Support Tiers

From QEMU
Revision as of 17:22, 30 October 2020 by DanielBerrange (talk | contribs) (Created page with " '''Note this article is a draft / work in progress / scratch-pad. ''' '''It is not official project policy at this time.''' == Definitions == QEMU features are classified...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Note this article is a draft / work in progress / scratch-pad.

It is not official project policy at this time.

Definitions

QEMU features are classified into a number of support tiers, which serve as guidance to users on the long term intentions wrt quality and supportability.

Tier 1
These are the top class features, suitable for use in a virtualization scenario and expected to be fully functional at all times.
MUST be built by automated CI systems prior to merge
MUST be tested by automated CI systems prior to merge
MUST be written to be robust against malicious guests
MUST have a nominated subsystem maintainer
Tier 2
These are generally good features, but have some caveats preventing them achieving the top tier. They may only be suitable for use in an emulation scenario and may not be functional except at time of release.
MUST be built by automated CI systems prior to merge
MAY NOT be tested by automated CI systems prior to merge
SHOULD be tested manually by maintainer prior to major releases
MAY NOT be written to be robust against malicious guests
MUST have a nominated subsystem maintainer
Tier 3
These features are generally not recommended to be used. They may be insecure and can be non-functional at any time, even in releases. They are liable to be removed entirely at any time.
MAY NOT be tested by automated CI systems prior to merge
MAY NOT be tested manually by maintainer prior to major releases
MAY NOT be written to be robust against malicious guests
MAY NOT have a nominated subsystem maintainer


Host architectures

x86_64 Tier 1
ppc64le Tier 2
s390x Tier 1


Guest architectures

Storage

Networking