Lines Matching refs:on
46 to make decisions based on fine-grained knowledge of the source of the data.
69 3. Deadlock-free teardown of state across multiple domains on domain destroy
149 1) Acquire a Read lock on L1.
150 2) Acquire a Read lock on L2.
155 1) Acquire a Read lock on L1.
156 2) Acquire a Write lock on L2.
161 inhibit operations on those rings until the locks are released.
165 1) Acquire a Write lock on L1.
169 operation can operate on Argo rings until the locks are released.
185 The majority of operations take a read-lock on this lock, allowing concurrent
188 The pointer d->argo on every domain is protected by this lock. A set of more
194 Points of write-locking on this lock:
201 * All of the notifications pending on those rings are cancelled.
202 * All of the notifications pending for this domain on wildcard rings owned
210 Enforcing that the write_lock is acquired on `L1_global_argo_rwlock` before
222 Holding a read lock on `rings_L2` protects the ring hash table and the elements
226 Holding a write lock on `rings_L2` protects all of the elements of all the
235 write on `rings_L2_rwlock`.
241 Is accessed by the hypervisor on behalf of:
253 Is accessed by the hypervisor on behalf of:
268 Is accessed by the hypervisor on behalf of:
286 A common model for reasoning about concurrent code focusses on accesses to
290 functions, when operating on data from multiple domains with concurrent
400 lock, acquiring this lock will stall operations on other active rings owned by
403 throughput due to the contention on the single per-domain lock. The granular
410 ### Teardown on domain destroy
416 gaining safe access to the data structures on each of the domains involved.
430 To perform the required state teardown on domain destruction, which can require
447 transmission rings, to improve resistance to data read attacks on