[igt-dev] [PATCH i-g-t 17/18] xe_mmap: skip VRAM tests if no VRAM is found

[Das, Nirmoy]

On 3/13/2023 7:48 AM, Zbigniew Kempczyński wrote:
> On Fri, Mar 10, 2023 at 11:52:15AM +0100, Das, Nirmoy wrote:
>> On 3/10/2023 9:23 AM, Zbigniew Kempczyński wrote:
>>> From: Mauro Carvalho Chehab <mchehab at kernel.org>
>>>
>>> Instead of producing an error, if there's no VRAM, the test
>>> should be skipped. There's a logic there that would be doing
>>> that, except that it doesn't really work :-)
>>>
>>> Signed-off-by: Mauro Carvalho Chehab <mchehab at kernel.org>
>>> Acked-by: Zbigniew Kempczyński <zbigniew.kempczynski at intel.com>
>>> ---
>>>    tests/xe/xe_mmap.c | 3 +--
>>>    1 file changed, 1 insertion(+), 2 deletions(-)
>>>
>>> diff --git a/tests/xe/xe_mmap.c b/tests/xe/xe_mmap.c
>>> index 9dce6adbe2..6b313a1895 100644
>>> --- a/tests/xe/xe_mmap.c
>>> +++ b/tests/xe/xe_mmap.c
>>> @@ -42,8 +42,7 @@ test_mmap(int fd, uint32_t flags)
>>>    	uint64_t mmo;
>>>    	void *map;
>>> -	if (flags & vram_memory(fd, 0))
>>> -		igt_require(xe_has_vram(fd));
>> This check should go to the igt_main instead.
> You mean to igt_subtest()? In igt_main() you can't use fd.

Ah okay.  Yes then we can't do much.

> IMO change is ok, if there's no region on xe fd flags passed == 0
> what will cause skip on require.


I am fine with the change. It was that xe_has_vram() seemed more readable.


Thanks,

Nirmoy

>
> --
> Zbigniew
>
>>
>> Nirmoy
>>
>>> +	igt_require_f(flags, "Device doesn't support such memory region\n");
>>>    	bo = xe_bo_create_flags(fd, 0, 4096, flags);
>>>    	mmo = xe_bo_mmap_offset(fd, bo);