drm/amdgpu: fix calltrace during kmd unload(v3)

issue:
kernel would report a warning from a double unpin
during the driver unloading on the CSB bo

why:
we unpin it during hw_fini, and there will be another
unpin in sw_fini on CSB bo.

fix:
actually we don't need to pin/unpin it during
hw_init/fini since it is created with kernel pinned,
we only need to fullfill the CSB again during hw_init
to prevent CSB/VRAM lost after S3

v2:
get_csb in init_rlc so hw_init() will make CSIB content
back even after reset or s3

v3:
use bo_create_kernel instead of bo_create_reserved for CSB
otherwise the bo_free_kernel() on CSB is not aligned and
would lead to its internal reserve pending there forever

take care of gfx7/8 as well
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c

@@ -124,13 +124,12 @@ int amdgpu_gfx_rlc_init_sr(struct amdgpu_device *adev, u32 dws)
  */
 int amdgpu_gfx_rlc_init_csb(struct amdgpu_device *adev)
 {
-	volatile u32 *dst_ptr;
 	u32 dws;
 	int r;
 
 	/* allocate clear state block */
 	adev->gfx.rlc.clear_state_size = dws = adev->gfx.rlc.funcs->get_csb_size(adev);
-	r = amdgpu_bo_create_reserved(adev, dws * 4, PAGE_SIZE,
+	r = amdgpu_bo_create_kernel(adev, dws * 4, PAGE_SIZE,
 				      AMDGPU_GEM_DOMAIN_VRAM,
 				      &adev->gfx.rlc.clear_state_obj,
 				      &adev->gfx.rlc.clear_state_gpu_addr,
@@ -141,13 +140,6 @@ int amdgpu_gfx_rlc_init_csb(struct amdgpu_device *adev)
 		return r;
 	}
 
-	/* set up the cs buffer */
-	dst_ptr = adev->gfx.rlc.cs_ptr;
-	adev->gfx.rlc.funcs->get_csb_buffer(adev, dst_ptr);
-	amdgpu_bo_kunmap(adev->gfx.rlc.clear_state_obj);
-	amdgpu_bo_unpin(adev->gfx.rlc.clear_state_obj);
-	amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-
 	return 0;
 }
 

drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c

@@ -990,39 +990,6 @@ static int gfx_v10_0_rlc_init(struct amdgpu_device *adev)
 	return 0;
 }
 
-static int gfx_v10_0_csb_vram_pin(struct amdgpu_device *adev)
-{
-	int r;
-
-	r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, false);
-	if (unlikely(r != 0))
-		return r;
-
-	r = amdgpu_bo_pin(adev->gfx.rlc.clear_state_obj,
-			AMDGPU_GEM_DOMAIN_VRAM);
-	if (!r)
-		adev->gfx.rlc.clear_state_gpu_addr =
-			amdgpu_bo_gpu_offset(adev->gfx.rlc.clear_state_obj);
-
-	amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-
-	return r;
-}
-
-static void gfx_v10_0_csb_vram_unpin(struct amdgpu_device *adev)
-{
-	int r;
-
-	if (!adev->gfx.rlc.clear_state_obj)
-		return;
-
-	r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, true);
-	if (likely(r == 0)) {
-		amdgpu_bo_unpin(adev->gfx.rlc.clear_state_obj);
-		amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-	}
-}
-
 static void gfx_v10_0_mec_fini(struct amdgpu_device *adev)
 {
 	amdgpu_bo_free_kernel(&adev->gfx.mec.hpd_eop_obj, NULL, NULL);
@@ -1784,25 +1751,7 @@ static void gfx_v10_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
 
 static int gfx_v10_0_init_csb(struct amdgpu_device *adev)
 {
-	int r;
-
-	if (adev->in_gpu_reset) {
-		r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, false);
-		if (r)
-			return r;
-
-		r = amdgpu_bo_kmap(adev->gfx.rlc.clear_state_obj,
-				   (void **)&adev->gfx.rlc.cs_ptr);
-		if (!r) {
-			adev->gfx.rlc.funcs->get_csb_buffer(adev,
-					adev->gfx.rlc.cs_ptr);
-			amdgpu_bo_kunmap(adev->gfx.rlc.clear_state_obj);
-		}
-
-		amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-		if (r)
-			return r;
-	}
+	adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr);
 
 	/* csib */
 	WREG32_SOC15(GC, 0, mmRLC_CSIB_ADDR_HI,
@@ -3773,10 +3722,6 @@ static int gfx_v10_0_hw_init(void *handle)
 	int r;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	r = gfx_v10_0_csb_vram_pin(adev);
-	if (r)
-		return r;
-
 	if (!amdgpu_emu_mode)
 		gfx_v10_0_init_golden_registers(adev);
 
@@ -3864,7 +3809,6 @@ static int gfx_v10_0_hw_fini(void *handle)
 	}
 	gfx_v10_0_cp_enable(adev, false);
 	gfx_v10_0_enable_gui_idle_interrupt(adev, false);
-	gfx_v10_0_csb_vram_unpin(adev);
 
 	return 0;
 }

drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c

@@ -4553,6 +4553,8 @@ static int gfx_v7_0_hw_init(void *handle)
 
 	gfx_v7_0_constants_init(adev);
 
+	/* init CSB */
+	adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr);
 	/* init rlc */
 	r = adev->gfx.rlc.funcs->resume(adev);
 	if (r)

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c

@@ -1321,39 +1321,6 @@ static int gfx_v8_0_rlc_init(struct amdgpu_device *adev)
 	return 0;
 }
 
-static int gfx_v8_0_csb_vram_pin(struct amdgpu_device *adev)
-{
-	int r;
-
-	r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, false);
-	if (unlikely(r != 0))
-		return r;
-
-	r = amdgpu_bo_pin(adev->gfx.rlc.clear_state_obj,
-			AMDGPU_GEM_DOMAIN_VRAM);
-	if (!r)
-		adev->gfx.rlc.clear_state_gpu_addr =
-			amdgpu_bo_gpu_offset(adev->gfx.rlc.clear_state_obj);
-
-	amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-
-	return r;
-}
-
-static void gfx_v8_0_csb_vram_unpin(struct amdgpu_device *adev)
-{
-	int r;
-
-	if (!adev->gfx.rlc.clear_state_obj)
-		return;
-
-	r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, true);
-	if (likely(r == 0)) {
-		amdgpu_bo_unpin(adev->gfx.rlc.clear_state_obj);
-		amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-	}
-}
-
 static void gfx_v8_0_mec_fini(struct amdgpu_device *adev)
 {
 	amdgpu_bo_free_kernel(&adev->gfx.mec.hpd_eop_obj, NULL, NULL);
@@ -3916,6 +3883,7 @@ static void gfx_v8_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
 
 static void gfx_v8_0_init_csb(struct amdgpu_device *adev)
 {
+	adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr);
 	/* csib */
 	WREG32(mmRLC_CSIB_ADDR_HI,
 			adev->gfx.rlc.clear_state_gpu_addr >> 32);
@@ -4836,10 +4804,6 @@ static int gfx_v8_0_hw_init(void *handle)
 	gfx_v8_0_init_golden_registers(adev);
 	gfx_v8_0_constants_init(adev);
 
-	r = gfx_v8_0_csb_vram_pin(adev);
-	if (r)
-		return r;
-
 	r = adev->gfx.rlc.funcs->resume(adev);
 	if (r)
 		return r;
@@ -4957,8 +4921,6 @@ static int gfx_v8_0_hw_fini(void *handle)
 		pr_err("rlc is busy, skip halt rlc\n");
 	amdgpu_gfx_rlc_exit_safe_mode(adev);
 
-	gfx_v8_0_csb_vram_unpin(adev);
-
 	return 0;
 }
 

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

@@ -1685,39 +1685,6 @@ static int gfx_v9_0_rlc_init(struct amdgpu_device *adev)
 	return 0;
 }
 
-static int gfx_v9_0_csb_vram_pin(struct amdgpu_device *adev)
-{
-	int r;
-
-	r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, false);
-	if (unlikely(r != 0))
-		return r;
-
-	r = amdgpu_bo_pin(adev->gfx.rlc.clear_state_obj,
-			AMDGPU_GEM_DOMAIN_VRAM);
-	if (!r)
-		adev->gfx.rlc.clear_state_gpu_addr =
-			amdgpu_bo_gpu_offset(adev->gfx.rlc.clear_state_obj);
-
-	amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-
-	return r;
-}
-
-static void gfx_v9_0_csb_vram_unpin(struct amdgpu_device *adev)
-{
-	int r;
-
-	if (!adev->gfx.rlc.clear_state_obj)
-		return;
-
-	r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, true);
-	if (likely(r == 0)) {
-		amdgpu_bo_unpin(adev->gfx.rlc.clear_state_obj);
-		amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
-	}
-}
-
 static void gfx_v9_0_mec_fini(struct amdgpu_device *adev)
 {
 	amdgpu_bo_free_kernel(&adev->gfx.mec.hpd_eop_obj, NULL, NULL);
@@ -2405,6 +2372,7 @@ static void gfx_v9_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
 
 static void gfx_v9_0_init_csb(struct amdgpu_device *adev)
 {
+	adev->gfx.rlc.funcs->get_csb_buffer(adev, adev->gfx.rlc.cs_ptr);
 	/* csib */
 	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmRLC_CSIB_ADDR_HI),
 			adev->gfx.rlc.clear_state_gpu_addr >> 32);
@@ -3696,10 +3664,6 @@ static int gfx_v9_0_hw_init(void *handle)
 
 	gfx_v9_0_constants_init(adev);
 
-	r = gfx_v9_0_csb_vram_pin(adev);
-	if (r)
-		return r;
-
 	r = adev->gfx.rlc.funcs->resume(adev);
 	if (r)
 		return r;
@@ -3781,8 +3745,6 @@ static int gfx_v9_0_hw_fini(void *handle)
 	gfx_v9_0_cp_enable(adev, false);
 	adev->gfx.rlc.funcs->stop(adev);
 
-	gfx_v9_0_csb_vram_unpin(adev);
-
 	return 0;
 }