[Mesa-dev] [PATCH 03/10] r600g: Add start_compute_cs atom to struct r600_context

The start_compute_cs atom initializes some config and context registers to the values needed for running compute shaders. When a compute shader is dispatched, this atom is emitted after the start_cs_cmd atom, which initializes registers that are common to both 3D and compute. --- src/gallium/driv

[Mesa-dev] [PATCH 04/10] r600g/compute: Move LOOP_CONST initialization to start_compute_cs atom

--- src/gallium/drivers/r600/evergreen_compute.c | 30 + 1 files changed, 16 insertions(+), 14 deletions(-) diff --git a/src/gallium/drivers/r600/evergreen_compute.c b/src/gallium/drivers/r600/evergreen_compute.c index eaffb75..3ec9d9b 100644 --- a/src/gallium/drivers/r

[Mesa-dev] [PATCH 05/10] r600g: Add pkt_flag parameter to r600_context_block_emit_dirty()

This allows the shader type bit to be set in the pm4 header when emitting registers for compute shaders. --- src/gallium/drivers/r600/r600.h |2 +- src/gallium/drivers/r600/r600_hw_context.c | 14 +- src/gallium/drivers/r600/r600_state_common.c |2 +- 3 files c

[Mesa-dev] [PATCH 06/10] r600g: Add r600_context_pipe_state_emit()

This function is used when dispatching compute shader in order to avoid mixing compute and 3D registers in the context's dirty list. This allows the compute code to resuse 3D functions like evergreen_cb, which return a struct r600_pipe_state and still have control over when and how the register wr

[Mesa-dev] [PATCH 07/10] r600g: Add is_rat flag to r600_resource_texture

--- src/gallium/drivers/r600/r600_resource.h |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/gallium/drivers/r600/r600_resource.h b/src/gallium/drivers/r600/r600_resource.h index 13fce00..d1f3a47 100644 --- a/src/gallium/drivers/r600/r600_resource.h +++ b/src/gallium/

[Mesa-dev] [PATCH 08/10] r600g: Use a texture as the underlying resource for compute_memory_pool

This the first step towards being able to use evergreen_cb to bind RATs. --- src/gallium/drivers/r600/compute_memory_pool.c | 34 +-- .../drivers/r600/evergreen_compute_internal.c | 21 2 files changed, 37 insertions(+), 18 deletions(-) diff --git a/src/g

[Mesa-dev] [PATCH 09/10] r600g: Add support for RATs in evergreen_cb()

--- src/gallium/drivers/r600/evergreen_state.c | 14 +++--- 1 files changed, 11 insertions(+), 3 deletions(-) diff --git a/src/gallium/drivers/r600/evergreen_state.c b/src/gallium/drivers/r600/evergreen_state.c index f0fdd2b..aeb92a5 100644 --- a/src/gallium/drivers/r600/evergreen_stat

[Mesa-dev] [PATCH 10/10] r600g/compute: Use evergreen_cb() for binding RATs

--- src/gallium/drivers/r600/compute_memory_pool.c |2 +- src/gallium/drivers/r600/evergreen_compute.c | 17 +++- .../drivers/r600/evergreen_compute_internal.c | 94 ++-- src/gallium/drivers/r600/evergreen_state.c |2 +- src/gallium/drivers/r600/r

Re: [Mesa-dev] [PATCH 02/10] r600g: Add pkt_flag member to struct r600_command_buffer

On Wed, Jun 27, 2012 at 07:17:37PM +0200, Marek Olšák wrote: > On Wed, Jun 27, 2012 at 5:37 PM, Tom Stellard wrote: > > Some packets require the shader type bit (bit 1) to be set when > > used for compute shaders.  The pkt_flag will be initialized to > > RADEON_CP_PACKET3

[Mesa-dev] [PATCH] clover: Increment an object's reference count in ret_object()

We need to increment the reference count for objects, like cl_event, that the user is responsible for destroying when they are returned from the API. Otherwise, the object will be destroyed when clover is done with it, even though the user will still have a reference to it. For example: 1. clEnq

Re: [Mesa-dev] [PATCH] clover: Add a function internalizer pass before LTO

I'll commit this tomorrow unless there are objections. -Tom On Thu, Jun 21, 2012 at 02:05:24PM -0400, Tom Stellard wrote: > The function internalizer pass marks non-kernel functions as internal, > which enables optimizations like function inlining and global dead-code &g

Re: [Mesa-dev] [PATCH] Use probed $LLVM_CONFIG value instead of hard-coding it

On Wed, Jun 27, 2012 at 10:56:38PM +0200, Luca Tettamanti wrote: > It makes it easier to use a non-default llvm-config binary. > Reviewed-by: Tom Stellard > Signed-off-by: Luca Tettamanti > --- > configure.ac |4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-)

Re: [Mesa-dev] [PATCH] clover: Increment an object's reference count in ret_object()

On Thu, Jun 28, 2012 at 01:09:20AM +0200, Francisco Jerez wrote: > Tom Stellard writes: > > > We need to increment the reference count for objects, like cl_event, > > that the user is responsible for destroying when they are returned from > > the API. Otherwise, the

[Mesa-dev] [PATCH] clover: Add a function internalizer pass before LTO v2

The function internalizer pass marks non-kernel functions as internal, which enables optimizations like function inlining and global dead-code elimination. v2: - Pass vector arguments by const reference --- .../state_trackers/clover/llvm/invocation.cpp | 59 --- 1 files c

Re: [Mesa-dev] [PATCH] clover: Increment an object's reference count in ret_object()

On Thu, Jun 28, 2012 at 03:47:23PM +0200, Francisco Jerez wrote: > Tom Stellard writes: > > > On Thu, Jun 28, 2012 at 01:09:20AM +0200, Francisco Jerez wrote: > >> Tom Stellard writes: > >> > >> > We need to increment the reference count for objec

[Mesa-dev] [PATCH] r600g: Update number of gprs when adding a vertex instruction

--- src/gallium/drivers/r600/r600_asm.c |4 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/src/gallium/drivers/r600/r600_asm.c b/src/gallium/drivers/r600/r600_asm.c index b0cda3a..2341bec 100644 --- a/src/gallium/drivers/r600/r600_asm.c +++ b/src/gallium/drivers/r600/r600

[Mesa-dev] [PATCH] clover: Handle NULL devs argument in clBuildProgram

If devs is NULL, then the kernel should be compiled for all devices associated with the program's context. --- src/gallium/state_trackers/clover/api/program.cpp | 16 ++-- 1 files changed, 10 insertions(+), 6 deletions(-) diff --git a/src/gallium/state_trackers/clover/api/program.cp

Re: [Mesa-dev] [PATCH] clover: Handle NULL devs argument in clBuildProgram

On Fri, Jun 29, 2012 at 12:18:15AM +0200, Francisco Jerez wrote: > Tom Stellard writes: > > > If devs is NULL, then the kernel should be compiled for all devices > > associated with the program's context. > > --- > > src/gallium/state_tra

[Mesa-dev] Mesa master branch: forced update

I just fetched from the master branch of the fdo mesa repo and was greeted with a "forced update" message, and the gitweb interface shows several days of history are missing from the master branch. olv appears to be the last user to modify the master branch: tstellar@annarchy:~$ ls -l /git/mesa/

Re: [Mesa-dev] [PATCH 13/22] r600g: move CB_TARGET_MASK setup into new cb_misc_state

e changes look fine to me. I just tested and there are no regressions. I'll change the compute code to use cb_misc_state when I get a change. Reviewed-by: Tom Stellard > --- > src/gallium/drivers/r600/evergreen_compute.c |3 +-- > .../drivers/r600/evergreen_compute_inte

[Mesa-dev] [PATCH 1/2] r600g: Unify 3D and compute vertex buffer emission

--- src/gallium/drivers/r600/evergreen_compute.c | 47 +--- .../drivers/r600/evergreen_compute_internal.c | 81 .../drivers/r600/evergreen_compute_internal.h |1 - src/gallium/drivers/r600/evergreen_state.c | 30 +-- src/gallium/dr

[Mesa-dev] [PATCH 2/2] r600g: Emit vertex buffers using the same method as constant buffers

--- src/gallium/drivers/r600/evergreen_compute.c |8 +++-- src/gallium/drivers/r600/evergreen_state.c | 45 - src/gallium/drivers/r600/r600_buffer.c |5 ++- src/gallium/drivers/r600/r600_hw_context.c |2 +- src/gallium/drivers/r600/r600_pipe.h

[Mesa-dev] [PATCH] configure: Check xcb version when X11 pkgconfig exists

Commit 6882381a2efbdf06b7002d11468c94b9964c2bc8 added a dependency on a newer version of xcb, but the version check wasn't added in all the necessary places. --- configure.ac |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/configure.ac b/configure.ac index 00a1d5d..0d38db5

Re: [Mesa-dev] [PATCH] Fix linking gallium drivers and with dricore after defadf2b1 (v2)

n configure but $(TOP) and $(LIB_DIR) are interpolated > later (where they are known, but VERSION isn't) > Just tested and this fixes the build, please commit ASAP. > Signed-off-by: Jon TURNEY Signed-off-by: Tom Stellard > --- > configure.ac |

Re: [Mesa-dev] [PATCH] radeon/llvm: fix CR/LF

Push, thanks. Sorry it took so long. -Tom On Tue, Jun 19, 2012 at 07:50:22PM +0200, Andreas Boll wrote: > --- > src/gallium/drivers/radeon/AMDILSIDevice.h |2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/src/gallium/drivers/radeon/AMDILSIDevice.h > b/src/gallium/

[Mesa-dev] Convert a few gallium libs automake

Hi, Here are some patches that convert a few more gallium libs to automake. These can also be found in the gallium-automake branch here: cgit.freedesktop.org/~tstellar/mesa/ -Tom ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.fr

[Mesa-dev] [PATCH 1/6] gallium: Add common automake include file

--- src/gallium/Automake.inc |7 +++ 1 files changed, 7 insertions(+), 0 deletions(-) create mode 100644 src/gallium/Automake.inc diff --git a/src/gallium/Automake.inc b/src/gallium/Automake.inc new file mode 100644 index 000..ded579f --- /dev/null +++ b/src/gallium/Automake.inc @@ -

[Mesa-dev] [PATCH 2/6] r300g: Use gallium automake include file

--- src/gallium/drivers/r300/Makefile.am | 10 +++--- 1 files changed, 3 insertions(+), 7 deletions(-) diff --git a/src/gallium/drivers/r300/Makefile.am b/src/gallium/drivers/r300/Makefile.am index 3d08611..6c4bba7 100644 --- a/src/gallium/drivers/r300/Makefile.am +++ b/src/gallium/drivers

[Mesa-dev] [PATCH 3/6] r300g: Build a libtool archive

--- src/gallium/drivers/r300/Makefile.am | 18 +++--- 1 files changed, 11 insertions(+), 7 deletions(-) diff --git a/src/gallium/drivers/r300/Makefile.am b/src/gallium/drivers/r300/Makefile.am index 6c4bba7..aaa2e27 100644 --- a/src/gallium/drivers/r300/Makefile.am +++ b/src/galliu

[Mesa-dev] [PATCH 4/6] radeon/winsys: Convert to automake

--- configure.ac |1 + src/gallium/winsys/radeon/drm/.gitignore |1 + src/gallium/winsys/radeon/drm/Makefile| 15 --- src/gallium/winsys/radeon/drm/Makefile.am | 18 ++ 4 files changed, 20 insertions(+), 15 deletions(-) cre

[Mesa-dev] [PATCH 5/6] trace: Convert to automake

--- configure.ac |1 + src/gallium/drivers/trace/.gitignore |2 ++ src/gallium/drivers/trace/Makefile| 13 - src/gallium/drivers/trace/Makefile.am | 17 + 4 files changed, 20 insertions(+), 13 deletions(-) create mode 100644 sr

[Mesa-dev] [PATCH 6/6] libgallium: Convert to automake

--- configure.ac |2 + src/gallium/auxiliary/.gitignore |1 + src/gallium/auxiliary/Makefile| 42 - src/gallium/auxiliary/Makefile.am | 47 + 4 files changed, 50 insertions(+), 42 deletions(-)

Re: [Mesa-dev] [PATCH 1/2] radeon/llvm: fix compiling when llvm is active, but opencl isn't

On Mon, Jul 16, 2012 at 12:14:27PM +0200, Christian König wrote: > Signed-off-by: Christian König > --- > src/gallium/drivers/r600/r600_llvm.c |4 > 1 file changed, 4 insertions(+) > > diff --git a/src/gallium/drivers/r600/r600_llvm.c > b/src/gallium/drivers/r600/r600_llvm.c > index 5e

Re: [Mesa-dev] [PATCH 1/2] radeon/llvm: fix compiling when llvm is active, but opencl isn't

On Mon, Jul 16, 2012 at 03:46:35PM +0200, Christian König wrote: > On 16.07.2012 15:32, Tom Stellard wrote: > >On Mon, Jul 16, 2012 at 12:14:27PM +0200, Christian König wrote: > >>Signed-off-by: Christian König > >>--- > >> src/gallium/drivers/r600/r600_llvm.c

Re: [Mesa-dev] wider registers in llvmpipe

> 1.7.10.2 > I think we need to add the util prefix to the u_cpu_detect.h include in lp_bld_quad.c. This is what is done in all the other gallivm files. See attached patch. -Tom > ___ > mesa-dev mailing list > mesa-dev@lists.freede

[Mesa-dev] [PATCH] configure.ac: Add libLLVMMCJIT to the LLVM_LDFLAGS

This is neccessary for linking the llvmpipe tests. It appears this dependency was introduced by the "wider native register" changes. --- configure.ac |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/configure.ac b/configure.ac index 9227474..068f632 100644 --- a/configure

Re: [Mesa-dev] [PATCH] configure.ac: Further LLVM fixups.

On Tue, Jul 17, 2012 at 07:00:48PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > * Also add mcjit in the non-OpenCL case. > * Replace hardcoded llvm-config with $LLVM_CONFIG everywhere. > > Signed-off-by: Michel Dänzer I just pushed a fix for mcjit, but the rest is: Reviewed-by: Tom S

[Mesa-dev] [PATCH] configure.ac: Add --with-llvm-prefix option

This option allows you to specify the llvm install prefix. It is useful for switching between different versions of LLVM. --- configure.ac | 13 - 1 files changed, 12 insertions(+), 1 deletions(-) diff --git a/configure.ac b/configure.ac index 1d60957..10216a3 100644 --- a/configur

[Mesa-dev] [PATCH] pipe_loader: Try to make a connection with the X server before probing pciids

When X is running it is neccesary for pipe_loader to authenticate with DRM, in order to be able to use the device. This makes it possible to run OpenCL programs while X is running. --- configure.ac |6 ++ .../auxiliary/pipe-loader/pipe_loader_drm.c

[Mesa-dev] [PATCH] pipe_loader: Try to connect with the X server before probing pciids v2

When X is running it is neccesary for pipe_loader to authenticate with DRM, in order to be able to use the device. This makes it possible to run OpenCL programs while X is running. v2: - Fix C++ style comments - Drop Xlib-xcb dependency - Close the X connection when done - Split auth code

[Mesa-dev] r600g compute cleanups

Hi, Here are few patches that cleanup state handling in the r600g compute code. I'm trying to move away from using struct evergreen_compute_resource for state handling and instead use atoms to make it easier for the compute and 3D code to coexist. -Tom ___

[Mesa-dev] [PATCH 1/5] r600g: Add helper functions for emitting compute SET_CONTEXT packets

--- src/gallium/drivers/r600/evergreend.h |5 - src/gallium/drivers/r600/r600_pipe.h | 18 ++ 2 files changed, 18 insertions(+), 5 deletions(-) diff --git a/src/gallium/drivers/r600/evergreend.h b/src/gallium/drivers/r600/evergreend.h index 6c4873c..55061d4 100644 ---

[Mesa-dev] [PATCH 2/5] r600g: Atomize compute shader state

--- src/gallium/drivers/r600/evergreen_compute.c | 168 +- src/gallium/drivers/r600/evergreen_compute.h |2 + src/gallium/drivers/r600/evergreen_state.c |1 + src/gallium/drivers/r600/r600_pipe.h |7 +- 4 files changed, 92 insertions(+), 86 deletions(

[Mesa-dev] [PATCH 3/5] r600g: Emit CB_TARGET_MASK in compute_emit_cs()

--- src/gallium/drivers/r600/evergreen_compute.c |9 + 1 files changed, 5 insertions(+), 4 deletions(-) diff --git a/src/gallium/drivers/r600/evergreen_compute.c b/src/gallium/drivers/r600/evergreen_compute.c index dd9b341..3822694 100644 --- a/src/gallium/drivers/r600/evergreen_comp

[Mesa-dev] [PATCH 4/5] r600g: Initialize VGT_PRIMITIVE_TYPE in the start_cs_cmd atom

The value of this register will always be DI_PT_POINTLIST for compute shaders. --- src/gallium/drivers/r600/evergreen_compute.c |6 +- 1 files changed, 5 insertions(+), 1 deletions(-) diff --git a/src/gallium/drivers/r600/evergreen_compute.c b/src/gallium/drivers/r600/evergreen_compute.c

[Mesa-dev] [PATCH 5/5] r600g: Emit dispatch state for compute directly to the cs

We no longer rely on an evergreen_compute_resource for emitting dispatch state. --- src/gallium/drivers/r600/evergreen_compute.c | 104 +++- src/gallium/drivers/r600/evergreen_compute.h |1 - .../drivers/r600/evergreen_compute_internal.c | 19 .../drive

Re: [Mesa-dev] [PATCH] radeon/llvm: fix calculation of max register number

On Tue, Jul 31, 2012 at 06:40:11PM +0200, Christian König wrote: > Signed-off-by: Christian König Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeon/SICodeEmitter.cpp |2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/src/gall

Re: [Mesa-dev] [PATCH 3/6] radeon/llvm: fix fp immediates on SI

On Wed, Aug 01, 2012 at 11:28:24PM +0200, Christian König wrote: > I don't know if this is a good idea, but it > fixes the problem at hand. > > Signed-off-by: Christian König Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeon/S

Re: [Mesa-dev] [PATCH 5/6] radeonsi: fix TEX writemask

On Thu, Aug 02, 2012 at 11:33:43AM +0200, Christian König wrote: > On 02.08.2012 11:21, Michel Dänzer wrote: > >On Don, 2012-08-02 at 11:05 +0200, Christian König wrote: > >>On 02.08.2012 07:51, Michel Dänzer wrote: > >>>On Mit, 2012-08-01 at 23:28 +0200, Christian König wrote: > Using the writ

Re: [Mesa-dev] [PATCH 2/2] radeonsi: Handle TGSI TXP opcode.

On Thu, Aug 02, 2012 at 05:37:03PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > Signed-off-by: Michel Dänzer Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeonsi/radeonsi_shader.c | 26 > ++-- > 1 file changed, 24 inser

Re: [Mesa-dev] [PATCH 1/2] radeonsi: Handle TGSI DIV opcode.

On Thu, Aug 02, 2012 at 05:37:02PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > Signed-off-by: Michel Dänzer Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeon/SIInstructions.td |5 + > 1 file changed, 5 insertions(+) > > diff --git a/sr

Re: [Mesa-dev] [PATCH 2/2] r600g: Fix instruction group merge when there are predicated insts.

On Wed, Aug 1, 2012 at 4:52 PM, Vincent Lejeune wrote: > --- > src/gallium/drivers/r600/r600_asm.c | 18 ++ > 1 file changed, 18 insertions(+) > > diff --git a/src/gallium/drivers/r600/r600_asm.c > b/src/gallium/drivers/r600/r600_asm.c > index 0f28824..b0194bf 100644 > --- a/src/

Re: [Mesa-dev] [PATCH 0/4] Convert Nouveau gallium drivers to automake

On Fri, Aug 03, 2012 at 01:53:06AM +0200, Lucas Stach wrote: > Trivial series to convert Nouveau gallium drivers to automake > in order to generate Makefiles. > > This is just the general direction in which the MESA buildsystem > is moving. Aside from this it makes the "make" output a lot more > r

Re: [Mesa-dev] [PATCH 1/5] radeon/llvm: add support to fetch temps as vectors

On Mon, Aug 13, 2012 at 01:29:45PM +0200, Christian König wrote: > Necessary for texture fetches with temp regs as source on SI. > Reviewed-by: Tom Stellard > Signed-off-by: Christian König > --- > .../drivers/radeon/radeon_setup_tgsi_llvm.c| 12 +++- >

Re: [Mesa-dev] [PATCH] r600g-llvm: Crude fix for a race in initialization of the llvm backend

cc'ing the list this time. On Wed, Aug 15, 2012 at 07:51:46AM +0200, Mathias Fröhlich wrote: > > Hi, > > This change adds application global locking around compiling an r600 llvm > shader. This fixes a race condition/crash that I observe when shaders are > concurrently compiled from different

Re: [Mesa-dev] [PATCH] Remove 3D registers from compute command stream

Hi, In the future, could you use git format-patch to send patches. Overall, I think this looks OK, I just need to test it out a little bit. -Tom On Mon, Aug 13, 2012 at 09:05:29PM +0100, archibald wrote: > diff --git a/src/gallium/drivers/r600/evergreen_compute.c > b/src/gallium/drivers/r600/e

Re: [Mesa-dev] [PATCH] Remove 3D registers from compute command stream

On Wed, Aug 15, 2012 at 12:30:38PM -0400, Matt Harvey wrote: > Hey, I'd be interested in testing this to make sure it works, but I have a > HD4200 running r600g, and I don't think that card has opencl support at the > moment. Do you know if that would still be useful for testing, and what I > would

Re: [Mesa-dev] [PATCH] r600g-llvm: Crude fix for a race in initialization of the llvm backend

On Thu, Aug 16, 2012 at 07:44:53PM +0200, Mathias Fröhlich wrote: > > Hi, > > > This is likely caused by the fact that the same LLVM Context is used > > for all threads. Take a look at the comment starting at > > src/gallium/auxilary/gallivm/lp_bld_init.c:314 > > > > You should be able to reprod

Re: [Mesa-dev] [PATCH] r600g-llvm: Crude fix for a race in initialization of the llvm backend

On Thu, Aug 16, 2012 at 02:39:48PM -0400, Tom Stellard wrote: > On Thu, Aug 16, 2012 at 07:44:53PM +0200, Mathias Fröhlich wrote: > > > > Hi, > > > > > This is likely caused by the fact that the same LLVM Context is used > > > for all threads. Take a lo

Re: [Mesa-dev] [PATCH] Remove 3D registers from compute command stream

On Mon, Aug 13, 2012 at 09:05:29PM +0100, archibald wrote: > Hi list, > > Here is my attempt at solving the task "Remove 3D registers from > compute > command stream" on http://dri.freedesktop.org/wiki/R600ToDo. It's my > first attempt at a patch for mesa, so I'd appreciate any comments or > advic

Re: [Mesa-dev] [PATCH] r600g-llvm: Crude fix for a race in initialization of the llvm backend

On Sat, Aug 18, 2012 at 11:43:45AM +0200, Mathias Fröhlich wrote: > > Tom, > > On Thursday, August 16, 2012 14:50:57 you wrote: > > Actually, I just noticed that the LLVM docs say that only one thread can > > access the target registry at a time: > > http://llvm.org/docs/doxygen/html/structllvm_1

Re: [Mesa-dev] [PATCH] radeon-llvm: Start multithreaded before using llvm.

Pushed. Thanks for investigating this and coming up with a clean solution. -Tom On Sat, Aug 18, 2012 at 11:43:51AM +0200, Mathias Fröhlich wrote: > This is required to make some of llvm's api calls > thread save. In particular the PassRegistry, which is > implicitly accessed while compiling shad

Re: [Mesa-dev] [PATCH 1/3] radon/llvm: br_cc f32 now lowered without cast

On Tue, Aug 21, 2012 at 12:41:07AM +0200, Vincent Lejeune wrote: Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeon/R600ISelLowering.cpp | 33 > ++--- > 1 file changed, 24 insertions(+), 9 deletions(-) > > diff --git a/src/gall

Re: [Mesa-dev] [PATCH 2/3] radeon/llvm: support setcc on f32

On Tue, Aug 21, 2012 at 12:41:08AM +0200, Vincent Lejeune wrote: A few small comments inline, but otherwise: Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeon/R600ISelLowering.cpp | 36 > ++--- > 1 file changed, 27 insertions(+), 9 deletions(-) >

Re: [Mesa-dev] [PATCH 3/3] radeon/llvm: custom lowering for FP_TO_UINT when dst is i1 (bool)

On Tue, Aug 21, 2012 at 12:41:09AM +0200, Vincent Lejeune wrote: > --- > src/gallium/drivers/radeon/R600ISelLowering.cpp | 27 > + > src/gallium/drivers/radeon/R600ISelLowering.h | 6 -- > 2 files changed, 31 insertions(+), 2 deletions(-) > > diff --git a/src/galli

Re: [Mesa-dev] [PATCH 4/4] radeonsi: Use FP16 shader export format when necessary / possible.

On Fri, Aug 24, 2012 at 02:53:01PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > Fixes piglit fbo-blending-formats. > > Signed-off-by: Michel Dänzer Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeon/SIInstructions.td |4 +- > sr

Re: [Mesa-dev] r300g broken

> >Please run "git clean -fdX" in the Mesa repository and then > >reconfigure and rebuild. > > Almost the same, even with this change: r300/compiler: Use variable > lists in the rename_regs pass > > When I run opera with HW enabled in valgrind I see something like this: > > r300: DRM version: 2.1

Re: [Mesa-dev] [PATCH] radeon/llvm: Handle TGSI KIL opcode for SI.

On Tue, Aug 28, 2012 at 04:26:43PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > Fixes piglit fp-kil with radeonsi. > > Signed-off-by: Michel Dänzer > --- > src/gallium/drivers/radeon/SIISelLowering.cpp | 35 > + > src/gallium/drivers/radeon/SIISelLowering.h

Re: [Mesa-dev] [PATCH 2/2] radeon/llvm: Handle TGSI KIL opcode for SI.

On Tue, Aug 28, 2012 at 08:14:08PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > Fixes piglit fp-kil and glBitmap() with radeonsi. > > Signed-off-by: Michel Dänzer Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeon/SIIS

Re: [Mesa-dev] [PATCH 1/2] radeon/llvm: Basic support for SI EXEC register.

On Tue, Aug 28, 2012 at 08:14:07PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > Reviewed-by: Tom Stellard > Signed-off-by: Michel Dänzer > --- > src/gallium/drivers/radeon/AMDGPUAsmPrinter.cpp |3 +++ > src/gallium/drivers/radeon/SIGenRegisterInfo.pl |7

Re: [Mesa-dev] [PATCH 00/32] First pass at automaking Gallium

On Wed, Aug 29, 2012 at 11:16:52PM -0700, Matt Turner wrote: > Here's the first pass at automaking Gallium. It converts the winsys > and all the driver directories but radeon. > I can work on converting the radeon directory to automake. -Tom > Compile tested. Files checked to confirm that they'

Re: [Mesa-dev] [PATCH 07/32] libgallium: Convert to automake

On Wed, Aug 29, 2012 at 11:16:59PM -0700, Matt Turner wrote: > From: Tom Stellard > I'm getting some build warnings when building the gallivm files: In file included from /usr/local/llvm/3.1/release/include/llvm/ExecutionEngine/JITEventListener.h:18:0, f

[Mesa-dev] [PATCH] radeonsi: Handle TGSI_SEMANTIC_PSIZE

The relevant POINT_SIZE registers are being set using the pipe_rasterizer_state, so we just need to tell the shader compiler which export type to use. This fixes several of the glean glsl tests. --- src/gallium/drivers/radeonsi/radeonsi_shader.c |1 + 1 files changed, 1 insertions(+), 0 delet

Re: [Mesa-dev] radeonsi: first successful piglit run

On Thu, Aug 30, 2012 at 07:00:44PM +0200, Christian König wrote: > Hi everybody, > > with the following patchset the radeonsi driver finally > completes a run of nearly all tests in Piglits "quick-driver" > profile without an GPU hang or X server crash. > > The only exceptions are: > 1. glx/glx-t

Re: [Mesa-dev] [PATCH] radeon/llvm: Convert to Automake

On Thu, Aug 30, 2012 at 10:47:55AM -0700, Matt Turner wrote: > On Thu, Aug 30, 2012 at 9:55 AM, wrote: > > From: Tom Stellard > > > > --- > > Hi Matt, > > > > This patch applies to your automake-gallium branch. > > Thanks a bunch! > > &g

Re: [Mesa-dev] radeonsi: first successful piglit run

On Thu, Aug 30, 2012 at 07:00:44PM +0200, Christian König wrote: > Hi everybody, > > with the following patchset the radeonsi driver finally > completes a run of nearly all tests in Piglits "quick-driver" > profile without an GPU hang or X server crash. > Which kernel tree are you using? -Tom

Re: [Mesa-dev] [PATCH] radeon/llvm: Convert to Automake

On Thu, Aug 30, 2012 at 02:11:22PM -0700, Matt Turner wrote: > On Thu, Aug 30, 2012 at 1:19 PM, Tom Stellard wrote: > > On Thu, Aug 30, 2012 at 10:47:55AM -0700, Matt Turner wrote: > >> On Thu, Aug 30, 2012 at 9:55 AM, wrote: > >> > From: Tom Stellard >

Re: [Mesa-dev] [PATCH] radeon/llvm: Convert to Automake

On Fri, Aug 31, 2012 at 01:26:33PM -0700, Matt Turner wrote: > On Fri, Aug 31, 2012 at 6:04 AM, Tom Stellard wrote: > > On Thu, Aug 30, 2012 at 02:11:22PM -0700, Matt Turner wrote: > >> On Thu, Aug 30, 2012 at 1:19 PM, Tom Stellard wrote: > >> > On Thu, Aug 30,

[Mesa-dev] [PATCH] configure.ac: Don't link gallium drivers with libdricore

From: Tom Stellard --- Is there any reason gallium drivers should be linking with libdricore? They seem to work fine without it. configure.ac |1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/configure.ac b/configure.ac index 82fe225..6e67ec0 100644 --- a/configure.ac

Re: [Mesa-dev] [PATCH] radeon/llvm: swap wrong OPCODE_IS_*_ZERO_* opcode and use

On Tue, Sep 04, 2012 at 04:49:25PM +0200, Vincent Lejeune wrote: > --- > src/gallium/drivers/radeon/AMDGPUInstrInfo.h| 4 ++-- > src/gallium/drivers/radeon/R600ISelLowering.cpp | 4 ++-- > 2 files changed, 4 insertions(+), 4 deletions(-) > Reviewed-by: Tom Stellard >

Re: [Mesa-dev] [PATCH 3/4] radeon/llvm: custom lowering for FP_TO_UINT when dst is i1 (bool)

On Tue, Sep 04, 2012 at 05:04:29PM +0200, Vincent Lejeune wrote: > Conflicts: > src/gallium/drivers/radeon/R600ISelLowering.cpp > --- > src/gallium/drivers/radeon/R600ISelLowering.cpp | 20 > src/gallium/drivers/radeon/R600ISelLowering.h | 6 -- > 2 files changed,

Re: [Mesa-dev] [PATCH 4/4] radeon/llvm: do not convert f32 operand of select_cc node

On Tue, Sep 04, 2012 at 05:04:30PM +0200, Vincent Lejeune wrote: > --- > src/gallium/drivers/radeon/R600ISelLowering.cpp | 40 > - > 1 file changed, 20 insertions(+), 20 deletions(-) > > diff --git a/src/gallium/drivers/radeon/R600ISelLowering.cpp > b/src/gallium/drivers

Re: [Mesa-dev] [PATCH 1/4] radon/llvm: br_cc f32 now lowered without cast

you can remove the 'conflicts:' line from the commit messages. With those changes, the entire series is: Reviewed-by: Tom Stellard > diff --git a/src/gallium/drivers/radeon/R600ISelLowering.cpp > b/src/gallium/drivers/radeon/R600ISelLowering.cpp > index fec9d4e..60c8ade

Re: [Mesa-dev] [PATCH] radeon/llvm: Convert to Automake

On Tue, Sep 04, 2012 at 11:10:52AM -0700, Matt Turner wrote: > On Fri, Aug 31, 2012 at 1:41 PM, Tom Stellard wrote: > > On Fri, Aug 31, 2012 at 01:26:33PM -0700, Matt Turner wrote: > >> On Fri, Aug 31, 2012 at 6:04 AM, Tom Stellard wrote: > >> > On Thu, Aug 30,

Re: [Mesa-dev] [PATCH 1/4] radeon/llvm: Extend SI EXEC register support.

On Thu, Sep 06, 2012 at 01:00:00PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > Add 32 bit lo and hi variants, and binary encodings. > > Signed-off-by: Michel Dänzer > --- > src/gallium/drivers/radeon/MCTargetDesc/SIMCCodeEmitter.cpp |3 +++ > src/gallium/drivers/radeon/SIGenRegis

Re: [Mesa-dev] [PATCH 2/4] radeon/llvm: SI shader vector instructions implicitly use the EXEC register.

e optimizer was incorrectly marking some instructions dead. It's possible there was something else I was doing that was wrong, but in any case I think we should keep an eye on this for potential bugs. Either way, this change won't hurt anything. Reviewed-by: Tom Stellard > Sign

Re: [Mesa-dev] [PATCH 3/4] radeon/llvm: Add intrinsic for enabling whole quad mode in SI pixel shaders.

On Thu, Sep 06, 2012 at 01:00:02PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > Reviewed-by: Tom Stellard > Signed-off-by: Michel Dänzer > --- > src/gallium/drivers/radeon/SIISelLowering.cpp | 13 + > src/gallium/drivers/radeon/SIISelLowering.h |2

Re: [Mesa-dev] [PATCH 1/4 v2] radeon/llvm: Extend SI EXEC register support.

sub-registers. Is this what you meant, Tom? Yes, looks good. Reviewed-by: Tom Stellard > > src/gallium/drivers/radeon/MCTargetDesc/SIMCCodeEmitter.cpp |3 +++ > src/gallium/drivers/radeon/SIGenRegisterInfo.pl |6 -- > 2 files changed, 7 insertions(+), 2 del

Re: [Mesa-dev] [PATCH 4/4] radeonsi: Enable whole quad for pixel shaders.

On Thu, Sep 06, 2012 at 01:00:03PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > Fixes wrong mipmap level being sampled at some triangle edges. > > Signed-off-by: Michel Dänzer Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeonsi/radeonsi_shader.c |

Re: [Mesa-dev] [PATCH 1/2] radeon/llvm: Match fexp2 for SI.

On Fri, Sep 07, 2012 at 12:27:35PM +0200, Michel Dänzer wrote: > From: Michel Dänzer > > Signed-off-by: Michel Dänzer For the series: Reviewed-by: Tom Stellard > --- > src/gallium/drivers/radeon/SIInstructions.td |4 +++- > 1 file changed, 3 insertions(+), 1 deletion(

[Mesa-dev] radeonsi: Flow Control preperation patches

Hi, The series is a prerequisite for flow control support on radeonsi. Flow control for If-Then-Else blocks is mostly implemented, but I'm still trying to debug some GPU hangs that it is causing. -Tom ___ mesa-dev mailing list mesa-dev@lists.freedeskt

[Mesa-dev] [PATCH 01/10] radeon/llvm: Add SHADER_TYPE instruction

From: Tom Stellard This allows the program to specify the type of shader being compiled (e.g. PXEL, VERTEX, etc.) --- src/gallium/drivers/radeon/AMDGPU.h|9 + src/gallium/drivers/radeon/AMDGPUInstructions.td |7 +++ src/gallium/drivers/radeon

[Mesa-dev] [PATCH 02/10] radeonsi: Pass shader type to the compiler

From: Tom Stellard --- src/gallium/drivers/radeonsi/radeonsi_shader.c | 12 1 files changed, 12 insertions(+), 0 deletions(-) diff --git a/src/gallium/drivers/radeonsi/radeonsi_shader.c b/src/gallium/drivers/radeonsi/radeonsi_shader.c index 9fc3446..0994fac 100644 --- a/src

[Mesa-dev] [PATCH 03/10] radeonsi: Add missing interpolation mode to check for enabled modes

From: Tom Stellard At least one interpolation mode must be enable, but the code that checks this was not checking for perspective center. --- src/gallium/drivers/radeonsi/si_state_draw.c |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/gallium/drivers/radeonsi

[Mesa-dev] [PATCH 04/10] radeonsi: Move interpolation mode check into the compiler

From: Tom Stellard The compiler needs to know which interpolation modes are enabled, so it knows which values will be preloaded into the VGPRs. --- src/gallium/drivers/radeon/SIAssignInterpRegs.cpp | 13 - src/gallium/drivers/radeonsi/si_state_draw.c | 20

[Mesa-dev] [PATCH 05/10] radeon/llvm: Coding style fixes

From: Tom Stellard --- src/gallium/drivers/radeon/AMDGPUAsmPrinter.cpp|2 +- src/gallium/drivers/radeon/SIAssignInterpRegs.cpp | 56 ++-- .../drivers/radeon/SIMachineFunctionInfo.cpp |2 +- src/gallium/drivers/radeon/SIMachineFunctionInfo.h |2 +- 4

[Mesa-dev] [PATCH 06/10] radeonsi: Handle position input parameter for pixel shaders

From: Tom Stellard --- src/gallium/drivers/radeon/SIInstructions.td | 20 src/gallium/drivers/radeon/SIIntrinsics.td |2 ++ src/gallium/drivers/radeonsi/radeonsi_shader.c | 16 +++- src/gallium/drivers/radeonsi/si_state_draw.c | 14

[Mesa-dev] [PATCH 07/10] radeon/llvm: Ignore special registers when calculating reg count

From: Tom Stellard --- src/gallium/drivers/radeon/AMDGPUAsmPrinter.cpp |2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/src/gallium/drivers/radeon/AMDGPUAsmPrinter.cpp b/src/gallium/drivers/radeon/AMDGPUAsmPrinter.cpp index 07f443d..392791c 100644 --- a/src/gallium

[Mesa-dev] [PATCH 08/10] radeon/llvm: Add register encoding for VCC

From: Tom Stellard --- .../radeon/MCTargetDesc/SIMCCodeEmitter.cpp|1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/gallium/drivers/radeon/MCTargetDesc/SIMCCodeEmitter.cpp b/src/gallium/drivers/radeon/MCTargetDesc/SIMCCodeEmitter.cpp index 438d2ac..ca4b579

[Mesa-dev] [PATCH 09/10] radeon/llvm: Assert if we try to encode an unknown register

From: Tom Stellard --- src/gallium/drivers/radeon/SIGenRegisterInfo.pl |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/gallium/drivers/radeon/SIGenRegisterInfo.pl b/src/gallium/drivers/radeon/SIGenRegisterInfo.pl index 3813eb5..48bd567 100644 --- a/src/gallium

<    2   3   4   5   6   7   8   9   10   11   >