Matt Turner <matts...@gmail.com> writes: > Ken did this earlier, and this is just me reimplementing his patch a > little differently.
Heh, it seems a little mean to Ken to revert his patch only to implement
almost the same thing a few commits later. I suggest you squash the
revert and this patch into PATCHv2 9 before pushing. Either way the end
result looks reasonable so the three patches (4, 9, 10) are:
Reviewed-by: Francisco Jerez <curroje...@riseup.net>
Thanks.
> ---
> src/mesa/drivers/dri/i965/brw_schedule_instructions.cpp | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_schedule_instructions.cpp
> b/src/mesa/drivers/dri/i965/brw_schedule_instructions.cpp
> index befa9ff..8d92584 100644
> --- a/src/mesa/drivers/dri/i965/brw_schedule_instructions.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_schedule_instructions.cpp
> @@ -897,7 +897,8 @@ is_scheduling_barrier(const fs_inst *inst)
> {
> return inst->opcode == FS_OPCODE_PLACEHOLDER_HALT ||
> inst->is_control_flow() ||
> - inst->has_side_effects();
> + inst->eot ||
> + (inst->has_side_effects() && inst->opcode != FS_OPCODE_FB_WRITE);
> }
>
> void
> --
> 2.4.10
signature.asc
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
