Re: [Qemu-devel] [PATCH] iotests: Add test for dataplane mirroring

[Max Reitz]

On 2017-06-29 12:10, Kevin Wolf wrote:
> Am 29.06.2017 um 01:23 hat Max Reitz geschrieben:
>> Signed-off-by: Max Reitz <address@hidden>
>> ---
>> Depends on Stefan's "virtio: use ioeventfd in TCG and qtest mode" series
>> to work at all, and on "mirror: Fix inconsistent backing AioContext for
>> after mirroring" (in my block branch) so it does not fail.
>> ---
>>  tests/qemu-iotests/106     | 97 
>> ++++++++++++++++++++++++++++++++++++++++++++++
>>  tests/qemu-iotests/106.out | 14 +++++++
> 
> Your initiative to fill up the numbering hole is laudable, but are you
> intentionally using 106 for multiple patches of yours? ;-)

Well, as I said, I'm fine with each new test having test number 001 and
then moving it when applying. O:-)

Wasn't intentional, but well, I find it easier to find free test numbers
when applying patches anyway...

In this case, though, since the other user of 106 is already fully
reviewed, it's probably better to move this out of the way.

>>  tests/qemu-iotests/group   |  1 +
>>  3 files changed, 112 insertions(+)
>>  create mode 100755 tests/qemu-iotests/106
>>  create mode 100644 tests/qemu-iotests/106.out
>>
>> diff --git a/tests/qemu-iotests/106 b/tests/qemu-iotests/106
>> new file mode 100755
>> index 0000000..ad438b5
>> --- /dev/null
>> +++ b/tests/qemu-iotests/106
>> @@ -0,0 +1,97 @@
>> +#!/bin/bash
>> +#
>> +# Test case for mirroring with dataplane
>> +#
>> +# Copyright (C) 2017 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
>> +# 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, see <http://www.gnu.org/licenses/>.
>> +#
>> +
>> +# creator
>> address@hidden
>> +
>> +seq=$(basename $0)
>> +echo "QA output created by $seq"
>> +
>> +here=$PWD
>> +status=1    # failure is the default!
>> +
>> +_cleanup()
>> +{
>> +    _cleanup_qemu
>> +    _cleanup_test_img
>> +    _rm_test_img "$TEST_IMG.overlay0"
>> +    _rm_test_img "$TEST_IMG.overlay1"
>> +}
>> +trap "_cleanup; exit \$status" 0 1 2 3 15
>> +
>> +# get standard environment, filters and qemu instance handling
>> +. ./common.rc
>> +. ./common.filter
>> +. ./common.qemu
>> +
>> +_supported_fmt qcow2
>> +_supported_proto file
>> +_supported_os Linux
>> +
>> +IMG_SIZE=64K
>> +
>> +_make_test_img $IMG_SIZE
>> +TEST_IMG="$TEST_IMG.overlay0" _make_test_img -b "$TEST_IMG" $IMG_SIZE
>> +TEST_IMG="$TEST_IMG.overlay1" _make_test_img -b "$TEST_IMG" $IMG_SIZE
>> +
>> +# So that we actually have something to mirror and the job does not return
>> +# immediately (which may be bad because then we cannot know whether the
>> +# 'return' or the 'BLOCK_JOB_READY' comes first).
>> +$QEMU_IO -c 'write 0 64' "$TEST_IMG.overlay0" | _filter_qemu_io
> 
> 64 bytes? Unusual, but yes, why not. We probably don't test this too
> often. :-)

*cough* well, somehow dropped the k there, but yes...

>> +# We cannot use virtio-blk here because that does not actually set the 
>> attached
>> +# BB's AioContext in qtest mode
> 
> Why that? I don't see any qtest special casing in the virtio-blk code,
> so is this intentional?

I don't know. But that's at least what I discovered when putting debug
code into bdrv_open_backing_file(); the overlay was still attached to
the main qemu AioContext.

>> +_launch_qemu \
>> +    -object iothread,id=iothr \
>> +    -blockdev 
>> node-name=source,driver=$IMGFMT,file.driver=file,file.filename="$TEST_IMG.overlay0"
>>  \
>> +    -device virtio-scsi,id=scsi-bus,iothread=iothr \
>> +    -device scsi-hd,bus=scsi-bus.0,drive=source
>> +
>> +_send_qemu_cmd $QEMU_HANDLE \
>> +    "{ 'execute': 'qmp_capabilities' }" \
>> +    'return'
>> +
>> +_send_qemu_cmd $QEMU_HANDLE \
>> +    "{ 'execute': 'drive-mirror',
>> +       'arguments': {
>> +           'job-id': 'mirror',
>> +           'device': 'source',
>> +           'target': '$TEST_IMG.overlay1',
>> +           'mode':   'existing',
>> +           'sync':   'top'
>> +       } }" \
>> +    'BLOCK_JOB_READY'
>> +
>> +# The backing BDS should be assigned the overlay's AioContext
>> +_send_qemu_cmd $QEMU_HANDLE \
>> +    "{ 'execute': 'block-job-complete',
>> +       'arguments': { 'device': 'mirror' } }" \
>> +    'BLOCK_JOB_COMPLETED'
>> +
>> +_send_qemu_cmd $QEMU_HANDLE \
>> +    "{ 'execute': 'quit' }" \
>> +    'return'
>> +
>> +wait=yes _cleanup_qemu
>> +
>> +# success, all done
>> +echo '*** done'
>> +rm -f $seq.full
>> +status=0
> 
> The actual test looks good to me.

OK, thanks for reviewing. I'll just send a v2 because I'd like to fix
the 64 (at least change it to 42 so people don't think it's by accident!
*cough*) and doing both that and giving it a different test number at
the same time seems a bit much.

Or I can just follow your model which is to send it and keep it in my
branch at the same time. :-)

Max

signature.asc
Description: OpenPGP digital signature