[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: oskit-mach boot failure report


Here is the info you requested,

$ sym-trace -o kernel
001fffba 0013ac4c 0013d23c 0012c4c5
Generating backtrace for object file kernel
0x01fffba in (0x01fff90) panic()
0x013ac4c in (0x013a9e0) user_trap()
0x013d23c in (0x013d236) _take_trap()
0x012c4c5 in (0x012c4b0) kmsg_init()

$ sym-trace -o kernel
001fffba 001174d4 0011733f
Generating backtrace for object file kernel
0x01fffba in (0x01fff90) panic()
0x01174d4 in (0x0117410) thread_dispatch()
0x011733f in (0x0117320) thread_continue()

> Welcome to GNUmach 1.2.91-OSKit!
> user trap, type 2, code = 14295c
> Dump of trap-state at 0x0d6d7e80:
> EAX 0000000e EBX 0000001 ECX 0dadc000 EDX 0dadc000
> ESI 20584145 EDI 0d6d7eb4 EBP 30303030 ESP 00303030
> EIP 3733 EFLAGS 30332050
> CS 3733 SS 3033 DS fff8 ES 0000 FS 7f6c GS 025a
> trapno 540028976, error 37205346, from user mode
> trap
> Backtrace: fp = d6d7f68
>  001fffba 0013ac4c 0013d23c 0012c4c5 3431203dthread_dispatch
> Backtrace: fp = d6d9f48
>  001fffba 001174d4 0011733f
> _exit(1) called; rebooting...
> Press a key to reboot

Please translate these addresses into source locations in your kernel
build.  The oskit's sym-trace script is handy for this.

Done, see above.
> Ok lets see, here is some hardware info:

None of this is immediately relevant to problems oskit-mach if you can boot
other oskit kernels that use your devices.  You've tried that, right?

Yes. They boot just fine, although the SMP kernel produces a string of garbled
text right before shutting down.
> Could it be that OSKit-mach is trying to go SMP and is failing? I didn't
> do anything special when compiling OSKit or OSKit-mach.

You have to use --enable-cpus=2 (or a higher number) to compile in any SMP
support.  You would have noticed because there are #warning messages in the
oskit-mach code when compiled for SMP (because the support isn't finished).
So that is not your problem here.

No, I didn't do that because you said you were still working on it.

An unrelated issue, I got my cable modem working with Hurd by following the
examples in the settrans faq.

Note: there is a typo in the FAQ. It says to use socket/2 but the example
uses socket/1. It had more socket/1's than 2's in the faq and I killed socket/1
which wiped out my auth server. I managed to fix it up but it was aggravating.
Ether works ok, but DSELECT seems to barf after downloading about 4Mb,
I just hit CTRL-C and restarted [I]nstall ( many times ) until I was done. I was
using the apt method for dselect.

Still haven't gotten X to start. I think I need to start a translator of some kind but
haven't investigated the situation yet.

Otherwise things look good.

- Doug

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply to: