2003-06-21 20:29:57

by Peter Braam

[permalink] [raw]
Subject: 2.5.72 as VMware guest doesn't boot

Hi,

I'm trying to boot a 2.5.72 guest in VMware workstation 3.2.

It uncompresess the kernel, says it is booting linux, and that's the
end of the console messages.

I have the feeling that I have possibly grossly misconfigured the
kernel. Does someone have a working .config file?

Just in case it's something more interesting, kgdb (which hasn't hit
it's initial breakpoint yet) shows:

(gdb) target remote /dev/ttys0
Remote debugging using /dev/ttys0
0x3c0a2965 in ?? ()
warning: shared library handler failed to enable breakpoint
(gdb) bt
#0 0x3c0a2965 in ?? ()
#1 0xc010c2fa in do_IRQ (regs=
{ebx = 67598, ecx = 0, edx = 20831, esi = 636928, edi = -1072672768,
= -1070014472, eax = -1054257152, xds = 123, xes = 123, orig_eax = -256,
= -1070013788, xcs = 96, eflags = 646, esp = -1070012999, xss =
0481085})
at arch/i386/kernel/irq.c:503
#2 0xc010ac70 in common_interrupt () at arch/i386/kernel/entry.S:426
(gdb) c
Continuing.
Can't send signals to this remote system. SIGEMT not sent.

Program received signal SIGEMT, Emulation trap.
0x3c0a2965 in ?? ()
(gdb)

Thanks for any help.

- Peter -


Subject: Re: 2.5.72 as VMware guest doesn't boot

Did you compile the VGA Console in the kernel ?

Joe

2003-06-21 22:31:40

by J.C. Wren

[permalink] [raw]
Subject: Kernel facilities for tracking file accesses

Does any facility exist in the 2.4 and up kernels for logging *every* open,
read, write, seek, close, etc call? I'm trying to debug a problem in a
package I don't have source to, and I'm trying to prove that it's not
applying a configuration path option to files it's opening.

Ideally, I'd like a kernel module I can load that I can apply a regexp or
limited pattern matchng to, and will log a selected group of operations as
defined by a bitmask or other configuration option. I would prefer something
that monitors the entire system, rather than trying to sandbox this
particular program (it runs as a daemon).

Do the kernels have facilties for such tracking?

--John

2003-06-21 22:57:40

by Samuel Thibault

[permalink] [raw]
Subject: Re: Kernel facilities for tracking file accesses

On sat 21 jun 2003 18:45:41 GMT, J.C. Wren wrote:
> Does any facility exist in the 2.4 and up kernels for logging *every* open,
> read, write, seek, close, etc call?

I'm currently working on Robert D. Russell's "Fast Kernel Tracing", see
his homepage http://www.cs.unh.edu/~rdr/ for the paper & distribution.
This is for kernel 2.4.0-test9, though.

The most recent version I'm working on is for 2.4.21, should support
on-the-fly log flushing and work on smp (not much tested though, it
began to work yesterday :) ). It might be ported to 2.5 without too
much work.

Hoping this might meet your needs,
Samuel Thibault

2003-06-22 03:55:38

by Chris Friesen

[permalink] [raw]
Subject: Re: Kernel facilities for tracking file accesses

J.C. Wren wrote:
> Does any facility exist in the 2.4 and up kernels for logging *every* open,
> read, write, seek, close, etc call?

Can't you just use strace against the daemon?

> I would prefer something
> that monitors the entire system, rather than trying to sandbox this
> particular program (it runs as a daemon).

How about the Linux Trace Toolkit?

Chris

--
Chris Friesen | MailStop: 043/33/F10
Nortel Networks | work: (613) 765-0557
3500 Carling Avenue | fax: (613) 765-2986
Nepean, ON K2H 8E9 Canada | email: [email protected]