[igt-dev] [RFC incomplete PATCH i-g-t] lib/igt_kms: Reduce probing and allow initial hotplug during igt_display_init()

[Maarten Lankhorst]

Op 23-03-18 om 14:03 schreef Ville Syrjälä:
> On Fri, Mar 23, 2018 at 01:48:28PM +0100, Maarten Lankhorst wrote:
>> When booting with fbdev disabled, all outputs are set to UNKNOWNCONNECTION.
>> In this case we have to reprobe, but in all others we can safely assume that
>> the output connector status is valid.
>>
>> Enumerate all outputs, and for good measure recheck after we probed, to add
>> and DP-MST we may have missed.
>>
>> Signed-off-by: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
>> Cc: Lyude <lyude at redhat.com>
>> Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan at intel.com>
>> ---
>> Unfortunately we lack a way to determine whether the probing will enumerate
>> additional DP-MST displays, so the only way right now to test DP-MST would be
>> to add a sleep(5) in igt_display_check_hotplug.
>>
>> Any thoughts how we can make this more reliable for DP-MST?
> When do we fill out the debugfs mst topology information?
I have no idea..

drm_dp_mst_link_probe_work seems to be called from irq, I don't think we initially do discovery, maybe we should..