Date: Wed, 31 Mar 2004 15:45:39 -0500 From: John Baldwin <jhb@FreeBSD.org> To: Scott Long <scottl@freebsd.org> Cc: Dan Nelson <dnelson@allantgroup.com> Subject: Re: panic: mutex Giant owned at ../../../kern/kern_thread.c:1341 Message-ID: <200403311545.39264.jhb@FreeBSD.org> In-Reply-To: <40675694.3080609@freebsd.org> References: <20040328211018.GN3446@dan.emsphone.com> <20040328222031.GA19463@dan.emsphone.com> <40675694.3080609@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sunday 28 March 2004 05:49 pm, Scott Long wrote:
> Dan Nelson wrote:
> > In the last episode (Mar 28), Dan Nelson said:
> >>Got this running an ibcs2_coff binary on today's kernel:
> >>
> >>(kgdb) where
> >>#6 0xc7c1dd5b in ibcs2_wait (td=0xc7ef5540, uap=0x0)
> >> at /usr/src/sys/i386/ibcs2/ibcs2_misc.c:164
> >
> > I think this is due to ibcs2_wait being left out when wait4() was
> > converted to giant-free (giant-allergic actually). The following seems
> > to fix it:
> >
> > Index: syscalls.master
> > ===================================================================
> > RCS file: /mnt/emssrv5/home/ncvs/src/sys/i386/ibcs2/syscalls.master,v
> > retrieving revision 1.17
> > diff -u -r1.17 syscalls.master
> > --- syscalls.master 6 Feb 2004 20:20:07 -0000 1.17
> > +++ syscalls.master 28 Mar 2004 22:00:40 -0000
> > @@ -37,7 +37,7 @@
> > 4 MNOPROTO { int write(int fd, char *buf, u_int nbytes); }
> > 5 STD { int ibcs2_open(char *path, int flags, int mode); }
> > 6 MNOPROTO { int close(int fd); }
> > -7 STD { int ibcs2_wait(int a1, int a2, int a3); }
> > +7 MSTD { int ibcs2_wait(int a1, int a2, int a3); }
> > 8 STD { int ibcs2_creat(char *path, int mode); }
> > 9 NOPROTO { int link(char *path, char *link); }
> > 10 STD { int ibcs2_unlink(char *path); }
>
> I agree that this is appropriate. However, ibcs2_wait() might need to
> have some proc locks sprinkled into it. It would be good to have
> someone like John Baldwin comment on it.
It's safe, I just forgot to tag it along with the other wait functions when I
added kern_wait() it seems.
--
John Baldwin <jhb@FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve" = http://www.FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200403311545.39264.jhb>