Skip to contentSkip to conversation

xfstests: fallocate + read/write tests, ext4 regression tests

PATCH In review Eric Sandeen 2 participants 4 messages Jun 16, 2009 – Jun 21, 2009

lore
Share & tools
lore
#0 lore raw

PATCH 3/3 xfstests: fallocate + read/write tests, ext4 regression tests

New test to test basic mixed fallocate + read & write, includes a couple regression tests for bugs that ext4 hit. Uses xfs_io to generate fallocate calls, so requires git xfsprogs and very recent glibc at this point.

Ext4 folks, this is hopefully a reasonable example of how to add a new test. :)

---

214 new file +131 -0
new file mode 100755
index 0000000..98cfc3a
@@ -0,0 +1,131 @@
+#! /bin/sh
+# FS QA Test No. 214
+#
+# Basic unwritten extent sanity checks
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2009 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
+#
+#-----------------------------------------------------------------------
+#
+# creator
[email protected]
+
+seq=`basename $0`
+echo "QA output created by $seq"
+
+_cleanup()
+{
+	cd /
+	rm -f $tmp.*
+}
+
+here=`pwd`
+tmp=$TEST_DIR/$$
+status=1	# failure is the default!
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+# real QA test starts here
+# generic, but xfs_io's fallocate must work
+_supported_fs generic
+# only Linux supports fallocate
+_supported_os Linux
+
+[ -n "$XFS_IO_PROG" ] || _notrun "xfs_io executable not found"
+
+rm -f $seq.full
+
+testio=`$XFS_IO_PROG -F -f -c "falloc 0 1m" $TEST_DIR/$tmp.io 2>&1`
+
+# Old xfs_io doesn't have fallocate support
+echo $testio | grep -q "not found" && \
+	_notrun "xfs_io fallocate support is missing"
+# Old glibc, old kernels, and some filesystems don't have fallocate support
+echo $testio | grep -q "Operation not supported" && \
+	_notrun "xfs_io fallocate command failed (old kernel? wrong fs?)"
+
+# Ok, off we go.
+
+# Super-trivial; preallocate a region and read it; get 0s.
+echo "=== falloc & read  ==="
+$XFS_IO_PROG -F -f 		\
+	-c 'falloc 0 4096'	\
+	-c 'pread -v 0 4096'	\
+	$TEST_DIR/ouch | _filter_xfs_io_unique
+rm -f $TEST_DIR/ouch
+
+# Preallocate a chunk, write 1 byte, read it all back.
+# Should get no stale data.  Early ext4 bug.
+
+echo "=== falloc, write beginning, read ==="
+$XFS_IO_PROG -F -f 		\
+	-c 'falloc 0 512'	\
+	-c 'pwrite 0 1'		\
+	-c 'pread -v 0 512'	\
+	$TEST_DIR/ouch | _filter_xfs_io_unique
+rm -f $TEST_DIR/ouch
+
+# Same but write in the middle of the region
+echo "=== falloc, write middle, read ==="
+$XFS_IO_PROG -F -f 		\
+	-c 'falloc 0 512'	\
+	-c 'pwrite 256 1'	\
+	-c 'pread -v 0 512'	\
+	$TEST_DIR/ouch | _filter_xfs_io_unique
+rm -f $TEST_DIR/ouch
+
+# Same but write the end of the region
+echo "=== falloc, write end, read ==="
+$XFS_IO_PROG -F -f		\
+	-c 'falloc 0 512'	\
+	-c 'pwrite 511 1'	\
+	-c 'pread -v 0 512'	\
+	$TEST_DIR/ouch | _filter_xfs_io_unique
+rm -f $TEST_DIR/ouch
+
+# Reported by IBM on ext4.
+#
+# Fixed by commit a41f20716975910d9beb90b7efc61107901492b8
+#
+# The file was previously preallocated, and then initialized the middle of
+# the preallocation area using Direct IO write, then overwrite part of
+# initialized area. Later after truncate the file (to the middle of the
+# initialized data), the initialized data *before* the new file size was
+# gone after remount the filesystem.
+
+echo "=== falloc, write, sync, truncate, read ==="
+# Allocate, write, sync, truncate (buffered)
+$XFS_IO_PROG -F -f				\
+	-c 'falloc         0x0     0x65C00'	\
+	-c 'pwrite -S 0xAA 0x12000 0x10000'	\
+	-c 'fsync'				\
+	-c 'truncate 0x16000'			\
+	$TEST_DIR/ouch | _filter_xfs_io_unique
+
+# now do a direct read and see what's on-disk
+$XFS_IO_PROG -F -f -d				\
+	-c 'pread -v 0 0x16000'			\
+	$TEST_DIR/ouch | _filter_xfs_io_unique
+
+rm -f $TEST_DIR/ouch
+
+# success, all done
+status=0
+exit
214.out new file +41 -0

41 lines not on this page · raw message

group +1 -0
index 578fa05..c0ad793 100644
@@ -322,3 +322,4 @@ prealloc
 211 auto aio quick
 212 auto aio quick
 213 rw auto prealloc quick
+214 rw auto prealloc quick

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

Reviewed-by on v1 3/3
#1 lore raw
Quoted 4 lines · Eric Sandeen

New test to test basic mixed fallocate + read & write, includes a couple regression tests for bugs that ext4 hit. Uses xfs_io to generate fallocate calls, so requires git xfsprogs and very recent glibc at this point.

Looks good to me, and again does the correct notrun for my 32-bit VM. Guess I need to do the raw syscall wireup i xfsprogs eventually..

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#2 lore raw
Quoted 8 lines · Christoph Hellwig
Quoted 4 lines · Eric Sandeen

New test to test basic mixed fallocate + read & write, includes a couple regression tests for bugs that ext4 hit. Uses xfs_io to generate fallocate calls, so requires git xfsprogs and very recent glibc at this point.

Looks good to me, and again does the correct notrun for my 32-bit VM. Guess I need to do the raw syscall wireup i xfsprogs eventually..

I don't mind doing it either ... how many arches? :)

-Eric

Quoted 3 lines

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#3 lore raw
Quoted 11 lines · Eric Sandeen
Quoted 8 lines · Christoph Hellwig
Quoted 4 lines · Eric Sandeen

New test to test basic mixed fallocate + read & write, includes a couple regression tests for bugs that ext4 hit. Uses xfs_io to generate fallocate calls, so requires git xfsprogs and very recent glibc at this point.

Looks good to me, and again does the correct notrun for my 32-bit VM. Guess I need to do the raw syscall wireup i xfsprogs eventually..

I don't mind doing it either ... how many arches? :)

Most important is x86 I think. arm might be the other really important candidate.

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs