[PATCH v3] drm/i915/gvt: Adding ppgtt to GVT GEM context after shadow pdps settled.

[Zhenyu Wang]

On 2019.07.04 16:45:06 +0800, Colin Xu wrote:
> Windows guest can't run after force-TDR with host log:
> ...
> gvt: vgpu 1: workload shadow ppgtt isn't ready
> gvt: vgpu 1: fail to dispatch workload, skip
> ...
> 
> The error is raised by set_context_ppgtt_from_shadow(), when it checks
> and found the shadow_mm isn't marked as shadowed.
> 
> In work thread before each submission, a shadow_mm is set to shadowed in:
> shadow_ppgtt_mm()
> <-intel_vgpu_pin_mm()
> <-prepare_workload()
> <-dispatch_workload()
> <-workload_thread()
> However checking whether or not shadow_mm is shadowed is prior to it:
> set_context_ppgtt_from_shadow()
> <-dispatch_workload()
> <-workload_thread()
> 
> In normal case, create workload will check the existence of shadow_mm,
> if not it will create a new one and marked as shadowed. If already exist
> it will reuse the old one. Since shadow_mm is reused, checking of shadowed
> in set_context_ppgtt_from_shadow() actually always see the state set in
> creation, but not the state set in intel_vgpu_pin_mm().
> 
> When force-TDR, all engines are reset, since it's not dmlr level, all
> ppgtt_mm are invalidated but not destroyed. Invalidation will mark all
> reused shadow_mm as not shadowed but still keeps in ppgtt_mm_list_head.
> If workload submission phase those shadow_mm are reused with shadowed
> not set, then set_context_ppgtt_from_shadow() will report error.
> 
> Pin for context after shadow_mm pinned and shadow pdps settled.
> 
> Fixes: 4f15665ccbba (drm/i915: Add ppgtt to GVT GEM context)
> 
> v2:
> Move set_context_ppgtt_from_shadow() after prepare_workload(). (zhenyu)
> v3:
> Move set_context_ppgtt_from_shadow() after shadow pdps updated.(zhenyu)
> 
> Cc: stable at vger.kernel.org
> Signed-off-by: Colin Xu <colin.xu at intel.com>

Applied, thanks!

> ---
>  drivers/gpu/drm/i915/gvt/scheduler.c | 27 ++++++++++++---------------
>  1 file changed, 12 insertions(+), 15 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c
> index 196b4155a309..9f3fd7d96a69 100644
> --- a/drivers/gpu/drm/i915/gvt/scheduler.c
> +++ b/drivers/gpu/drm/i915/gvt/scheduler.c
> @@ -364,16 +364,13 @@ static void release_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
>  	wa_ctx->indirect_ctx.shadow_va = NULL;
>  }
>  
> -static int set_context_ppgtt_from_shadow(struct intel_vgpu_workload *workload,
> -					 struct i915_gem_context *ctx)
> +static void set_context_ppgtt_from_shadow(struct intel_vgpu_workload *workload,
> +					  struct i915_gem_context *ctx)
>  {
>  	struct intel_vgpu_mm *mm = workload->shadow_mm;
>  	struct i915_ppgtt *ppgtt = i915_vm_to_ppgtt(ctx->vm);
>  	int i = 0;
>  
> -	if (mm->type != INTEL_GVT_MM_PPGTT || !mm->ppgtt_mm.shadowed)
> -		return -EINVAL;
> -
>  	if (mm->ppgtt_mm.root_entry_type == GTT_TYPE_PPGTT_ROOT_L4_ENTRY) {
>  		px_dma(ppgtt->pd) = mm->ppgtt_mm.shadow_pdps[0];
>  	} else {
> @@ -384,8 +381,6 @@ static int set_context_ppgtt_from_shadow(struct intel_vgpu_workload *workload,
>  			px_dma(pd) = mm->ppgtt_mm.shadow_pdps[i];
>  		}
>  	}
> -
> -	return 0;
>  }
>  
>  static int
> @@ -614,6 +609,8 @@ static void release_shadow_batch_buffer(struct intel_vgpu_workload *workload)
>  static int prepare_workload(struct intel_vgpu_workload *workload)
>  {
>  	struct intel_vgpu *vgpu = workload->vgpu;
> +	struct intel_vgpu_submission *s = &vgpu->submission;
> +	int ring = workload->ring_id;
>  	int ret = 0;
>  
>  	ret = intel_vgpu_pin_mm(workload->shadow_mm);
> @@ -622,8 +619,16 @@ static int prepare_workload(struct intel_vgpu_workload *workload)
>  		return ret;
>  	}
>  
> +	if (workload->shadow_mm->type != INTEL_GVT_MM_PPGTT ||
> +	    !workload->shadow_mm->ppgtt_mm.shadowed) {
> +		gvt_vgpu_err("workload shadow ppgtt isn't ready\n");
> +		return -EINVAL;
> +	}
> +
>  	update_shadow_pdps(workload);
>  
> +	set_context_ppgtt_from_shadow(workload, s->shadow[ring]->gem_context);
> +
>  	ret = intel_vgpu_sync_oos_pages(workload->vgpu);
>  	if (ret) {
>  		gvt_vgpu_err("fail to vgpu sync oos pages\n");
> @@ -674,7 +679,6 @@ static int dispatch_workload(struct intel_vgpu_workload *workload)
>  {
>  	struct intel_vgpu *vgpu = workload->vgpu;
>  	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
> -	struct intel_vgpu_submission *s = &vgpu->submission;
>  	struct i915_request *rq;
>  	int ring_id = workload->ring_id;
>  	int ret;
> @@ -685,13 +689,6 @@ static int dispatch_workload(struct intel_vgpu_workload *workload)
>  	mutex_lock(&vgpu->vgpu_lock);
>  	mutex_lock(&dev_priv->drm.struct_mutex);
>  
> -	ret = set_context_ppgtt_from_shadow(workload,
> -					    s->shadow[ring_id]->gem_context);
> -	if (ret < 0) {
> -		gvt_vgpu_err("workload shadow ppgtt isn't ready\n");
> -		goto err_req;
> -	}
> -
>  	ret = intel_gvt_workload_req_alloc(workload);
>  	if (ret)
>  		goto err_req;
> -- 
> 2.22.0
> 

-- 
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/intel-gvt-dev/attachments/20190705/ad4597ca/attachment.sig>