~paelzer/ubuntu/+source/libvirt:lp-1899180-flaky-smoke-lxc-FOCAL

Last commit made on 2020-12-08
Get this branch:
git clone -b lp-1899180-flaky-smoke-lxc-FOCAL https://git.launchpad.net/~paelzer/ubuntu/+source/libvirt
Only Christian Ehrhardt  can upload to this branch. If you are Christian Ehrhardt  please log in for upload directions.

Branch merges

Branch information

Name:
lp-1899180-flaky-smoke-lxc-FOCAL
Repository:
lp:~paelzer/ubuntu/+source/libvirt

Recent commits

dc7889e... by Christian Ehrhardt 

changelog: Improve flaky smoke-lxc test (LP: #1899180)

Signed-off-by: Christian Ehrhardt <email address hidden>

32cdace... by Christian Ehrhardt 

d/t/control, d/t/smoke-lxc: retry service restart and skip test if failing; Was flaky on some release/architectures (LP: #1899180)

Signed-off-by: Christian Ehrhardt <email address hidden>

9180f76... by Christian Ehrhardt 

d/t/smoke-lxc: retry check_domain being flaky on LP-Infra (LP: #1899180)

Signed-off-by: Christian Ehrhardt <email address hidden>

84b7458... by Christian Ehrhardt 

changelog: add named types and definitions for EPYC-Rome chips (LP: #1887490)

Signed-off-by: Christian Ehrhardt <email address hidden>

dc322c5... by Christian Ehrhardt 

d/p/ubuntu/lp-1887490-*: add named types and definitions for EPYC-Rome chips (LP: #1887490)

This backports a set of things:
- the actual changes for EPYC Rome that we need
- self-test that affect qemu 4.2 (=Focal) to ensure integrity at build time
- related other updates HW-feature-bit definitions needed by the EPYC Rome
  (those we usually SRU as well every one in a while as once requested, so
  doing them here as dependency of Rome is just as fine)
- finally the build system changed and we needed to modify Makefiles
  to reflect what upstream did in meson (Groovy already had this).
- we were not adding the feature to specify smp-dies nor the i8042 attribute
  of qemu on this SRU as it would change behavior with new qemu (for these
  users should go to >=20.10 please)
- I did not include the changes related to the Cooperlake model.
  Different to the others - that would not just be bit-definitions, named-types
  and tests but also logic changes. Those would introduce extra regression
  risk and only will be considered once a case outweighting them was made.
  See https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1887490/comments/17

I know it seems huge, but the actual non-self-test changes are only:
 src/cpu_map/index.xml | 1
 src/cpu_map/x86_EPYC-Rome.xml | 85
 src/cpu_map/x86_features.xml | 60
 src/Makefile.in | 1
 src/cpu_map/Makefile.inc.am | 1

Signed-off-by: Christian Ehrhardt <email address hidden>

2a8bae3... by Christian Ehrhardt 

changelog: fix up LP bug reference

Signed-off-by: Christian Ehrhardt <email address hidden>

0a9f552... by Christian Ehrhardt 

changelog: avoid stale libvirt capability cache (LP: 1874647)

Signed-off-by: Christian Ehrhardt <email address hidden>

f2a03c7... by Christian Ehrhardt 

avoid stale libvirt capability cache (LP: 1874647)
  - when host cpu changes (e.g. nested with different configuration)
  - when s390x protvirt or AMD SEV changes

Signed-off-by: Christian Ehrhardt <email address hidden>

66ba2c2... by Marc Deslauriers

6.0.0-0ubuntu8.3 (patches unapplied)

Imported using git-ubuntu import.

3538203... by Christian Ehrhardt 

changelog: enable attr support to store XATTR labels (LP: #691590)

Signed-off-by: Christian Ehrhardt <email address hidden>