[Mesa-dev] [PATCH] glsl_to_tgsi: v2 Invalidate and revalidate uniform backing storage

[Ian Romanick]

On 01/02/2012 08:27 AM, Alex Deucher wrote:
> On Thu, Dec 22, 2011 at 9:35 AM, Vadim Girlin<vadimgirlin at gmail.com>  wrote:
>> If glUniform1i and friends are going to dump data directly in
>> driver-allocated, the pointers have to be updated when the storage
>> moves.  This should fix the regressions seen with commit 7199096.
>>
>> I'm not sure if this is the only place that needs this treatment.  I'm
>> a little uncertain about the various functions in st_glsl_to_tgsi that
>> modify the TGSI IR and try to propagate changes about that up to the
>> gl_program.  That seems sketchy to me.
>>
>> Signed-off-by: Ian Romanick<ian.d.romanick at intel.com>
>>
>> v2:
>>
>> Revalidate when shader_program is not NULL.
>> Update the pointers for all _LinkedShaders.
>> Init glsl_to_tgsi_visitor::shader_program to NULL in the
>> get_pixel_transfer_visitor&  get_bitmap_visitor.
>>
>> Signed-off-by: Vadim Girlin<vadimgirlin at gmail.com>
>> ---
>>
>> Based on the patch from Ian Romanick:
>> http://lists.freedesktop.org/archives/mesa-dev/2011-November/014675.html
>>
>> Fixes uniform regressions with r600g (and probably other drivers)
>> after commit 719909698c67c287a393d2380278e7b7495ae018
>
>
> Are there any objections to pushing this patch?

It seems okay to me.  I'll assume that it has received some testing and 
fixes things? :)

> Alex
>
>>
>> Tested on evergreen with r600.tests: no regressions.
>>
>>   src/mesa/state_tracker/st_glsl_to_tgsi.cpp |   25 +++++++++++++++++++++++++
>>   1 files changed, 25 insertions(+), 0 deletions(-)
>>
>> diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>> index 77aa0d1..fce92bb 100644
>> --- a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>> +++ b/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
>> @@ -3708,6 +3708,7 @@ get_pixel_transfer_visitor(struct st_fragment_program *fp,
>>     /* Copy attributes of the glsl_to_tgsi_visitor in the original shader. */
>>     v->ctx = original->ctx;
>>     v->prog = prog;
>> +   v->shader_program = NULL;
>>     v->glsl_version = original->glsl_version;
>>     v->native_integers = original->native_integers;
>>     v->options = original->options;
>> @@ -3837,6 +3838,7 @@ get_bitmap_visitor(struct st_fragment_program *fp,
>>     /* Copy attributes of the glsl_to_tgsi_visitor in the original shader. */
>>     v->ctx = original->ctx;
>>     v->prog = prog;
>> +   v->shader_program = NULL;
>>     v->glsl_version = original->glsl_version;
>>     v->native_integers = original->native_integers;
>>     v->options = original->options;
>> @@ -4550,6 +4552,15 @@ st_translate_program(
>>     t->pointSizeOutIndex = -1;
>>     t->prevInstWrotePointSize = GL_FALSE;
>>
>> +   if (program->shader_program) {
>> +      for (i = 0; i<  program->shader_program->NumUserUniformStorage; i++) {
>> +         struct gl_uniform_storage *const storage =
>> +&program->shader_program->UniformStorage[i];
>> +
>> +         _mesa_uniform_detach_all_driver_storage(storage);
>> +      }
>> +   }
>> +
>>     /*
>>      * Declare input attributes.
>>      */
>> @@ -4776,6 +4787,20 @@ st_translate_program(
>>                         t->insn[t->labels[i].branch_target]);
>>     }
>>
>> +   if (program->shader_program) {
>> +      /* This has to be done last.  Any operation the can cause
>> +       * prog->ParameterValues to get reallocated (e.g., anything that adds a
>> +       * program constant) has to happen before creating this linkage.
>> +       */
>> +      for (unsigned i = 0; i<  MESA_SHADER_TYPES; i++) {
>> +         if (program->shader_program->_LinkedShaders[i] == NULL)
>> +            continue;
>> +
>> +         _mesa_associate_uniform_storage(ctx, program->shader_program,
>> +               program->shader_program->_LinkedShaders[i]->Program->Parameters);
>> +      }
>> +   }
>> +
>>   out:
>>     if (t) {
>>        FREE(t->insn);
>> --
>> 1.7.7.4
>>
>> _______________________________________________
>> mesa-dev mailing list
>> mesa-dev at lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/mesa-dev