On 03/05/16 19:57, Jordan Justen wrote:
> On 2016-05-03 05:21:50, Samuel Iglesias Gonsálvez wrote:
>> From: Iago Toral Quiroga <ito...@igalia.com>
>>
>> This can happen if the register already has a non-zero subreg_offset
>> when byte_offset() is called.
>> ---
>> src/mesa/drivers/dri/i965/brw_ir_fs.h | 4 ++++
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/src/mesa/drivers/dri/i965/brw_ir_fs.h
>> b/src/mesa/drivers/dri/i965/brw_ir_fs.h
>> index 2f72326..8faca08 100644
>> --- a/src/mesa/drivers/dri/i965/brw_ir_fs.h
>> +++ b/src/mesa/drivers/dri/i965/brw_ir_fs.h
>> @@ -109,6 +109,10 @@ byte_offset(fs_reg reg, unsigned delta)
>> assert(delta == 0);
>> }
>> reg.subreg_offset += delta % 32;
>> + if (reg.subreg_offset >= 32) {
>> + reg.reg_offset += reg.subreg_offset / 32;
>> + reg.subreg_offset %= 32;
>> + }
>
> What if at the top of the function, we have
>
> reg.subreg_offset += delta;
>
> Then, in the middle we use reg.subreg_offset instead of delta for
> adjusting the register offset.
>
> Then, at the end:
>
> reg.subreg_offset = reg.subreg_offset % 32;
>
Yes, it would work. I am going to adjust the UNIFORM case to do the same
but with 4 instead of 32.
Sam
> -Jordan
>
>> return reg;
>> }
>>
>> --
>> 2.5.0
>>
>> _______________________________________________
>> mesa-dev mailing list
>> mesa-dev@lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
