The series makes the grant table locking for fine-grained and add
per-VCPU maptrack free lists, which greatly improves scalability.
The series builds on the original series by Matt Wilson and Christoph
Egger from Amazon.
The per-VCPU maptrack free lists makes one of our aggregate intrahost
network throughput benchmarks increases from 15 Gbit/s to 75 Gbit/s,
when compared to just Amazon's original patches.
v6:
* Remove most uses of the grant table lock.
* Make the grant table lock a spin lock again (there were only
writers left after the above)
* Add per-VCPU maptrack free lists.
v5:
* Addressed locking issue pointed out by Jan Beulich
* Fixed git rebase merge issue introduced in v4
(acquiring locking twice)
* Change for ()-loop in grant_map_exists
* Coding style fixes
v4:
* Coding style nits from Jan Beulich
* Fixup read locks pointed out by Jan Beulich
* renamed double_gt_(un)lock to double_maptrack_(un)lock
per request from Jan Beulich
* Addressed ASSERT()'s from Jan Beulich
* Addressed locking issues in unmap_common pointed out
by Jan Beulich
v3:
* Addressed gnttab_swap_grant_ref() comment from Andrew Cooper
v2:
* Add arm part per request from Julien Grall
David
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
