[Top][All Lists]

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

Re: [Qemu-devel] [PATCH] qemu-img convert: Fix copy length for the itera

From: Max Reitz
Subject: Re: [Qemu-devel] [PATCH] qemu-img convert: Fix copy length for the iteration
Date: Wed, 11 Feb 2015 09:39:07 -0500
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.4.0

On 2015-02-11 at 06:08, Kevin Wolf wrote:
qemu-img converts already checks the block status of the next sectors to
copy and skips them if it knows that they don't need to be copied (e.g.
because they are zero and the destination is zero-initialised).

It doesn't, however, limit the number of sectors read to the number of
allocated sectors when copying data yet. This means that when using -B,
some data is needlessly copied from the backing file into the overlay.
For images without backing files, it means that more data is read than
is necessary, even though the zero buffer detection will still keep
those sectors sparse.

This patch changes qemu-img to respect the number of allocated sectors
and limit its requests accordingly when processing data, which fixes -B
to expose the expected behaviour and is a performance optimisation for
other cases.

Reported-by: Max Reitz <address@hidden>
Signed-off-by: Kevin Wolf <address@hidden>
  qemu-img.c                 |  1 +
  tests/qemu-iotests/122     | 56 ++++++++++++++++++++++++++++++++++++++++++++++
  tests/qemu-iotests/122.out | 11 +++++++++
  tests/qemu-iotests/group   |  1 +
  4 files changed, 69 insertions(+)
  create mode 100755 tests/qemu-iotests/122
  create mode 100644 tests/qemu-iotests/122.out

diff --git a/qemu-img.c b/qemu-img.c
index e148af8..b94afe2 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -1809,6 +1809,7 @@ restart:
n = MIN(nb_sectors, bufsectors);
+            n = MIN(n, sector_num_next_status - sector_num);

This needs to be done only if sector_num_next_status > sector_num, otherwise we'll run into an infinite loop.

Apparently I accidentally broke -n for qemu-img convert, but once I get it working again, you can see that this command:

$ qemu-img convert -n -O raw test.qcow2 nbd://localhost:10809

does not terminate (because out_baseimg is NULL and has_zero_init is false due to nbd not supporting that, which means that sector_num_next_status is never set to anything but 0).

Other than that, this looks good to me.


              /* round down request length to an aligned sector, but
               * do not bother doing this on short requests. They happen
diff --git a/tests/qemu-iotests/122 b/tests/qemu-iotests/122
new file mode 100755
index 0000000..0106e8a
--- /dev/null
+++ b/tests/qemu-iotests/122
@@ -0,0 +1,56 @@
+# Test image sparseness after qemu-img convert
+# Copyright (C) 2015 Red Hat, Inc.
+# 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; either version 2 of the License, or
+# (at your option) any later version.
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# 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, see <http://www.gnu.org/licenses/>.
+# creator
+seq="$(basename $0)"
+echo "QA output created by $seq"
+status=1       # failure is the default!
+       _cleanup_test_img
+trap "_cleanup; exit \$status" 0 1 2 3 15
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+_supported_fmt qcow2
+_supported_proto generic
+_supported_os Linux
+TEST_IMG=$TEST_IMG.base _make_test_img 64M
+_make_test_img -b $TEST_IMG.base
+$QEMU_IO -c "write -P 0x11 0 64M" $TEST_IMG.base 2>&1 | _filter_qemu_io | 
+$QEMU_IO -c "write -P 0x22 0 3M" $TEST_IMG 2>&1 | _filter_qemu_io | 
+$QEMU_IMG convert -O $IMGFMT -B $TEST_IMG.base $TEST_IMG $TEST_IMG.orig
+$QEMU_IMG map $TEST_IMG.orig | _filter_qemu_img_map
+# success, all done
+echo '*** done'
+rm -f $seq.full
diff --git a/tests/qemu-iotests/122.out b/tests/qemu-iotests/122.out
new file mode 100644
index 0000000..6e4a760
--- /dev/null
+++ b/tests/qemu-iotests/122.out
@@ -0,0 +1,11 @@
+QA output created by 122
+Formatting 'TEST_DIR/t.IMGFMT.base', fmt=IMGFMT size=67108864
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 
+wrote 67108864/67108864 bytes at offset 0
+64 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+wrote 3145728/3145728 bytes at offset 0
+3 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+Offset          Length          File
+0               0x300000        TEST_DIR/t.IMGFMT.orig
+0x300000        0x3d00000       TEST_DIR/t.IMGFMT.base
+*** done
diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
index 4b2b93b..28c78e5 100644
--- a/tests/qemu-iotests/group
+++ b/tests/qemu-iotests/group
@@ -117,3 +117,4 @@
  113 rw auto quick
  114 rw auto quick
  116 rw auto quick
+122 rw auto quick

reply via email to

[Prev in Thread] Current Thread [Next in Thread]