Date: Sat, 20 Apr 2013 10:57:38 -0700 From: Marcel Moolenaar <marcel@xcllnt.net> To: John Baldwin <jhb@FreeBSD.org> Cc: Jerome Ibanes <jibanes@gmail.com>, freebsd-ia64@freebsd.org Subject: Re: panic: mutex sched lock 0 not owned at /usr/src/sys/kern/sched_ule.c:2341 Message-ID: <C2708F0E-BEA8-489A-B2C3-7D0466D7DA91@xcllnt.net> In-Reply-To: <201304031540.59183.jhb@freebsd.org> References: <CAB%2B41mFJh0kBNJYBwPGedSe7Gy9PhjGDHcNARSV78tBj2zRVUQ@mail.gmail.com> <CAB%2B41mEJ%2BhfEM7sfW=cTbGVYAq1r%2B5w3wO2K_p2Lh261a7tJAw@mail.gmail.com> <CAB%2B41mHKC-n_c=iGwfmH%2B2YsJGCbejarvcTYzos6fcc==zO-3g@mail.gmail.com> <201304031540.59183.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Apr 3, 2013, at 12:40 PM, John Baldwin <jhb@FreeBSD.org> wrote: > On Tuesday, March 26, 2013 4:18:24 pm Jerome Ibanes wrote: >> John, Marcel, @List, >>=20 >>> On most platforms the APs don't enable interrupts until their first = context >>> switch into their idle thread completes. It seems they are enabled = by >>> accident here? >>=20 >> Is it possible to disable this when booting from the iso? >=20 > Originally I said no, but you can try disabling SMP altogether by = setting=20 > kern.smp.disabled=3D1' from the loader prompt. It may actually trigger a bug in the kernel on ia64... The problem is that we're still collecting CPU information and create PCPU structures for them. With SMP disabled this means yields some inconsistency due to partial initialization. All IIRC... --=20 Marcel Moolenaar marcel@xcllnt.net
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?C2708F0E-BEA8-489A-B2C3-7D0466D7DA91>