On Wed, Mar 25, 2020 at 10:55:28AM -0700, Jacob Pan wrote: > Each IOASID or set could have multiple users with its own HW context > to maintain. Often times access to the HW context requires thread context. > For example, consumers of IOASIDs can register notification blocks to > sync up its states. Having an atomic notifier is not feasible for these > update operations. > > This patch converts allocator lock from spinlock to mutex in preparation > for IOASID notifier.
Unfortunately this doesn't work for SVA, which needs to call ioasid_free() from the RCU callback of mmu_notifier_put(), which cannot sleep. We're relying on MMU notifers this way to ensure that there is a single IOASID per mm. Thanks, Jean _______________________________________________ iommu mailing list [email protected] https://lists.linuxfoundation.org/mailman/listinfo/iommu
