| Message ID | 20250204134707.2018526-1-hch@lst.de (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [1/2] xfs/614: query correct direct I/O alignment | expand |
On Tue, Feb 04, 2025 at 02:46:54PM +0100, Christoph Hellwig wrote: > When creating XFS file systems on files, mkfs will query the file system > for the minimum alignment, which can be larger than that of the > underlying device. Do the same to link the right output file. > > Signed-off-by: Christoph Hellwig <hch@lst.de> > --- > tests/xfs/614 | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/tests/xfs/614 b/tests/xfs/614 > index 0f8952e50b9a..06cc2384f38c 100755 > --- a/tests/xfs/614 > +++ b/tests/xfs/614 > @@ -26,7 +26,7 @@ _require_loop > $MKFS_XFS_PROG 2>&1 | grep -q concurrency || \ > _notrun "mkfs does not support concurrency options" > > -test_dev_lbasize=$(blockdev --getss $TEST_DEV) > +test_dev_lbasize=$($here/src/min_dio_alignment $TEST_DIR $TEST_DEV) Hmmmm... I have a patch with a similar aim in my dev tree that determines the lba size from whatever mkfs decides is the sector size: # Figure out what sector size mkfs will use to format, which might be dependent # upon the directio write geometry of the test filesystem. loop_file=$TEST_DIR/$seq.loop rm -f "$loop_file" truncate -s 16M "$loop_file" $MKFS_XFS_PROG -f -N "$loop_file" | _filter_mkfs 2>$tmp.mkfs >/dev/null . $tmp.mkfs seqfull=$0 _link_out_file "lba${sectsz}" What do you think of that approach? --D > seqfull=$0 > _link_out_file "lba${test_dev_lbasize}" > > -- > 2.45.2 > >
On Tue, Feb 04, 2025 at 09:12:58AM -0800, Darrick J. Wong wrote: > Hmmmm... I have a patch with a similar aim in my dev tree that > determines the lba size from whatever mkfs decides is the sector size: > > # Figure out what sector size mkfs will use to format, which might be dependent > # upon the directio write geometry of the test filesystem. > loop_file=$TEST_DIR/$seq.loop > rm -f "$loop_file" > truncate -s 16M "$loop_file" > $MKFS_XFS_PROG -f -N "$loop_file" | _filter_mkfs 2>$tmp.mkfs >/dev/null > . $tmp.mkfs > seqfull=$0 > _link_out_file "lba${sectsz}" > > What do you think of that approach? That sound sensible. Where is that patch, it doesn't seem to be in the realtime-reflink branch that I'm usually working against.
On Wed, Feb 05, 2025 at 04:44:22PM +0100, Christoph Hellwig wrote: > On Tue, Feb 04, 2025 at 09:12:58AM -0800, Darrick J. Wong wrote: > > Hmmmm... I have a patch with a similar aim in my dev tree that > > determines the lba size from whatever mkfs decides is the sector size: > > > > # Figure out what sector size mkfs will use to format, which might be dependent > > # upon the directio write geometry of the test filesystem. > > loop_file=$TEST_DIR/$seq.loop > > rm -f "$loop_file" > > truncate -s 16M "$loop_file" > > $MKFS_XFS_PROG -f -N "$loop_file" | _filter_mkfs 2>$tmp.mkfs >/dev/null > > . $tmp.mkfs > > seqfull=$0 > > _link_out_file "lba${sectsz}" > > > > What do you think of that approach? > > That sound sensible. Where is that patch, it doesn't seem to be > in the realtime-reflink branch that I'm usually working against. It's one of the few zoned changes I have in the fstests branch: https://git.kernel.org/pub/scm/linux/kernel/git/djwong/xfstests-dev.git/log/?h=zoned_2025-02-04 --D
On Wed, Feb 05, 2025 at 07:51:11AM -0800, Darrick J. Wong wrote: > It's one of the few zoned changes I have in the fstests branch: > > https://git.kernel.org/pub/scm/linux/kernel/git/djwong/xfstests-dev.git/log/?h=zoned_2025-02-04 Oh, you probably want to pull this in to make your life easier: http://git.infradead.org/?p=users/hch/xfstests-dev.git;a=shortlog;h=refs/heads/xfs-zoned-rebase Note that the dio size mismatch also happens when using a 4k LBA rt device with a 512 byte LBA data device without any zoned code.
diff --git a/tests/xfs/614 b/tests/xfs/614 index 0f8952e50b9a..06cc2384f38c 100755 --- a/tests/xfs/614 +++ b/tests/xfs/614 @@ -26,7 +26,7 @@ _require_loop $MKFS_XFS_PROG 2>&1 | grep -q concurrency || \ _notrun "mkfs does not support concurrency options" -test_dev_lbasize=$(blockdev --getss $TEST_DEV) +test_dev_lbasize=$($here/src/min_dio_alignment $TEST_DIR $TEST_DEV) seqfull=$0 _link_out_file "lba${test_dev_lbasize}"
When creating XFS file systems on files, mkfs will query the file system for the minimum alignment, which can be larger than that of the underlying device. Do the same to link the right output file. Signed-off-by: Christoph Hellwig <hch@lst.de> --- tests/xfs/614 | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)