[PATCH v3] drm/i915/gvt: add a spin_lock to protect the current_workload

[Zhenyu Wang]

On 2017.02.16 12:54:09 +0000, Dong, Chuanxiao wrote:
> > > +	if (unlikely(!workload ||
> > 
> > My previous point is that if to fix NULL pointer reference oops, just check
> > workload as above should be enough. And looking more into this seems like
> > we don't align request complete with its workload properly that might lead
> > to inconsistent state, e.g the current_workload of scheduler might not
> > belong to notified request. And we don't have sane handling of error
> > workload, e.g might be already freed but later got notification on it. We
> > should manage workload lifecycle better aligned with request fence.
> 
> Thanks Zhenyu for the suggestion. Just had a discussion with Zhi, currently the i915 driver has a reply mechanism after a GPU reset which made the gvt workload lifecycle not aligned with request. So to better manage the workload lifecycle with the request fence may also need to investigate from i915 side as well. This will take some time. So for a short term solution to resolve the kernel NULL pointer, we just need to check the workload. Do you agree to use this short term solution to resolve the kernel NULL pointer issue?
> 

yes, as I stated above.

-- 
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: 163 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/intel-gvt-dev/attachments/20170217/d35ba746/attachment.sig>