[Mesa-dev] [PATCH] i965/eu: use simd8 when exec_size != EXECUTE_16

Francisco Jerez currojerez at riseup.net
Wed Jun 1 18:44:29 UTC 2016


Alejandro Piñeiro <apinheiro at igalia.com> writes:

> Among other thigs, fix a gpu hang when using INTEL_DEBUG=shader_time
> for any shader.
>
> Signed-off-by: Jason Ekstrand <jason at jlekstrand.net>
> Signed-off-by: Alejandro Piñeiro <apinheiro at igalia.com>
> ---
>
> This is the change suggested by Jason (so I added him as signed-off). I just
> run a full piglit run to check that there isn't any regression.
>
>  src/mesa/drivers/dri/i965/brw_eu_emit.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_eu_emit.c b/src/mesa/drivers/dri/i965/brw_eu_emit.c
> index 2538f0d..8218f9c 100644
> --- a/src/mesa/drivers/dri/i965/brw_eu_emit.c
> +++ b/src/mesa/drivers/dri/i965/brw_eu_emit.c
> @@ -2909,7 +2909,7 @@ brw_set_dp_untyped_atomic_message(struct brw_codegen *p,
>  
>     if (devinfo->gen >= 8 || devinfo->is_haswell) {
>        if (brw_inst_access_mode(devinfo, p->current) == BRW_ALIGN_1) {
> -         if (brw_inst_exec_size(devinfo, p->current) == BRW_EXECUTE_8)
> +         if (brw_inst_exec_size(devinfo, p->current) != BRW_EXECUTE_16)
>              msg_control |= 1 << 4; /* SIMD8 mode */
>  
>           brw_inst_set_dp_msg_type(devinfo, insn,
> @@ -2922,7 +2922,7 @@ brw_set_dp_untyped_atomic_message(struct brw_codegen *p,
>        brw_inst_set_dp_msg_type(devinfo, insn,
>                                 GEN7_DATAPORT_DC_UNTYPED_ATOMIC_OP);
>  
> -      if (brw_inst_exec_size(devinfo, p->current) == BRW_EXECUTE_8)
> +      if (brw_inst_exec_size(devinfo, p->current) != BRW_EXECUTE_16)

LGTM,

Reviewed-by: Francisco Jerez <currojerez at riseup.net>

>           msg_control |= 1 << 4; /* SIMD8 mode */
>     }
>  
> -- 
> 2.7.4
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 212 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20160601/0ba6ceb5/attachment.sig>


More information about the mesa-dev mailing list