Re: [PATCH for review] [69/145] x86_64: Disable DAC on VIA PCI bridges

[Muli Ben-Yehuda <muli@il.ibm.com>]

On Fri, Aug 11, 2006 at 08:51:53AM +0200, Andi Kleen wrote:

> I've haven't decided yet. I put it out for review for now at least.
> 
> I got various reports of the VIA bridges having trouble with DAC over the
> years, but usually when I asked for confirmation the reporters disappeared.
> I finally did the patch now because with cheap 2GB DIMMs VIA systems
> with 4GB (which gives some memory over 4GB) are becomming more common.
> 
> But again the last reporters disappeared this time.
> 
> I will probably not sent it off before final confirmation again.

Ok. I'll give it and the rest of the patches a spin on my systems with
and without Calgary on Sunday.

> > >  int dma_supported(struct device *dev, u64 mask)
> > >  {
> > >  	if (dma_ops->dma_supported)
> > >  		return dma_ops->dma_supported(dev, mask);
> > 
> > I just checked, no ops has a dma_supported method... should we remove
> > it?
> 
> The if()? Possible.

the .dma_supported member of dma_ops. If no one is using it, I don't
see a point in keeping it there - we can always reintroduce it when an
IOMMU implementation that needs it comes along.

> > > +	    if (!strncmp(p, "nodac", 5))
> > > +		    allow_dac = -1;
> > 
> > Why <0? we usually set 1 for enabled and 0 for disabled.
> 
> For hardware workarounds it is usually best to have three values:
> 
> - Force disabled
> - Default based on black/white list 
>   The black/white list is not active when != 0
> - Force enabled
> 
> I tend to use -1/0/1 for this. 

I understand, makes sense. Thanks.

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