<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Mon, Jan 8, 2018 at 4:57 AM, Iago Toral Quiroga <<a href="mailto:itoral@igalia.com" target="_blank">itoral@igalia.com</a>> wrote: <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">From the Vulkan spec with KHX extensions: "If queries are used while executing a render pass instance that has multiview enabled, the query uses N consecutive query indices in the query pool (starting at query) where N is the number of bits set in the view mask in the subpass the query is used in. How the numerical results of the query are distributed among the queries is implementation-dependent. For example, some implementations may write each view's results to a distinct query, while other implementations may write the total result to the first query and write zero to the other queries. However, the sum of the results in all the queries must accurately reflect the total result of the query summed over all views. Applications can sum the results from all the queries to compute the total result." In our case we only really emit a single query (in the first query index) that stores the aggregated result for all views, but we still need to manage availability for all the other query indices involved, even if we don't actually use them. This is relevant when clients call vkGetQueryPoolResults and pass all N queries to retrieve the results. In that scenario, without this patch, we will never see queries other than the first being available since we never emit them. Fixes test failures in some work-in-progress CTS multiview+query tests. --- src/intel/vulkan/genX_query.c | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) diff --git a/src/intel/vulkan/genX_query.c b/src/intel/vulkan/genX_query.c index 7683d0d1e3..231c605b6b 100644 --- a/src/intel/vulkan/genX_query.c +++ b/src/intel/vulkan/genX_query.c @@ -462,6 +462,24 @@ void genX(CmdEndQuery)( default: unreachable(""); } + + /* When multiview is active the spec requires that N consecutive query + * indices are used, where N is the number of active views in the subpass. + * The spec allows that we only write the results to one of the queries + * but we still need to manage result availability for all the query indices. + * Since we only emit a single query for all active views in the + * first index, mark the other query indices as being already available + * with result 0. + */ + if (!cmd_buffer->state.subpass || !cmd_buffer->state.subpass->view_mask) + return; </blockquote><div> </div><div>I think this would be better as just an if instead of an early return. </div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> + + uint32_t num_queries = _mesa_bitcount(cmd_buffer->state.subpass->view_mask); + for (uint32_t i = 1; i < num_queries; i++) { + uint64_t *slot = pool->bo.map + (query + i) * pool->stride; + slot[0] = 1; + memset(&slot[1], 0, sizeof(uint64_t) * pool->stride); </blockquote><div> </div><div>We can't set this from the CPU, we need to emit an MI_STORE_DATA_IMM and call emit_query_availability. </div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> + } } #define TIMESTAMP 0x2358 @@ -504,6 +522,24 @@ void genX(CmdWriteTimestamp)( } emit_query_availability(cmd_buffer, &pool->bo, offset); + + /* When multiview is active the spec requires that N consecutive query + * indices are used, where N is the number of active views in the subpass. + * The spec allows that we only write the results to one of the queries + * but we still need to manage result availability for all the query indices. + * Since we only emit a single query for all active views in the + * first index, mark the other query indices as being already available + * with result 0. + */ + if (!cmd_buffer->state.subpass || !cmd_buffer->state.subpass->view_mask) + return; + + uint32_t num_queries = _mesa_bitcount(cmd_buffer->state.subpass->view_mask); + for (uint32_t i = 1; i < num_queries; i++) { + uint64_t *slot = pool->bo.map + (query + i) * pool->stride; + slot[0] = 1; + memset(&slot[1], 0, sizeof(uint64_t) * pool->stride); </blockquote><div> </div><div>Same comments here. </div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> + } } #if GEN_GEN > 7 || GEN_IS_HASWELL -- 2.11.0 </blockquote></div> </div></div>