Messages in this thread | | | Date | Fri, 3 Oct 2014 11:58:05 -0400 | From | Dave Jones <> | Subject | Re: pipe/page fault oddness. |
| |
On Fri, Oct 03, 2014 at 08:43:57AM -0700, Linus Torvalds wrote: > On Thu, Oct 2, 2014 at 10:00 PM, Sasha Levin <sasha.levin@oracle.com> wrote: > > > > For the record, I tweaked the environment to put some more pressure on the > > scheduler and found out what broke (which is not related to this thread at > > all). > > Ok. It's probably still worth testing Mel's patches, since that's what > goes into 3.17. But it would be interesting to eventually go back to > stability-testing the protnone set, since it *looked* like it was > working well aside from this issue. I might decide to just do it > during the 3.18 merge window..
FYI, I've been trying to reproduce that initial bug all week, and haven't seen a single reoccurance of it, just other bugs. Kind of a pain in the ass for confirming this numa stuff was indeed the cause, but that's been true of so many of the more obscure bugs trinity has shaken out.
Dave
| |