1 files changed, 53 insertions, 29 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
index 47b47d21f464..9fec28d8a5fc 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
@@ -1319,12 +1319,12 @@ static void gfx_v12_0_alloc_ip_dump(struct amdgpu_device *adev)
 	}
 }
 
-static int gfx_v12_0_sw_init(void *handle)
+static int gfx_v12_0_sw_init(struct amdgpu_ip_block *ip_block)
 {
 	int i, j, k, r, ring_id = 0;
 	unsigned num_compute_rings;
 	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(12, 0, 0):
@@ -1346,6 +1346,12 @@ static int gfx_v12_0_sw_init(void *handle)
 		break;
 	}
 
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	default:
+		adev->gfx.enable_cleaner_shader = false;
+		break;
+	}
+
 	/* recalculate compute rings to use based on hardware configuration */
 	num_compute_rings = (adev->gfx.mec.num_pipe_per_mec *
 			     adev->gfx.mec.num_queue_per_pipe) / 2;
@@ -1460,6 +1466,10 @@ static int gfx_v12_0_sw_init(void *handle)
 
 	gfx_v12_0_alloc_ip_dump(adev);
 
+	r = amdgpu_gfx_sysfs_isolation_shader_init(adev);
+	if (r)
+		return r;
+
 	return 0;
 }
 
@@ -1492,10 +1502,10 @@ static void gfx_v12_0_rlc_autoload_buffer_fini(struct amdgpu_device *adev)
 			(void **)&adev->gfx.rlc.rlc_autoload_ptr);
 }
 
-static int gfx_v12_0_sw_fini(void *handle)
+static int gfx_v12_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]);
@@ -1519,6 +1529,8 @@ static int gfx_v12_0_sw_fini(void *handle)
 
 	gfx_v12_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);
@@ -2601,7 +2613,7 @@ static int gfx_v12_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) &
@@ -3513,10 +3525,10 @@ static void gfx_v12_0_init_golden_registers(struct amdgpu_device *adev)
 	}
 }
 
-static int gfx_v12_0_hw_init(void *handle)
+static int gfx_v12_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;
 
 	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO) {
 		if (adev->gfx.imu.funcs && (amdgpu_dpm > 0)) {
@@ -3603,9 +3615,9 @@ static int gfx_v12_0_hw_init(void *handle)
 	return r;
 }
 
-static int gfx_v12_0_hw_fini(void *handle)
+static int gfx_v12_0_hw_fini(struct amdgpu_ip_block *ip_block)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	struct amdgpu_device *adev = ip_block->adev;
 	uint32_t tmp;
 
 	amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0);
@@ -3643,14 +3655,14 @@ static int gfx_v12_0_hw_fini(void *handle)
 	return 0;
 }
 
-static int gfx_v12_0_suspend(void *handle)
+static int gfx_v12_0_suspend(struct amdgpu_ip_block *ip_block)
 {
-	return gfx_v12_0_hw_fini(handle);
+	return gfx_v12_0_hw_fini(ip_block);
 }
 
-static int gfx_v12_0_resume(void *handle)
+static int gfx_v12_0_resume(struct amdgpu_ip_block *ip_block)
 {
-	return gfx_v12_0_hw_init(handle);
+	return gfx_v12_0_hw_init(ip_block);
 }
 
 static bool gfx_v12_0_is_idle(void *handle)
@@ -3664,11 +3676,11 @@ static bool gfx_v12_0_is_idle(void *handle)
 		return true;
 }
 
-static int gfx_v12_0_wait_for_idle(void *handle)
+static int gfx_v12_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 */
@@ -3695,9 +3707,9 @@ static uint64_t gfx_v12_0_get_gpu_clock_counter(struct amdgpu_device *adev)
 	return clock;
 }
 
-static int gfx_v12_0_early_init(void *handle)
+static int gfx_v12_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_v12_0_gfx_funcs;
 
@@ -3717,9 +3729,9 @@ static int gfx_v12_0_early_init(void *handle)
 	return gfx_v12_0_init_microcode(adev);
 }
 
-static int gfx_v12_0_late_init(void *handle)
+static int gfx_v12_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);
@@ -5022,8 +5034,6 @@ static void gfx_v12_0_emit_mem_sync(struct amdgpu_ring *ring)
 
 static void gfx_v12_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);
@@ -5034,13 +5044,19 @@ static void gfx_v12_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 void gfx_v12_ip_print(void *handle, struct drm_printer *p)
+static void gfx_v12_0_ring_emit_cleaner_shader(struct amdgpu_ring *ring)
 {
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	/* 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 void gfx_v12_ip_print(struct amdgpu_ip_block *ip_block, struct drm_printer *p)
+{
+	struct amdgpu_device *adev = ip_block->adev;
 	uint32_t i, j, k, reg, index = 0;
 	uint32_t reg_count = ARRAY_SIZE(gc_reg_list_12_0);
 
@@ -5102,9 +5118,9 @@ static void gfx_v12_ip_print(void *handle, struct drm_printer *p)
 	}
 }
 
-static void gfx_v12_ip_dump(void *handle)
+static void gfx_v12_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_12_0);
 
@@ -5297,7 +5313,8 @@ static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_gfx = {
 		3 + /* CNTX_CTRL */
 		5 + /* HDP_INVL */
 		8 + 8 + /* FENCE x2 */
-		8, /* gfx_v12_0_emit_mem_sync */
+		8 + /* gfx_v12_0_emit_mem_sync */
+		2, /* gfx_v12_0_ring_emit_cleaner_shader */
 	.emit_ib_size =	4, /* gfx_v12_0_ring_emit_ib_gfx */
 	.emit_ib = gfx_v12_0_ring_emit_ib_gfx,
 	.emit_fence = gfx_v12_0_ring_emit_fence,
@@ -5318,6 +5335,9 @@ static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_gfx = {
 	.soft_recovery = gfx_v12_0_ring_soft_recovery,
 	.emit_mem_sync = gfx_v12_0_emit_mem_sync,
 	.reset = gfx_v12_0_reset_kgq,
+	.emit_cleaner_shader = gfx_v12_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_v12_0_ring_funcs_compute = {
@@ -5336,7 +5356,8 @@ static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_compute = {
 		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
 		2 + /* gfx_v12_0_ring_emit_vm_flush */
 		8 + 8 + 8 + /* gfx_v12_0_ring_emit_fence x3 for user fence, vm fence */
-		8, /* gfx_v12_0_emit_mem_sync */
+		8 + /* gfx_v12_0_emit_mem_sync */
+		2, /* gfx_v12_0_ring_emit_cleaner_shader */
 	.emit_ib_size =	7, /* gfx_v12_0_ring_emit_ib_compute */
 	.emit_ib = gfx_v12_0_ring_emit_ib_compute,
 	.emit_fence = gfx_v12_0_ring_emit_fence,
@@ -5353,6 +5374,9 @@ static const struct amdgpu_ring_funcs gfx_v12_0_ring_funcs_compute = {
 	.soft_recovery = gfx_v12_0_ring_soft_recovery,
 	.emit_mem_sync = gfx_v12_0_emit_mem_sync,
 	.reset = gfx_v12_0_reset_kcq,
+	.emit_cleaner_shader = gfx_v12_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_v12_0_ring_funcs_kiq = {