[PATCH] Split amd64_buster test on efi and pcbios

[Anton Mikanovich]

We do "init" once for both efi and pcbios run modes.

This causes unexpected behaviour when bitbake fails to run new qemu
process right after the previous was killed.

This could be also solved by small (about 0.1 sec) delay between
both `vm_start` calls in the test, but it's more correct to use
separate test for each qemu execution.

Signed-off-by: Anton Mikanovich <ami...@ilbers.de>
---
testsuite/citest.py | 3 +++
1 file changed, 3 insertions(+)

diff --git a/testsuite/citest.py b/testsuite/citest.py
index 53881d25..3a8dd69e 100755
--- a/testsuite/citest.py
+++ b/testsuite/citest.py
@@ -660,6 +660,9 @@ class VmBootTestFull(CIBaseTest):
self.init()
# test efi boot
self.vm_start('amd64', 'buster', image='isar-image-ci')
+
+ def test_amd64_buster_pcbios(self):
+ self.init()
# test pcbios boot
self.vm_start('amd64', 'buster', True, image='isar-image-ci')

--
2.34.1

[Anton Mikanovich]

28/08/2025 09:16, Anton Mikanovich wrote:
> We do "init" once for both efi and pcbios run modes.
>
> This causes unexpected behaviour when bitbake fails to run new qemu
> process right after the previous was killed.
>
> This could be also solved by small (about 0.1 sec) delay between
> both `vm_start` calls in the test, but it's more correct to use
> separate test for each qemu execution.
>
> Signed-off-by: Anton Mikanovich <ami...@ilbers.de>

Applied to next.