fix the issue > > ext4: fix NULL pointer dereference in ext4_mark_inode_dirty() > > Signed-off-by: Eryu Guan <eguan@redhat.com> > --- Fails with: @@ -1,2 +1,6 @@ QA output created by 338 Silence is golden +specified blocksize 1024 is less than device physical sector size 4096 +switching to logical sector size 512 +mkfs.xfs: /dev/mapper/error-test appears to contain an existing filesystem (xfs). +mkfs.xfs: Use the -f option to force overwrite. And then it failed to clean up properly and caused all sorts of subsequent problems. -Dave.
On Wed, Mar 23, 2016 at 01:53:27PM +1100, Dave Chinner wrote: > On Tue, Mar 15, 2016 at 04:12:14PM +0800, Eryu Guan wrote: > > This is a test that performs simple I/O on dm error device, which > > returns EIO on all I/O request. > > > > This is motivated by an ext4 bug that crashes kernel on error path when > > trying to update atime. Following kernel patch should fix the issue > > > > ext4: fix NULL pointer dereference in ext4_mark_inode_dirty() > > > > Signed-off-by: Eryu Guan <eguan@redhat.com> > > --- > > Fails with: > > @@ -1,2 +1,6 @@ > QA output created by 338 > Silence is golden > +specified blocksize 1024 is less than device physical sector size 4096 > +switching to logical sector size 512 > +mkfs.xfs: /dev/mapper/error-test appears to contain an existing filesystem (xfs). > +mkfs.xfs: Use the -f option to force overwrite. > > And then it failed to clean up properly and caused all sorts of > subsequent problems. Test passed for me, seems it has something to do with the "physical sector size 4096" device. I'll look into it. Thanks for the review! Eryu -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Mar 23, 2016 at 11:25:31AM +0800, Eryu Guan wrote: > On Wed, Mar 23, 2016 at 01:53:27PM +1100, Dave Chinner wrote: > > On Tue, Mar 15, 2016 at 04:12:14PM +0800, Eryu Guan wrote: > > > This is a test that performs simple I/O on dm error device, which > > > returns EIO on all I/O request. > > > > > > This is motivated by an ext4 bug that crashes kernel on error path when > > > trying to update atime. Following kernel patch should fix the issue > > > > > > ext4: fix NULL pointer dereference in ext4_mark_inode_dirty() > > > > > > Signed-off-by: Eryu Guan <eguan@redhat.com> > > > --- > > > > Fails with: > > > > @@ -1,2 +1,6 @@ > > QA output created by 338 > > Silence is golden > > +specified blocksize 1024 is less than device physical sector size 4096 > > +switching to logical sector size 512 > > +mkfs.xfs: /dev/mapper/error-test appears to contain an existing filesystem (xfs). > > +mkfs.xfs: Use the -f option to force overwrite. > > > > And then it failed to clean up properly and caused all sorts of > > subsequent problems. > > Test passed for me, seems it has something to do with the "physical > sector size 4096" device. I'll look into it. Thanks for the review! It fails because "_mkfs_dev $DMERROR_DEV" refuses to create new fs without "-f" option, has nothing to do with the 4k sector device. It passed for me is because I add "-f" mkfs option to my local.config for xfs sections, so _mkfs_dev passed. I'll send v3 to fix this. And the test fails to do cleanups on failure because "dmsetup remove error-test" reports device is busy. Adding a "$UDEV_SETTLE_PROG" call before "dmsetup remove error-test" in common/dmerror fixes the issue for me. I'll send another patch to fix it. Thanks, Eryu -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/tests/generic/338 b/tests/generic/338 new file mode 100755 index 0000000..235549a --- /dev/null +++ b/tests/generic/338 @@ -0,0 +1,80 @@ +#! /bin/bash +# FS QA Test 338 +# +# Test I/O on dm error device. +# +# Motivated by an ext4 bug that crashes kernel on error path when trying to +# update atime. +# +#----------------------------------------------------------------------- +# Copyright (c) 2016 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.* + _dmerror_cleanup +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter +. ./common/dmerror + +# remove previous $seqres.full before test +rm -f $seqres.full + +# real QA test starts here +_supported_fs generic +_supported_os Linux +_require_scratch +_require_dm_target error +# If SCRATCH_DEV is not a valid block device, FSTYP cannot be mkfs'ed either +_require_block_device $SCRATCH_DEV + +echo "Silence is golden" + +_dmerror_init +_mkfs_dev $DMERROR_DEV + +# Use strictatime mount option here to force atime updates, which could help +# trigger the NULL pointer dereference on ext4 more easily +_dmerror_mount "-o strictatime" +_dmerror_load_error_table + +# flush dmerror block device buffers and drop all caches, force reading from +# error device +blockdev --flushbufs $DMERROR_DEV +echo 3 > /proc/sys/vm/drop_caches + +# do some test I/O +ls -l $SCRATCH_MNT >>$seqres.full 2>&1 +$XFS_IO_PROG -fc "pwrite 0 1M" $SCRATCH_MNT/testfile >>$seqres.full 2>&1 + +# no panic no hang, success, all done +status=0 +exit diff --git a/tests/generic/338.out b/tests/generic/338.out new file mode 100644 index 0000000..3482cf4 --- /dev/null +++ b/tests/generic/338.out @@ -0,0 +1,2 @@ +QA output created by 338 +Silence is golden diff --git a/tests/generic/group b/tests/generic/group index 727648c..8818827 100644 --- a/tests/generic/group +++ b/tests/generic/group @@ -340,3 +340,4 @@ 335 auto quick metadata 336 auto quick metadata 337 auto quick metadata +338 auto quick rw
This is a test that performs simple I/O on dm error device, which returns EIO on all I/O request. This is motivated by an ext4 bug that crashes kernel on error path when trying to update atime. Following kernel patch should fix the issue ext4: fix NULL pointer dereference in ext4_mark_inode_dirty() Signed-off-by: Eryu Guan <eguan@redhat.com> --- v2: - use SCRATCH_DEV directly instead of loop device and call blockdev --flushbufs $SCRATCH_DEV before drop caches (suggested by Dave) tests/generic/338 | 80 +++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/338.out | 2 ++ tests/generic/group | 1 + 3 files changed, 83 insertions(+) create mode 100755 tests/generic/338 create mode 100644 tests/generic/338.out