[PATCH v1 02/10] drm/amdgpu: Add cp queues support fro gfx10 in ipdump
Sunil Khatri
sunil.khatri at amd.com
Wed May 22 07:29:42 UTC 2024
Add support to dump registers of all instances of
cp queue registers of gfx10 to devcoredump.
Signed-off-by: Sunil Khatri <sunil.khatri at amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h | 1 +
drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c | 116 +++++++++++++++++++++++-
2 files changed, 113 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
index a28462643b00..77ab5d0fd592 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
@@ -436,6 +436,7 @@ struct amdgpu_gfx {
/* IP reg dump */
uint32_t *ip_dump_core;
+ uint32_t *ip_dump_cp_queues;
};
struct amdgpu_gfx_ras_reg_entry {
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 73149150fe2e..ab378d9a74c8 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -381,6 +381,49 @@ static const struct amdgpu_hwip_reg_entry gc_reg_list_10_1[] = {
SOC15_REG_ENTRY_STR(GC, 0, mmGRBM_STATUS_SE3)
};
+static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_10[] = {
+ /* compute registers */
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_VMID),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PERSISTENT_STATE),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PIPE_PRIORITY),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_QUEUE_PRIORITY),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_QUANTUM),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_BASE),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_BASE_HI),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_RPTR),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_CONTROL),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_BASE_ADDR),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_BASE_ADDR_HI),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_RPTR),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_IB_CONTROL),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_DEQUEUE_REQUEST),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_BASE_ADDR),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_BASE_ADDR_HI),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_CONTROL),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_RPTR),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_WPTR),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_EVENTS),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_BASE_ADDR_LO),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_BASE_ADDR_HI),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_CONTROL),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CNTL_STACK_OFFSET),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CNTL_STACK_SIZE),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_WG_STATE_OFFSET),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_CTX_SAVE_SIZE),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_GDS_RESOURCE_STATE),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_ERROR),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_EOP_WPTR_MEM),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_LO),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_PQ_WPTR_HI),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_CNTL_STACK_OFFSET),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_CNTL_STACK_DW_CNT),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_SUSPEND_WG_STATE_OFFSET),
+ SOC15_REG_ENTRY_STR(GC, 0, mmCP_HQD_DEQUEUE_STATUS)
+};
+
static const struct soc15_reg_golden golden_settings_gc_10_1[] = {
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCB_HW_CONTROL_4, 0xffffffff, 0x00400014),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCGTT_CPF_CLK_CTRL, 0xfcff8fff, 0xf8000100),
@@ -4595,10 +4638,11 @@ static int gfx_v10_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
hw_prio, NULL);
}
-static void gfx_v10_0_alloc_dump_mem(struct amdgpu_device *adev)
+static void gfx_v10_0_alloc_ip_dump(struct amdgpu_device *adev)
{
uint32_t reg_count = ARRAY_SIZE(gc_reg_list_10_1);
uint32_t *ptr;
+ uint32_t inst;
ptr = kcalloc(reg_count, sizeof(uint32_t), GFP_KERNEL);
if (ptr == NULL) {
@@ -4607,6 +4651,19 @@ static void gfx_v10_0_alloc_dump_mem(struct amdgpu_device *adev)
} else {
adev->gfx.ip_dump_core = ptr;
}
+
+ /* Allocate memory for compute queue registers for all the instances */
+ reg_count = ARRAY_SIZE(gc_cp_reg_list_10);
+ inst = adev->gfx.mec.num_mec * adev->gfx.mec.num_pipe_per_mec *
+ adev->gfx.mec.num_queue_per_pipe;
+
+ ptr = kcalloc(reg_count * inst, sizeof(uint32_t), GFP_KERNEL);
+ if (ptr == NULL) {
+ DRM_ERROR("Failed to allocate memory for GFX CP IP Dump\n");
+ adev->gfx.ip_dump_cp_queues = NULL;
+ } else {
+ adev->gfx.ip_dump_cp_queues = ptr;
+ }
}
static int gfx_v10_0_sw_init(void *handle)
@@ -4761,7 +4818,7 @@ static int gfx_v10_0_sw_init(void *handle)
gfx_v10_0_gpu_early_init(adev);
- gfx_v10_0_alloc_dump_mem(adev);
+ gfx_v10_0_alloc_ip_dump(adev);
return 0;
}
@@ -4816,6 +4873,7 @@ static int gfx_v10_0_sw_fini(void *handle)
gfx_v10_0_free_microcode(adev);
kfree(adev->gfx.ip_dump_core);
+ kfree(adev->gfx.ip_dump_cp_queues);
return 0;
}
@@ -9289,7 +9347,7 @@ static void gfx_v10_0_emit_mem_sync(struct amdgpu_ring *ring)
static void gfx_v10_ip_print(void *handle, struct drm_printer *p)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
- uint32_t i;
+ uint32_t i, j, k, reg, index = 0;
uint32_t reg_count = ARRAY_SIZE(gc_reg_list_10_1);
if (!adev->gfx.ip_dump_core)
@@ -9299,12 +9357,36 @@ static void gfx_v10_ip_print(void *handle, struct drm_printer *p)
drm_printf(p, "%-50s \t 0x%08x\n",
gc_reg_list_10_1[i].reg_name,
adev->gfx.ip_dump_core[i]);
+
+ /* print compute queue registers for all instances */
+ if (!adev->gfx.ip_dump_cp_queues)
+ return;
+
+ reg_count = ARRAY_SIZE(gc_cp_reg_list_10);
+ drm_printf(p, "\nnum_mec: %d num_pipe: %d num_queue: %d\n",
+ adev->gfx.mec.num_mec,
+ adev->gfx.mec.num_pipe_per_mec,
+ adev->gfx.mec.num_queue_per_pipe);
+
+ for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+ for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+ for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+ drm_printf(p, "\nmec %d, pipe %d, queue %d\n", i, j, k);
+ for (reg = 0; reg < reg_count; reg++) {
+ drm_printf(p, "%-50s \t 0x%08x\n",
+ gc_cp_reg_list_10[reg].reg_name,
+ adev->gfx.ip_dump_cp_queues[index + reg]);
+ }
+ index += reg_count;
+ }
+ }
+ }
}
static void gfx_v10_ip_dump(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
- uint32_t i;
+ uint32_t i, j, k, reg, index = 0;
uint32_t reg_count = ARRAY_SIZE(gc_reg_list_10_1);
if (!adev->gfx.ip_dump_core)
@@ -9314,6 +9396,32 @@ static void gfx_v10_ip_dump(void *handle)
for (i = 0; i < reg_count; i++)
adev->gfx.ip_dump_core[i] = RREG32(SOC15_REG_ENTRY_OFFSET(gc_reg_list_10_1[i]));
amdgpu_gfx_off_ctrl(adev, true);
+
+ /* dump compute queue registers for all instances */
+ if (!adev->gfx.ip_dump_cp_queues)
+ return;
+
+ reg_count = ARRAY_SIZE(gc_cp_reg_list_10);
+ amdgpu_gfx_off_ctrl(adev, false);
+ mutex_lock(&adev->srbm_mutex);
+ for (i = 0; i < adev->gfx.mec.num_mec; i++) {
+ for (j = 0; j < adev->gfx.mec.num_pipe_per_mec; j++) {
+ for (k = 0; k < adev->gfx.mec.num_queue_per_pipe; k++) {
+ /* ME0 is for GFX so start from 1 for CP */
+ nv_grbm_select(adev, 1 + i, j, k, 0);
+
+ for (reg = 0; reg < reg_count; reg++) {
+ adev->gfx.ip_dump_cp_queues[index + reg] =
+ RREG32(SOC15_REG_ENTRY_OFFSET(
+ gc_cp_reg_list_10[reg]));
+ }
+ index += reg_count;
+ }
+ }
+ }
+ nv_grbm_select(adev, 0, 0, 0, 0);
+ mutex_unlock(&adev->srbm_mutex);
+ amdgpu_gfx_off_ctrl(adev, true);
}
static const struct amd_ip_funcs gfx_v10_0_ip_funcs = {
--
2.34.1
More information about the amd-gfx
mailing list