[Mesa-dev] [PATCH] i965: fix dma_buf import with non-zero offset.

[Eric Anholt]

Gwenole Beauchesne <gb.devel at gmail.com> writes:

> Hi,
>
> 2014-03-11 19:06 GMT+01:00 Eric Anholt <eric at anholt.net>:
>> Gwenole Beauchesne <gb.devel at gmail.com> writes:
>>
>>> Fix eglCreateImage() from a packed dma_buf surface with a non-zero offset
>>> to pixels data. In particular, this fixes support for planar YUV surfaces
>>> when they are individually mapped on a per-plane basis, i.e. when the
>>> OES_EGL_image_external is not used and user application wants to use its
>>> own shader code for composition, or processing on individual plane (OCL).
>>>
>>> Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne at intel.com>
>>> ---
>>>  src/mesa/drivers/dri/i965/intel_screen.c |    9 +++++++++
>>>  1 file changed, 9 insertions(+)
>>>
>>> diff --git a/src/mesa/drivers/dri/i965/intel_screen.c b/src/mesa/drivers/dri/i965/intel_screen.c
>>> index 05cf6b1..81a8330 100644
>>> --- a/src/mesa/drivers/dri/i965/intel_screen.c
>>> +++ b/src/mesa/drivers/dri/i965/intel_screen.c
>>> @@ -651,6 +651,7 @@ intel_create_image_from_fds(__DRIscreen *screen,
>>>  {
>>>     struct intel_screen *intelScreen = screen->driverPrivate;
>>>     struct intel_image_format *f;
>>> +   uint32_t mask_x, mask_y;
>>>     __DRIimage *image;
>>>     int i, index;
>>>
>>> @@ -684,6 +685,14 @@ intel_create_image_from_fds(__DRIscreen *screen,
>>>        image->strides[index] = strides[index];
>>>     }
>>>
>>> +   if (f->nplanes == 1) {
>>> +      image->offset = image->offsets[0];
>>> +      intel_region_get_tile_masks(image->region, &mask_x, &mask_y, false);
>>> +      if (image->offset & mask_x)
>>> +         _mesa_warning(NULL,
>>> +                       "intel_create_image_from_fds: offset not on tile boundary");
>>> +   }
>>
>> I'm confused here.  offset is a byte offset, right?  If so, then a tile
>> boundary is 4096, not the byte width of a row in a tile.
>>
>> Are you planning on supporting non-4096-aligned imports?  I'm hoping
>> not.
>
> You are right.
>
> Would the following be OK then?
>
> image->offset = image->offsets[0];
> if (image->region->tiling != I915_TILING_NONE && (image->offset & 0xfff))
>    _mesa_warning(NULL, "intel_create_image_from_fds: offset not on
> tile boundary");
>
> (and we might also fix from_planar() along the way?)

Looks good.  Long term, it would be nice if we could throw proper errors
of some sort from these functions, but using _mesa_warning() isn't new
here.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 818 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/mesa-dev/attachments/20140311/afa91f01/attachment.pgp>