Re: [PATCH] migration: savevm_state_insert_handler: constant-time element insertion

2019-10-18 Thread Juan Quintela
Scott Cheloha wrote: > On Thu, Oct 17, 2019 at 10:43:08AM +0200, Juan Quintela wrote: >> Scott Cheloha wrote: >> >> > Registering a SaveStateEntry object via savevm_state_insert_handler() >> > is an O(n) operation because the list is a priority queue maintained by >> > walking the list from head

Re: [PATCH] migration: savevm_state_insert_handler: constant-time element insertion

2019-10-17 Thread Scott Cheloha
On Thu, Oct 17, 2019 at 10:43:08AM +0200, Juan Quintela wrote: > Scott Cheloha wrote: > > > Registering a SaveStateEntry object via savevm_state_insert_handler() > > is an O(n) operation because the list is a priority queue maintained by > > walking the list from head to tail to find a suitable i

Re: [PATCH] migration: savevm_state_insert_handler: constant-time element insertion

2019-10-17 Thread Juan Quintela
Scott Cheloha wrote: Hi > Registering a SaveStateEntry object via savevm_state_insert_handler() > is an O(n) operation because the list is a priority queue maintained by > walking the list from head to tail to find a suitable insertion point. > > This adds considerable overhead for VMs with many

[PATCH] migration: savevm_state_insert_handler: constant-time element insertion

2019-10-16 Thread Scott Cheloha
Registering a SaveStateEntry object via savevm_state_insert_handler() is an O(n) operation because the list is a priority queue maintained by walking the list from head to tail to find a suitable insertion point. This adds considerable overhead for VMs with many such objects. For instance, ppc64