> > No. Dumping to on-disk partition is _very_ risky, because it is easy
> > to miss the right partition. Floppy is much safer because it is much
> > harder to hit harddisk this way.
>
> In theory, when a system is "oopsing", it is unsafe to do anything.
>
> In practise, I agree with you that writing to a harddisk is more
> dangerous than writing to a floppy.
>
> On the other hand, having a system control (ioctl on a file?, sysctl?,
> /proc?) that tells the system: "Please use that for crash dumps", and
> requires the file to be appropriately sized and preallocated would, in
> my opinion, be safe enough. Once the system thinks: "this is going the
> wrong way", it should write out the messages buffer to the
> pre-determined blocks.
>
> If those pre-determined block numbers are wrong (random pointer), then
> it is most likely that lots of other stuff was broken first, so that
> writing to disk is no longer possible.
If you took a look at that floppy writer patch, it was _very_
clever. It basically took machine down to real mode and then done I/O
using bios. I think that floppy-oopser patch is safe (as long as ROM
is ROM :-).
Anything what stores block numbers relies on our i/o drivers to
work. Going through bios looks much more reliable.
Pavel
-- I'm really pavel@ucw.cz. Look at http://195.113.31.123/~pavel. Pavel Hi! I'm a .signature virus! Copy me into your ~/.signature, please!- 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/