[igt-dev] [PATCH i-g-t 0/2] Add test for hotplug validation in simulation

[Joshi, Kunal1]

Hello Kamil,

On 10/4/2023 4:06 PM, Kamil Konieczny wrote:
> Hi Juha-Pekka,
>
> On 2023-10-04 at 11:15:25 +0300, Juha-Pekka Heikkila wrote:
>> On 4.10.2023 8.48, Modem, Bhanuprakash wrote:
>>> Forgot to add Kamil to the TO/CC list.
>>>
>>> On Tue-03-10-2023 02:42 pm, Modem, Bhanuprakash wrote:
>>>> Hi Kunal,
>>>>
>>>> On Sun-01-10-2023 04:00 pm, Kunal Joshi wrote:
>>>>> Test to validate hotplug in simulation environment we
>>>>> have scratch pad register 0x4f080 from which we can
>>>>> trigger uevents such as HPD, VBLANK..etc, Idea is to
>>>>> have port mapping placed in igtrcthrough which test
>>>>> can write to 0x4f080 and check if we get hotplug
>>>> In first glance, do we really need to upstream this test? I am
>>>> pretty sure, no one (including Intel) is going to use this test from
>>>> upstream repo (We don't have any simulation environment in public
>>>> CI).
>>>>
>>>> So, we can't merge this.
>>>>
>>>> + Kamil, thoughts?
>>>>
>> Not sure why this wouldn't fit in upstream? It is valid igt test, will be of
>> interest only for Intel guys hence in in intel directory and when in
>> upstream it will not be causing issues during backports. We anyway already
>> have more generic tests taking into account if they're being run in
>> simulated environments.
>>
>> /Juha-Pekka
>>
> imho this should not be upstreamed as any user outside Intel
> will not be able to run it. Why it cannot be a subtest with
> some generic test like kms_hotplug.c with only timeout increased
> for simulation? Such test may be in upstream if it will work
> with existing hardware in CI.
>
> Regards,
> Kamil

This test will only work in presi environment, for testing hotplug in 
hardware we have kms_chamelium_hpd,
But we can have it in upstream as someday we may require to run upstream 
igt's on presi.

Any thoughts bhanu and JP?

Thanks and Regards
Kunal Joshi