Re: [discuss] [PATCH 4/4] x86_64 ioapic: Improve the heuristics for when check_timer fails.

[Andi Kleen]

> We just got a completely different bug reported that was confirmed to be 
> caused by Andi's patch:
>    AMD64/ATI : timer is running twice as fast as it should [1]

I have such a machine that showed this problem and when I wrote the patch I 
tested it on it (and on a couple of others of course). No twice as fast on 
my testing.

In fact there are two types of ATI machines: ones that have a BIOS workaround
for the original Linux issue and ones that don't. Keeping both
happy is not easy.

So I'm somewhat dubious on that. Where is that report?

> 
> My whole point is that for 2.6.20, we can live with simply reverting 
> Andi's commit.

I agree. It's more problematical than I expected. Reverting is 
the best option right now.

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