summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c117
1 files changed, 77 insertions, 40 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index d3e8be82a172..894fc04201c3 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -293,14 +293,20 @@ static void gfx_v11_0_update_perf_clk(struct amdgpu_device *adev,
static void gfx11_kiq_set_resources(struct amdgpu_ring *kiq_ring, uint64_t queue_mask)
{
+ struct amdgpu_device *adev = kiq_ring->adev;
+ u64 shader_mc_addr;
+
+ /* Cleaner shader MC address */
+ shader_mc_addr = adev->gfx.cleaner_shader_gpu_addr >> 8;
+
amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
amdgpu_ring_write(kiq_ring, PACKET3_SET_RESOURCES_VMID_MASK(0) |
PACKET3_SET_RESOURCES_UNMAP_LATENTY(0xa) | /* unmap_latency: 0xa (~ 1s) */
PACKET3_SET_RESOURCES_QUEUE_TYPE(0)); /* vmid_mask:0 queue_type:0 (KIQ) */
amdgpu_ring_write(kiq_ring, lower_32_bits(queue_mask)); /* queue mask lo */
amdgpu_ring_write(kiq_ring, upper_32_bits(queue_mask)); /* queue mask hi */
- amdgpu_ring_write(kiq_ring, 0); /* gws mask lo */
- amdgpu_ring_write(kiq_ring, 0); /* gws mask hi */
+ amdgpu_ring_write(kiq_ring, lower_32_bits(shader_mc_addr)); /* cleaner shader addr lo */
+ amdgpu_ring_write(kiq_ring, upper_32_bits(shader_mc_addr)); /* cleaner shader addr hi */
amdgpu_ring_write(kiq_ring, 0); /* oac mask */
amdgpu_ring_write(kiq_ring, 0); /* gds heap base:0, gds heap size:0 */
}
@@ -483,8 +489,6 @@ static void gfx_v11_0_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel,
static void gfx_v11_ring_insert_nop(struct amdgpu_ring *ring, uint32_t num_nop)
{
- int i;
-
/* Header itself is a NOP packet */
if (num_nop == 1) {
amdgpu_ring_write(ring, ring->funcs->nop);
@@ -495,8 +499,7 @@ static void gfx_v11_ring_insert_nop(struct amdgpu_ring *ring, uint32_t num_nop)
amdgpu_ring_write(ring, PACKET3(PACKET3_NOP, min(num_nop - 2, 0x3ffe)));
/* Header is at index 0, followed by num_nops - 1 NOP packet's */
- for (i = 1; i < num_nop; i++)
- amdgpu_ring_write(ring, ring->funcs->nop);
+ amdgpu_ring_insert_nop(ring, num_nop - 1);
}
static int gfx_v11_0_ring_test_ring(struct amdgpu_ring *ring)
@@ -1536,11 +1539,11 @@ static void gfx_v11_0_alloc_ip_dump(struct amdgpu_device *adev)
}
}
-static int gfx_v11_0_sw_init(void *handle)
+static int gfx_v11_0_sw_init(struct amdgpu_ip_block *ip_block)
{
int i, j, k, r, ring_id = 0;
int xcc_id = 0;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
case IP_VERSION(11, 0, 0):
@@ -1575,6 +1578,11 @@ static int gfx_v11_0_sw_init(void *handle)
break;
}
+ switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+ default:
+ adev->gfx.enable_cleaner_shader = false;
+ }
+
/* Enable CG flag in one VF mode for enabling RLC safe mode enter/exit */
if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 0, 3) &&
amdgpu_sriov_is_pp_one_vf(adev))
@@ -1700,6 +1708,10 @@ static int gfx_v11_0_sw_init(void *handle)
gfx_v11_0_alloc_ip_dump(adev);
+ r = amdgpu_gfx_sysfs_isolation_shader_init(adev);
+ if (r)
+ return r;
+
return 0;
}
@@ -1732,10 +1744,10 @@ static void gfx_v11_0_rlc_autoload_buffer_fini(struct amdgpu_device *adev)
(void **)&adev->gfx.rlc.rlc_autoload_ptr);
}
-static int gfx_v11_0_sw_fini(void *handle)
+static int gfx_v11_0_sw_fini(struct amdgpu_ip_block *ip_block)
{
int i;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
for (i = 0; i < adev->gfx.num_gfx_rings; i++)
amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
@@ -1749,6 +1761,8 @@ static int gfx_v11_0_sw_fini(void *handle)
amdgpu_gfx_kiq_fini(adev, 0);
}
+ amdgpu_gfx_cleaner_shader_sw_fini(adev);
+
gfx_v11_0_pfp_fini(adev);
gfx_v11_0_me_fini(adev);
gfx_v11_0_rlc_fini(adev);
@@ -1759,6 +1773,8 @@ static int gfx_v11_0_sw_fini(void *handle)
gfx_v11_0_free_microcode(adev);
+ amdgpu_gfx_sysfs_isolation_shader_fini(adev);
+
kfree(adev->gfx.ip_dump_core);
kfree(adev->gfx.ip_dump_compute_queues);
kfree(adev->gfx.ip_dump_gfx_queues);
@@ -1893,8 +1909,10 @@ static void gfx_v11_0_init_compute_vmid(struct amdgpu_device *adev)
soc21_grbm_select(adev, 0, 0, 0, 0);
mutex_unlock(&adev->srbm_mutex);
- /* Initialize all compute VMIDs to have no GDS, GWS, or OA
- acccess. These should be enabled by FW for target VMIDs. */
+ /*
+ * Initialize all compute VMIDs to have no GDS, GWS, or OA
+ * access. These should be enabled by FW for target VMIDs.
+ */
for (i = adev->vm_manager.first_kfd_vmid; i < AMDGPU_NUM_VMID; i++) {
WREG32_SOC15_OFFSET(GC, 0, regGDS_VMID0_BASE, 2 * i, 0);
WREG32_SOC15_OFFSET(GC, 0, regGDS_VMID0_SIZE, 2 * i, 0);
@@ -3555,7 +3573,7 @@ static int gfx_v11_0_cp_gfx_resume(struct amdgpu_device *adev)
WREG32_SOC15(GC, 0, regCP_RB0_WPTR, lower_32_bits(ring->wptr));
WREG32_SOC15(GC, 0, regCP_RB0_WPTR_HI, upper_32_bits(ring->wptr));
- /* set the wb address wether it's enabled or not */
+ /* set the wb address whether it's enabled or not */
rptr_addr = ring->rptr_gpu_addr;
WREG32_SOC15(GC, 0, regCP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
WREG32_SOC15(GC, 0, regCP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) &
@@ -3593,7 +3611,7 @@ static int gfx_v11_0_cp_gfx_resume(struct amdgpu_device *adev)
ring->wptr = 0;
WREG32_SOC15(GC, 0, regCP_RB1_WPTR, lower_32_bits(ring->wptr));
WREG32_SOC15(GC, 0, regCP_RB1_WPTR_HI, upper_32_bits(ring->wptr));
- /* Set the wb address wether it's enabled or not */
+ /* Set the wb address whether it's enabled or not */
rptr_addr = ring->rptr_gpu_addr;
WREG32_SOC15(GC, 0, regCP_RB1_RPTR_ADDR, lower_32_bits(rptr_addr));
WREG32_SOC15(GC, 0, regCP_RB1_RPTR_ADDR_HI, upper_32_bits(rptr_addr) &
@@ -4568,10 +4586,13 @@ static void gfx_v11_0_disable_gpa_mode(struct amdgpu_device *adev)
WREG32_SOC15(GC, 0, regCPG_PSP_DEBUG, data);
}
-static int gfx_v11_0_hw_init(void *handle)
+static int gfx_v11_0_hw_init(struct amdgpu_ip_block *ip_block)
{
int r;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
+
+ amdgpu_gfx_cleaner_shader_init(adev, adev->gfx.cleaner_shader_size,
+ adev->gfx.cleaner_shader_ptr);
if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
if (adev->gfx.imu.funcs) {
@@ -4665,9 +4686,9 @@ static int gfx_v11_0_hw_init(void *handle)
return r;
}
-static int gfx_v11_0_hw_fini(void *handle)
+static int gfx_v11_0_hw_fini(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0);
@@ -4703,14 +4724,14 @@ static int gfx_v11_0_hw_fini(void *handle)
return 0;
}
-static int gfx_v11_0_suspend(void *handle)
+static int gfx_v11_0_suspend(struct amdgpu_ip_block *ip_block)
{
- return gfx_v11_0_hw_fini(handle);
+ return gfx_v11_0_hw_fini(ip_block);
}
-static int gfx_v11_0_resume(void *handle)
+static int gfx_v11_0_resume(struct amdgpu_ip_block *ip_block)
{
- return gfx_v11_0_hw_init(handle);
+ return gfx_v11_0_hw_init(ip_block);
}
static bool gfx_v11_0_is_idle(void *handle)
@@ -4724,11 +4745,11 @@ static bool gfx_v11_0_is_idle(void *handle)
return true;
}
-static int gfx_v11_0_wait_for_idle(void *handle)
+static int gfx_v11_0_wait_for_idle(struct amdgpu_ip_block *ip_block)
{
unsigned i;
u32 tmp;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
for (i = 0; i < adev->usec_timeout; i++) {
/* read MC_STATUS */
@@ -4774,12 +4795,12 @@ int gfx_v11_0_request_gfx_index_mutex(struct amdgpu_device *adev,
return 0;
}
-static int gfx_v11_0_soft_reset(void *handle)
+static int gfx_v11_0_soft_reset(struct amdgpu_ip_block *ip_block)
{
u32 grbm_soft_reset = 0;
u32 tmp;
int r, i, j, k;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
amdgpu_gfx_rlc_enter_safe_mode(adev, 0);
@@ -4905,10 +4926,10 @@ static int gfx_v11_0_soft_reset(void *handle)
return gfx_v11_0_cp_resume(adev);
}
-static bool gfx_v11_0_check_soft_reset(void *handle)
+static bool gfx_v11_0_check_soft_reset(struct amdgpu_ip_block *ip_block)
{
int i, r;
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
struct amdgpu_ring *ring;
long tmo = msecs_to_jiffies(1000);
@@ -4929,12 +4950,13 @@ static bool gfx_v11_0_check_soft_reset(void *handle)
return false;
}
-static int gfx_v11_0_post_soft_reset(void *handle)
+static int gfx_v11_0_post_soft_reset(struct amdgpu_ip_block *ip_block)
{
+ struct amdgpu_device *adev = ip_block->adev;
/**
* GFX soft reset will impact MES, need resume MES when do GFX soft reset
*/
- return amdgpu_mes_resume((struct amdgpu_device *)handle);
+ return amdgpu_mes_resume(adev);
}
static uint64_t gfx_v11_0_get_gpu_clock_counter(struct amdgpu_device *adev)
@@ -4995,9 +5017,9 @@ static void gfx_v11_0_ring_emit_gds_switch(struct amdgpu_ring *ring,
(1 << (oa_size + oa_base)) - (1 << oa_base));
}
-static int gfx_v11_0_early_init(void *handle)
+static int gfx_v11_0_early_init(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
adev->gfx.funcs = &gfx_v11_0_gfx_funcs;
@@ -5018,9 +5040,9 @@ static int gfx_v11_0_early_init(void *handle)
return gfx_v11_0_init_microcode(adev);
}
-static int gfx_v11_0_late_init(void *handle)
+static int gfx_v11_0_late_init(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
int r;
r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0);
@@ -6639,9 +6661,9 @@ static int gfx_v11_0_reset_kcq(struct amdgpu_ring *ring, unsigned int vmid)
return amdgpu_ring_test_ring(ring);
}
-static void gfx_v11_ip_print(void *handle, struct drm_printer *p)
+static void gfx_v11_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
uint32_t i, j, k, reg, index = 0;
uint32_t reg_count = ARRAY_SIZE(gc_reg_list_11_0);
@@ -6703,9 +6725,9 @@ static void gfx_v11_ip_print(void *handle, struct drm_printer *p)
}
}
-static void gfx_v11_ip_dump(void *handle)
+static void gfx_v11_ip_dump(struct amdgpu_ip_block *ip_block)
{
- struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_device *adev = ip_block->adev;
uint32_t i, j, k, reg, index = 0;
uint32_t reg_count = ARRAY_SIZE(gc_reg_list_11_0);
@@ -6769,6 +6791,13 @@ static void gfx_v11_ip_dump(void *handle)
amdgpu_gfx_off_ctrl(adev, true);
}
+static void gfx_v11_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
+{
+ /* Emit the cleaner shader */
+ amdgpu_ring_write(ring, PACKET3(PACKET3_RUN_CLEANER_SHADER, 0));
+ amdgpu_ring_write(ring, 0); /* RESERVED field, programmed to zero */
+}
+
static const struct amd_ip_funcs gfx_v11_0_ip_funcs = {
.name = "gfx_v11_0",
.early_init = gfx_v11_0_early_init,
@@ -6818,7 +6847,8 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_gfx = {
5 + /* HDP_INVL */
22 + /* SET_Q_PREEMPTION_MODE */
8 + 8 + /* FENCE x2 */
- 8, /* gfx_v11_0_emit_mem_sync */
+ 8 + /* gfx_v11_0_emit_mem_sync */
+ 2, /* gfx_v11_0_ring_emit_cleaner_shader */
.emit_ib_size = 4, /* gfx_v11_0_ring_emit_ib_gfx */
.emit_ib = gfx_v11_0_ring_emit_ib_gfx,
.emit_fence = gfx_v11_0_ring_emit_fence,
@@ -6841,6 +6871,9 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_gfx = {
.soft_recovery = gfx_v11_0_ring_soft_recovery,
.emit_mem_sync = gfx_v11_0_emit_mem_sync,
.reset = gfx_v11_0_reset_kgq,
+ .emit_cleaner_shader = gfx_v11_0_ring_emit_cleaner_shader,
+ .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use,
+ .end_use = amdgpu_gfx_enforce_isolation_ring_end_use,
};
static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = {
@@ -6861,7 +6894,8 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = {
SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
2 + /* gfx_v11_0_ring_emit_vm_flush */
8 + 8 + 8 + /* gfx_v11_0_ring_emit_fence x3 for user fence, vm fence */
- 8, /* gfx_v11_0_emit_mem_sync */
+ 8 + /* gfx_v11_0_emit_mem_sync */
+ 2, /* gfx_v11_0_ring_emit_cleaner_shader */
.emit_ib_size = 7, /* gfx_v11_0_ring_emit_ib_compute */
.emit_ib = gfx_v11_0_ring_emit_ib_compute,
.emit_fence = gfx_v11_0_ring_emit_fence,
@@ -6879,6 +6913,9 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = {
.soft_recovery = gfx_v11_0_ring_soft_recovery,
.emit_mem_sync = gfx_v11_0_emit_mem_sync,
.reset = gfx_v11_0_reset_kcq,
+ .emit_cleaner_shader = gfx_v11_0_ring_emit_cleaner_shader,
+ .begin_use = amdgpu_gfx_enforce_isolation_ring_begin_use,
+ .end_use = amdgpu_gfx_enforce_isolation_ring_end_use,
};
static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_kiq = {
generated by cgit v1.2.3 (git 2.0.4) at 2025-11-08 02:06:53 +0000