AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

radv: Chain cmd buffers in queue code when possible, not in winsys.

Browse Source 
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22220>
This commit is contained in:
Timur Kristóf
2023-03-30 18:54:35 +02:00
committed by Marge Bot
parent c876e99aa4
commit 6aa518ea86
1 changed files with 21 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/amd/vulkan/radv_queue.c
+21 -13
View File
@@ -1582,7 +1582,6 @@ static VkResult
radv_queue_submit_normal(struct radv_queue *queue, struct vk_queue_submit *submission)
{
struct radeon_winsys_ctx *ctx = queue->hw_ctx;
bool can_patch = true;
bool use_ace = false;
bool use_perf_counters = false;
VkResult result;
@@ -1596,9 +1595,6 @@ radv_queue_submit_normal(struct radv_queue *queue, struct vk_queue_submit *submi
return result;
if (use_ace) {
/* TODO: chaining with gang submit. */
can_patch = false;
result = radv_update_gang_preambles(queue);
if (result != VK_SUCCESS)
return result;
@@ -1652,9 +1648,6 @@ radv_queue_submit_normal(struct radv_queue *queue, struct vk_queue_submit *submi
/* RADV only supports perf counters on the GFX queue currently. */
assert(radv_queue_ring(queue) == AMD_IP_GFX);
/* Disallow chaining the lock/unlock CS. */
can_patch = false;
if (!perf_ctr_lock_cs || !perf_ctr_unlock_cs) {
result = VK_ERROR_OUT_OF_HOST_MEMORY;
goto fail;
@@ -1694,6 +1687,9 @@ radv_queue_submit_normal(struct radv_queue *queue, struct vk_queue_submit *submi
.uses_shadow_regs = queue->state.uses_shadow_regs,
};
const bool chaining_en = !(queue->device->instance->debug_flags & RADV_DEBUG_NO_IBS) &&
queue->device->physical_device->rad_info.gfx_level >= GFX7;
for (uint32_t j = 0, advance; j < cmd_buffer_count; j += advance) {
advance = MIN2(max_cs_submission, cmd_buffer_count - j);
const bool last_submit = j + advance == cmd_buffer_count;
@@ -1708,22 +1704,34 @@ radv_queue_submit_normal(struct radv_queue *queue, struct vk_queue_submit *submi
if (use_perf_counters)
cs_array[num_submitted_cs++] = perf_ctr_lock_cs;
struct radeon_cmdbuf *chainable = NULL;
struct radeon_cmdbuf *chainable_ace = NULL;
/* Add CS from submitted command buffers. */
for (unsigned c = 0; c < advance; ++c) {
struct radv_cmd_buffer *cmd_buffer =
(struct radv_cmd_buffer *)submission->command_buffers[j + c];
assert(cmd_buffer->vk.level == VK_COMMAND_BUFFER_LEVEL_PRIMARY);
const bool can_chain_next = chaining_en && !(cmd_buffer->usage_flags &
VK_COMMAND_BUFFER_USAGE_SIMULTANEOUS_USE_BIT);
/* ACE needs to be first because the last CS must match the queue's IP type. */
if (radv_cmd_buffer_needs_ace(cmd_buffer)) {
queue->device->ws->cs_unchain(cmd_buffer->ace_internal.cs);
cs_array[num_submitted_cs++] = cmd_buffer->ace_internal.cs;
if (!chainable_ace ||
!queue->device->ws->cs_chain(chainable_ace, cmd_buffer->ace_internal.cs, false))
cs_array[num_submitted_cs++] = cmd_buffer->ace_internal.cs;
chainable_ace = can_chain_next ? cmd_buffer->ace_internal.cs : NULL;
submit_ace = true;
}
can_patch &= !(cmd_buffer->usage_flags & VK_COMMAND_BUFFER_USAGE_SIMULTANEOUS_USE_BIT);
queue->device->ws->cs_unchain(cmd_buffer->cs);
cs_array[num_submitted_cs++] = cmd_buffer->cs;
if (!chainable ||
!queue->device->ws->cs_chain(chainable, cmd_buffer->cs, queue->state.uses_shadow_regs))
cs_array[num_submitted_cs++] = cmd_buffer->cs;
chainable = can_chain_next ? cmd_buffer->cs : NULL;
cs_idx = num_submitted_cs - 1;
}
@@ -1740,9 +1748,9 @@ radv_queue_submit_normal(struct radv_queue *queue, struct vk_queue_submit *submi
submit.cs_count = num_submitted_cs;
submit.preamble_count = submit_ace ? num_preambles : num_1q_preambles;
result = queue->device->ws->cs_submit(
ctx, &submit, j == 0 ? wait_count : 0, waits,
last_submit ? submission->signal_count : 0, submission->signals, can_patch);
result = queue->device->ws->cs_submit(ctx, &submit, j == 0 ? wait_count : 0, waits,
last_submit ? submission->signal_count : 0,
submission->signals, false);
if (result != VK_SUCCESS)
goto fail;