On 18/10/17 09:37, Miklós Máté wrote:
On 16/10/17 01:59, Timothy Arceri wrote:
On 16/10/17 04:10, Miklós Máté wrote:
Hi,
I'd like to ask you to revert this change.
As Ian Romanick pointed out this makes ATI_fs behave like ARB_fp,
however there is a major difference between the two: with ATI_fs
there is no way of knowing the texture targets until the draw call.
When an ATI_fs is created, st_init_atifs_prog() sets every texture
target to TEXTURE_2D_BIT, and st_fixup_atifs() sets the correct one,
but unfortunately _mesa_update_texture_state() is called between
them. After this patch update_program_texture_state() validates the
texture targets to match the bound texture units, and thus rejects
sampling from cube maps. This results in broken rendering in Knights
of the Old Republic.
Before reverting anything are you able to create a piglit test that
reproduces the issue, so that this doesn't get broken again in future?
I've been meaning to create piglit tests for ATI_fs ever since I
implemented it for gallium, I just couldn't find the time to do so. This
regression a great incentive to start working on it.
That would be great. Thanks Miklós.
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
