Cyril Brulebois, le Thu 21 Aug 2014 02:03:48 +0200, a écrit : > Steven Chamberlain <ste...@pyro.eu.org> (2014-08-21): > > On 21/08/14 00:23, Steven Chamberlain wrote: > > > Using qemu-system-i386 with the -smp 2 flag, I started seeing weird > > > issues early in d-i: > > > > It seems more likely this was triggered by -enable-kvm (on a wheezy > > linux-amd64 host). > > > > Without that flag it looks reliable so far, even with -smp 2 (but I'm > > worried it could be non-deterministic). > > While this might be a bug in the emulation layer indeed, another view > would be that -enable-kvm means stuff go quicker and you might have > higher chances of hitting race conditions.
Without -enable-kvm, one only gets software simulation, which exposes extremely little concurrency indeed: there is no real concurrent execution at all, and just periodic execution switch from one virtual CPU to another. Samuel -- To UNSUBSCRIBE, email to debian-bsd-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20140821001142.gg21...@type.youpi.perso.aquilenet.fr
