qemu-devel
[Top][All Lists]
Advanced

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

Re: [PATCH for-5.1 v2 2/2] iotests: add test for unaligned granularity b


From: Stefan Reiter
Subject: Re: [PATCH for-5.1 v2 2/2] iotests: add test for unaligned granularity bitmap backup
Date: Mon, 10 Aug 2020 17:35:19 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.9.0

On 8/10/20 5:11 PM, Max Reitz wrote:
(Note: When submitting a patch series with multiple patches, our
guidelines require a cover letter:
https://wiki.qemu.org/Contribute/SubmitAPatch#Include_a_meaningful_cover_letter

But not too important now.)


Sorry, remembered for next time. Thanks for applying the patches!

On 10.08.20 11:55, Stefan Reiter wrote:
Start a VM with a 4097 byte image attached, add a 4096 byte granularity
dirty bitmap, mark it dirty, and then do a backup.

This used to run into an assert and fail, check that it works as
expected and also check the created image to ensure that misaligned
backups in general work correctly.

Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
---

I saw Andrey's big series covering iotest 303 so I went for 304.

Works for me.

Never submitted
one before so I hope that's okay, if not feel free to renumber it.

Yep, if there’s a clash I tend to just renumber it when applying it.

  tests/qemu-iotests/304     | 68 ++++++++++++++++++++++++++++++++++++++
  tests/qemu-iotests/304.out |  2 ++
  tests/qemu-iotests/group   |  1 +
  3 files changed, 71 insertions(+)
  create mode 100755 tests/qemu-iotests/304
  create mode 100644 tests/qemu-iotests/304.out

diff --git a/tests/qemu-iotests/304 b/tests/qemu-iotests/304
new file mode 100755
index 0000000000..9a3b0224fa
--- /dev/null
+++ b/tests/qemu-iotests/304
@@ -0,0 +1,68 @@
+#!/usr/bin/env python3
+#
+# Tests dirty-bitmap backup with unaligned bitmap granularity
+#
+# Copyright (c) 2020 Proxmox Server Solutions
+#
+# 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/>.
+#
+# owner=s.reiter@proxmox.com
+
+import iotests
+from iotests import qemu_img_create, qemu_img_log, file_path
+
+iotests.script_initialize(supported_fmts=['qcow2'],
+                          supported_protocols=['file'])
+
+test_img = file_path('test.qcow2')
+target_img = file_path('target.qcow2')
+
+# unaligned by one byte
+image_len = 4097
+bitmap_granularity = 4096
+
+qemu_img_create('-f', iotests.imgfmt, test_img, str(image_len))
+
+# create VM and add dirty bitmap
+vm = iotests.VM().add_drive(test_img)
+vm.launch()
+
+vm.qmp('block-dirty-bitmap-add', **{
+    'node': 'drive0',
+    'name': 'bitmap0',
+    'granularity': bitmap_granularity
+})
+
+# mark entire bitmap as dirty
+vm.hmp_qemu_io('drive0', 'write -P0x16 0 4096');
+vm.hmp_qemu_io('drive0', 'write -P0x17 4097 1');

s/4097/4096/?

(4097 works, too, because of something somewhere aligning up the 4097 to
512 byte sectors, I suppose, but I don’t think it’s the address you want
here)


You're right, it seems counting is hard. I'll take you up on the offer from your other mail, you can fix this please :)

+
+# do backup and wait for completion
+vm.qmp('drive-backup', **{
+    'device': 'drive0',
+    'sync': 'full',
+    'target': target_img,
+    'bitmap': 'bitmap0',
+    'bitmap-mode': 'on-success'

The bitmap is unnecessary, isn’t it?  I.e., if I drop the
block-dirty-bitmap-add call and the bitmap* parameters here, I still get
an assertion failure without patch 1.

Not that it really matters, it’s just that this makes it look like less
of an issue than it actually is.  (Which is why I’d drop the bitmap
stuff in case there’s no actual reason for it.)


Oh my, I just realized that I misunderstood the root cause then. I mean, the fix is fine, I see it now, but you're right, no dirty bitmap needed - you can remove that as well if you want.

+})
+
+event = vm.event_wait(name='BLOCK_JOB_COMPLETED',
+                      match={'data': {'device': 'drive0'}},
+                      timeout=5.0)

(By the way, “vm.run_job('drive0', auto_dismiss=True)” would have worked
as well.  But since the backup job just needs waiting for a single
event, I suppose it doesn’t matter.  Just a hint in case you start
writing more iotests in the future.)

Max





reply via email to

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