On Tue, Apr 09, 2013 at 09:22:58AM -0500, Suthikulpanit, Suravee wrote:
> The reason I implemented the "per-thread IOMMU handling" and the
> "workaround" together in one patch
> is because it simplifies the synchronization of clearing and
> checking the interrupt enabling bits.
> In the previous implementation, we could end up having multiple
> threads trying to access the status register
> at the same time.
This is the reason for the iommu->lock spinlock in this code at the
moment.
Joerg
_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu
