[2.0.3] Mysterious kernel hang

G Sumner Hayes (sumner+@cmu.edu)
Wed, 10 Jul 1996 12:40:58 -0400 (EDT)

Hi,

I was just reading mail when the kernel mysteriously hung.
<Alt-F1>...<ALT-F6> failed to switch consoles, the mouse (I was
running GPM) did nothing. <SHIFT-SCRLOCK> and <CTRL-SCRLOCK> revealed
nothing. Hmm. Walk across the room to another machine and telnet in
to mine - works fine. Try <CTRL-ALT-DEL> on console. No dice.
Shutdown -r now remotely and post to linux-kernel. . .

I was running 2 tcshs, telnet, vi, and standard system daemons (inetd,
kerneld, httpd, etc). No X, no DOSemu, no SVGALib - none of the
things that generally grab the console. I had run xf86quake a half
hour earlier, but had since left X and quake and was on console. The
1.3.x and 1.99.x series worked fine on this machine, as did 2.0.0 and
1.2.13. The machine had been up a couple of days running large
compiles (and multiple XEmacsen) it's a development platform for 3
full-time programmers; luckily nobody had started work yet.

The only thing out of the ordinary is that the kernel was compiled
with Pentium GCC 2.7.2pl8. I've been using that for months without
problems, but it could be the culprit. (-O2 -fno-strength-reduce,
standard -malign=2 statements, -mpentium)

% uname -a
Linux collegium.innova.com 2.0.3 #5 Sat Jul 6 17:30:26 EDT 1996 i686

Hardware: PPro 180mHz, NE2000 clone, ps2aux mouse, Adaptec 2940UW SCSI
onboard.

Software: Running Kerneld + modules, Linux kernel 2.0.3, libc 5.2.18,
Red Hat 3.0.3(picasso) heavily upgraded to support kernel 2.0.

TTFN,

Sumner