drm/amdgpu: restrict debugfs register access under SR-IOV

Under bare metal, there is no more else to take
care of the GPU register access through MMIO.
Under Virtualization, to access GPU register is
implemented through KIQ during run-time due to
world-switch.

Therefore, under SR-IOV user can only access
debugfs to r/w GPU registers when meets all
three conditions below.
- amdgpu_gpu_recovery=0
- TDR happened
- in_gpu_reset=0

v2: merge amdgpu_virt_can_access_debugfs() into
    amdgpu_virt_enable_access_debugfs()

v3: drop ret variable in amdgpu_virt_enable_access_debugfs()
    and directly return result
Signed-off-by: Yintian Tao <yttao@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

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

@@ -152,11 +152,16 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	if (use_bank) {
 		if ((sh_bank != 0xFFFFFFFF && sh_bank >= adev->gfx.config.max_sh_per_se) ||
 		    (se_bank != 0xFFFFFFFF && se_bank >= adev->gfx.config.max_shader_engines)) {
 			pm_runtime_mark_last_busy(adev->ddev->dev);
 			pm_runtime_put_autosuspend(adev->ddev->dev);
+			amdgpu_virt_disable_access_debugfs(adev);
 			return -EINVAL;
 		}
 		mutex_lock(&adev->grbm_idx_mutex);
@@ -207,6 +212,7 @@ static int  amdgpu_debugfs_process_reg_op(bool read, struct file *f,
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 
@@ -255,6 +261,10 @@ static ssize_t amdgpu_debugfs_regs_pcie_read(struct file *f, char __user *buf,
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	while (size) {
 		uint32_t value;
 
@@ -263,6 +273,7 @@ static ssize_t amdgpu_debugfs_regs_pcie_read(struct file *f, char __user *buf,
 		if (r) {
 			pm_runtime_mark_last_busy(adev->ddev->dev);
 			pm_runtime_put_autosuspend(adev->ddev->dev);
+			amdgpu_virt_disable_access_debugfs(adev);
 			return r;
 		}
 
@@ -275,6 +286,7 @@ static ssize_t amdgpu_debugfs_regs_pcie_read(struct file *f, char __user *buf,
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 
@@ -304,6 +316,10 @@ static ssize_t amdgpu_debugfs_regs_pcie_write(struct file *f, const char __user
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	while (size) {
 		uint32_t value;
 
@@ -311,6 +327,7 @@ static ssize_t amdgpu_debugfs_regs_pcie_write(struct file *f, const char __user
 		if (r) {
 			pm_runtime_mark_last_busy(adev->ddev->dev);
 			pm_runtime_put_autosuspend(adev->ddev->dev);
+			amdgpu_virt_disable_access_debugfs(adev);
 			return r;
 		}
 
@@ -325,6 +342,7 @@ static ssize_t amdgpu_debugfs_regs_pcie_write(struct file *f, const char __user
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 
@@ -354,6 +372,10 @@ static ssize_t amdgpu_debugfs_regs_didt_read(struct file *f, char __user *buf,
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	while (size) {
 		uint32_t value;
 
@@ -362,6 +384,7 @@ static ssize_t amdgpu_debugfs_regs_didt_read(struct file *f, char __user *buf,
 		if (r) {
 			pm_runtime_mark_last_busy(adev->ddev->dev);
 			pm_runtime_put_autosuspend(adev->ddev->dev);
+			amdgpu_virt_disable_access_debugfs(adev);
 			return r;
 		}
 
@@ -374,6 +397,7 @@ static ssize_t amdgpu_debugfs_regs_didt_read(struct file *f, char __user *buf,
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 
@@ -403,6 +427,10 @@ static ssize_t amdgpu_debugfs_regs_didt_write(struct file *f, const char __user
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	while (size) {
 		uint32_t value;
 
@@ -410,6 +438,7 @@ static ssize_t amdgpu_debugfs_regs_didt_write(struct file *f, const char __user
 		if (r) {
 			pm_runtime_mark_last_busy(adev->ddev->dev);
 			pm_runtime_put_autosuspend(adev->ddev->dev);
+			amdgpu_virt_disable_access_debugfs(adev);
 			return r;
 		}
 
@@ -424,6 +453,7 @@ static ssize_t amdgpu_debugfs_regs_didt_write(struct file *f, const char __user
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 
@@ -453,6 +483,10 @@ static ssize_t amdgpu_debugfs_regs_smc_read(struct file *f, char __user *buf,
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	while (size) {
 		uint32_t value;
 
@@ -461,6 +495,7 @@ static ssize_t amdgpu_debugfs_regs_smc_read(struct file *f, char __user *buf,
 		if (r) {
 			pm_runtime_mark_last_busy(adev->ddev->dev);
 			pm_runtime_put_autosuspend(adev->ddev->dev);
+			amdgpu_virt_disable_access_debugfs(adev);
 			return r;
 		}
 
@@ -473,6 +508,7 @@ static ssize_t amdgpu_debugfs_regs_smc_read(struct file *f, char __user *buf,
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 
@@ -502,6 +538,10 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	while (size) {
 		uint32_t value;
 
@@ -509,6 +549,7 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 		if (r) {
 			pm_runtime_mark_last_busy(adev->ddev->dev);
 			pm_runtime_put_autosuspend(adev->ddev->dev);
+			amdgpu_virt_disable_access_debugfs(adev);
 			return r;
 		}
 
@@ -523,6 +564,7 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 
@@ -651,16 +693,24 @@ static ssize_t amdgpu_debugfs_sensor_read(struct file *f, char __user *buf,
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	r = amdgpu_dpm_read_sensor(adev, idx, &values[0], &valuesize);
 
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
-	if (r)
+	if (r) {
+		amdgpu_virt_disable_access_debugfs(adev);
 		return r;
+	}
 
-	if (size > valuesize)
+	if (size > valuesize) {
+		amdgpu_virt_disable_access_debugfs(adev);
 		return -EINVAL;
+	}
 
 	outsize = 0;
 	x = 0;
@@ -673,6 +723,7 @@ static ssize_t amdgpu_debugfs_sensor_read(struct file *f, char __user *buf,
 		}
 	}
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return !r ? outsize : r;
 }
 
@@ -720,6 +771,10 @@ static ssize_t amdgpu_debugfs_wave_read(struct file *f, char __user *buf,
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	/* switch to the specific se/sh/cu */
 	mutex_lock(&adev->grbm_idx_mutex);
 	amdgpu_gfx_select_se_sh(adev, se, sh, cu);
@@ -734,16 +789,20 @@ static ssize_t amdgpu_debugfs_wave_read(struct file *f, char __user *buf,
 	pm_runtime_mark_last_busy(adev->ddev->dev);
 	pm_runtime_put_autosuspend(adev->ddev->dev);
 
-	if (!x)
+	if (!x) {
+		amdgpu_virt_disable_access_debugfs(adev);
 		return -EINVAL;
+	}
 
 	while (size && (offset < x * 4)) {
 		uint32_t value;
 
 		value = data[offset >> 2];
 		r = put_user(value, (uint32_t *)buf);
-		if (r)
+		if (r) {
+			amdgpu_virt_disable_access_debugfs(adev);
 			return r;
+		}
 
 		result += 4;
 		buf += 4;
@@ -751,6 +810,7 @@ static ssize_t amdgpu_debugfs_wave_read(struct file *f, char __user *buf,
 		size -= 4;
 	}
 
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 
@@ -805,6 +865,10 @@ static ssize_t amdgpu_debugfs_gpr_read(struct file *f, char __user *buf,
 	if (r < 0)
 		return r;
 
+	r = amdgpu_virt_enable_access_debugfs(adev);
+	if (r < 0)
+		return r;
+
 	/* switch to the specific se/sh/cu */
 	mutex_lock(&adev->grbm_idx_mutex);
 	amdgpu_gfx_select_se_sh(adev, se, sh, cu);
@@ -840,6 +904,7 @@ static ssize_t amdgpu_debugfs_gpr_read(struct file *f, char __user *buf,
 
 err:
 	kfree(data);
+	amdgpu_virt_disable_access_debugfs(adev);
 	return result;
 }
 

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

@@ -33,6 +33,7 @@ static void amdgpu_job_timedout(struct drm_sched_job *s_job)
 	struct amdgpu_ring *ring = to_amdgpu_ring(s_job->sched);
 	struct amdgpu_job *job = to_amdgpu_job(s_job);
 	struct amdgpu_task_info ti;
+	struct amdgpu_device *adev = ring->adev;
 
 	memset(&ti, 0, sizeof(struct amdgpu_task_info));
 
@@ -49,10 +50,13 @@ static void amdgpu_job_timedout(struct drm_sched_job *s_job)
 	DRM_ERROR("Process information: process %s pid %d thread %s pid %d\n",
 		  ti.process_name, ti.tgid, ti.task_name, ti.pid);
 
-	if (amdgpu_device_should_recover_gpu(ring->adev))
+	if (amdgpu_device_should_recover_gpu(ring->adev)) {
 		amdgpu_device_gpu_recover(ring->adev, job);
-	else
+	} else {
 		drm_sched_suspend_timeout(&ring->sched);
+		if (amdgpu_sriov_vf(adev))
+			adev->virt.tdr_debug = true;
+	}
 }
 
 int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,

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

@@ -334,3 +334,27 @@ void amdgpu_detect_virtualization(struct amdgpu_device *adev)
 			adev->virt.caps |= AMDGPU_PASSTHROUGH_MODE;
 	}
 }
+
+bool amdgpu_virt_can_access_debugfs(struct amdgpu_device *adev)
+{
+	return amdgpu_sriov_is_debug(adev) ? true : false;
+}
+
+int amdgpu_virt_enable_access_debugfs(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev))
+		return 0;
+
+	if (amdgpu_virt_can_access_debugfs(adev))
+		adev->virt.caps &= ~AMDGPU_SRIOV_CAPS_RUNTIME;
+	else
+		return -EPERM;
+
+	return 0;
+}
+
+void amdgpu_virt_disable_access_debugfs(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev))
+		adev->virt.caps |= AMDGPU_SRIOV_CAPS_RUNTIME;
+}

drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h

@@ -265,6 +265,7 @@ struct amdgpu_virt {
 	uint32_t gim_feature;
 	uint32_t reg_access_mode;
 	int req_init_data_ver;
+	bool tdr_debug;
 };
 
 #define amdgpu_sriov_enabled(adev) \
@@ -296,6 +297,8 @@ static inline bool is_virtual_machine(void)
 
 #define amdgpu_sriov_is_pp_one_vf(adev) \
 	((adev)->virt.gim_feature & AMDGIM_FEATURE_PP_ONE_VF)
+#define amdgpu_sriov_is_debug(adev) \
+	((!adev->in_gpu_reset) && adev->virt.tdr_debug)
 
 bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev);
 void amdgpu_virt_init_setting(struct amdgpu_device *adev);
@@ -314,4 +317,8 @@ int amdgpu_virt_fw_reserve_get_checksum(void *obj, unsigned long obj_size,
 					unsigned int chksum);
 void amdgpu_virt_init_data_exchange(struct amdgpu_device *adev);
 void amdgpu_detect_virtualization(struct amdgpu_device *adev);
+
+bool amdgpu_virt_can_access_debugfs(struct amdgpu_device *adev);
+int amdgpu_virt_enable_access_debugfs(struct amdgpu_device *adev);
+void amdgpu_virt_disable_access_debugfs(struct amdgpu_device *adev);
 #endif