Brian Paul [Wed, 13 Jan 2010 23:46:53 +0000 (16:46 -0700)]
gallium: remove forward struct decl
Other struct types (like pipe_texture) are referenced before they're
defined so no reason to special-case this one.
Jesse Barnes [Thu, 14 Jan 2010 00:01:51 +0000 (19:01 -0500)]
DRI2: don't send DRI2 requests the server doesn't know about
At init time, NULL out the hooks that send the server new requests.
DRI2 never supported these extensions, so this preserves old behavior on
old servers with new Mesa.
Jesse Barnes [Wed, 13 Jan 2010 23:42:55 +0000 (18:42 -0500)]
GLX/DRI2: pass X drawable ID instead
These happened to be the same in my testing since I was using apps that
depend on the GLX 1.2 behavior where X and GLX drawables are
interchangeable.
Jesse Barnes [Wed, 13 Jan 2010 20:48:14 +0000 (15:48 -0500)]
GLX: use glproto CFLAGS etc. when building
Roll the glproto check into the dri2proto and libdrm check so we get the
proper CFLAGS and LDFLAGS for the package.
Corbin Simpson [Wed, 13 Jan 2010 09:41:31 +0000 (01:41 -0800)]
r300g: Account for CS space used per atom.
Oh yeah, those atoms are startin' to pay off. The main obstacle now
for OA playability is the absurdly low default mouse sensitivity, IMO.
Not totally smooth yet, but getting there.
Corbin Simpson [Wed, 13 Jan 2010 09:07:49 +0000 (01:07 -0800)]
r300g: Two dirty state optimizations.
Corbin Simpson [Wed, 13 Jan 2010 08:57:03 +0000 (00:57 -0800)]
r300g: DCE in emit.
Corbin Simpson [Wed, 13 Jan 2010 06:22:22 +0000 (22:22 -0800)]
r300g: Lower the immd mode threshold.
Corbin Simpson [Wed, 13 Jan 2010 05:52:00 +0000 (21:52 -0800)]
r300g: Fix is_r400 status for rs4xx and rc4xx.
As suggested by agd5f.
Corbin Simpson [Wed, 13 Jan 2010 05:50:14 +0000 (21:50 -0800)]
r300g: Switch to immediate mode for tiny renders.
Useful for e.g. blitter.
Marek Olšák [Fri, 8 Jan 2010 13:09:57 +0000 (14:09 +0100)]
r300g: emit blend LTE/GTE thresholds on >=RV350
As per classic r300.
Marek Olšák [Sun, 10 Jan 2010 06:09:56 +0000 (07:09 +0100)]
r300g: update the texture initialization so that it respects tiling
This is still work-in-progress and tiling is not enabled by default.
Marek Olšák [Sun, 10 Jan 2010 01:29:02 +0000 (02:29 +0100)]
r300g: add emission of texture tiling bits
Marek Olšák [Fri, 8 Jan 2010 13:08:02 +0000 (14:08 +0100)]
r300g: add is_r400 flag
r4xx has some additional fragment shader registers compared to r3xx.
Marek Olšák [Sat, 9 Jan 2010 18:23:21 +0000 (19:23 +0100)]
r300g: silence a warning
Marek Olšák [Tue, 12 Jan 2010 22:55:37 +0000 (23:55 +0100)]
r300g: only emit scissor when needed
Reverting some bits from
ce1c493ff8fad4b62e2b66f06636ac6560a6e0ad.
Given the latest fixes, it's not needed to always emit scissor, really.
Marek Olšák [Tue, 12 Jan 2010 22:23:05 +0000 (23:23 +0100)]
Revert "r300g: Placate kernel checker by explicitly disabling depth test."
This reverts commit
e920ee23b32f6031a7b8527b540566e7ada6af8a.
Assuming the FB state is not dirty and the DSA state is, then the depth test
will be re-enabled, making the kernel checker angry. Reverting the commit
fixes piglit/glsl-bug-22603.
Marek Olšák [Tue, 12 Jan 2010 20:50:10 +0000 (21:50 +0100)]
r300g: mark all states as dirty after flush
It fixes almost all regressions introduced lately.
Marcin Slusarz [Tue, 12 Jan 2010 16:39:23 +0000 (17:39 +0100)]
nv50: fix 2 off by one memory leaks (nv50_miptree_level->image_offset)
Marcin Slusarz [Tue, 12 Jan 2010 16:38:52 +0000 (17:38 +0100)]
nv50: fix memory leak on nv50_pc free
José Fonseca [Tue, 12 Jan 2010 12:15:24 +0000 (12:15 +0000)]
gallium: Rename ALIGN_STACK -> PIPE_ALIGN_STACK for consistency.
José Fonseca [Tue, 12 Jan 2010 11:47:37 +0000 (11:47 +0000)]
gallium: Simplify PIPE_ALIGN_VAR.
gcc allows pre-fix variable attributes.
Suggested by Ian Romanick.
José Fonseca [Sun, 10 Jan 2010 12:58:11 +0000 (12:58 +0000)]
gallium: Generalize the alignment macros to other compilers and any alignment.
Corbin Simpson [Tue, 12 Jan 2010 10:09:07 +0000 (02:09 -0800)]
r300g: Always emit scissors.
Save some code, improve FPS, and fix piglit tests. Everybody wins.
Chia-I Wu [Tue, 12 Jan 2010 09:55:53 +0000 (17:55 +0800)]
st/egl_g3d: Get the stapi from the driver.
egl_g3d_get_st was called because the driver argument was not given
before. It can be fixed now.
Chia-I Wu [Tue, 12 Jan 2010 09:44:03 +0000 (17:44 +0800)]
egl: Add _EGLDriver as the first argument to GetProcAddress.
The rest of the driver API has it as the first argument. It should be
there so that a driver has access to itself.
Chia-I Wu [Tue, 12 Jan 2010 09:49:43 +0000 (17:49 +0800)]
egl_xdri: Fix build breakage.
driScreen->swapBuffers takes 3 more arguments since
daf7fe69f7bd0caa955d30b43fc35b7ce0069b6b.
Chia-I Wu [Tue, 12 Jan 2010 06:01:56 +0000 (14:01 +0800)]
st/egl_g3d: Improve comments to the native display interface.
Document the future of __GLcontextModes. Improve the description of
surface validation and native_flush_frontbuffer.
Corbin Simpson [Tue, 12 Jan 2010 04:52:19 +0000 (20:52 -0800)]
r300g: Add back dirty state check.
From
07ea7e6c80. This is the only questionable part of that commit,
AFAICT...
Corbin Simpson [Tue, 12 Jan 2010 04:41:57 +0000 (20:41 -0800)]
r300g: Placate kernel checker by explicitly disabling depth test.
This probably should disable stencil, too, if the kernel cares enough.
Note: When atomized, framebuffer setup should go towards the end anyway,
but it *must* follow these test setups anyway.
Corbin Simpson [Tue, 12 Jan 2010 04:27:54 +0000 (20:27 -0800)]
r300g: Avoid segfault on binding null RS state.
Brian Paul [Tue, 12 Jan 2010 04:17:49 +0000 (21:17 -0700)]
st/mesa: removed unused var
Brian Paul [Tue, 12 Jan 2010 02:04:11 +0000 (19:04 -0700)]
st/mesa: comments, whitespace
Chia-I Wu [Sun, 10 Jan 2010 15:08:39 +0000 (23:08 +0800)]
winsys/drm: Use egl_g3d to build EGL drivers.
The new EGL drivers use egl_g3d and respect EGL_DISPLAYS. They are
named after the display supported and the DRM name.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sun, 3 Jan 2010 11:24:04 +0000 (19:24 +0800)]
st/egl_g3d: Add support for KMS native display.
The KMS native display implements the modeset interface using DRM
modesetting.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Mon, 4 Jan 2010 04:01:38 +0000 (12:01 +0800)]
st/egl_g3d: Add support for EGL_MESA_screen_surface.
This is implemented through the modeset interface.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sun, 3 Jan 2010 13:27:22 +0000 (21:27 +0800)]
st/egl_g3d: Update the native display interface for modesetting.
The new interface can be used to implement EGL_MESA_screen_surface.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sun, 10 Jan 2010 17:23:01 +0000 (01:23 +0800)]
st/egl_g3d: New EGL state tracker that uses Gallium.
This new (intermediate) EGL state tracker is the base work for EGL
drivers that uses Gallium. It makes it easier to support new window
systems.
Currently, there is support only for X11. This driver supports multiple
APIs (OpenVG, OpenGL, ...) and supports hardware acceleration through
winsys/drm.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sat, 2 Jan 2010 14:06:42 +0000 (22:06 +0800)]
winsys/drm: Add name field to drm_api.
The name can be used for driver selection. It has the same name as the
DRI driver does right now.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Mon, 4 Jan 2010 04:56:04 +0000 (12:56 +0800)]
egl: EGL_SCREEN_BIT_MESA is a valid bit.
Update _eglValidateConfig so that it passes the test.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Mon, 4 Jan 2010 05:32:50 +0000 (13:32 +0800)]
progs/egl: Fix screen surface demos.
Set the screen surface size to the mode size, as the spec requires the
screen surface size to be larger than the mode size. Besides, bind the
API to OpenGL as they are written in it.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sat, 2 Jan 2010 13:57:16 +0000 (21:57 +0800)]
progs/egl: glGenTextures is called too early in xeglbindtex.
It should be called after eglMakeCurrent.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sat, 2 Jan 2010 13:56:13 +0000 (21:56 +0800)]
progs: Check the number of configs returned by eglChooseConfig.
A successful eglChooseConfig call does not imply there are valid
configs.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sat, 2 Jan 2010 13:51:46 +0000 (21:51 +0800)]
progs: Add missing EGL_RENDERABLE_TYPE for EGL demos.
The default value of EGL_RENDERABLE_TYPE is EGL_OPENGL_ES_BIT. The
proper values should be specified if we are not using OpenGL ES.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sat, 2 Jan 2010 13:32:33 +0000 (21:32 +0800)]
st/mesa: Remove st_api.c.
st_api.c is supposed to define st_api_OpenGL to advertise OpenGL
support. However, the linker discards the symbol because it has no
user. It is better to leave this to other state trackers that link to
libmesagallium.a.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Mon, 11 Jan 2010 04:03:07 +0000 (12:03 +0800)]
st/vega: Make it compatible with st_public.h.
It missed a stub for st_get_proc_address, and st_make_current should
return a boolean.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sat, 2 Jan 2010 13:25:34 +0000 (21:25 +0800)]
st/vega: Advertise OpenVG support.
Define st_api_OpenVG to advertise OpenVG support.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Jesse Barnes [Mon, 11 Jan 2010 22:28:10 +0000 (17:28 -0500)]
GLX: check glproto version
This was a silent dependency before, but now we really need 1.4.11 for
INTEL_swap_event support, so check for it.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Jesse Barnes [Mon, 11 Jan 2010 20:23:29 +0000 (15:23 -0500)]
Merge branch 'master' of ssh://people.freedesktop.org/~jbarnes/mesa
Conflicts due to DRI1 removal:
src/mesa/drivers/dri/intel/intel_context.c
src/mesa/drivers/dri/intel/intel_screen.c
Chia-I Wu [Sun, 10 Jan 2010 16:13:41 +0000 (00:13 +0800)]
st/xlib: Mark GLX functions as public.
Mark all functions start with glX as public.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sun, 10 Jan 2010 16:06:45 +0000 (00:06 +0800)]
st: Mark functions in st_public.h and vg_tracker.h as public.
These functions are the API of Gallium state tracker, and are used by
EGL.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sun, 10 Jan 2010 16:00:52 +0000 (00:00 +0800)]
gallium: Add PUBLIC macro for function visibility.
As the default build has -fvisibility=hidden, add a macro to control the
visibility.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Sun, 10 Jan 2010 16:03:21 +0000 (00:03 +0800)]
st/vega: Fix breakage from -fvisibility=hidden.
Mark VG and VGU functions as public.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
José Fonseca [Mon, 11 Jan 2010 14:00:31 +0000 (14:00 +0000)]
python: Integrate with llvmpipe if possible.
Temporary. We should soon have the ability to integrate with any driver
via glx extensions.
José Fonseca [Mon, 11 Jan 2010 13:57:12 +0000 (13:57 +0000)]
python: Mask out PIPE_TEXTURE_USAGE_DISPLAY_TARGET usage.
As displaying surfaces directly isn't supported by the python state
tracker.
José Fonseca [Mon, 11 Jan 2010 13:56:09 +0000 (13:56 +0000)]
python: Prefer PIPE_FORMAT_Z32_UNORM.
José Fonseca [Mon, 11 Jan 2010 13:55:24 +0000 (13:55 +0000)]
python: Set PIPE_TEXTURE_USAGE_RENDER_TARGET instead of DISPLAY_TARGET.
As the python state tracker is not integrated with window system an
cannot present surfaces directly.
Corbin Simpson [Mon, 11 Jan 2010 13:39:56 +0000 (05:39 -0800)]
util: Set flatshade for blitter.
It was bugging me. +1 FPS on softpipe.
Corbin Simpson [Mon, 11 Jan 2010 13:31:54 +0000 (05:31 -0800)]
radeong: Check for DONTBLOCK before flushing, remove dead code.
Keith Whitwell [Mon, 11 Jan 2010 10:56:41 +0000 (10:56 +0000)]
util/blitter: remove unused variable
Luca Barbieri [Sun, 10 Jan 2010 20:04:21 +0000 (12:04 -0800)]
Fix compressed texture loads for non-minimal pitches
The current glCompressedTexImage support in the state tracker assumes
that compressed textures have minimal pitch.
However, in some cases this is not true, such as for mipmaps of non-POT
compressed textures on nVidia hardware.
This patch adds a check and does a memcpy for each line instead of the
whole image in that case.
Signed-off-by: Keith Whitwell <keithw@vmware.com>
Tweaks for C90 compilation.
Corbin Simpson [Mon, 11 Jan 2010 07:38:05 +0000 (23:38 -0800)]
r300g: Don't bother testing for empty scissors.
If somebody goes through that much effort, they probably intended it.
So humor them. :3
Corbin Simpson [Mon, 11 Jan 2010 07:30:01 +0000 (23:30 -0800)]
util: Apply Brian's suggested blit improvements.
copypix works just fine.
Dave Airlie [Thu, 7 Jan 2010 01:45:48 +0000 (11:45 +1000)]
gallium: u_blitter add overlapping blit support.
the xorg state tracker really wants the driver to handle overlapping
blits, and r300 uses u_blitter for blits. This patch adds overlapping
blit support via a temporary surface when its required.
Corbin Simpson [Mon, 11 Jan 2010 07:02:48 +0000 (23:02 -0800)]
r300g: Atomize scissors.
Argfl. Some of this code is so questionable.
Corbin Simpson [Mon, 11 Jan 2010 06:32:58 +0000 (22:32 -0800)]
r300g: Move TCL bypass switch to main context.
Corbin Simpson [Mon, 11 Jan 2010 02:07:28 +0000 (18:07 -0800)]
r300g: Atomize viewport.
Goddammit, some of these hax are really annoying.
Corbin Simpson [Mon, 11 Jan 2010 01:51:40 +0000 (17:51 -0800)]
r300g: Move ROPCNTL to the top of the emit order.
According to the docs, this decreases stalls, and indeed we get a tiny
bit more glxgears from it.
Corbin Simpson [Mon, 11 Jan 2010 01:42:06 +0000 (17:42 -0800)]
r300g: Don't avoid rewriting ZTOP.
Simplify the code. Added a comment to keep me from doing it again.
Corbin Simpson [Sun, 10 Jan 2010 22:29:32 +0000 (14:29 -0800)]
r300g: Atomize rasterizer.
I want to stab things now.
Marcin Kościelnicki [Mon, 11 Jan 2010 05:15:38 +0000 (05:15 +0000)]
nouveau: Make winsys recognise the original NV50.
Dave Airlie [Mon, 11 Jan 2010 03:21:47 +0000 (13:21 +1000)]
radeon: fix prediction for r100 inline vert/elt emits.
On r100 we emit the indices inline so we need to account
for that in the emission size.
Dave Airlie [Mon, 11 Jan 2010 03:21:02 +0000 (13:21 +1000)]
radeon: fix bug in realloc code.
This bug was fixed in libdrm ages ago, port to non-kms
Ben Skeggs [Mon, 11 Jan 2010 01:41:47 +0000 (11:41 +1000)]
st/dri: update dri2 drawables when viewport is changed
Fixes gnome-shell on nouveau, as well as window resize with various
other applications.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Corbin Simpson [Sun, 10 Jan 2010 19:49:25 +0000 (11:49 -0800)]
r300g: Atomize ZTOP.
Also do state-change checks. ZTOP's too important to not check.
Corbin Simpson [Sun, 10 Jan 2010 19:08:14 +0000 (11:08 -0800)]
r300g: Atomize DSA.
Also a bit of ztop.
Corbin Simpson [Sun, 10 Jan 2010 18:36:57 +0000 (10:36 -0800)]
r300g: Atomize UCP.
Meh.
Corbin Simpson [Sun, 10 Jan 2010 18:26:15 +0000 (10:26 -0800)]
r300g: Atomize blend color.
Corbin Simpson [Sun, 10 Jan 2010 17:35:55 +0000 (09:35 -0800)]
r300g: Start using atoms.
No benefits yet.
José Fonseca [Sun, 10 Jan 2010 18:37:07 +0000 (18:37 +0000)]
llvmpipe: Update instructions.
Explain how to build for windows. Both MSVC 9 and cross MinGW supported.
Stop documenting LLVM 2.5 as supported. It still supported at the moment
but it will soon stop being.
Chia-I Wu [Mon, 21 Dec 2009 03:13:18 +0000 (11:13 +0800)]
egl: Clean up the Makefile rules.
This allows libEGL to be built as a static library and removes libX11
from the dependencies.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Chia-I Wu [Tue, 5 Jan 2010 13:39:15 +0000 (21:39 +0800)]
egl_softpipe: Remove the unnecessary call to _eglSwapBuffers.
The call to _eglSwapBuffers is unnecessary, and the function is missing
when -fvisibility=hidden. Also remove the extraneous braces and indent
the block.
José Fonseca [Sun, 10 Jan 2010 11:20:11 +0000 (11:20 +0000)]
llvmpipe: Yet another hack to get release LLVM static libraries to link in debug builds.
José Fonseca [Sun, 10 Jan 2010 10:51:42 +0000 (10:51 +0000)]
scons: LLVM <-> MSVC integration tweaks.
José Fonseca [Sun, 10 Jan 2010 10:44:25 +0000 (10:44 +0000)]
gallium: Remove dead code.
José Fonseca [Sun, 10 Jan 2010 10:36:35 +0000 (10:36 +0000)]
gallium: Provide alternative stdint.h and stdbool.h C99 headers, instead of ad-hoc definitions.
Everybody is using the C99's integer types. Everybody except Microsoft,
which in turns means everybody is including their own definitions of C99
integer types for MSVC, causing duplicate definitions when linking two
projects. This is the case of building Gallium and LLVM with MSVC.
Shipping alternative stdint.h and stdbool.h headers for MSVC allows us
to share a single definition. It also removes clutter from the Gallium
headers.
Maciej Cencora [Sat, 9 Jan 2010 18:46:09 +0000 (19:46 +0100)]
r300: minor accelerated blit fixes
Maciej Cencora [Sat, 9 Jan 2010 17:12:55 +0000 (18:12 +0100)]
r300: fallback on depth buffer blits
Depth buffer accelerated blits aren't implemented yet.
Kristian Høgsberg [Fri, 8 Jan 2010 17:35:47 +0000 (12:35 -0500)]
intel/DRI2: add DRI2flushExtension support with invalidate hook
Needed to support the SwapBuffers code properly.
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
Jesse Barnes [Mon, 4 Jan 2010 21:26:17 +0000 (16:26 -0500)]
xdemos/glsync: handle no sync method better
Print out count, finish rendering, etc. Makes the -sn option more useful.
Jesse Barnes [Mon, 4 Jan 2010 18:13:57 +0000 (13:13 -0500)]
xdemos/glsync: check glXMakeCurrent return value
Since this program is used for testing, catching this case can be helpful.
Jesse Barnes [Tue, 3 Nov 2009 18:48:17 +0000 (10:48 -0800)]
xdemos/glsync: Make glsync test draw a rectangle
Doing simple buffer clears isn't enough to actually allocate render buffers, we
need to do real drawing.
Jesse Barnes [Mon, 2 Nov 2009 19:20:08 +0000 (11:20 -0800)]
xdemos/glsync: Add swap interval support to glsync test
Jesse Barnes [Thu, 12 Nov 2009 16:48:07 +0000 (16:48 +0000)]
DRI2/GLX: add INTEL_swap_event support
Add event support for the GLX swap buffers event, along with DRI2 protocol
support for generating GLX swap buffers events in the direct rendered case.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Jesse Barnes [Tue, 10 Nov 2009 21:28:01 +0000 (13:28 -0800)]
DRI2: add SwapInterval support
Add support for the DRI2SwapInterval protocol request. This allows
direct rendered clients to control their swap interval per the
SGI_swap_control extension.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Jesse Barnes [Wed, 16 Sep 2009 06:23:09 +0000 (23:23 -0700)]
DRI2: add OML_sync_control support
Add OML_sync_control support, along with a simple program for testing
it. This means adding support for the DRI2GetMSC, DRI2WaitMSC and
DRI2WaitSBC requests.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Jesse Barnes [Fri, 17 Jul 2009 07:30:29 +0000 (03:30 -0400)]
DRI2: add SwapBuffers support
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>
Christoph Bumiller [Fri, 8 Jan 2010 16:07:40 +0000 (17:07 +0100)]
nv50: handle TGSI_OPCODE_UMAD,UMUL,NOT and fix SAD
Christoph Bumiller [Fri, 8 Jan 2010 15:04:11 +0000 (16:04 +0100)]
nv50: free src temp_temps on emit
Christoph Bumiller [Fri, 8 Jan 2010 16:15:52 +0000 (17:15 +0100)]
nv50: try to honor sprite coord mode