Re: [VIA Support] was: [BUG FIX] Make x86_32 uni-processor Atomic ops, Atomic

[Michael S. Zick]

On Wed May 27 2009, Harald Welte wrote:
> On Wed, May 27, 2009 at 07:18:08AM -0500, Michael S. Zick wrote:
> > On Sun May 24 2009, Harald Welte wrote:
> > > 
> > > Once I understand it in full detail, I can talk to the right people inside
> > > CentaurLabs (VIA's CPU division).  
> > > 
> > > If somebody (optionally) can phrase a precise technical question that I can
> > > directly forward to somebody with low-level x86 knowledge but no Linux background,
> > > it would definitely help speeding up the process.
> > > 
> > 
> > What is the PCI Cache Line size in the CX700?  In the CN896?
> 
> The chipset documentation doesn't say anything about that, I'd have to inquire
> inside VIA.  I doubt any difference between CX700/CN896.
> 
> Also, setting the PCI config space register to a too small cache line size
> (such as 32) on a system that supports more (say 64) doesn't really cause any
> problems, but just reduces performance - as far as I know.
> 
> Setting it too big will cause trouble.  But since 32 is the default and
> only on AMD and Intel CPU's it is increased, I see no issue here either.
> 

Since the system chip sets where designed for use with the processor -
I am going to poke it up to the processor cache line size - just for fun.

If our assumptions are correct (I do agree with your statements myself) -
then all that will happen is we halve the number of cache line flushes. ;)
If not, perhaps we get another test case data point to consider.

Thanks for the quick reply;
Mike

--
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/