Skip to contentSkip to conversation

xfstests: pwrite hang when writing from mmaped buffer of the same page

PATCH Superseded Zhong, Xin 2 participants 2 messages Dec 14, 2010

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Zhong, Xin

author
#0 lore raw

PATCH v2 1/1 xfstests: pwrite hang when writing from mmaped buffer of the same page

The problem is found in meego testing on btrfs:
http://bugs.meego.com/show_bug.cgi?id=6672

---
 248                       |   58 +++++++++++++++++++++++++++++++++
 248.out                   |    2 +
 group                     |    1 +
 src/Makefile              |    2 +-
 src/pwrite_mmap_blocked.c |   78 +++++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 140 insertions(+), 1 deletions(-)
 create mode 100755 248
 create mode 100644 248.out
 create mode 100644 src/pwrite_mmap_blocked.c
248 new file +58 -0
new file mode 100755
index 0000000..a4664d9
@@ -0,0 +1,58 @@
+#! /bin/bash
+# FS QA Test No. 248
+#
+# Test for pwrite hang problem when writing from mmaped buffer of the same page 
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2010 YOUR NAME HERE.  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"
+
+here=`pwd`
+tmp=/tmp/$$
+status=1	# failure is the default!
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+_cleanup()
+{
+    cd /
+    rm -f $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+# real QA test starts here
+
+# Modify as appropriate.
+_supported_fs generic
+_supported_os Linux
+
+TESTFILE=$TEST_DIR/test_file
+TEST_PROG=$here/src/pwrite_mmap_blocked
+
+$TEST_PROG $TESTFILE
+rm -f $TESTFILE
+
+# success, all done
+status=0
+exit
248.out new file +2 -0
new file mode 100644
index 0000000..d02c1ac
@@ -0,0 +1,2 @@
+QA output created by 248
+pwrite 1 bytes from 2 to 3
group +1 -0
index 0f94dd9..ddda26c 100644
@@ -361,3 +361,4 @@ deprecated
 245 auto quick dir
 246 auto quick rw
 247 auto quick rw
+248 other quick rw 
src/Makefile +1 -1
index b827bd0..47d7334 100644
@@ -17,7 +17,7 @@ LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \
 	preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \
 	locktest unwritten_mmap bulkstat_unlink_test t_stripealign \
 	bulkstat_unlink_test_modified t_dir_offset t_futimens t_immutable \
-	stale_handle
+	stale_handle pwrite_mmap_blocked
 
 SUBDIRS =
 
src/pwrite_mmap_blocked.c new file +78 -0

78 lines not on this page · raw message

Signature · 6 lines
1.6.2.2

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
#1 lore raw
Quoted 22 lines · Xin

The problem is found in meego testing on btrfs:
http://bugs.meego.com/show_bug.cgi?id=6672

---
 248                       |   58 +++++++++++++++++++++++++++++++++
 248.out                   |    2 +
 group                     |    1 +
 src/Makefile              |    2 +-
 src/pwrite_mmap_blocked.c |   78 +++++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 140 insertions(+), 1 deletions(-)
 create mode 100755 248
 create mode 100644 248.out
 create mode 100644 src/pwrite_mmap_blocked.c
248 new file +0 -0
new file mode 100755
index 0000000..a4664d9

248 line 7

Quoted 7 lines
248 new file +7 -0
@@ -0,0 +1,58 @@
+#! /bin/bash
+# FS QA Test No. 248
+#
+# Test for pwrite hang problem when writing from mmaped buffer of the same page 
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2010 YOUR NAME HERE.  All Rights Reserved.
                        ^^^^^^^^^^^^^^

Needs updating, I think. ;)

While I'm here, a two minute review...

....

Quoted 11 lines
248 new file +11 -0
+_cleanup()
+{
+    cd /
+    rm -f $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+# real QA test starts here

248 line 54

Quoted 10 lines
248 new file +10 -0
+
+# Modify as appropriate.
+_supported_fs generic
+_supported_os Linux
+
+TESTFILE=$TEST_DIR/test_file
+TEST_PROG=$here/src/pwrite_mmap_blocked
+
+$TEST_PROG $TESTFILE
+rm -f $TESTFILE

This rm should be in the _cleanup() function so that it is removed even when the test is aborted.

.....

group deprecated · line 364

Quoted 9 lines
group +1 -0
index 0f94dd9..ddda26c 100644
@@ -361,3 +361,4 @@ deprecated
 245 auto quick dir
 246 auto quick rw
 247 auto quick rw
+248 other quick rw 

Why not the auto group?

group

Quoted 4 lines
group +4 -0
+char *progname;
+loff_t size;
+int fd;
+void *mapped_mem;

None of these need to be global.

Quoted 6 lines
group +6 -0
+
+
+int main(int argc, char *argv[])
+{
+	int ret;
+

group

Quoted 10 lines
group +10 -0
+	progname = argv[0];
+	size = 5;
+	fd = open(argv[1], O_RDWR|O_TRUNC|O_CREAT, 0666);
+	if (fd < 0) {
+		fprintf(stderr, "%s: Cannot open `%s': %s\n",
+			progname, argv[1], strerror(errno));
+		exit(1);
+	}
+
+	char *cc = "01234";

Don't mix declarations in the the code. Better would be:

	int ret;
	char *cc = "01234";

	progname = argv[0];
	size = strlen(cc);
	....

group

Quoted 2 lines
group +2 -0
+	if ((ret = pwrite(fd, (const char *)cc,
+				size, 0)) != size) {
	ret = pwrite(fd, cc, size, 0);
	if (ret != size) {

group

Quoted 5 lines
group +5 -0
+		fprintf(stderr, "%s: pwrite returned %d\n",
+			__FUNCTION__, ret);
+		perror("pwrite");
+		exit(1);
+	}

No need for both fprintf and perror calls.

group

Quoted 11 lines
group +11 -0
+
+	mapped_mem = mmap(0, size, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0);
+	if (mapped_mem == MAP_FAILED) {
+		perror("mmap");
+		exit(1);
+	}
+	loff_t amount = 1;
+	loff_t from = 2;
+	loff_t to = 3;
+	printf("pwrite %Ld bytes from %Ld to %Ld\n", amount, from, to);
+	if ((ret = pwrite(fd, (char *)mapped_mem + from, amount, to)) != amount) {
	ret = pwrite(fd, (char *)mapped_mem + from, amount, to));
	if (ret != amount) {

group

Quoted 8 lines
group +8 -0
+		printf("%s: pwrite returned %d, not %Ld\n",
+					__FUNCTION__, ret, amount);
+		if (errno == EFAULT) {
+			printf("pwrite: EFAULT\n");
+		} else if (ret < 0) {
+			perror("pwrite");
+			exit(1);
+		}

A single call to perror() will be sufficient here...

group

Quoted 5 lines
group +5 -0
+	}
+	munmap(mapped_mem,0);
+	close(fd);
+	exit(0);
+}

Cheers,

Dave.

Signature · 7 lines
Dave Chinner
[email protected]

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