Il 31/07/2014 15:05, Frederic Konrad ha scritto: >>> >> >> I think that our first patch set part (icount fixes and icount based >> clock) is fully reviewed and can be upstreamed, >> but since the reverse execution mechanism itself is very similar >> between our series and Pavels, while Pavel’s >> set goes further in terms of addressing IO replay, I suggest Pavel >> uses our icount counter with his mechanism. >> >> In addition maybe this series can be split eg: >> the migration related series/the mechanism itself so it will be easier >> for me to review your mechanism? > > Is that making sense?
Yes, definitely. If you rebase reverse execution on top of branch uq/master of git://git.kernel.org/pub/scm/virt/kvm/qemu-kvm.git I can include it in the same pull request as icount alignment (which is now ready for inclusion in 2.2). Paolo