On 27.06.25 11:49, Sunil Khatri wrote:
> Add a debugfs file under the client directory which shares
> the root page table base address of the VM.
>
> This address could be used to dump the pagetable for debug
> memory issues.
>
> Signed-off-by: Sunil Khatri <sunil.kha...@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 2 +-
> drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 50 ++++++++++++++++++++++++-
> drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 4 +-
> 3 files changed, 53 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> index 195ed81d39ff..27aa1b551dbf 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
> @@ -1395,7 +1395,7 @@ int amdgpu_driver_open_kms(struct drm_device *dev,
> struct drm_file *file_priv)
> if (r)
> goto error_pasid;
>
> - r = amdgpu_vm_init(adev, &fpriv->vm, fpriv->xcp_id);
> + r = amdgpu_vm_init(adev, &fpriv->vm, fpriv->xcp_id, file_priv);
> if (r)
> goto error_pasid;
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
> index 3911c78f8282..93843c9bc9ed 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
> @@ -2520,12 +2520,58 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)
> get_task_comm(vm->task_info->process_name, current->group_leader);
> }
>
> +#if defined(CONFIG_DEBUG_FS)
> +static int amdgpu_pt_info_read(struct seq_file *m, void *unused)
> +{
> + struct drm_file *file;
> + struct amdgpu_fpriv *fpriv;
> + struct amdgpu_bo *root_bo;
> + int r;
> +
> + file = m->private;
> + if (!file)
> + return -EINVAL;
> +
> + fpriv = file->driver_priv;
> + if (!fpriv && !fpriv->vm.root.bo)
> + return -ENODEV;
> +
> + root_bo = amdgpu_bo_ref(fpriv->vm.root.bo);
> + r = amdgpu_bo_reserve(root_bo, true);
> + if (r) {
> + amdgpu_bo_unref(&root_bo);
> + return -EINVAL;
> + }
> +
> + seq_printf(m, "gpu_address: 0x%llx\n",
> amdgpu_bo_gpu_offset(fpriv->vm.root.bo));
> +
> + amdgpu_bo_unreserve(root_bo);
> + amdgpu_bo_unref(&root_bo);
This chunk here as helper in amdgpu_vm.c, move the rest of the code into
amdgpu_debugfs.c.
> +
> + return 0;
> +}
> +
> +static int amdgpu_pt_info_open(struct inode *inode, struct file *file)
> +{
> + return single_open(file, amdgpu_pt_info_read, inode->i_private);
> +}
> +
> +static const struct file_operations amdgpu_pt_info_fops = {
> + .owner = THIS_MODULE,
> + .open = amdgpu_pt_info_open,
> + .read = seq_read,
> + .llseek = seq_lseek,
> + .release = single_release,
> +};
> +#endif
> +
> /**
> * amdgpu_vm_init - initialize a vm instance
> *
> * @adev: amdgpu_device pointer
> * @vm: requested vm
> * @xcp_id: GPU partition selection id
> + * @file: drm_file
> *
> * Init @vm fields.
> *
> @@ -2533,7 +2579,7 @@ void amdgpu_vm_set_task_info(struct amdgpu_vm *vm)
> * 0 for success, error for failure.
> */
> int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
> - int32_t xcp_id)
> + int32_t xcp_id, struct drm_file *file)
> {
> struct amdgpu_bo *root_bo;
> struct amdgpu_bo_vm *root;
> @@ -2609,6 +2655,8 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct
> amdgpu_vm *vm,
> if (r)
> DRM_DEBUG("Failed to create task info for VM\n");
>
> + debugfs_create_file("vm_pagetable_info", 0444, file->debugfs_client,
> file,
> + &amdgpu_pt_info_fops);
Put that into amdgpu_debugfs.c instead of amdgpu_vm.c.
Apart from that looks good to me,
Christian.
> amdgpu_bo_unreserve(vm->root.bo);
> amdgpu_bo_unref(&root_bo);
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
> index f3ad687125ad..555afaf867c4 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
> @@ -487,7 +487,9 @@ int amdgpu_vm_set_pasid(struct amdgpu_device *adev,
> struct amdgpu_vm *vm,
> u32 pasid);
>
> long amdgpu_vm_wait_idle(struct amdgpu_vm *vm, long timeout);
> -int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, int32_t
> xcp_id);
> +int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, int32_t
> xcp_id,
> + struct drm_file *file);
> +
> int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm);
> void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm);
> int amdgpu_vm_lock_pd(struct amdgpu_vm *vm, struct drm_exec *exec,
