| Age | Commit message (Collapse) | Author |
|---|
|
The pull constants require sending out to an overworked shared unit
and waiting for a response, while push constants are nicely loaded in
for us at thread dispatch time. By putting things we access in every
VS invocation there, ETQW performance improved by 2.5% +/- 1.6% (n=6).
|
|
Everything has been constant-sized until now, but constant buffer
handling changes will make us want some additional variable sized
array.
|
|
Previously for frame throttling we would wait on the first batch after
a swap before emitting another swap, because we had no hook after a
swap was emitted. This meant that if an app managed to squeeze
everything it for a frame had into one batch, it would lock-step with
the GPU. With the swapbuffers changes, we now have the entrypoint we
want.
This takes the WoW intro screen from 25% GPU idle and visibly jerky to
4-5% GPU idle and rather smooth. Other apps such as OpenArena have
run into this problem as well.
|
|
|
|
based on Maciej's r300 patch.
|
|
|
|
as per 46dc6fd3ed5ef96cda53641a97bc68c3bc104a9f
|
|
Only enabled with KMS.
|
|
Only enabled with KMS.
|
|
The codepaths in the function were almost entirely different.
|
|
|
|
Fixes piglit vp-arl-constant-array-huge-overwritten.
|
|
|
|
the new relocation for CB_COLOR0_FRAG & CB_COLOR0_TILE add 4
dwords to the default command stream. Increase the prediction
default size to take this into account
|
|
FRAG & TILE buffer are unused but still they need
to be associated with a valid relocation so that
userspace can't try to abuse them to overwritte
GART and then try to write anywhere in system
memory.
|
|
rtype enums are different, DST_REG_OUTPUT got SRC_REG_CONSTANT in some
shaders and produced invalid output/hang
as TEX output is temp register always set out src to SRC_REG_TEMPORARY
|
|
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
Reviewed-by: Corbin Simpson <MostAwesomeDude@gmail.com>
|
|
|
|
- fill in more src/dst formats
- disable depth copies for now
- set proper data formats in render target regs
- fill in additional default state
The swizzles on some of the less used mesa formats
are probably wrong.
|
|
|
|
Unfinished.
|
|
Conflicts due to DRI1 removal:
src/mesa/drivers/dri/intel/intel_context.c
src/mesa/drivers/dri/intel/intel_screen.c
|
|
On r100 we emit the indices inline so we need to account
for that in the emission size.
|
|
This bug was fixed in libdrm ages ago, port to non-kms
|
|
|
|
Depth buffer accelerated blits aren't implemented yet.
|
|
Needed to support the SwapBuffers code properly.
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
|
|
Support the new DRI2 protocol request, DRI2SwapBuffers, in both direct
and indirect rendering context. This request allows the display server
to optimize back->front swaps (e.g. through page flipping) and allows us
to more easily support other GLX features like swap interval and the OML
sync extension in DRI2.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
|
|
Conflicts:
src/mesa/drivers/dri/i965/brw_wm_emit.c
|
|
|
|
|
|
|
|
R600_OUT_BATCH_RELOC doesn't really use offset so set it
in TEX_RESOURCE2
+ typo fix
|
|
|
|
The ordering of texture components was
backwards for the floating point textures.
Signed-off-by: Pierre Ossman <pierre@ossman.eu>
|
|
Check whether the index is within bounds before accessing the array.
Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Brian Paul <brianp@vmware.com>
|
|
Seems to be the only way to stay fully portable.
|
|
|
|
Conflicts:
configs/default
src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c
src/mesa/main/version.h
|
|
Invalid assertion found by Roel Kluin <roel.kluin@gmail.com>
|
|
This can never be true.
Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
|
|
|
|
* remove-intel-dri1:
intel: intelScreenContext() is no longer used
intel: Remove remaining dri2.enabled tests
intel: Drop more cliprect bookkeeping
intel: Remove struct intel_framebuffer
intel: Remove client-side vblank code
intel: Drop intelWindowMoved()
intel: Drop batchbuffer cliprect_mode tracking
intel: Drop DRI1 static regions
intel: Use depth buffer from ctx.DrawBuffer in copypix_src_region()
intel: Drop LOCK/UNLOCK_HARDWARE()
intel: Drop DRI1 SwapBuffer implementation
intel: Drop DRI1 CopySubBuffer implementation
intel: Drop DRI1 support
Push __driDriverExtensions out of dri_util.c and into the drivers
Remove leftover __DRI{screen,drawable,context}Private references
Check for libdrm_$chipset.pc when needed
|
|
R600_OUT_BATCH_RELOC doesn't really use offset so set it
in TEX_RESOURCE2
+ typo fix
|
|
|
|
When we have integer-valued texture formats, the texture border color
must also store integer and uint values.
With GL 3.0, the new glTexParameterIiv() and glTexParameterIuiv() functions
can set the border color to int or uint values.
|
|
Conflicts:
docs/relnotes.html
src/gallium/drivers/llvmpipe/lp_tex_sample_c.c
src/gallium/drivers/r300/r300_cs.h
src/mesa/drivers/dri/i965/brw_wm_surface_state.c
src/mesa/main/enums.c
|
|
It was OK before because we proceed to clamp the value to hardware
limits, but given that other use of MaxLevel has been a trap, let's
avoid it.
|
|
Fixes piglit fbo-copyteximage.
|
|
If we ever had a non-tile-aligned tiled renderbuffer, the math was all
off. Use the existing x,y coordinates instead of trying to
reconstruct them from an incorrectly-calculated offset value.
|
