[Mesa-dev] [PATCH] i965: Make intelEmitCopyBlit not truncate large strides.

[Kenneth Graunke]

On Sunday, January 22, 2017 9:07:53 AM PST Jason Ekstrand wrote:
> I think there are places where we can drop some stride checks.  We should
> do that too.  Thanks!
> 
> I thought we didn't see any blits last night.  How does this fix it?

I'm not sure...I went home, turned off Z16, and ran

./glcts --deqp-surface-height=64 --deqp-gl-config-name=rgba8888 --deqp-surface-type=fbo --deqp-case=GL45-CTS.gtf30.GL3Tests.depth_texture.depth_texture_copyteximage --deqp-surface-width=-1 --deqp-use-fbo=GL_RGBA8,GL_DEPTH24_STENCIL8,max,64

and it was definitely using the BLT path.  I do hit CPU maps, but they
come from glGetTex[Sub?]Image().  Maybe we were seeing that, and that's
why the data appeared to be totally broken - we were actually looking at
the result of the BLT gone wrong?

--Ken
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part.
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20170122/f9aa9ce4/attachment.sig>