On Thu, Jun 05, 2025 at 12:16:48PM -0300, Jason Gunthorpe wrote:

On Wed, Jun 04, 2025 at 09:11:07PM -0700, Nicolin Chen wrote:

quoted

I found the entire ictx would be locked by iommufd_access_create(),
then the release fop couldn't even get invoked to destroy objects.

Yes, that makes sense..

It looks to me like you can safely leave ictx as NULL instead of
adding a flag? That would be nicer than leaving a unrefcounted
pointer floating around..

Hmm, there are a few iommufd_get_object calls using access->ictx
in iommufd_access_attach() and iommufd_access_destroy().

We could have a set of internal access APIs to leave access->ictx
as NULL, as an internal caller has an ictx to pass in. It's going
to be a larger change though..

Thanks
Nicolin