<div dir="ltr"><div>The kernel could report the true alignment from the ioctl instead of 0.</div><div><br></div><div>Marek<br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Sep 23, 2022 at 1:31 AM Christian König <<a href="mailto:christian.koenig@amd.com">christian.koenig@amd.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Am 23.09.22 um 07:28 schrieb lepton:<br>
> On Thu, Sep 22, 2022 at 10:14 PM Christian König<br>
> <<a href="mailto:christian.koenig@amd.com" target="_blank">christian.koenig@amd.com</a>> wrote:<br>
>> Am 23.09.22 um 01:04 schrieb Lepton Wu:<br>
>>> Since size has been aligned to PAGE_SIZE already, just align it<br>
>>> to PAGE_SIZE so later the buffer can be used as a texture in mesa<br>
>>> after <a href="https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fcgit.freedesktop.org%2Fmesa%2Fmesa%2Fcommit%2F%3Fid%3Df7a4051b8&data=05%7C01%7Cchristian.koenig%40amd.com%7C645f6878a7bd487588b708da9d246c4c%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637995077041120091%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=NMEAl8TByDLQFWW1d%2FaJfiGrXc4mpwL5dxNH0M0QH84%3D&reserved=0" rel="noreferrer" target="_blank">https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fcgit.freedesktop.org%2Fmesa%2Fmesa%2Fcommit%2F%3Fid%3Df7a4051b8&data=05%7C01%7Cchristian.koenig%40amd.com%7C645f6878a7bd487588b708da9d246c4c%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637995077041120091%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=NMEAl8TByDLQFWW1d%2FaJfiGrXc4mpwL5dxNH0M0QH84%3D&reserved=0</a><br>
>>> Otherwise, si_texture_create_object will fail at line<br>
>>> "buf->alignment < tex->surface.alignment"<br>
>> I don't think that those Mesa checks are a good idea in the first place.<br>
>><br>
>> The alignment value is often specified as zero when it doesn't matter<br>
>> because the minimum alignment can never be less than the page size.<br>
> Are you suggesting to change those mesa checks?<br>
<br>
Yes, the minimum alignment of allocations is always 4096 because that's <br>
the page size of the GPU.<br>
<br>
> While that can be<br>
> done, I still think a kernel side "fix" is still<br>
> useful since it doesn't hurt while can fix issues for some versions of mesa.<br>
<br>
No, we have tons of places where we don't specify and alignment for <br>
buffers because it never mattered. I certainly don't want to fix all of <br>
those.<br>
<br>
Regards,<br>
Christian.<br>
<br>
>> Christian.<br>
>><br>
>>> Signed-off-by: Lepton Wu <<a href="mailto:ytht.net@gmail.com" target="_blank">ytht.net@gmail.com</a>><br>
>>> ---<br>
>>> drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 2 +-<br>
>>> 1 file changed, 1 insertion(+), 1 deletion(-)<br>
>>><br>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c<br>
>>> index 8ef31d687ef3b..8dca0c920d3ce 100644<br>
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c<br>
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c<br>
>>> @@ -928,7 +928,7 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,<br>
>>> args->size = ALIGN(args->size, PAGE_SIZE);<br>
>>> domain = amdgpu_bo_get_preferred_domain(adev,<br>
>>> amdgpu_display_supported_domains(adev, flags));<br>
>>> - r = amdgpu_gem_object_create(adev, args->size, 0, domain, flags,<br>
>>> + r = amdgpu_gem_object_create(adev, args->size, PAGE_SIZE, domain, flags,<br>
>>> ttm_bo_type_device, NULL, &gobj);<br>
>>> if (r)<br>
>>> return -ENOMEM;<br>
<br>
</blockquote></div>