We've had some regressions (https://bugzilla.kernel.org/show_bug.cgi?id=210983 , although this happened only with silent streaming enabled) caused by lack of coverage for display audio hotplug cases.
One idea to test for these cases is to use the i915 EDID injection capability as is done in following test case code for i915:
https://gitlab.freedesktop.org/drm/igt-gpu-tools/-/blob/master/tests/kms_hdmi_inject.c
We could share this logic in SOF tests and have scenarios where monitor is repeatedly connected and disconnected and ensure driver reports the status correctly via ALSA mixer, and that driver remains operational. To check for resource leakage, a large number of hotplug events should be executed.
We've had some regressions (https://bugzilla.kernel.org/show_bug.cgi?id=210983 , although this happened only with silent streaming enabled) caused by lack of coverage for display audio hotplug cases.
One idea to test for these cases is to use the i915 EDID injection capability as is done in following test case code for i915:
https://gitlab.freedesktop.org/drm/igt-gpu-tools/-/blob/master/tests/kms_hdmi_inject.c
We could share this logic in SOF tests and have scenarios where monitor is repeatedly connected and disconnected and ensure driver reports the status correctly via ALSA mixer, and that driver remains operational. To check for resource leakage, a large number of hotplug events should be executed.