From: bsa@kf8nh.wariat.org (Brandon S. Allbery) Subject: Re: possible bug in mount() Date: Sun, 18 Apr 1993 01:43:22 GMT
In article <1993Apr17.233557.10285@klaava.Helsinki.FI> torvalds@klaava.Helsinki.FI (Linus Torvalds) writes:
>In article <1993Apr17.220003.3133@kf8nh.wariat.org> bsa@kf8nh.wariat.org (Brandon S. Allbery) writes:
>I'd like to point out that what you saw wasn't even close to a panic -
Okay, sorry. My experience has always been that kernel GP faults are
immediately followed by panics... even once or twice under Linux, but with
older kernels.
In any case, I still think it's improper for the kernel to get into trouble
over bad user-mode actions (within limits; some things can't be caught, but
most of those are reserved for root), but I'm not complaining about the kernel
so much as HJ's assumption that the fix is just to use prototypes in user-mode
code. I've seen programs which scramble their own memory and then execute it
(jump through an NFG function pointer) crash systems in the past, and there is
even a program available which deliberately tries to do that to test the
resilience of a kernel. (I hear that most commercial *ixes fail the test....)
++Brandon
-- Brandon S. Allbery bsa@kf8nh.wariat.orgIt's not too late to turn back from the "Gates" of Hell... Linux: the FREE 32-bit operating system, available NOW. Why waaaaaait for NT?