On Thu, 11 Jun 1998, Mike Montour wrote:
> It's still a problem in the same sense as the Intel 'f00f' bug: a usermode
> program can cause a complete system lockup. It may not be a big deal
> for home systems like mine, but it would be a problem for any multi-user
> system running on a K6.
Well, if a user maliciously exploited this CPU erratum in order to crash
the system (as could much more easily be done with the Intel f00f bug),
then it is a problem. But usually users aren't as close to being
"experts" as we are, and they're rarely malicious, at least not on systems
they have accounts on that don't belong to them.
Also, like the f00f bug, this probably won't be a problem for a
normally running system, even a server. Only one very specific
condition causes the CPU bug, and it's one that very rarely, if
ever, would be encountered in normal running, IMHO.
> The markings on my chip are:
> AMD-K6-233ANR
> 3.2Vcore / 3.3V I/O
> C 9815DPDW
>
> It's running in an ASUS TX97-XE board, at recommended clock settings.
Good choice of motherboard. I've built a couple for friends using that
board and it's extremely stable, although a little more pricey than
some other mobo models or manufacturers. But Asus *means*quality* :^)
9815, I assume, means it was produced the 15th week of this year.
I've got my heat sink attached to the CPU with thermal grease so I
can't easily remove it and check my date :^( But the AMD web site
says the "C" means it's a revision C CPU, so it's a very recent one,
that's for sure.
Could you send us the output of /proc/cpuinfo, please? Also, you might
try running the program cpuspd3e that you can download from the AMD
web site (http://www.amd.com/K6/k6docs/cpuspd3e.exe) on your chip and
posting the results. I haven't run this on my new CPU yet, but I
understand it gives you lots of helpful info about your CPU, like
stepping/rev number, as well as its speed.
You could also try compiling the crashme program and running it with
the seed values mentioned in some of the posts to linux-kernel to
see if you can reproduce their hard lock-ups.
Clemmitt Sigler
Va. Tech Physics Dept.