Re: Linux 2.2.11pre2 proposed patch

BOSZORMENYI Zoltan (zboszor@mol.hu)
Tue, 20 Jul 1999 13:27:20 +0100


Alan Cox wrote:
>
> > And if the bugfix itself causes the file corruption then
> > no matter where you enable the bugfix and which tool you use...
>
> The bug fix that was included turned on NOLOCK - thats _known_ to have side
> effects.

That was in linux-2.2.8, and yes, someone reported to
me that he had strange effects with his Banshee with NOLOCK on,
which was solved by turning it off. Some other BM cards may have
suffered from that, too.

That's why it was replaced with something else in 2.2.10 and if
you look at http://wauug.erols.com/~balsa/linux/cyrix/index.html
and search for the November '97 news, you can find
exactly this fix which does not turn on NOLOCK. Instead,
(as the writer of that page, Andrew D. Balsa writes)
it serializes the execution of the xchg instruction
and it is the official(?) fix from Cyrix.

This fix was tested by me and by the above mentioned
Banshee owner and

1. it did not cause strange effects with his Banshee
2. it made our machines invulnerable against the "coma bug"

Best regards,
Zoltan Boszormenyi

--
Microsoft: It's where you don't want to go today.

- 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/