Messages in this thread | | | Date | Mon, 23 May 2011 13:41:06 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH v4] x86: UV uv_tlb.c cleanup |
| |
* Ingo Molnar <mingo@elte.hu> wrote:
> > * Cliff Wickman <cpw@sgi.com> wrote: > > > One logic difference: > > Breaking uv_init_per_cpu() into get_cpu_topology() and summarize_uvhub_sockets() > > actually corrects a possible bug for the theoretical case in which there is a > > missing socket on a blade. An 'smaster' (socket master) had been expected for > > every socket. > > Please, as i asked in the previous mail, do not add logic changes to the > cleanup patch! Please keep it in a separate patch, after the big cleanup patch. > > It will be very hard to debug if you find a regression in that cleanup patch: > since it also changes logic you do not know whether it's due to changed logic > or one of the cleanups having an unintended side-effect ...
Btw, once you've fixed that side-effect i think we can apply your current set of cleanups already: it's big enough already and starting to be too big for review, and the improvements are very clear. Then we could work on fixing the remaining cleanliness issues.
Thanks,
Ingo
| |