On 15/02/2017 16:33, Chris Wilson wrote:
On Wed, Feb 15, 2017 at 04:06:34PM +0000, Tvrtko Ursulin wrote:+static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset) +{ + static const u32 pc6[6] = { GFX_OP_PIPE_CONTROL(6), 0, 0, 0, 0, 0 }; + + memcpy(batch, pc6, sizeof(pc6)); + + batch[1] = flags; + batch[2] = offset; + + return batch + 6;godbolt would seem to say it is best to use static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset) { batch[0] = GFX_OP_PIPE_CONTROL(6); batch[1] = flags; batch[2] = offset; batch[3] = 0; batch[4] = 0; batch[5] = 0; return batch + 6; }
Yeah agreed, it was a bit silly. I falsely remember it had quite good effects on the optimisation gcc was able to do but couldn't repro that.
How about though replacing the last three assignments with memset(&batch[3], 0, 3 * sizeof(u32))? That is indeed helpful on 64-bit.
Regards, Tvrtko _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
