Date: Mon, 22 Dec 2014 18:21:47 -0800 From: Neel Natu <neelnatu@gmail.com> To: Julian Elischer <julian@freebsd.org> Cc: "freebsd-virtualization@freebsd.org" <virtualization@freebsd.org> Subject: Re: bhyve failure Message-ID: <CAFgRE9EYP%2BvJrR2ygQS2MtucDq7MATmQ7u75sxmpVsHZZgi8Rw@mail.gmail.com> In-Reply-To: <CAFgRE9FPnq3x12pD02eJ0UUtqxdDBokWNoK-drV-VRskffXP_Q@mail.gmail.com> References: <54773CF8.2020306@freebsd.org> <CAFgRE9H56-LHUD%2BPHe9Hv9_37Sbn1SBjGRUhZHtHz77Bzzbs5g@mail.gmail.com> <54789699.7030507@freebsd.org> <CAFgRE9FPnq3x12pD02eJ0UUtqxdDBokWNoK-drV-VRskffXP_Q@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Julian, On Sat, Nov 29, 2014 at 6:31 PM, Neel Natu <neelnatu@gmail.com> wrote: > Hi Julian, > > On Fri, Nov 28, 2014 at 7:36 AM, Julian Elischer <julian@freebsd.org> wrote: >> On 11/28/14, 3:12 AM, Neel Natu wrote: >>> >>> Hi Julian, >>> >>> On Thu, Nov 27, 2014 at 7:02 AM, Julian Elischer <julian@freebsd.org> >>> wrote: >>>> >>>> I added WITNESS to a kernel I've been running and during boot I got this >>>> result: >>>> >>>> vm exit[0] >>>> reason VMX >>>> rip 0xffffffff8078c1a7 >>>> inst_length 7 >>>> status 0 >>>> exit_reason 2 >>> >>> That's a triple fault (from sys/amd64/vmm/intel/vmcs.h) >>> >>> #define EXIT_REASON_TRIPLE_FAULT 2 >>> >>> >>>> qualification 0x0000000000000000 >>>> inst_type 0 >>>> inst_error 0 >>>> Abort trap (core dumped) >>>> >>>> anyone have a decoder ring? >>>> >>> A few questions to narrow this down: >>> >>> Did you add WITNESS to the guest kernel or the host kernel when this >>> happened? >> >> yes I added it to the guest kernel.. I suspect it ay be related to this, >> especially since there was a kernel module loaded that was not recompiled. >> I'm trying to ascertain now whether a withness kernel requires special >> handling for its modules. >> >> >>> >>> Is this readily reproducible? >> >> don't know yet. >> > > Its interesting that the guest went all the way to a triple fault > instead of stopping at the double fault handler. > > I'll be happy to track this down if there is a reliably repro case > (guest binaries would be sufficient). > I just committed a change that should let you trace every exception seen by the guest: https://svnweb.freebsd.org/changeset/base/276098 It will let you trace all the exceptions leading up to the triple fault via ktr(4). best Neel > best > Neel > >>> >>> Can you share the host and guest revision numbers? >> >> >> guest is custom sources based on stable-10 on the same day as 10.0 was >> released.basically whatever svn revision# was on 10.0 but on the stable >> branch. >> host is >> FreeBSD 10.1-BETA1 #0 r271671: Tue Sep 16 09:31:23 PDT 2014 >> >> >> >> >>> >>> best >>> Neel >>> >>>> julian >>>> _______________________________________________ >>>> freebsd-virtualization@freebsd.org mailing list >>>> http://lists.freebsd.org/mailman/listinfo/freebsd-virtualization >>>> To unsubscribe, send any mail to >>>> "freebsd-virtualization-unsubscribe@freebsd.org" >> >>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAFgRE9EYP%2BvJrR2ygQS2MtucDq7MATmQ7u75sxmpVsHZZgi8Rw>