Skip to contentSkip to conversation

xfstests 299: test write to the last block of a sparse file

PATCH In review Eryu Guan 2 participants 3 messages Feb 22, 2013 – Feb 26, 2013 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Eryu Guan

author
#1 lore raw

PATCH 2/2 xfstests 299: test write to the last block of a sparse file

Write to the last block of a sparse file in extent format on ext4 would hit BUG_ON() on unpatched kernel.

Regression test for commit:
f17722f ext4: Fix max file size and logical block counting of extent format file

---

Note that the second dd could triger a BUG_ON() on ext4/3.8 kernel in ext4_es_remove_extent().

I sent a patch for this issue, please see
http://www.spinics.net/lists/linux-ext4/msg36784.html

 299     | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 299.out |  1 +
 group   |  1 +
 3 files changed, 64 insertions(+)
 create mode 100644 299
 create mode 100644 299.out
299 new file +62 -0
new file mode 100644
index 0000000..9b52243
@@ -0,0 +1,62 @@
+#! /bin/bash
+# FS QA Test No. 299
+#
+# Write to the last block of a sparse file in extent format on ext4 would
+# hit BUG_ON() on unpatched kernel.
+#
+# Regression test for commit:
+# f17722f ext4: Fix max file size and logical block counting of extent format file
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2013 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"
+
+status=1	# failure is the default!
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+_cleanup()
+{
+	cd /
+	rm -f $testfile
+}
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+testfile=$TEST_DIR/testfile.$seq
+
+# real QA test starts here
+_supported_fs generic
+_supported_os Linux
+
+block_size=`stat -f $TEST_DEV | grep "Block size" | cut -d " " -f3`
+
+# Create sparse file
+dd if=/dev/zero of=$testfile bs=$block_size count=1 seek=$((2**32 - 2)) >/dev/null 2>&1
+sync
+# Write to the last block
+dd if=/dev/zero of=$testfile bs=$block_size count=1 seek=$((2**32 - 1)) >/dev/null 2>&1
+
+# Got here without hitting BUG_ON(), test passed
+status=0
+exit
299.out new file +1 -0
new file mode 100644
index 0000000..0b41c6f
@@ -0,0 +1 @@
+QA output created by 299
group +1 -0
index 3039af4..a993729 100644
@@ -417,3 +417,4 @@ deprecated
 296 dump auto quick
 297 auto freeze
 298 auto quick metadata
+299 auto quick
Signature · 6 lines
1.8.0.2

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
#2 lore raw
Quoted 70 lines · Eryu Guan

Write to the last block of a sparse file in extent format on ext4 would hit BUG_ON() on unpatched kernel.

Regression test for commit:
f17722f ext4: Fix max file size and logical block counting of extent format file

---

Note that the second dd could triger a BUG_ON() on ext4/3.8 kernel in ext4_es_remove_extent().

I sent a patch for this issue, please see
http://www.spinics.net/lists/linux-ext4/msg36784.html

 299     | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 299.out |  1 +
 group   |  1 +
 3 files changed, 64 insertions(+)
 create mode 100644 299
 create mode 100644 299.out
299 new file +42 -0
new file mode 100644
index 0000000..9b52243
@@ -0,0 +1,62 @@
+#! /bin/bash
+# FS QA Test No. 299
+#
+# Write to the last block of a sparse file in extent format on ext4 would
+# hit BUG_ON() on unpatched kernel.
+#
+# Regression test for commit:
+# f17722f ext4: Fix max file size and logical block counting of extent format file
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2013 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"
+
+status=1	# failure is the default!
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+_cleanup()
+{
+	cd /
+	rm -f $testfile
+}
+
+# get standard environment, filters and checks

299 line 52

Quoted 10 lines
299 new file +10 -0
+. ./common.rc
+. ./common.filter
+
+testfile=$TEST_DIR/testfile.$seq
+
+# real QA test starts here
+_supported_fs generic
+_supported_os Linux
+
+block_size=`stat -f $TEST_DEV | grep "Block size" | cut -d " " -f3`

stat -f -c %s

299 line 58

Quoted 6 lines
299 new file +5 -0
+# Create sparse file
+dd if=/dev/zero of=$testfile bs=$block_size count=1 seek=$((2**32 - 2)) >/dev/null 2>&1
+sync
 
+# Write to the last block
+dd if=/dev/zero of=$testfile bs=$block_size count=1 seek=$((2**32 - 1)) >/dev/null 2>&1

That doesn't write to the last block - that writes one block beyond the EOF. And it also truncates the file first, so what is triggering the bug? the truncate of the previous file, or the writing of the file? If it is writing of the file, then what's the point of the first write?

Also, I'd prefer that xfs_io is used instead of dd because then the operations being performed are explicit. i.e, the above it this:

offset=$(((2**32 - 2) * $block_size))
XFS_IO_PROG -f -c "pwrite $offset $block_size" \
-c fsync $testfile | filter_io

XFS_IO_PROG -f -c "truncate 0" \
		-c "pwrite $((offset + $block_size)) $block_size" \
		-c fsync $testfile | filter_io

And it's not clear from either the test of the comments which of the 3 operations is actaully the one that causes the bug to be triggered....

Cheers,

Dave.

Signature · 7 lines
Dave Chinner
[email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
author
#3 lore raw
Quoted 75 lines · Dave Chinner
Quoted 74 lines · Eryu Guan

Write to the last block of a sparse file in extent format on ext4 would hit BUG_ON() on unpatched kernel.

Regression test for commit:
f17722f ext4: Fix max file size and logical block counting of extent format file

---

Note that the second dd could triger a BUG_ON() on ext4/3.8 kernel in ext4_es_remove_extent().

I sent a patch for this issue, please see
http://www.spinics.net/lists/linux-ext4/msg36784.html

 299     | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 299.out |  1 +
 group   |  1 +
 3 files changed, 64 insertions(+)
 create mode 100644 299
 create mode 100644 299.out
299 new file +46 -0
new file mode 100644
index 0000000..9b52243
@@ -0,0 +1,62 @@
+#! /bin/bash
+# FS QA Test No. 299
+#
+# Write to the last block of a sparse file in extent format on ext4 would
+# hit BUG_ON() on unpatched kernel.
+#
+# Regression test for commit:
+# f17722f ext4: Fix max file size and logical block counting of extent format file
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2013 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"
+
+status=1	# failure is the default!
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+_cleanup()
+{
+	cd /
+	rm -f $testfile
+}
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+testfile=$TEST_DIR/testfile.$seq
Quoted 8 lines

299 line 52

Quoted 6 lines
299 new file +6 -0
+
+# real QA test starts here
+_supported_fs generic
+_supported_os Linux
+
+block_size=`stat -f $TEST_DEV | grep "Block size" | cut -d " " -f3`

stat -f -c %s

This is definitely better, thanks!

Quoted 20 lines

299 line 58

Quoted 7 lines
299 new file +5 -0
+# Create sparse file
+dd if=/dev/zero of=$testfile bs=$block_size count=1 seek=$((2**32 - 2)) >/dev/null 2>&1
+sync
 
+# Write to the last block
+dd if=/dev/zero of=$testfile bs=$block_size count=1 seek=$((2**32 - 1)) >/dev/null 2>&1

That doesn't write to the last block - that writes one block beyond the EOF. And it also truncates the file first, so what is triggering the bug? the truncate of the previous file, or the writing of the file? If it is writing of the file, then what's the point of the first write?

Also, I'd prefer that xfs_io is used instead of dd because then the operations being performed are explicit. i.e, the above it this:

offset=$(((2**32 - 2) * $block_size))
XFS_IO_PROG -f -c "pwrite $offset $block_size" \
-c fsync $testfile | filter_io

Quoted 7 lines
XFS_IO_PROG -f -c "truncate 0" \
		-c "pwrite $((offset + $block_size)) $block_size" \
		-c fsync $testfile | filter_io

And it's not clear from either the test of the comments which of the 3 operations is actaully the one that causes the bug to be triggered....

Thanks for your review! I'll adopt xfs_io and try to make the comments more informative in next version.

Thanks,

Eryu

Quoted 7 lines

Cheers,

Dave.

Signature · 2 lines
Dave Chinner
[email protected]

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