> According to Dr. Werner Fink:
> > What's about mistakes? This interface is very easy and
> > therefore it _is_ very dangerous, isn't it?
>
> I agree. Perhaps the device should require that writing take place
> at an offset greater than zero ("dd oskip=1k"?), [...]
i'm just a kernel neophyte here, but since we're talking about hacks
(;-)...
how about a sort of side-device that's used as a guard? this reminds
me of the big paper-cutter machine i used in high school print shop --
one had to press the left-button with the left hand and the
right-button with the right hand, or the blade wouldn't come down. so
there could be a /dev/bios and /dev/biosguard, the latter of which has
to be open (for writing, of course) in order for a dangerous write to
the former. like this:
fd1 = open("/dev/biosguard", O_WRONLY);
fd2 = open("/dev/bios", O_WRONLY);
write(fd2, newbioscode, newbioscodelen);
close(fd2);
close(fd1);
i'm assuming the bios driver could know which process has opened
/dev/bios and make sure that same process has also opened
/dev/biosguard.
obviously, this could work under /proc, too.
john
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu