Debugging problems on amd64-current
jrisom at gmail.com
Fri Aug 24 00:12:09 PDT 2007
As some may remember, about a month ago I asked a question about my dvd
drive, apparently an sata atapi drive. Updating to a 7.0-current/amd64
kernel did allow it to be recognized and I've been able to use it. But
recently I'd been noticing a kernel crash I could create. My computer
is headless so I didn't get any dump(savecore saved nothing). If I
used gdb and stepped after a segfault, it'd crash the system. Pressing
the power button did nothing, and pressing and holding it was required
to get it to do something. Although I knew there was a problem, I
figured I'd update my world to match my kernel(a few apps didn't work
with the mismatch but those I needed did). It went smoothly, until I
tried running gdb.
I currently do some work on parrot(http://parrotcode.org) and have
worked on the amd64 jit for it(I prefer asm over c apparently). To do
this, I have to use a debugging to be able to look at the generated
machine code. Before, I could debug but just had to mind myself. Now
I can't debug the program at all. Here's a ps snippet while running
gdb(after running parrot).
7316 p0 R 0:19.17 gdb parrot
7320 p0 TX+ 0:08.62 /bin/tcsh -c exec
/usr/home/jisom/parrot/parrot -j t/compilers
7323 p0 Z+ 0:00.01 <defunct>
It does not matter what options I pass to 'run' but the program will
not run. The program itself will run just fine. I tried a quick test
of using gdb on ls and it worked just fine. At the same time, some
other change seems to have broken the jit I'd done, and I also can't
debug it. So presently my options seem to be "not see my dvd drive"
with a 6-stable system, "dvd drive works but risky debugging and some
stuff broken" with a 7-current kernel and a 6-stable system, or "dvd
drive works but can't debug at all" with a 7-current system. So far,
the middle option seemed the best. I'm using a GENERIC kernel not for
safety because I didn't deal with figuring out all the options needed
for the hardware I'd use. Anyway, here's some of the relevant
information that might be needed.
[freebsd:~] jisom% uname -a
FreeBSD freebsd 7.0-CURRENT FreeBSD 7.0-CURRENT #2: Wed Aug 22 07:44:19
CDT 2007 root at freebsd:/usr/obj/usr/src/sys/GENERIC amd64
[freebsd:~] jisom% dmesg
Copyright (c) 1992-2007 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD Foundation.
FreeBSD 7.0-CURRENT #2: Wed Aug 22 07:44:19 CDT 2007
root at freebsd:/usr/obj/usr/src/sys/GENERIC
WARNING: WITNESS option enabled, expect reduced performance.
Timecounter "i8254" frequency 1193182 Hz quality 0
CPU: AMD Athlon(tm) 64 Processor 3000+ (1999.79-MHz K8-class CPU)
Origin = "AuthenticAMD" Id = 0xf4a Stepping = 10
usable memory = 505294848 (481 MB)
avail memory = 485785600 (463 MB)
ACPI APIC Table: <VIAK8M AWRDACPI>
ioapic0 <Version 0.3> irqs 0-23 on motherboard
...[etc probably less relevant]...
If any more information is wanted. Since this is very disparaging to
me, I'm hoping for some resolution.
More information about the freebsd-hackers