4

v3dv: enable GPU jobs to signal multiple semaphores

 2 years ago
source link: https://gitlab.freedesktop.org/mesa/mesa/-/commit/fd973218a6c95d09cad8558712756a772ae78f66
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
neoserver,ios ssh client

Commit fd973218

authored 8 months ago by

Melissa Wen

Browse files

v3dv: enable GPU jobs to signal multiple semaphores 


In addition to keep a copy of wait semaphores, extend
v3dv_submit_info_semaphores to hold a copy of signal semaphores too.
With a copy of wait and signal semaphores, we can enable GPU jobs to
handle more than one wait and signal semaphores.

By now, we don't change the way as we signalling semaphores when all
jobs complete, i.e., we still use the master thread to signal
semaphores. In this context, no GPU job is actually in charge of
signalling, but the support for multiple signal semaphores is done
here.
Signed-off-by: Melissa Wen <[email protected]>
Reviewed-by: Iago Toral Quiroga <[email protected]>
Part-of: <!13178>

1 merge request!13178v3dv: add support to multiple wait and signal semaphores

Showing
......@@ -957,8 +957,12 @@ struct v3dv_copy_query_results_cpu_job_info {
struct v3dv_submit_info_semaphores {
/* List of semaphores to wait before running a job */
uint32_t sem_count;
VkSemaphore *sems;
uint32_t wait_sem_count;
VkSemaphore *wait_sems;
/* List of semaphores to signal when all jobs complete */
uint32_t signal_sem_count;
VkSemaphore *signal_sems;
};
struct v3dv_event_set_cpu_job_info {
......
......@@ -89,7 +89,7 @@ get_absolute_timeout(uint64_t timeout)
static VkResult
queue_submit_job(struct v3dv_queue *queue,
struct v3dv_job *job,
struct v3dv_submit_info_semaphores *wait_sems_info,
struct v3dv_submit_info_semaphores *sems_info,
pthread_t *wait_thread);
/* Waits for active CPU wait threads spawned before the current thread to
......@@ -297,15 +297,24 @@ copy_semaphores_info(struct v3dv_device *device,
if (!info_copy)
return NULL;
result = copy_semaphores(device, info->sems, info->sem_count,
&info_copy->sems, &info_copy->sem_count);
result = copy_semaphores(device, info->wait_sems, info->wait_sem_count,
&info_copy->wait_sems, &info_copy->wait_sem_count);
if (result != VK_SUCCESS)
goto fail;
result = copy_semaphores(device, info->signal_sems, info->signal_sem_count,
&info_copy->signal_sems,
&info_copy->signal_sem_count);
if (result != VK_SUCCESS)
goto fail;
return info_copy;
fail:
if (info_copy->wait_sem_count > 0)
vk_free(&device->vk.alloc, info_copy->wait_sems);
vk_free(&device->vk.alloc, info_copy);
return NULL;
}
......@@ -315,8 +324,12 @@ free_semaphores_info(struct v3dv_device *device,
{
assert(sems_info != NULL);
if (sems_info->sem_count > 0)
vk_free(&device->vk.alloc, sems_info->sems);
if (sems_info->wait_sem_count > 0)
vk_free(&device->vk.alloc, sems_info->wait_sems);
if (sems_info->signal_sem_count > 0)
vk_free(&device->vk.alloc, sems_info->signal_sems);
vk_free(&device->vk.alloc, sems_info);
}
......@@ -410,7 +423,7 @@ spawn_event_wait_thread(struct v3dv_job *job, pthread_t *wait_thread)
static VkResult
handle_wait_events_cpu_job(struct v3dv_job *job,
struct v3dv_submit_info_semaphores *wait_sems_info,
struct v3dv_submit_info_semaphores *sems_info,
pthread_t *wait_thread)
{
assert(job->type == V3DV_JOB_TYPE_CPU_WAIT_EVENTS);
......@@ -438,10 +451,10 @@ handle_wait_events_cpu_job(struct v3dv_job *job,
/* As events can be signaled by the host, jobs after the event wait must
* still wait for semaphores, if any. So, whenever we spawn a wait thread,
* we keep a copy of wait semaphores (info->sems_info) to be used when
* we keep a copy of the semaphores (info->sems_info) to be used when
* submitting pending jobs in the wait thread context.
*/
info->sems_info = copy_semaphores_info(job->device, wait_sems_info);
info->sems_info = copy_semaphores_info(job->device, sems_info);
if (!info->sems_info)
return VK_ERROR_OUT_OF_HOST_MEMORY;
......@@ -691,17 +704,19 @@ set_multisync(struct drm_v3d_multi_sync *ms,
struct v3dv_device *device,
struct drm_v3d_sem *out_syncs,
struct drm_v3d_sem *in_syncs,
bool do_sem_signal,
bool serialize,
enum v3d_queue queue)
{
uint32_t out_sync_count = 0, in_sync_count = 0;
/* pSignalSemaphores are handled in the master thread (the reason for
* out_sync_count being initially zero), but we still want to always signal
* last_job_sync so we can serialize jobs when needed.
/* We only want to signal out semaphores for this submission upon
* completion of the last job involved with it. We still want to always
* signal last_job_sync so we can serialize jobs when needed.
*/
out_sync_count = 0;
out_syncs = set_syncs(device, &out_sync_count, NULL, device->last_job_sync);
out_sync_count = do_sem_signal ? sems_info->signal_sem_count : 0;
out_syncs = set_syncs(device, &out_sync_count, sems_info->signal_sems,
device->last_job_sync);
assert(out_sync_count > 0);
......@@ -712,8 +727,8 @@ set_multisync(struct drm_v3d_multi_sync *ms,
* it wait for completion of the last job submitted, so in that case we can
* skip waiting for any additional semaphores.
*/
in_sync_count = serialize ? 0 : sems_info->sem_count;
in_syncs = set_syncs(device, &in_sync_count, sems_info->sems,
in_sync_count = serialize ? 0 : sems_info->wait_sem_count;
in_syncs = set_syncs(device, &in_sync_count, sems_info->wait_sems,
(serialize ? device->last_job_sync : 0));
if (!in_syncs && in_sync_count) {
vk_free(&device->vk.alloc, out_syncs);
......@@ -788,7 +803,7 @@ handle_cl_job(struct v3dv_queue *queue,
* dependencies strictly through barriers.
*/
const bool needs_bcl_sync =
sems_info->sem_count > 0 || job->needs_bcl_sync;
sems_info->wait_sem_count > 0 || job->needs_bcl_sync;
const bool needs_rcl_sync = job->serialize && !needs_bcl_sync;
struct drm_v3d_sem *out_syncs = NULL, *in_syncs = NULL;
......@@ -799,7 +814,12 @@ handle_cl_job(struct v3dv_queue *queue,
*/
if (device->pdevice->caps.multisync) {
struct drm_v3d_multi_sync ms = { 0 };
set_multisync(&ms, sems_info, NULL, device, out_syncs, in_syncs,
/* We are processing all signal VkSemaphores together in the submit
* master thread and therefore we don't handle signal VkSemaphores in cl
* submission yet. For this reason, we set do_sem_signal to false in the
* multisync extension.
*/
set_multisync(&ms, sems_info, NULL, device, out_syncs, in_syncs, false,
job->serialize, needs_rcl_sync ? V3D_RENDER : V3D_BIN);
if (!ms.base.id)
return vk_error(device->instance, VK_ERROR_OUT_OF_HOST_MEMORY);
......@@ -845,7 +865,7 @@ handle_tfu_job(struct v3dv_queue *queue,
{
struct v3dv_device *device = queue->device;
const bool needs_sync = sems_info->sem_count || job->serialize;
const bool needs_sync = sems_info->wait_sem_count || job->serialize;
struct drm_v3d_sem *out_syncs = NULL, *in_syncs = NULL;
mtx_lock(&device->mutex);
......@@ -855,7 +875,12 @@ handle_tfu_job(struct v3dv_queue *queue,
*/
if (device->pdevice->caps.multisync) {
struct drm_v3d_multi_sync ms = { 0 };
set_multisync(&ms, sems_info, NULL, device, out_syncs, in_syncs,
/* We are processing all signal VkSemaphores together in the submit
* master thread and therefore we don't handle signal VkSemaphores in
* tfu jobs yet. For this reason, we set do_sem_signal to false in the
* multisync extension.
*/
set_multisync(&ms, sems_info, NULL, device, out_syncs, in_syncs, false,
job->serialize, V3D_TFU);
if (!ms.base.id)
return vk_error(device->instance, VK_ERROR_OUT_OF_HOST_MEMORY);
......@@ -904,7 +929,7 @@ handle_csd_job(struct v3dv_queue *queue,
assert(bo_idx == submit->bo_handle_count);
submit->bo_handles = (uintptr_t)(void *)bo_handles;
const bool needs_sync = sems_info->sem_count || job->serialize;
const bool needs_sync = sems_info->wait_sem_count || job->serialize;
struct drm_v3d_sem *out_syncs = NULL, *in_syncs = NULL;
mtx_lock(&queue->device->mutex);
......@@ -913,7 +938,12 @@ handle_csd_job(struct v3dv_queue *queue,
*/
if (device->pdevice->caps.multisync) {
struct drm_v3d_multi_sync ms = { 0 };
set_multisync(&ms, sems_info, NULL, device, out_syncs, in_syncs,
/* We are processing all signal VkSemaphores together in the submit
* master thread and therefore we don't handle signal VkSemaphores in
* csd jobs yet. For this reason, we set do_sem_signal to false in the
* multisync extension.
*/
set_multisync(&ms, sems_info, NULL, device, out_syncs, in_syncs, false,
job->serialize, V3D_CSD);
if (!ms.base.id)
return vk_error(device->instance, VK_ERROR_OUT_OF_HOST_MEMORY);
......@@ -952,18 +982,18 @@ handle_csd_job(struct v3dv_queue *queue,
static VkResult
queue_submit_job(struct v3dv_queue *queue,
struct v3dv_job *job,
struct v3dv_submit_info_semaphores *wait_sems_info,
struct v3dv_submit_info_semaphores *sems_info,
pthread_t *wait_thread)
{
assert(job);
switch (job->type) {
case V3DV_JOB_TYPE_GPU_CL:
return handle_cl_job(queue, job, wait_sems_info);
return handle_cl_job(queue, job, sems_info);
case V3DV_JOB_TYPE_GPU_TFU:
return handle_tfu_job(queue, job, wait_sems_info);
return handle_tfu_job(queue, job, sems_info);
case V3DV_JOB_TYPE_GPU_CSD:
return handle_csd_job(queue, job, wait_sems_info);
return handle_csd_job(queue, job, sems_info);
case V3DV_JOB_TYPE_CPU_RESET_QUERIES:
return handle_reset_query_cpu_job(job);
case V3DV_JOB_TYPE_CPU_END_QUERY:
......@@ -973,11 +1003,11 @@ queue_submit_job(struct v3dv_queue *queue,
case V3DV_JOB_TYPE_CPU_SET_EVENT:
return handle_set_event_cpu_job(job);
case V3DV_JOB_TYPE_CPU_WAIT_EVENTS:
return handle_wait_events_cpu_job(job, wait_sems_info, wait_thread);
return handle_wait_events_cpu_job(job, sems_info, wait_thread);
case V3DV_JOB_TYPE_CPU_COPY_BUFFER_TO_IMAGE:
return handle_copy_buffer_to_image_cpu_job(job);
case V3DV_JOB_TYPE_CPU_CSD_INDIRECT:
return handle_csd_indirect_cpu_job(queue, job, wait_sems_info);
return handle_csd_indirect_cpu_job(queue, job, sems_info);
case V3DV_JOB_TYPE_CPU_TIMESTAMP_QUERY:
return handle_timestamp_query_cpu_job(job);
default:
......@@ -1002,7 +1032,7 @@ queue_create_noop_job(struct v3dv_queue *queue)
static VkResult
queue_submit_noop_job(struct v3dv_queue *queue,
struct v3dv_submit_info_semaphores *wait_sems_info)
struct v3dv_submit_info_semaphores *sems_info)
{
/* VkQueue host access is externally synchronized so we don't need to lock
* here for the static variable.
......@@ -1013,25 +1043,24 @@ queue_submit_noop_job(struct v3dv_queue *queue,
return result;
}
return queue_submit_job(queue, queue->noop_job, wait_sems_info, NULL);
return queue_submit_job(queue, queue->noop_job, sems_info, NULL);
}
static VkResult
queue_submit_cmd_buffer(struct v3dv_queue *queue,
struct v3dv_cmd_buffer *cmd_buffer,
struct v3dv_submit_info_semaphores *wait_sems_info,
struct v3dv_submit_info_semaphores *sems_info,
pthread_t *wait_thread)
{
assert(cmd_buffer);
assert(cmd_buffer->status == V3DV_CMD_BUFFER_STATUS_EXECUTABLE);
if (list_is_empty(&cmd_buffer->jobs))
return queue_submit_noop_job(queue, wait_sems_info);
return queue_submit_noop_job(queue, sems_info);
list_for_each_entry_safe(struct v3dv_job, job,
&cmd_buffer->jobs, list_link) {
VkResult result = queue_submit_job(queue, job, wait_sems_info,
wait_thread);
VkResult result = queue_submit_job(queue, job, sems_info, wait_thread);
if (result != VK_SUCCESS)
return result;
}
......@@ -1118,9 +1147,11 @@ queue_submit_cmd_buffer_batch(struct v3dv_queue *queue,
/* Wrap wait semaphores info from VkSubmitInfo to use it whenever we need
* the data to submit all jobs in the same command buffer batch.
*/
struct v3dv_submit_info_semaphores wait_sems_info = {
.sem_count = pSubmit->waitSemaphoreCount,
.sems = (VkSemaphore *) pSubmit->pWaitSemaphores,
struct v3dv_submit_info_semaphores sems_info = {
.wait_sem_count = pSubmit->waitSemaphoreCount,
.wait_sems = (VkSemaphore *) pSubmit->pWaitSemaphores,
.signal_sem_count = pSubmit->signalSemaphoreCount,
.signal_sems = (VkSemaphore *) pSubmit->pSignalSemaphores,
};
/* Even if we don't have any actual work to submit we still need to wait
......@@ -1129,13 +1160,13 @@ queue_submit_cmd_buffer_batch(struct v3dv_queue *queue,
* to do anything special, it should not be a common case anyway.
*/
if (pSubmit->commandBufferCount == 0) {
result = queue_submit_noop_job(queue, &wait_sems_info);
result = queue_submit_noop_job(queue, &sems_info);
} else {
for (uint32_t i = 0; i < pSubmit->commandBufferCount; i++) {
pthread_t wait_thread;
struct v3dv_cmd_buffer *cmd_buffer =
v3dv_cmd_buffer_from_handle(pSubmit->pCommandBuffers[i]);
result = queue_submit_cmd_buffer(queue, cmd_buffer, &wait_sems_info,
result = queue_submit_cmd_buffer(queue, cmd_buffer, &sems_info,
&wait_thread);
/* We get VK_NOT_READY if we had to spawn a wait thread for the
......
Please register or sign in to comment

report

Recommend

About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK