drm/amdgpu: fix vulkan test performance drop and hang on VI
Subsystems (3)
PATCH 1/1 drm/amdgpu: fix vulkan test performance drop and hang on VI
caused by not program dynamic_cu_mask_addr in the KIQ MQD.
v2: create struct vi_mqd_allocation in FB which will contain
1. PM4 MQD structure.
2. Write Pointer Poll Memory.
3. Read Pointer Report Memory
4. Dynamic CU Mask.
5. Dynamic RB Mask.
--- drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 27 ++-- drivers/gpu/drm/amd/include/vi_structs.h | 268 +++++++++++++++++++++++++++++++ 2 files changed, 285 insertions(+), 10 deletions(-)
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +17 -10
@@ -40,7 +40,6 @@#include "bif/bif_5_0_d.h" #include "bif/bif_5_0_sh_mask.h" -#include "gca/gfx_8_0_d.h" #include "gca/gfx_8_0_enum.h" #include "gca/gfx_8_0_sh_mask.h"
@@ -2100,7 +2099,7 @@ static int gfx_v8_0_sw_init(void *handle)return r;/* create MQD for all compute queues as well as KIQ for SRIOV case */ - r =amdgpu_gfx_compute_mqd_sw_init (adev,sizeof (struct vi_mqd )); + r =amdgpu_gfx_compute_mqd_sw_init (adev,sizeof (struct vi_mqd_allocation ));if (r)return r;
@@ -4715,9 +4714,6 @@ static int gfx_v8_0_mqd_init(struct amdgpu_ring *ring)uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;uint32_t tmp; -/* init the mqd struct */ -memset (mqd,0 ,sizeof (struct vi_mqd )); - mqd->header =0xC0310800 ; mqd->compute_pipelinestat_enable =0x00000001 ; mqd->compute_static_thread_mgmt_se0 =0xffffffff ;
@@ -4725,7 +4721,12 @@ static int gfx_v8_0_mqd_init(struct amdgpu_ring *ring) mqd->compute_static_thread_mgmt_se2 =0xffffffff ; mqd->compute_static_thread_mgmt_se3 =0xffffffff ; mqd->compute_misc_reserved =0x00000003 ; - +if (!(adev->flags &AMD_IS_APU )) { + mqd->dynamic_cu_mask_addr_lo =lower_32_bits (ring->mqd_gpu_addr + + offsetof(struct vi_mqd_allocation , dyamic_cu_mask)); + mqd->dynamic_cu_mask_addr_hi =upper_32_bits (ring->mqd_gpu_addr + + offsetof(struct vi_mqd_allocation , dyamic_cu_mask)); + } eop_base_addr = ring->eop_gpu_addr >>8 ; mqd->cp_hqd_eop_base_addr_lo = eop_base_addr; mqd->cp_hqd_eop_base_addr_hi =upper_32_bits (eop_base_addr);
@@ -4900,7 +4901,7 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)if (adev->gfx.in_reset) {/* for GPU_RESET case */ /* reset MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) -memcpy (mqd, adev->gfx.mec.mqd_backup[mqd_idx],sizeof (*mqd)); +memcpy (mqd, adev->gfx.mec.mqd_backup[mqd_idx],sizeof (struct vi_mqd_allocation ));/* reset ring buffer */ ring->wptr =0 ;
@@ -4916,6 +4917,9 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)vi_srbm_select (adev,0 ,0 ,0 ,0 );mutex_unlock (&adev->srbm_mutex); }else { +memset ((void *)mqd,0 ,sizeof (struct vi_mqd_allocation )); + ((struct vi_mqd_allocation *)mqd)->dyamic_cu_mask =0xFFFFFFFF ; + ((struct vi_mqd_allocation *)mqd)->dyamic_rb_mask =0xFFFFFFFF ;mutex_lock (&adev->srbm_mutex);vi_srbm_select (adev, ring->me, ring->pipe, ring->queue,0 );gfx_v8_0_mqd_init (ring);
@@ -4929,7 +4933,7 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)mutex_unlock (&adev->srbm_mutex);if (adev->gfx.mec.mqd_backup[mqd_idx]) -memcpy (adev->gfx.mec.mqd_backup[mqd_idx], mqd,sizeof (*mqd)); +memcpy (adev->gfx.mec.mqd_backup[mqd_idx], mqd,sizeof (struct vi_mqd_allocation )); }return r;
@@ -4947,6 +4951,9 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)int mqd_idx = ring - &adev->gfx.compute_ring[0 ];if (!adev->gfx.in_reset && !adev->gfx.in_suspend) { +memset ((void *)mqd,0 ,sizeof (struct vi_mqd_allocation )); + ((struct vi_mqd_allocation *)mqd)->dyamic_cu_mask =0xFFFFFFFF ; + ((struct vi_mqd_allocation *)mqd)->dyamic_rb_mask =0xFFFFFFFF ;mutex_lock (&adev->srbm_mutex);vi_srbm_select (adev, ring->me, ring->pipe, ring->queue,0 );gfx_v8_0_mqd_init (ring);
@@ -4954,11 +4961,11 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)mutex_unlock (&adev->srbm_mutex);if (adev->gfx.mec.mqd_backup[mqd_idx]) -memcpy (adev->gfx.mec.mqd_backup[mqd_idx], mqd,sizeof (*mqd)); +memcpy (adev->gfx.mec.mqd_backup[mqd_idx], mqd,sizeof (struct vi_mqd_allocation )); }else if (adev->gfx.in_reset) {/* for GPU_RESET case */ /* reset MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) -memcpy (mqd, adev->gfx.mec.mqd_backup[mqd_idx],sizeof (*mqd)); +memcpy (mqd, adev->gfx.mec.mqd_backup[mqd_idx],sizeof (struct vi_mqd_allocation ));/* reset ring buffer */ ring->wptr =0 ;amdgpu_ring_clear_ring (ring);
drivers/gpu/drm/amd/include/vi_structs.h +268 -0
274 lines not on this page · raw message
Signature · 6 lines
1.9.1 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): 1499076777-5170-1-git-send-email-Rex.Zhu-5C7GfCeVMHo@public.gmane.org
Deucher, Alexander
Reviewed-by Alex DeucherQuoted 18 lines
-----Original Message-----
From: amd-gfx [mailto:[email protected]] On Behalf Of Rex Zhu
Sent: Monday, July 03, 2017 6:13 AM
To: [email protected]
Cc: Zhu, Rex
Subject: [PATCH] drm/amdgpu: fix vulkan test performance drop and hang on VIcaused by not program
dynamic_cu_mask_addrin the KIQ MQD.v2: create struct
vi_mqd_allocationin FB which will contain
1. PM4 MQD structure.
2. Write Pointer Poll Memory.
3. Read Pointer Report Memory
4. Dynamic CU Mask.
5. Dynamic RB Mask.
Quoted 2 lines
Change-Id I22c840f1bf8d365f7df33a27d6b11e1aea8f2958Signed-off-by Rex Zhu <[email protected]>
Quoted 414 lines
--- drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c | 27 ++-- drivers/gpu/drm/amd/include/vi_structs.h | 268 +++++++++++++++++++++++++++++++ 2 files changed, 285 insertions(+), 10 deletions(-)drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-2@@ -40,7 +40,6 @@#include "bif/bif_5_0_d.h" #include "bif/bif_5_0_sh_mask.h" -#include "gca/gfx_8_0_d.h" #include "gca/gfx_8_0_enum.h" #include "gca/gfx_8_0_sh_mask.h" @@ -2100,7 +2099,7 @@ static int gfx_v8_0_sw_init(void *handle)return r;/* create MQD for all compute queues as well as KIQ for SRIOV case */ - r =amdgpu_gfx_compute_mqd_sw_init (adev,sizeof (struct vi_mqd)); + r =amdgpu_gfx_compute_mqd_sw_init (adev,sizeof (struct vi_mqd_allocation ));drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0
-3if (r)return r;@@ -4715,9 +4714,6 @@ static int gfx_v8_0_mqd_init(struct amdgpu_ring *ring)uint64_t hqd_gpu_addr, wb_gpu_addr, eop_base_addr;uint32_t tmp; -/* init the mqd struct */ -memset (mqd,0 ,sizeof (struct vi_mqd )); - mqd->header =0xC0310800 ; mqd->compute_pipelinestat_enable =0x00000001 ;drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +2
-1mqd->compute_static_thread_mgmt_se0 =0xffffffff ;@@ -4725,7 +4721,12 @@ static int gfx_v8_0_mqd_init(struct amdgpu_ring *ring) mqd->compute_static_thread_mgmt_se2 =0xffffffff ; mqd->compute_static_thread_mgmt_se3 =0xffffffff ; mqd->compute_misc_reserved =0x00000003 ; - +if (!(adev->flags &AMD_IS_APU )) { + mqd->dynamic_cu_mask_addr_lo = lower_32_bits(ring-Quoted 1 line
mqd_gpu_addrdrivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-0+ + offsetof(struct
vi_mqd_allocation,dyamic_cu_mask));drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-0+ mqd->dynamic_cu_mask_addr_hi = upper_32_bits(ring-Quoted 1 line
mqd_gpu_addrdrivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-0+ + offsetof(struct
vi_mqd_allocation,dyamic_cu_mask));drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-0+ } eop_base_addr = ring->eop_gpu_addr >>8 ; mqd->cp_hqd_eop_base_addr_lo = eop_base_addr; mqd->cp_hqd_eop_base_addr_hi =
upper_32_bits(eop_base_addr);drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-1@@ -4900,7 +4901,7 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring)if (adev->gfx.in_reset) {/* for GPU_RESET case */ /* reset MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) -memcpy (mqd, adev- >gfx.mec.mqd_backup[mqd_idx], sizeof(*mqd)); +memcpy (mqd, adev- >gfx.mec.mqd_backup[mqd_idx], sizeof(struct vi_mqd_allocation));drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +3
-0/* reset ring buffer */ ring->wptr =0 ;@@ -4916,6 +4917,9 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring) vi_srbm_select(adev, 0, 0, 0, 0);mutex_unlock (&adev->srbm_mutex); }else { +memset ((void *)mqd,0 ,sizeof (struct vi_mqd_allocation )); + ((struct vi_mqd_allocation *)mqd)->dyamic_cu_mask =0xFFFFFFFF ; + ((struct vi_mqd_allocation *)mqd)->dyamic_rb_mask =0xFFFFFFFF ;drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-1mutex_lock (&adev->srbm_mutex);vi_srbm_select (adev, ring->me, ring->pipe, ring->queue,0 );gfx_v8_0_mqd_init (ring);@@ -4929,7 +4933,7 @@ static int gfx_v8_0_kiq_init_queue(struct amdgpu_ring *ring) mutex_unlock(&adev->srbm_mutex);if (adev->gfx.mec.mqd_backup[mqd_idx]) -memcpy (adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd)); +memcpy (adev->gfx.mec.mqd_backup[mqd_idx], mqd,sizeof (struct vi_mqd_allocation ));drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +3
-0}return r;@@ -4947,6 +4951,9 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)int mqd_idx = ring - &adev->gfx.compute_ring[0 ];if (!adev->gfx.in_reset && !adev->gfx.in_suspend) { +memset ((void *)mqd,0 ,sizeof (struct vi_mqd_allocation )); + ((struct vi_mqd_allocation *)mqd)->dyamic_cu_mask =0xFFFFFFFF ; + ((struct vi_mqd_allocation *)mqd)->dyamic_rb_mask =0xFFFFFFFF ;drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-2mutex_lock (&adev->srbm_mutex);vi_srbm_select (adev, ring->me, ring->pipe, ring->queue,0 );gfx_v8_0_mqd_init (ring);@@ -4954,11 +4961,11 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring) mutex_unlock(&adev->srbm_mutex);if (adev->gfx.mec.mqd_backup[mqd_idx]) -memcpy (adev->gfx.mec.mqd_backup[mqd_idx], mqd, sizeof(*mqd)); +memcpy (adev->gfx.mec.mqd_backup[mqd_idx], mqd,sizeof (struct vi_mqd_allocation )); }else if (adev->gfx.in_reset) {/* for GPU_RESET case */ /* reset MQD to a clean status */ if (adev->gfx.mec.mqd_backup[mqd_idx]) - memcpy(mqd, adev-Quoted 1 line
gfx.mec.mqd_backup[mqd_idx],
sizeof(*mqd));drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +1
-0+ memcpy(mqd, adev-Quoted 1 line
gfx.mec.mqd_backup[mqd_idx],
sizeof(struct vi_mqd_allocation));drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +0
-0/* reset ring buffer */ ring->wptr =0 ;amdgpu_ring_clear_ring (ring);drivers/gpu/drm/amd/include/vi_structs.h +268
-1@@ -195,6 +195,274 @@ struct vi_mqd {uint32_t compute_wave_restore_addr_lo;uint32_t compute_wave_restore_addr_hi;uint32_t compute_wave_restore_control; +uint32_t reserved9; +uint32_t reserved10; +uint32_t reserved11; +uint32_t reserved12; +uint32_t reserved13; +uint32_t reserved14; +uint32_t reserved15; +uint32_t reserved16; +uint32_t reserved17; +uint32_t reserved18; +uint32_t reserved19; +uint32_t reserved20; +uint32_t reserved21; +uint32_t reserved22; +uint32_t reserved23; +uint32_t reserved24; +uint32_t reserved25; +uint32_t reserved26; +uint32_t reserved27; +uint32_t reserved28; +uint32_t reserved29; +uint32_t reserved30; +uint32_t reserved31; +uint32_t reserved32; +uint32_t reserved33; +uint32_t reserved34; +uint32_t compute_user_data_0; +uint32_t compute_user_data_1; +uint32_t compute_user_data_2; +uint32_t compute_user_data_3; +uint32_t compute_user_data_4; +uint32_t compute_user_data_5; +uint32_t compute_user_data_6; +uint32_t compute_user_data_7; +uint32_t compute_user_data_8; +uint32_t compute_user_data_9; +uint32_t compute_user_data_10; +uint32_t compute_user_data_11; +uint32_t compute_user_data_12; +uint32_t compute_user_data_13; +uint32_t compute_user_data_14; +uint32_t compute_user_data_15; +uint32_t cp_compute_csinvoc_count_lo; +uint32_t cp_compute_csinvoc_count_hi; +uint32_t reserved35; +uint32_t reserved36; +uint32_t reserved37; +uint32_t cp_mqd_query_time_lo; +uint32_t cp_mqd_query_time_hi; +uint32_t cp_mqd_connect_start_time_lo; +uint32_t cp_mqd_connect_start_time_hi; +uint32_t cp_mqd_connect_end_time_lo; +uint32_t cp_mqd_connect_end_time_hi; +uint32_t cp_mqd_connect_end_wf_count; +uint32_t cp_mqd_connect_end_pq_rptr; +uint32_t cp_mqd_connect_endvi_sdma_mqd_pq_wptr; +uint32_t cp_mqd_connect_end_ib_rptr; +uint32_t reserved38; +uint32_t reserved39; +uint32_t cp_mqd_save_start_time_lo; +uint32_t cp_mqd_save_start_time_hi; +uint32_t cp_mqd_save_end_time_lo; +uint32_t cp_mqd_save_end_time_hi; +uint32_t cp_mqd_restore_start_time_lo; +uint32_t cp_mqd_restore_start_time_hi; +uint32_t cp_mqd_restore_end_time_lo; +uint32_t cp_mqd_restore_end_time_hi; +uint32_t disable_queue; +uint32_t reserved41; +uint32_t gds_cs_ctxsw_cnt0; +uint32_t gds_cs_ctxsw_cnt1; +uint32_t gds_cs_ctxsw_cnt2; +uint32_t gds_cs_ctxsw_cnt3; +uint32_t reserved42; +uint32_t reserved43; +uint32_t cp_pq_exe_status_lo; +uint32_t cp_pq_exe_status_hi; +uint32_t cp_packet_id_lo; +uint32_t cp_packet_id_hi; +uint32_t cp_packet_exe_status_lo; +uint32_t cp_packet_exe_status_hi; +uint32_t gds_save_base_addr_lo; +uint32_t gds_save_base_addr_hi; +uint32_t gds_save_mask_lo; +uint32_t gds_save_mask_hi; +uint32_t ctx_save_base_addr_lo; +uint32_t ctx_save_base_addr_hi; +uint32_t dynamic_cu_mask_addr_lo; +uint32_t dynamic_cu_mask_addr_hi; +uint32_t cp_mqd_base_addr_lo; +uint32_t cp_mqd_base_addr_hi; +uint32_t cp_hqd_active; +uint32_t cp_hqd_vmid; +uint32_t cp_hqd_persistent_state; +uint32_t cp_hqd_pipe_priority; +uint32_t cp_hqd_queue_priority; +uint32_t cp_hqd_quantum; +uint32_t cp_hqd_pq_base_lo; +uint32_t cp_hqd_pq_base_hi; +uint32_t cp_hqd_pq_rptr; +uint32_t cp_hqd_pq_rptr_report_addr_lo; +uint32_t cp_hqd_pq_rptr_report_addr_hi; +uint32_t cp_hqd_pq_wptr_poll_addr_lo; +uint32_t cp_hqd_pq_wptr_poll_addr_hi; +uint32_t cp_hqd_pq_doorbell_control; +uint32_t cp_hqd_pq_wptr; +uint32_t cp_hqd_pq_control; +uint32_t cp_hqd_ib_base_addr_lo; +uint32_t cp_hqd_ib_base_addr_hi; +uint32_t cp_hqd_ib_rptr; +uint32_t cp_hqd_ib_control; +uint32_t cp_hqd_iq_timer; +uint32_t cp_hqd_iq_rptr; +uint32_t cp_hqd_dequeue_request; +uint32_t cp_hqd_dma_offload; +uint32_t cp_hqd_sema_cmd; +uint32_t cp_hqd_msg_type; +uint32_t cp_hqd_atomic0_preop_lo; +uint32_t cp_hqd_atomic0_preop_hi; +uint32_t cp_hqd_atomic1_preop_lo; +uint32_t cp_hqd_atomic1_preop_hi; +uint32_t cp_hqd_hq_status0; +uint32_t cp_hqd_hq_control0; +uint32_t cp_mqd_control; +uint32_t cp_hqd_hq_status1; +uint32_t cp_hqd_hq_control1; +uint32_t cp_hqd_eop_base_addr_lo; +uint32_t cp_hqd_eop_base_addr_hi; +uint32_t cp_hqd_eop_control; +uint32_t cp_hqd_eop_rptr; +uint32_t cp_hqd_eop_wptr; +uint32_t cp_hqd_eop_done_events; +uint32_t cp_hqd_ctx_save_base_addr_lo; +uint32_t cp_hqd_ctx_save_base_addr_hi; +uint32_t cp_hqd_ctx_save_control; +uint32_t cp_hqd_cntl_stack_offset; +uint32_t cp_hqd_cntl_stack_size; +uint32_t cp_hqd_wg_state_offset; +uint32_t cp_hqd_ctx_save_size; +uint32_t cp_hqd_gds_resource_state; +uint32_t cp_hqd_error; +uint32_t cp_hqd_eop_wptr_mem; +uint32_t cp_hqd_eop_dones; +uint32_t reserved46; +uint32_t reserved47; +uint32_t reserved48; +uint32_t reserved49; +uint32_t reserved50; +uint32_t reserved51; +uint32_t reserved52; +uint32_t reserved53; +uint32_t reserved54; +uint32_t reserved55; +uint32_t iqtimer_pkt_header; +uint32_t iqtimer_pkt_dw0; +uint32_t iqtimer_pkt_dw1; +uint32_t iqtimer_pkt_dw2; +uint32_t iqtimer_pkt_dw3; +uint32_t iqtimer_pkt_dw4; +uint32_t iqtimer_pkt_dw5; +uint32_t iqtimer_pkt_dw6; +uint32_t iqtimer_pkt_dw7; +uint32_t iqtimer_pkt_dw8; +uint32_t iqtimer_pkt_dw9; +uint32_t iqtimer_pkt_dw10; +uint32_t iqtimer_pkt_dw11; +uint32_t iqtimer_pkt_dw12; +uint32_t iqtimer_pkt_dw13; +uint32_t iqtimer_pkt_dw14; +uint32_t iqtimer_pkt_dw15; +uint32_t iqtimer_pkt_dw16; +uint32_t iqtimer_pkt_dw17; +uint32_t iqtimer_pkt_dw18; +uint32_t iqtimer_pkt_dw19; +uint32_t iqtimer_pkt_dw20; +uint32_t iqtimer_pkt_dw21; +uint32_t iqtimer_pkt_dw22; +uint32_t iqtimer_pkt_dw23; +uint32_t iqtimer_pkt_dw24; +uint32_t iqtimer_pkt_dw25; +uint32_t iqtimer_pkt_dw26; +uint32_t iqtimer_pkt_dw27; +uint32_t iqtimer_pkt_dw28; +uint32_t iqtimer_pkt_dw29; +uint32_t iqtimer_pkt_dw30; +uint32_t iqtimer_pkt_dw31; +uint32_t reserved56; +uint32_t reserved57; +uint32_t reserved58; +uint32_t set_resources_header; +uint32_t set_resources_dw1; +uint32_t set_resources_dw2; +uint32_t set_resources_dw3; +uint32_t set_resources_dw4; +uint32_t set_resources_dw5; +uint32_t set_resources_dw6; +uint32_t set_resources_dw7; +uint32_t reserved59; +uint32_t reserved60; +uint32_t reserved61; +uint32_t reserved62; +uint32_t reserved63; +uint32_t reserved64; +uint32_t reserved65; +uint32_t reserved66; +uint32_t reserved67; +uint32_t reserved68; +uint32_t reserved69; +uint32_t reserved70; +uint32_t reserved71; +uint32_t reserved72; +uint32_t reserved73; +uint32_t reserved74; +uint32_t reserved75; +uint32_t reserved76; +uint32_t reserved77; +uint32_t reserved78; +uint32_t reserved_t[256 ]; +}; + +struct vi_mqd_allocation { +struct vi_mqd mqd; +uint32_t wptr_poll_mem; +uint32_t rptr_report_mem; +uint32_t dyamic_cu_mask; +uint32_t dyamic_rb_mask; +}; + +struct cz_mqd { +uint32_t header; +uint32_t compute_dispatch_initiator; +uint32_t compute_dim_x; +uint32_t compute_dim_y; +uint32_t compute_dim_z; +uint32_t compute_start_x; +uint32_t compute_start_y; +uint32_t compute_start_z; +uint32_t compute_num_thread_x; +uint32_t compute_num_thread_y; +uint32_t compute_num_thread_z; +uint32_t compute_pipelinestat_enable; +uint32_t compute_perfcount_enable; +uint32_t compute_pgm_lo; +uint32_t compute_pgm_hi; +uint32_t compute_tba_lo; +uint32_t compute_tba_hi; +uint32_t compute_tma_lo; +uint32_t compute_tma_hi; +uint32_t compute_pgm_rsrc1; +uint32_t compute_pgm_rsrc2; +uint32_t compute_vmid; +uint32_t compute_resource_limits; +uint32_t compute_static_thread_mgmt_se0; +uint32_t compute_static_thread_mgmt_se1; +uint32_t compute_tmpring_size; +uint32_t compute_static_thread_mgmt_se2; +uint32_t compute_static_thread_mgmt_se3; +uint32_t compute_restart_x; +uint32_t compute_restart_y; +uint32_t compute_restart_z; +uint32_t compute_thread_trace_enable; +uint32_t compute_misc_reserved; +uint32_t compute_dispatch_id; +uint32_t compute_threadgroup_id; +uint32_t compute_relaunch; +uint32_t compute_wave_restore_addr_lo; +uint32_t compute_wave_restore_addr_hi; +uint32_t compute_wave_restore_control;uint32_t reserved_39;uint32_t reserved_40;uint32_t reserved_41; --1.9.1
Quoted 4 lines
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx