| Age | Commit message (Collapse) | Author |
|---|
|
Try and fit user buffers into a small GART scratch area at validate time,
instead of going to a lot of effort to fit these (mostly) use-once-and-discard
objects into VRAM.
|
|
|
|
|
|
|
|
|
|
|
|
If an app never called glViewport, the viewport size was always 0 by 0 pixels.
Now pass initial size to st_create_framebuffer() and initialize the viewport
and scissor bounds in st_make_current().
This could also be fixed by ensuring the gl_framebuffers passed to
_mesa_make_current() were initialized to the right size. But that involves
allocating the renderbuffers/pipe_surfaces earlier and that runs into some
other issues ATM.
Also remove obsolete createRenderbuffers param to st_create_framebuffer().
|
|
It's annoying for several reasons, especially in its current form. May
possibly be reincarnated later (DDX depends on it these days), but in
not quite the same way.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
st_create_framebuffer() now takes pipe_formats for the color, depth, stencil
buffers.
This avoids a round-about chain of calls to pipe->is_format_supported() for
window renderbuffers (their format never changes).
Renderbuffer format selection code in st_format.c is simpler now too.
|
|
|
|
|
|
we did some small changes in the beginning of the gallium3d
lifecycle in i915tex which is not in master anymore and just
causes problems when doing merges. getting rid of the headache
by just nuking it here
|
|
Some things that worked before are now broken, there's an "XXX:" around one
of the culprits in the GL state tracker so hopefully it'll get fixed soon!
|
|
Conflicts:
src/mesa/pipe/Makefile
|
|
Add a 'CheckQuery()' driver callback to mesa to check query completion.
Make pipe_query an opaque type.
Rework softpipe queries, support overlapping occlusion queries.
|
|
|
|
|
|
Conflicts:
src/mesa/drivers/dri/Makefile.template
|
|
surface, etc.
Additional types may be added in the future.
|
|
These are useless or even harmful due to referencing symbols no longer
available in the Gallium build.
|
|
|
|
|
|
In some situations WAIT_RING would get called while the GPU was processing
data from outside the "master" ring, which caused dma.free to be updated
incorrectly and much fun was had.
WAIT_RING will now wait until it reads GET values from within the main ring
buffer before calculating free space.
|
|
|
|
|
|
Conflicts:
src/mesa/drivers/dri/intel_winsys/intel_winsys_pipe.c
src/mesa/pipe/Makefile
src/mesa/pipe/draw/draw_vertex_fetch.c
src/mesa/pipe/i915simple/i915_texture.c
src/mesa/pipe/softpipe/sp_texture.c
src/mesa/pipe/xlib/xm_winsys.c
src/mesa/state_tracker/st_cb_fbo.c
|
|
|
|
|
|
|
|
Unfortunately we still need to keep the forced flushes as mesa's vbo path
is still insane. But, at least this is a little less brutal. The bo code will
detect when it needs to do a sync.
|
|
|
|
|
|
|
|
This allows making a bunch of functions static, and removes a state
tracker dependency on driverfuncs.c
|
|
This makes debugging a _lot_ easier.
In gdb, "print format" used to display 613570600, now you see PIPE_FORMAT_A8R8G8B8_UNORM.
|
|
For example, replace PIPE_FORMAT_U_A8_R8_G8_B8 with PIPE_FORMAT_A8R8G8B8_UNORM
|
|
Directly use struct pipe_buffer_handle for storage and struct pipe_surface for
(un)mapping.
|
|
pipe->get_tex_surface() has to be used for access to texture image data.
|
|
|
|
This allows making a bunch of functions static, and removes a state
tracker dependency on driverfuncs.c
|
|
This makes debugging a _lot_ easier.
In gdb, "print format" used to display 613570600, now you see PIPE_FORMAT_A8R8G8B8_UNORM.
|
|
For example, replace PIPE_FORMAT_U_A8_R8_G8_B8 with PIPE_FORMAT_A8R8G8B8_UNORM
|
|
Directly use struct pipe_buffer_handle for storage and struct pipe_surface for
(un)mapping.
|
