On 23/02/15 19:42, Justin Cormack wrote: > ok so it eventually has to allocate memory while initialising the > sysctl nodes and that is what causes it to reschedule.
The malloc hypercall should never block and therefore never reschedule. Blocking, if it is allowed along the callpath, is handled in the rump kernel. Here "blocking" of course means blocking in rump kernel virtual time. If the malloc() or similar backend routine called by the hypercall implementation decides to e.g. do pageout, it will just appear like some instructions take a really long time, but, as opposed to rescheduling, will not cause incorrect execution in the rump kernel. I guess that should be documented better too. ------------------------------------------------------------------------------ Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server from Actuate! Instantly Supercharge Your Business Reports and Dashboards with Interactivity, Sharing, Native Excel Exports, App Integration & more Get technology previously reserved for billion-dollar corporations, FREE http://pubads.g.doubleclick.net/gampad/clk?id=190641631&iu=/4140/ostg.clktrk _______________________________________________ rumpkernel-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/rumpkernel-users
