Re: 2.1.120 broken on P5 SMP machines + fix

Carsten Gross (carsten@wohnheim.uni-ulm.de)
Sun, 6 Sep 1998 14:34:17 +0200

In article <19980906105830.55609@alpha.franken.de>,
tsbogend@alpha.franken.de (Thomas Bogendoerfer) writes:
> I'm not sure, but it looks like the F00F workaround gets only enabled for
> one CPU, when we do smp_begin() before check_bugs(). So it makes sense,
> when people see that the first start of a f00f trigger program gives
> illegal instruction while the second one hangs the machine. When we do

Thanks for the patch. I've applied it, because otherwise 2.1.120 won't boot
on my machine and gives an Oops during boot (My machine: dual P100 & HX
chipsset, SMP kernel).

with your patch, the f00f bug is shown in cpuinfo, and the workaround is
okay if check_bugs() is executed before smp_init():

python:~/ass-src/minitools>objdump --disassemble f00f-exploit
[...]
08048074 <_start>:
8048074: f0 0f c7 c8 lock cmpxchg8b %eax
[...]

You can execute this program without problems:

python:~/ass-src/minitools>./f00f-exploit
Illegal instruction (core dumped)
python:~/ass-src/minitools>./f00f-exploit
Illegal instruction (core dumped)
and so on...

Regards

Carsten

-- 
Linux, WinNT and MS-DOS. The Good, The Bad and The Ugly
Carsten Gross                     carsten@sol.wohnheim.uni-ulm.de
Wohnheim Heilmeyersteige        Sebastian Kneipp Weg 6, 89075 Ulm


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/faq.html