llvm-3.0svn revision 134525 removed the TargetMachine parameter from
createMCInstPrinter
Signed-off-by: Tobias Droste
---
src/gallium/auxiliary/gallivm/lp_bld_debug.cpp |5 +
1 files changed, 1 insertions(+), 4 deletions(-)
diff --git a/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp
b/
On 07/07/2011 06:13 PM, Dan McCabe wrote:
> Is there a good reason not to build the internal glsl_compiler (in
> mesa/src/glsl) and link it with builtin_function.o? It is currently
> being built with builtin_stubs.o.
Ian changed that in bf9850db to kludge around some kind of linker errors
he was g
SUB & LRP instructions should toggle NEG bit instead of setting it,
otherwise e.g. "SUB a,b,-1" is translated as "ADD a,b,-1"
Signed-off-by: Vadim Girlin
---
src/gallium/drivers/r600/r600_shader.c |9 +++--
1 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/src/gallium/driver
LOG instruction should use absolute values of source operand.
Signed-off-by: Vadim Girlin
---
src/gallium/drivers/r600/r600_shader.c | 26 ++
1 files changed, 18 insertions(+), 8 deletions(-)
diff --git a/src/gallium/drivers/r600/r600_shader.c
b/src/gallium/drivers/r6
On Thu, 2011-07-07 at 21:19 +0200, Henri Verbeet wrote:
> Not necessarily wrong, but I think "bc_src->neg = !bc_src->neg;" would
> be the more common way to write this.
I agree with your comments for both patches. I'll change them as you
suggested and resend. Thanks for reviewing.
__
Is there a good reason not to build the internal glsl_compiler (in
mesa/src/glsl) and link it with builtin_function.o? It is currently
being built with builtin_stubs.o.
I can understand why builtin_compiler uses builtin_stubs.o instead of
builtin_function.o (to avoid conflicts while building
There's no pretty way to avoid the overwriting of the src operands, so
just use a temporary destination and rely on the MOV optimization.
---
src/mesa/drivers/dri/i965/brw_wm_fp.c | 23 +++
1 files changed, 19 insertions(+), 4 deletions(-)
diff --git a/src/mesa/drivers/dri/i
On Thu, 7 Jul 2011 17:31:47 -0600, Brian Paul wrote:
> Fixes segfault when running cubemap demo on i945. This happened
> when intel_region_reference() was called in i915_set_draw_region()
> with depth_region=NULL.
Whoops, managed to drop that check in the cleanup. Thanks!
Reviewed-by: Eric An
---
tests/shaders/fp-lit-src-equals-dst.c |6 ++
1 files changed, 2 insertions(+), 4 deletions(-)
diff --git a/tests/shaders/fp-lit-src-equals-dst.c
b/tests/shaders/fp-lit-src-equals-dst.c
index 11d495d..e59b3b3 100644
--- a/tests/shaders/fp-lit-src-equals-dst.c
+++ b/tests/shaders/fp-li
---
tests/shaders/fp-lit-src-equals-dst.c | 33 ++---
1 files changed, 10 insertions(+), 23 deletions(-)
diff --git a/tests/shaders/fp-lit-src-equals-dst.c
b/tests/shaders/fp-lit-src-equals-dst.c
index 464dbcc..11d495d 100644
--- a/tests/shaders/fp-lit-src-equals-ds
We were stomping over the source for the body of the LIT instruction
when doing the MOV of 1.0 to the uninteresting channels.
---
src/mesa/drivers/dri/i965/brw_wm_fp.c | 22 +++---
1 files changed, 11 insertions(+), 11 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_wm_f
The 965 driver was failing while breaking it down into channel-wise
operations and overwriting the operands.
---
tests/all.tests |2 ++
tests/shaders/generic/fp-dst-aliasing-1.vpfp | 16
tests/shaders/generic/fp-dst-aliasing-2.vpfp | 16 +++
Fixes segfault when running cubemap demo on i945. This happened
when intel_region_reference() was called in i915_set_draw_region()
with depth_region=NULL.
---
src/mesa/drivers/dri/intel/intel_regions.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/src/mesa/drivers/dri
On 6 July 2011 15:03, Paul Berry wrote:
> On 6 July 2011 12:18, Ian Romanick wrote:
>>
>> One other question: does this series cause any piglit regressions on
>> platforms that require flow-control lowering (e.g., i915 or r300)?
>
> I don't know. I will try to get ahold of a test platform when I
https://bugs.freedesktop.org/show_bug.cgi?id=39017
Thomas Hellström changed:
What|Removed |Added
AssignedTo|mesa-dev@lists.freedesktop. |mar...@gmail.com
|o
On 7 July 2011 06:20, Vadim Girlin wrote:
> +static inline void r600_bc_src_toggle_neg(struct r600_bc_alu_src *bc_src)
> +{
> + bc_src->neg = 1 - bc_src->neg;
> +}
> +
Not necessarily wrong, but I think "bc_src->neg = !bc_src->neg;" would
be the more common way to write this.
___
On 7 July 2011 06:20, Vadim Girlin wrote:
> -static void r600_bc_src(struct r600_bc_alu_src *bc_src,
> +static inline void r600_bc_src(struct r600_bc_alu_src *bc_src,
This looks like an unrelated change. Personally I think "inline" is
best left up to the compiler to decide in the majority of cases
On 07/07/2011 09:56 AM, Paul Berry wrote:
> On 7 July 2011 01:16, Kenneth Graunke wrote:
>> What exactly is the difference between lowering returns in "main" and
>> lowering them in other subroutines? void vs. non-void is definitely
>> different, but I don't see why main is special. Looking at t
On 07/07/2011 09:41 AM, Paul Berry wrote:
> On 7 July 2011 01:13, Kenneth Graunke wrote:
>>>visit_block(&ir->body);
>>>
>>> + /* If the body ended in an unconditional return of non-void,
>>> + * then we don't need to lower it because an unconditional
>>> + * return of non-
On 7 July 2011 01:16, Kenneth Graunke wrote:
> What exactly is the difference between lowering returns in "main" and
> lowering them in other subroutines? void vs. non-void is definitely
> different, but I don't see why main is special. Looking at the code, it
> doesn't actually seem to use lowe
On Thu, 7 Jul 2011 00:49:13 -0700, Kenneth Graunke
wrote:
> Signed-off-by: Kenneth Graunke
> ---
> src/mesa/drivers/dri/i965/brw_misc_state.c | 15 +--
> src/mesa/drivers/dri/i965/brw_structs.h|9 -
> 2 files changed, 5 insertions(+), 19 deletions(-)
>
> diff --g
On Thu, 7 Jul 2011 00:49:14 -0700, Kenneth Graunke
wrote:
> This is a little different from most because it's a single DWord;
> there's no length field.
Remove the struct definition, too?
pgpM6vgcmoWSv.pgp
Description: PGP signature
___
mesa-dev mai
On 7 July 2011 01:13, Kenneth Graunke wrote:
>> visit_block(&ir->body);
>>
>> + /* If the body ended in an unconditional return of non-void,
>> + * then we don't need to lower it because an unconditional
>> + * return of non-void at the end of a function is what returns
>>
https://bugs.freedesktop.org/show_bug.cgi?id=38842
--- Comment #4 from Jonathan Kirkham 2011-07-07
09:30:59 PDT ---
The change I have submitted above works, i.e. configs are returned when these
attributes.
Unfortunately the values of the attributes has been effectively thrown away.
I do not kn
https://bugs.freedesktop.org/show_bug.cgi?id=38842
--- Comment #3 from Jonathan Kirkham 2011-07-07
09:22:52 PDT ---
Apologies the change should have read :
case GLX_MAX_PBUFFER_WIDTH:
parselist +=2; //skip
break;
case GLX_MAX_PBUFFER_HEIGHT:
pa
https://bugs.freedesktop.org/show_bug.cgi?id=38842
--- Comment #2 from Jonathan Kirkham 2011-07-07
09:21:33 PDT ---
This solves the immediate problem that the attributes are rejected, however, I
do not believe this change is solving the real problem.
In this patch if the attribute GLX_MAX_PBUF
On Thu, Jul 7, 2011 at 3:08 PM, Thomas Hellstrom wrote:
> When the state tracker adds a front buffer, nothing triggers a validate
> drawable call, since the state tracker manager is never notified.
>
> Force a validate drawable call by invalidating the framebuffer's stamp, so
> that the window sys
When the state tracker adds a front buffer, nothing triggers a validate
drawable call, since the state tracker manager is never notified.
Force a validate drawable call by invalidating the framebuffer's stamp, so
that the window system's renderbuffer (if any) is picked up.
This fixes bug 38988
ht
On Wed, Jul 6, 2011 at 10:43 PM, Marek Olšák wrote:
> On Wed, Jul 6, 2011 at 5:49 PM, Brian Paul wrote:
>> On 06/30/2011 06:29 PM, Marek Olšák wrote:
>>>
>>> Hi,
>>>
>>> this patch series implements ARB_depth_buffer_float in Mesa and Gallium.
>>> There is complete r600g/r600-r700 support in my pr
On 07/07/2011 01:12 AM, Zou, Nanhai wrote:
Hi,
What is the intention of switching struct-style command to OUT_BATCH
style?
Thanks
Zou Nanhai
Several reasons:
1) Better code generation.
Eric has told me that GCC generates pretty sub-optimal code for
packed-bit structs, and using OUT
On 07/05/2011 03:07 PM, Paul Berry wrote:
The visitor class in lower_jumps.cpp never removes or replaces the
instruction being visited, but it frequently alters or removes the
instructions that follow it. Therefore, in order to be safe, it needs
to iterate through exec_lists using foreach_list r
On 07/05/2011 03:07 PM, Paul Berry wrote:
Previously ir_reader was only able to handle return of non-void.
This patch is necessary in order to allow optimization passes to be
tested in isolation.
---
src/glsl/ir_reader.cpp | 24 +---
1 files changed, 13 insertions(+), 11
On 07/05/2011 03:07 PM, Paul Berry wrote:
No functional change.
---
src/glsl/lower_jumps.cpp | 336 --
1 files changed, 322 insertions(+), 14 deletions(-)
Wow. Just wow. Thanks for doing this.
Reviewed-by: Kenneth Graunke
_
On 07/05/2011 03:07 PM, Paul Berry wrote:
Previously, do_lower_jumps.cpp determined whether to lower return
statements in ir_lower_jumps_visitor::should_lower_jumps(). Moved
this logic to ir_lower_jumps_visitor::visit(ir_function_signature *),
so that it can be used in determining whether to low
Hi,
What is the intention of switching struct-style command to OUT_BATCH
style?
Thanks
Zou Nanhai
>>-Original Message-
>>From: mesa-dev-bounces+nanhai.zou=intel@lists.freedesktop.org
>>[mailto:mesa-dev-bounces+nanhai.zou=intel@lists.freedesktop.org] On
>>Behalf Of Kennet
On 07/05/2011 03:07 PM, Paul Berry wrote:
Previously, lower_jumps.cpp only lowered return statements that
appeared inside of an if statement.
Without this patch, lower_jumps.cpp might not lower certain return
statements, causing some back-ends to fail (as in bug #36669).
Fixes unit test test_lo
We simply emit these using OUT_BATCH and bitshifting, as it results in
better compiled code than packed structures. Since our documentation
is public, it's not terribly useful to keep these around for reference.
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_structs.h | 441 -
Also rename it from CMD_STATE_INSN_POINTER to CMD_STATE_SIP to match the
documentation.
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_defines.h|2 +-
src/mesa/drivers/dri/i965/brw_misc_state.c | 17 -
2 files changed, 5 insertions(+), 14 deletions(-)
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_misc_state.c | 14 --
1 files changed, 4 insertions(+), 10 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_misc_state.c
b/src/mesa/drivers/dri/i965/brw_misc_state.c
index fba292d..73d2746 100644
--- a/src/mes
This is a little different from most because it's a single DWord;
there's no length field.
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_misc_state.c | 14 --
1 files changed, 4 insertions(+), 10 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_misc_state
Signed-off-by: Kenneth Graunke
---
src/mesa/drivers/dri/i965/brw_misc_state.c | 15 +--
src/mesa/drivers/dri/i965/brw_structs.h|9 -
2 files changed, 5 insertions(+), 19 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_misc_state.c
b/src/mesa/drivers/dri/i96
This patch series converts a few of the remaining struct-style command
submissions to OUT_BATCH style and then removes a ton of unused structs.
Keeping the structures in the codebase is misleading, since (a) you might
presume that they're actually used for something (they aren't), and (b)
they may
42 matches
Mail list logo
