Date: Thu, 26 Sep 2002 19:52:40 -0400 (EDT) From: Andrew Gallatin <gallatin@cs.duke.edu> To: John Baldwin <jhb@FreeBSD.ORG> Cc: Wilko Bulte <wkb@freebie.xs4all.nl>, freebsd-alpha@FreeBSD.ORG Subject: RE: Now is the time to test 4.7-RC on your Alphas.. Message-ID: <15763.40392.718008.370407@grasshopper.cs.duke.edu> In-Reply-To: <XFMail.20020926110845.jhb@FreeBSD.org> References: <20020926170429.A14342@freebie.xs4all.nl> <XFMail.20020926110845.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
John Baldwin writes:
> 4.7-RC1 has been on ftp.freebsd.org for quite some time now. RC2 is
> building atm. RC1 doesn't have any packages though. It also has
> an FTP install available. I'll try and make sure the RC2 announce
> is bcc'd to alpha and not just -stable.
We have a serious problem with the new &**&*& binutils.
On my desktop (21264A, 640MB), loading all modules but linux.ko seems
to work fine. When I load linux.ko, I get a trap that looks like
this:
#0 0xfffffc00003ed460 in dumpsys () at ../../kern/kern_shutdown.c:486
#1 0xfffffc00003ecfa8 in boot (howto=256) at
../../kern/kern_shutdown.c:316
#2 0xfffffc00003ed870 in panic (fmt=0xfffffc000061da1c "trap")
at ../../kern/kern_shutdown.c:595
#3 0xfffffc00005ad4c0 in trap (a0=0x9010102464c457f, a1=0xfffffe0019c49e30, a2=0, entry=2, framep=0xfffffe0019c49c20)
at ../../alpha/alpha/trap.c:551
#4 0xfffffc000059f31c in XentMM ()
#5 0xfffffc00003f3f2c in sysctl_register_oid (oidp=0xfffffe00020cc000)
at ../../kern/kern_sysctl.c:102
<the rest from ddb, which actually works to get a stack trace..>
sysctl_find_oid_name()
sysctl_register_iod()
sysctl_register_set()
linker_file_register_sysctls()
linker_load_file()
kldload()
syscall()
The faulting address is always 0x9010102464c457f.
This is oidp->oid_parent:
(gdb) p oidp->oid_parent
$29 = (struct sysctl_oid_list *) 0x9010102464c457f
The module itself seems to be loaded OK. Eg, I can examine areas of
the modules text segment. It is "just" the bad sysctls..
This seems to be independant of the memory size (at least as far as
setting hw.physmem is concerned).
I CANNOT duplicate this on my crashbox (21264, 512M), only my desktop.
That makes it quite a bear to debug.
Drew
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?15763.40392.718008.370407>