currently, i'm working on a short writeup showing how to use gdb to
debug a running kernel and, when i first wrote this a while back, it
was for a 32-bit system and things worked just fine.
in short, i fired up gdb using the standard kernel-oriented
invocation with:
$ sudo gdb /tmp/vmlinux /proc/kcore
where i copied the vmlinux file for the running kernel into /tmp to
save keystrokes. and that's the standard way to use gdb to debug the
kernel, but here's where it starts to get weird.
on a 32-bit system, if i wanted to print the current value of
jiffies, i used:
(gdb) p jiffies_64
however, since this *is* a 64-bit system, i apparently have no such
symbol, i would just use:
(gdb) p __jiffies
$12 = 4294937296
(gdb) p __jiffies
$13 = 4294937296
(gdb) p __jiffies
$14 = 4294937296
...
and as kernel-oriented gdb users know, you'll keep getting the same
answer over and over until you flush the gdb cache and reread
/proc/kcore:
(gdb) core-file /proc/kcore
but even after i do that, i'll get the same value. why? is there
some really fundamental difference between debugging a 32-bit vs
64-bit kernel?
and this is also weird:
(gdb) p loops_per_jiffy
$16 = 4096
huh? that value is way too low for this system. on my old 32-bit
system, that value would be around 2 million or so.
none of this makes sense. am i missing something fundamental? this
is the first time i've tried this on a 64-bit system.
rday
--
========================================================================
Robert P. J. Day Waterloo, Ontario, CANADA
Top-notch, inexpensive online Linux/OSS/kernel courses
http://crashcourse.ca
Twitter: http://twitter.com/rpjday
LinkedIn: http://ca.linkedin.com/in/rpjday
========================================================================
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to [email protected]
Please read the FAQ at http://kernelnewbies.org/FAQ
