Re: [PATCH v2] generic/728: Add a test for xattr ctime updates

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]



On Fri, May 05, 2023 at 03:39:29PM -0700, Darrick J. Wong wrote:
> On Fri, May 05, 2023 at 01:24:27PM -0400, Anna Schumaker wrote:
> > From: Anna Schumaker <Anna.Schumaker@xxxxxxxxxx>
> > 
> > The NFS client wasn't updating ctime after a setxattr request. This is a
> > test written while fixing the bug.
> > 
> > Signed-off-by: Anna Schumaker <Anna.Schumaker@xxxxxxxxxx>
> > 
> > ---
> > v2:
> >  - Move test to generic/
> >  - Address comments from the mailing list
> > ---
> >  tests/generic/728     | 42 ++++++++++++++++++++++++++++++++++++++++++
> >  tests/generic/728.out |  2 ++
> >  2 files changed, 44 insertions(+)
> >  create mode 100755 tests/generic/728
> >  create mode 100644 tests/generic/728.out
> > 
> > diff --git a/tests/generic/728 b/tests/generic/728
> > new file mode 100755
> > index 000000000000..ab2414c151db
> > --- /dev/null
> > +++ b/tests/generic/728
> > @@ -0,0 +1,42 @@
> > +#! /bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (c) 2023 Netapp Inc., All Rights Reserved.
> > +#
> > +# FS QA Test 728
> > +#
> > +# Test a bug where the NFS client wasn't sending a post-op GETATTR to the
> > +# server after setting an xattr, resulting in `stat` reporting a stale ctime.
> > +#
> > +. ./common/preamble
> > +_begin_fstest auto quick attr
> > +
> > +# Import common functions
> > +. ./common/attr
> > +
> > +# real QA test starts here
> > +_supported_fs generic
> > +_require_test
> > +_require_attrs
> > +
> > +rm -rf $TEST_DIR/testfile
> > +touch $TEST_DIR/testfile
> > +
> > +
> > +_check_xattr_op()
> > +{
> > +  what=$1
> > +  shift 1
> > +
> > +  before_ctime=$(stat -c %z $TEST_DIR/testfile)
> > +  $SETFATTR_PROG $* $TEST_DIR/testfile
> > +  after_ctime=$(stat -c %z $TEST_DIR/testfile)
> 
> What happens if the filesystem's timestamp granularity isn't sufficient
> to capture a difference across a setxattr operation?  Most of the time
> we only update the (fast) timestamps at every clock tick, right?

OK, by talking with Darrick, I think "sleep 2s" might avoid most problem
of timestamp granularity at here. (If this change don't affect your original
test target). Any thoughts?

> 
> --D
> 
> > +
> > +  test "$before_ctime" != "$after_ctime" || echo "Expected ctime to change after $what."
> > +}
> > +
> > +_check_xattr_op setxattr -n user.foobar -v 123
> > +_check_xattr_op removexattr -x user.foobar
> > +
> > +echo "Silence is golden"
> > +status=0
> > +exit
> > diff --git a/tests/generic/728.out b/tests/generic/728.out
> > new file mode 100644
> > index 000000000000..ab39f45fe5da
> > --- /dev/null
> > +++ b/tests/generic/728.out
> > @@ -0,0 +1,2 @@
> > +QA output created by 728
> > +Silence is golden
> > -- 
> > 2.40.1
> > 
> 




[Index of Archives]     [Linux Filesystems Development]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux