[igt-dev] [PATCH 2/2] i915/i915_fb_tiling: Check if device supports tiling

[Vanshidhar Konda]

On Tue, Nov 19, 2019 at 02:44:28PM -0800, Brian Welty wrote:
>
>
>On 11/19/2019 2:33 PM, Vanshidhar Konda wrote:
>> On Tue, Nov 19, 2019 at 02:23:41PM -0800, Brian Welty wrote:
>>>
>>>
>>> On 11/15/2019 3:16 PM, Vanshidhar Konda wrote:
>>>> Skip this test if the platform does not support setting tiling for frame
>>>> buffer object.
>>>>
>>>
>>> Looking over some other tests, I came across a libdrm function that uses SET_TILING,
>>> drm_intel_bo_alloc_tiled().
>>>
>>> Any chance you can expand your patch series to fix tests using that function as well?
>>
>> The change to igt doesn't apply directly to libdrm. libdrm change will
>> have to come seperately.
>
>Yes, libdrm changes would be separate.
>I only meant that until that happens, the igts could be fixed to use igt_require
>before the offending libdrm function calls.  Just as you have done here....
>so just as you have identified not to call igt_create_fb(), there are few libdrm
>functions that should not be used by igts either (at least not unless 'legacy tiling'
>is supported).

Oh yeah. I think I should do that. I'm trying to find the easiest way
to gather a list of all the tests that fail with this signature and
disable/skip them all with the same patch.

I'm currently working on hacking the BAT tests so that we can call
VLK-4351 finished. We probably need another JIRA item to work on all
the other tests that need to be fixed/skipped.

Vanshi

>
>>
>>> It is not quite as easy, I believe you maybe shouldn't disable them all but should
>>> maybe conditionally use drm_intel_bo_alloc_tiled() or drm_intel_bo_alloc().
>>> Not sure.
>>
>> I was asked to put FIXME in the patches where I'm skipping execution of these tests.
>> I was thinking that since the GET/SET_TILING feature doesn't exist anymore the only
>> solution was to just remove these tests. I'm going to start a discussion on this to
>> finalize if the change is just a FIXME or if that is the final change.
>>
>> Once we close on that discussion, I think we can fix on whether these methods
>> should fail with an error and the tests calling these need to be fixed.
>> For now things still seem to be in a limbo.
>>
>> Vanshi
>>
>>>
>>> Thanks,
>>> -Brian
>>>
>>>
>>>> Signed-off-by: Vanshidhar Konda <vanshidhar.r.konda at intel.com>
>>>> Reviewed-by: Ashutosh Dixit <ashutosh.dixit at intel.com>
>>>> ---
>>>>  tests/i915/i915_fb_tiling.c | 2 ++
>>>>  1 file changed, 2 insertions(+)
>>>>
>>>> diff --git a/tests/i915/i915_fb_tiling.c b/tests/i915/i915_fb_tiling.c
>>>> index 7d5c3f1f..4ec84962 100644
>>>> --- a/tests/i915/i915_fb_tiling.c
>>>> +++ b/tests/i915/i915_fb_tiling.c
>>>> @@ -32,6 +32,8 @@ igt_simple_main
>>>>      struct igt_fb fb;
>>>>      int ret;
>>>>
>>>> +    igt_require(gem_has_legacy_hw_tiling(drm_fd));
>>>> +
>>>>      igt_create_fb(drm_fd, 512, 512, DRM_FORMAT_XRGB8888,
>>>>                LOCAL_I915_FORMAT_MOD_X_TILED, &fb);
>>>>
>>>>