[Mesa-dev] [PATCH] mesa: Fix setup of LocalParams array.

[Brian Paul]

On 11/18/2013 04:25 PM, Brian Paul wrote:
> On 11/18/2013 11:07 AM, Eric Anholt wrote:
>> i965 passed piglit, but swrast and gallium both segfaulted without this.
>> i965 happened to work because it never ran _mesa_load_state_parameters()
>> on the new program before the test called glProgramLocalParameter(),
>> which
>> was allocating a LocalParams array for the fallback path.
>>
>> Bugzilla:
>> https://urldefense.proofpoint.com/v1/url?u=https://bugs.freedesktop.org/show_bug.cgi?id%3D71734&k=oIvRg1%2BdGAgOoM1BIlLLqw%3D%3D%0A&r=lGQMzzTgII0I7jefp2FHq7WtZ%2BTLs8wadB%2BiIj9xpBY%3D%0A&m=n3QosB0LfKkjHz3sK6d%2F4MBWkl%2BZgLi1HA1PNmcAH4Q%3D%0A&s=8ce86f72ab8d603739e1c773d926cb29e6f912ee659e0e7181ee3da13da6ec7b
>>
>> ---
>>   src/mesa/program/arbprogparse.c | 1 +
>>   1 file changed, 1 insertion(+)
>>
>> diff --git a/src/mesa/program/arbprogparse.c
>> b/src/mesa/program/arbprogparse.c
>> index 5b96650..5d92176 100644
>> --- a/src/mesa/program/arbprogparse.c
>> +++ b/src/mesa/program/arbprogparse.c
>> @@ -127,6 +127,7 @@ _mesa_parse_arb_fragment_program(struct
>> gl_context* ctx, GLenum target,
>>      if (program->Base.Parameters)
>>         _mesa_free_parameter_list(program->Base.Parameters);
>>      program->Base.Parameters    = prog.Parameters;
>> +   program->Base.LocalParams = prog.LocalParams;
>>
>>      /* Append fog instructions now if the program has "OPTION
>> ARB_fog_exp"
>>       * or similar.  We used to leave this up to drivers, but it appears
>>
>
> The same change needs to be done in the _mesa_parse_arb_vertex_program()
> function too:
>
> @@ -205,6 +206,7 @@ _mesa_parse_arb_vertex_program(struct gl_context
> *ctx, GLenu
>      if (program->Base.Parameters)
>         _mesa_free_parameter_list(program->Base.Parameters);
>      program->Base.Parameters = prog.Parameters;
> +   program->Base.LocalParams = prog.LocalParams;
>
>   #if DEBUG_VP
>      printf("____________Vertex program %u __________\n",
> program->Base.Id);

And yes, it's lame that the same code is in two places.  I could fix 
that up in a follow-on patch.

-Brian