Re: x86/mce: machine check warning during poweroff

[Greg KH]

On Mon, Jan 16, 2012 at 06:27:16PM +0000, Luck, Tony wrote:
> > The "correct" way to fix this up would be to have a per-cpu structure
> > for all of the different mce things that are created in this driver
> > (struct device, struct mce, exception counts, work queues, polling
> > banks, etc.), but that seems pretty messy, and I imagine some of these
> > want to stay as-is for some performance issues.  As I don't know this
> > code at all, I'm a bit leary to make that kind of change.
> 
> If you get so many machine checks that you care about the performance
> of the handler - you may be worrying about the wrong things.
> 
> I'm more concerned about maintainability of the code. Seto-san has
> submitted many patches re-grouping the functions inside mce.c into
> functional areas - keeping the data structures separated makes
> sense - especially if there is some goal of splitting mce.c into
> separate files.

Ok, I'll leave that alone, and just focus on the struct device stuff, as
that I know can't be performance critical :)

thanks,

greg k-h
--
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/