Date: Mon, 23 Jul 2007 12:52:32 +0200 From: Attilio Rao <attilio@FreeBSD.org> To: Teufel <bsd@kuehlbox.de> Cc: emulation@FreeBSD.org, current@FreeBSD.org Subject: Re: kqemu crash (page fault) with -current Message-ID: <46A48870.2080202@FreeBSD.org> In-Reply-To: <469B3491.2000402@kuehlbox.de> References: <200707131834.27131.h.schmalzbauer@omnisec.de> <4697CCEB.9080707@FreeBSD.org> <200707132155.43783.h.schmalzbauer@omnisec.de> <4698E7C4.9080001@FreeBSD.org> <469B3491.2000402@kuehlbox.de>
next in thread | previous in thread | raw e-mail | index | archive | help
Teufel wrote: > Hi Attilio, > > well, I am not Harry, but I have downloaded also your patch and compiled > it with yesterdays CURRENT CVS, launched a portsnap fetch update and > build kqemu + qemu from the ports without any changes to the default > system. > Loaded kqemu.ko, and started a win2k3 image with qemu -kernel-kqemu. It > runs now for about 10 hours. No crashes so far. But I did not try a > unpatched kernel.previously. > Just wanted to let you know that here is a 7-CURRENT running from > yesterday with kqemu in kernel and usermode under SMP. Any tests I > should do? Ok, so a fix for this (even if slightly different from what you have tested) has been committed yesterday night. Now kqemu should not care if kernel is compiled or not with KSE and can be compiled cleanly with or without KSE. As an addictional note, I can say I think SMP breakage for kqemu is due to the option SMP not passed as compilation directive since SMP offers another compile time discriminant for the ABI. In particular, if a module is compiled without SMP on a SMP kernel it will use racy synchronization primitives. This means we should force at least modules compilation to always use SMP (I'm not sure if it alredy happens honestly, I just hope so). Thanks, Attilio
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46A48870.2080202>