Merge ~bryce/ubuntu/+source/virglrenderer:merge-v0.10.4-1-lunar into ubuntu/+source/virglrenderer:debian/sid

Proposed by Bryce Harrington
Status: Merged
Merge reported by: git-ubuntu bot
Merged at revision: b2faf6dfaf9dc1f963fa53aa9d6c4f48e0f803ca
Proposed branch: ~bryce/ubuntu/+source/virglrenderer:merge-v0.10.4-1-lunar
Merge into: ubuntu/+source/virglrenderer:debian/sid
Diff against target: 115 lines (+64/-2)
3 files modified
debian/changelog (+54/-0)
debian/control (+2/-1)
debian/rules (+8/-1)
Reviewer Review Type Date Requested Status
Christian Ehrhardt  (community) Needs Fixing
Canonical Server Reporter Pending
Canonical Server Core Reviewers Pending
Canonical Server Pending
git-ubuntu import Pending
Review via email: mp+435207@code.launchpad.net

Description of the change

All changes drop, so this is rather just a sync proposal.

There are no test cases for the package, but I've uploaded to a PPA to verify the build:
  https://launchpad.net/~bryce/+archive/ubuntu/virglrenderer-merge-v0.10.4-1

For us this is a major version update from 0.9 to 0.10, comprising about a year's worth of fairly active development upstream. It brings in a bunch of fixes including the items we've been carrying patches for.

Debian actually has two versions in sid right now, 0.10.3 and 0.10.4. The latter was just uploaded yesterday so I think it's still just in migration, rather than there being update excuse issues. However even if there are I think we should just return to staying in sync with Debian at this point.

To post a comment you must log in.
Revision history for this message
Christian Ehrhardt  (paelzer) wrote :

Old delta mapped to commits:
673f4d0c1 -> 0.10.0
95e581fd1 -> 0.10.0
b05bb61f4 -> 0.10.0

yep, all changes seem applied upstream and therefore this can be a sync.

review: Approve
Revision history for this message
Christian Ehrhardt  (paelzer) wrote :

However, while it looks good here, this recent update
- https://bugs.launchpad.net/ubuntu/+source/virglrenderer/+bug/1993453/comments/2
- https://salsa.debian.org/debian/virglrenderer/-/merge_requests/5
suggests otherwise

But having HW accel there would actually be great, we do not have the skill to support such graphic heavy things for the serve rteam. But it was mentioned it was in main before, maybe the former owning team is willing to continue owning it?
Please dig who that was and ask them.

Revision history for this message
Bryce Harrington (bryce) wrote :

On Fri, Jan 06, 2023 at 07:03:01AM -0000, Christian Ehrhardt  wrote:
> However, while it looks good here, this recent update
> - https://bugs.launchpad.net/ubuntu/+source/virglrenderer/+bug/1993453/comments/2
> - https://salsa.debian.org/debian/virglrenderer/-/merge_requests/5
> suggests otherwise
>
> But having HW accel there would actually be great, we do not have the skill to support such graphic heavy things for the serve rteam. But it was mentioned it was in main before, maybe the former owning team is willing to continue owning it?
> Please dig who that was and ask them.
> --
> https://code.launchpad.net/~bryce/ubuntu/+source/virglrenderer/+git/virglrenderer/+merge/435207
> You are the owner of ~bryce/ubuntu/+source/virglrenderer:merge-v0.10.4-1-lunar.

Thanks for the review Christian.

The video acceleration functionality is new in 0.10, so since we
currently carry 0.9, it's not a regression if we use Jeremy's changes to
drop, it at least for the near term while bugs are shaking out in
testing.

libva left main way back in Trusty, and is related part to the
X.org/Wayland stack. I don't think there was really an "owning team"
for it back then, but since it interfaces between desktop and
kernel/graphics it could fall under either team. I can reach out to
Timo Aaltonen who may be better able to advise.

98b7c57... by Bryce Harrington

  * d/rules: Disable video acceleration on Ubuntu until libva gets
    repromoted

Pulls in one of the commits from this Debian MR:
https://salsa.debian.org/debian/virglrenderer/-/merge_requests/5

b2faf6d... by Bryce Harrington

changelog

Revision history for this message
Bryce Harrington (bryce) wrote :

Timo and Jeremy confirmed there does not seem to be any activity around getting libva into main, just a general desire, and it's not clear what team would champion a MIR. Given that server team isn't the right owner for it, I think we should leave it disabled until others take initiative on it. If it becomes requested by our users we can revisit.

Revision history for this message
Christian Ehrhardt  (paelzer) wrote :

Thanks Bryce, and sorry for the back and forth that was needed here.
I've planned a candidate that we can revisit for 23.10 and until then can check opinions/responsibilities in further discussion. For this merge it should be ok as-is now.

Except that you currently have an empty 0.10.4-1ubuntu2 on top which is wrong.
Furthermore following the usual "change done on merge" style I'd expect "* d/rules: Disable video acceleration on Ubuntu until libva gets repromoted." as

* Added changes:
  - d/rules: Disable video acceleration on Ubuntu until libva gets repromoted.

Once you resolved those two I'd be +1, so consider uploading unless there is more.

review: Needs Fixing
Revision history for this message
Bryce Harrington (bryce) wrote :

Thanks, branch updated with those changes, and package uploaded to lunar:

Vcs-Git: https://git.launchpad.net/~bryce/ubuntu/+source/virglrenderer
Vcs-Git-Commit: b2faf6dfaf9dc1f963fa53aa9d6c4f48e0f803ca
Vcs-Git-Ref: refs/heads/merge-v0.10.4-1-lunar

$ dput ubuntu ../virglrenderer_0.10.4-1ubuntu1_source.changes
D: Setting host argument.
Checking signature on .changes
Checking signature on .dsc
Uploading to ubuntu (via ftp to upload.ubuntu.com):
  Uploading virglrenderer_0.10.4-1ubuntu1.dsc: done.
  Uploading virglrenderer_0.10.4.orig.tar.bz2: done.
  Uploading virglrenderer_0.10.4-1ubuntu1.debian.tar.xz: done.
  Uploading virglrenderer_0.10.4-1ubuntu1_source.buildinfo: done.
  Uploading virglrenderer_0.10.4-1ubuntu1_source.changes: done.
Successfully uploaded packages.

Preview Diff

[H/L] Next/Prev Comment, [J/K] Next/Prev File, [N/P] Next/Prev Hunk
1diff --git a/debian/changelog b/debian/changelog
2index 3625bc9..9dfecd2 100644
3--- a/debian/changelog
4+++ b/debian/changelog
5@@ -1,3 +1,24 @@
6+virglrenderer (0.10.4-1ubuntu1) lunar; urgency=medium
7+
8+ * Merge with Debian unstable (LP: #1993453).
9+ * Added changes:
10+ - d/rules: Disable video acceleration on Ubuntu until libva gets repromoted.
11+ * Dropped:
12+ - d/p/lp-1950941*: fix out of bounds check
13+ (LP #1950941)
14+ [Fixed in upstream release 0.10.0]
15+ - out-of-bounds write in read_transfer_data()
16+ + d/p/CVE-2022-0135.patch: Add test to resource OOB write
17+ and fix it in src/vrend_renderer.c, tests/test_fuzzer_formats.c.
18+ [Fixed in upstream release 0.10.0]
19+ - info leak in vrend_resource_alloc_buffer()
20+ + d/p/CVE-2022-0175.patch: clear memory when allocating a
21+ host-backed memory resource in src/vrend_renderer.c,
22+ tests/test_virgl_transfer.c.
23+ [Fixed in upstream release 0.10.0]
24+
25+ -- Bryce Harrington <bryce@canonical.com> Thu, 05 Jan 2023 06:49:25 -0800
26+
27 virglrenderer (0.10.4-1) unstable; urgency=medium
28
29 * New upstream version 0.10.4
30@@ -54,6 +75,27 @@ virglrenderer (0.10.0-1) unstable; urgency=medium
31 * Enable video backend and add dependency accordingly
32 -- Gert Wollny <gewo@debian.org> Fri, 02 Sep 2022 17:36:05 +0200
33
34+virglrenderer (0.9.1-1~exp1ubuntu2) jammy; urgency=medium
35+
36+ * SECURITY UPDATE: out-of-bounds write in read_transfer_data()
37+ - debian/patches/CVE-2022-0135.patch: Add test to resource OOB write
38+ and fix it in src/vrend_renderer.c, tests/test_fuzzer_formats.c.
39+ - CVE-2022-0135
40+ * SECURITY UPDATE: info leak in vrend_resource_alloc_buffer()
41+ - debian/patches/CVE-2022-0175.patch: clear memory when allocating a
42+ host-backed memory resource in src/vrend_renderer.c,
43+ tests/test_virgl_transfer.c.
44+ - CVE-2022-0175
45+
46+ -- Marc Deslauriers <marc.deslauriers@ubuntu.com> Mon, 28 Feb 2022 14:19:07 -0500
47+
48+virglrenderer (0.9.1-1~exp1ubuntu1) jammy; urgency=medium
49+
50+ * Merge with Debian unstable (LP: #1959175). Remaining changes:
51+ - d/p/lp-1950941*: fix out of bounds check (LP #1950941)
52+
53+ -- Utkarsh Gupta <utkarsh@ubuntu.com> Fri, 25 Feb 2022 02:23:43 +0530
54+
55 virglrenderer (0.9.1-1~exp1) experimental; urgency=medium
56
57 * New upstream version 0.9.1
58@@ -67,6 +109,18 @@ virglrenderer (0.9.0-1~exp1) experimental; urgency=medium
59
60 -- Gert Wollny <gewo@debian.org> Tue, 06 Apr 2021 15:24:55 +0200
61
62+virglrenderer (0.8.2-5ubuntu1) jammy; urgency=medium
63+
64+ * d/p/lp-1950941*: fix out of bounds check (LP: #1950941)
65+
66+ -- Christian Ehrhardt <christian.ehrhardt@canonical.com> Thu, 02 Dec 2021 08:34:48 +0100
67+
68+virglrenderer (0.8.2-5build1) impish; urgency=medium
69+
70+ * No-change rebuild to build packages with zstd compression.
71+
72+ -- Matthias Klose <doko@ubuntu.com> Thu, 07 Oct 2021 12:26:08 +0200
73+
74 virglrenderer (0.8.2-5) unstable; urgency=medium
75
76 [ Christian Ehrhardt ]
77diff --git a/debian/control b/debian/control
78index dc591e6..e7b4a0e 100644
79--- a/debian/control
80+++ b/debian/control
81@@ -1,7 +1,8 @@
82 Source: virglrenderer
83 Section: libs
84 Priority: optional
85-Maintainer: Gert Wollny <gewo@debian.org>
86+Maintainer: Ubuntu Developers <ubuntu-devel-discuss@lists.ubuntu.com>
87+XSBC-Original-Maintainer: Gert Wollny <gewo@debian.org>
88 Build-Depends: check,
89 debhelper-compat (= 13),
90 libdrm-dev,
91diff --git a/debian/rules b/debian/rules
92index 188f483..d10b629 100755
93--- a/debian/rules
94+++ b/debian/rules
95@@ -4,12 +4,19 @@ include /usr/share/dpkg/buildflags.mk
96 export DEB_BUILD_MAINT_OPTIONS = hardening=+all
97 export CK_TIMEOUT_MULTIPLIER=100
98
99+# libva is not currently in Ubuntu main
100+ifeq ($(shell dpkg-vendor --query vendor),Ubuntu)
101+VIDEO := -Dvideo=false
102+else
103+VIDEO := -Dvideo=true
104+endif
105+
106 %:
107 dh $@ --buildsystem=meson
108
109 configure-opts = \
110 -Dtests=true \
111- -Dvideo=true \
112+ $(VIDEO) \
113
114 ifneq (armel, ${DEB_HOST_ARCH})
115 # vulkan requires 32bit atomic ints, armel lacks one (#1021861)

Subscribers

People subscribed via source and target branches