<div dir="auto"><div><br><div class="gmail_extra"><br><div class="gmail_quote">On 7 Sep. 2017 6:34 pm, "Nicolai Hähnle" <<a href="mailto:nhaehnle@gmail.com">nhaehnle@gmail.com</a>> wrote:<br type="attribution"><blockquote class="quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="quoted-text">On <a href="tel:07.09.2017%2009" value="+61709201709" target="_blank">07.09.2017 09</a>:58, Bas Nieuwenhuizen wrote:<br> <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> I'm not really happy with this, what happens if a game actually uses<br> e.g. indirect count in a secondary cmd buf?<br> </blockquote> <br></div> Note that some packets that require fetching data in the CP cannot be run in secondary command buffers, because the data fetching path overlaps with that used for fetching IB2. I thought this only affects SET_PREDICATION and OCCLUSION_QUERY though.<br></blockquote></div></div></div><div dir="auto"><br></div><div dir="auto"><br></div><div dir="auto">Can you check it, it may be multi indirect is busted on some firmwares, this seems cik specific.</div><div dir="auto"><br></div><div dir="auto">Also is set predication a problem for all Pred setting types?</div><div dir="auto"><br></div><div dir="auto">Dave.</div><div dir="auto"><div class="gmail_extra"><div class="gmail_quote"><blockquote class="quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"></blockquote></div></div></div><div dir="auto"><div class="gmail_extra"><div class="gmail_quote"><blockquote class="quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="elided-text"> <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> <br> That said this patch seems correct, so<br> <br> Reviewed-by: Bas Nieuwenhuizen <<a href="mailto:bas@basnieuwenhuizen.nl" target="_blank">bas@basnieuwenhuizen.nl</a>><br> <br> You may want to nominate this for stable (and provide a backport<br> probably..)<br> <br> On Thu, Sep 7, 2017, at 05:03, Dave Airlie wrote:<br> <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> From: Dave Airlie <<a href="mailto:airlied@redhat.com" target="_blank">airlied@redhat.com</a>><br> <br> This fixes some observed hangs on CIK GPUs.<br> <br> Signed-off-by: Dave Airlie <<a href="mailto:airlied@redhat.com" target="_blank">airlied@redhat.com</a>><br> ---<br> src/amd/vulkan/radv_cmd_buffer<wbr>.c | 37<br> +++++++++++++++++++++++-------<wbr>-------<br> 1 file changed, 23 insertions(+), 14 deletions(-)<br> <br> diff --git a/src/amd/vulkan/radv_cmd_buff<wbr>er.c<br> b/src/amd/vulkan/radv_cmd_buff<wbr>er.c<br> index b372123..bc4aeb3 100644<br> --- a/src/amd/vulkan/radv_cmd_buff<wbr>er.c<br> +++ b/src/amd/vulkan/radv_cmd_buff<wbr>er.c<br> @@ -2834,20 +2834,29 @@ radv_cs_emit_indirect_draw_pac<wbr>ket(struct<br> radv_cmd_buffer *cmd_buffer,<br> uint32_t base_reg = cmd_buffer->state.pipeline->gr<wbr>aphics.vtx_base_sgpr;<br> assert(base_reg);<br> - radeon_emit(cs, PKT3(indexed ? PKT3_DRAW_INDEX_INDIRECT_MULTI :<br> - PKT3_DRAW_INDIRECT_MULTI,<br> - 8, false));<br> - radeon_emit(cs, 0);<br> - radeon_emit(cs, (base_reg - SI_SH_REG_OFFSET) >> 2);<br> - radeon_emit(cs, ((base_reg + 4) - SI_SH_REG_OFFSET) >> 2);<br> - radeon_emit(cs, (((base_reg + 8) - SI_SH_REG_OFFSET) >> 2) |<br> - S_2C3_DRAW_INDEX_ENABLE(draw_<wbr>id_enable) |<br> - S_2C3_COUNT_INDIRECT_ENABLE(!<wbr>!count_va));<br> - radeon_emit(cs, draw_count); /* count */<br> - radeon_emit(cs, count_va); /* count_addr */<br> - radeon_emit(cs, count_va >> 32);<br> - radeon_emit(cs, stride); /* stride */<br> - radeon_emit(cs, di_src_sel);<br> + if (draw_count == 1 && !count_va && !draw_id_enable) {<br> + radeon_emit(cs, PKT3(indexed ? PKT3_DRAW_INDEX_INDIRECT :<br> + PKT3_DRAW_INDIRECT, 3, false));<br> + radeon_emit(cs, 0);<br> + radeon_emit(cs, (base_reg - SI_SH_REG_OFFSET) >> 2);<br> + radeon_emit(cs, ((base_reg + 4) - SI_SH_REG_OFFSET) >><br> 2);<br> + radeon_emit(cs, di_src_sel);<br> + } else {<br> + radeon_emit(cs, PKT3(indexed ?<br> PKT3_DRAW_INDEX_INDIRECT_MULTI :<br> + PKT3_DRAW_INDIRECT_MULTI,<br> + 8, false));<br> + radeon_emit(cs, 0);<br> + radeon_emit(cs, (base_reg - SI_SH_REG_OFFSET) >> 2);<br> + radeon_emit(cs, ((base_reg + 4) - SI_SH_REG_OFFSET) >><br> 2);<br> + radeon_emit(cs, (((base_reg + 8) - SI_SH_REG_OFFSET) >><br> 2) |<br> + S_2C3_DRAW_INDEX_ENABLE(draw_<wbr>id_enable) |<br> + S_2C3_COUNT_INDIRECT_ENABLE(!<wbr>!count_va));<br> + radeon_emit(cs, draw_count); /* count */<br> + radeon_emit(cs, count_va); /* count_addr */<br> + radeon_emit(cs, count_va >> 32);<br> + radeon_emit(cs, stride); /* stride */<br> + radeon_emit(cs, di_src_sel);<br> + }<br> }<br> static void<br> -- <br> 2.9.4<br> <br> ______________________________<wbr>_________________<br> mesa-dev mailing list<br> <a href="mailto:mesa-dev@lists.freedesktop.org" target="_blank">mesa-dev@lists.freedesktop.org</a><br> <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/mesa-dev</a><br> </blockquote> ______________________________<wbr>_________________<br> mesa-dev mailing list<br> <a href="mailto:mesa-dev@lists.freedesktop.org" target="_blank">mesa-dev@lists.freedesktop.org</a><br> <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/mesa-dev</a><br> <br> </blockquote> <br> <br></div><font color="#888888"> -- <br> Lerne, wie die Welt wirklich ist,<br> Aber vergiss niemals, wie sie sein sollte.</font><div class="elided-text"><br> ______________________________<wbr>_________________<br> mesa-dev mailing list<br> <a href="mailto:mesa-dev@lists.freedesktop.org" target="_blank">mesa-dev@lists.freedesktop.org</a><br> <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/mesa-dev</a><br> </div></blockquote></div><br></div></div></div>