If the size is lower than the alignment, we must use the alignment to
select the bucket.
Otherwise, the selected bucket won't be able to satisfy our request
and will fail.
Signed-off-by: José Fonseca <jfonseca@vmware.com>
The cache manager stores buffers with a reference count that dropped to 0.
pipe_reference asserts in this case on debug builds,
so use pipe_reference_init instead.
Signed-off-by: José Fonseca <jfonseca@vmware.com>
It was added to skip checking EGLDisplay, EGLSurface, and etc. It is
never defined and the spec does not allow the checks to be skipped.
Remove it for good.
This fixes a regression with Lightsmark, where more compact TGSI from Mesa
was causing a zero mask MOV to be emitted for shadow map compare, causing
problems in some backends.
Add a few more assertions to catch cases like this.
Fixes crash in Homeworld2 which tries to create a 14MB buffer, because we
now avoid creating GMR buffers larger than 8MB to ensure progress given
we have only a 16MB pool.
Do this by extending the nearest range to cover the new range.
This fixes an access fault in Call of Duty which was doing many disjoint
glBufferSubData calls.
This fixes incorrect Z position of glBitmap, glDraw/CopyPixels for the
svga driver. Now we use 0.5, 0.5 as is typical for ordinary 3D rendering.
(cherry picked from commit bcd561c667)
The effect of this was that all objects were aligned to 128 bytes
on all generations, rather than just gen2.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This reverts commit fff5be8e7b.
Probably went too soon with this, dileX reported OA not working for him
it works here fine, but the optimisations I wanted aren't working properly
yet so I'll fix that now.
Signed-off-by: Dave Airlie <airlied@redhat.com>
This creates a cleaner winsys and drop the simple screen stuff.
It makes r300g use pb_bufmgr structs and adds usage of the cached
bufmgr for vertex/index buffers.
It also avoids mapping too often.
I'm not 100% sure this is perfect but it won't find its own bugs.
Signed-off-by: Dave Airlie <airlied@redhat.com>
The path for VAP_ALT_NUM_VERTICES is also in place (and tested)
but not enabled by default due to the missing support of this reg
in the upstream kernel.
Also, a non-zero BUFFER_BASE in the INDX_BUFFER packet3 hangs the machine.
Am I missing something? Because of this, only draw_arrays can render
more than 65535 vertices without the use of VAP_ALT_NUM_VERTICES.
The surface_format field is 9 bits wide which has a maximum value
of 511. BRW_SURFACEFORMAT_INVALID has a value of 4095 which exceeds
the width of the surface_format field.
This patch asserts that the format is not BRW_SURFACEFORMAT_INVALID
before assignment to the surface_format field.
