diff options
author | Igor Mammedov <imammedo@redhat.com> | 2019-05-02 16:51:59 +0200 |
---|---|---|
committer | Michael S. Tsirkin <mst@redhat.com> | 2019-05-20 18:40:02 -0400 |
commit | 3dc018749d2942fdb81676ab2b5948984c961a01 (patch) | |
tree | 0c1f309cefe18e8231216dc083a9e5f06d6ad731 /hw | |
parent | ce513b46151398fc4dad853dbfc744ee3e2b7a35 (diff) | |
download | qemu-3dc018749d2942fdb81676ab2b5948984c961a01.zip qemu-3dc018749d2942fdb81676ab2b5948984c961a01.tar.gz qemu-3dc018749d2942fdb81676ab2b5948984c961a01.tar.bz2 |
tests: acpi: allow to override default accelerator
By default test cases were run with 'kvm:tcg' accelerators to speed up
tests execution. While it works for x86, were change of accelerator
doesn't affect ACPI tables, the approach doesn't works for ARM usecase
though.
In arm/virt case, KVM mode requires using 'host' cpu model, which
isn't available in TCG mode. That could be worked around with 'max'
cpu model, which works both for KVM and TCG. However in KVM mode it
is necessary to specify matching GIC version, which also could use
'max' value to automatically pick GIC version suitable for host's CPU.
Depending on host cpu type, different GIC versions would be used,
which in turn leads to different ACPI tables (APIC) generated.
As result while comparing with reference blobs, test would fail if
host's GIC version won't match the version on the host where
reference blobs where generated.
Let's keep testing simple for now and allow ARM tests run in TCG only
mode. To do so introduce 'accel' parameter in test configuration, so
test case could override default "kvm:tcg" with accelerator of choice.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <1556808723-226478-12-git-send-email-imammedo@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Diffstat (limited to 'hw')
0 files changed, 0 insertions, 0 deletions