Re: 2.1.63 - testing Pentium bug workaround..

Troels Arvin (root@pc248.studmed.ku.dk)
Thu, 13 Nov 1997 00:38:33 +0100


About the pentium-bug fix in 2.1.63:

When syslogd/klogd is not running, the bug cannot be reproduced.

When syslogd/klogd is running, running a pentium-bug trigger results in
the following situation:

If I'm running X, everything locks up.

If I'm not running X, the system also stops working after displaying
the following two lines on the console:
<c2800fc8/c2800f8>
<handler c01098c0... ...done>

Pressing ctrl-scrollock reveals the following lines (among others):
syslogd -11 R current 0 134 1 143 20
klogd 10 R C1E5DF30 0 143 1 164 134
pentiumbug 21 R C008B050 0 520 496

About my system:
Intel Pentium 133 (as far as I remember)
Redhat 4.9.1 (glibc2)
Kernel compiled with egcs-2.90.14p 971023 (I know that I should be
using a 'stable' gcc for this to be of general interest, but I still
haven't found out how to have both compilers on my system)

By the way:
What's the nature of the bug-workaround? Does it bring down
performance? (I haven't had time for benchmark tests, and I'm too
un-used to my current system configuration to be able to feel a
difference.)