Re: [PATCH 2/2] Suppress false "mtrr all empty" warning message when running as VMware guest

[Yan Li <elliot.li.tech@gmail.com>]

On Wed, Sep 24, 2008 at 02:54:14PM +0200, Laurent Vivier wrote:
> Le mercredi 24 septembre 2008 à 20:24 +0800, Yan Li a écrit :
> > +#ifdef CONFIG_VMWARE_GUEST_DETECT
> > +		/* the "mtrr all blank" warning will be deferred until
> > +		 * after DMI scanning and we know the machine is not a
> > +		 * VMware guest
> > +		 */
> > +		printk(KERN_INFO "CPU MTRRs all blank\n");
> > +		return EMTRR_ALL_BLANK;
> > +#else
> >  		WARN(!kvm_para_available(), KERN_WARNING
> >  				"WARNING: strange, CPU MTRRs all blank?\n");
> >  		return 0;
> > +#endif
> >  	}
> 
> perhaps something like:
> 
> #ifdef CONFIG_VMWARE_GUEST_DETECT
>  		WARN(!kvm_para_available() && !is_vmware_guest(), KERN_WARNING
> 				"WARNING: strange, CPU MTRRs all blank?\n");
> #else
>  		WARN(!kvm_para_available(), KERN_WARNING
>   				"WARNING: strange, CPU MTRRs all blank?\n");
> ...
> 
> ???

As stated in the comments: the warning has to be deferred. The reason
is that "is_vmware_guest()" replies on DMI and can only be used after
dmi_scan().  These mtrr codes here are run very early during
setup_arch() when the DMI is not available. So when I've detected all
MTRRs are blank, I returned a status code to setup_arch(), who will
call "is_vmware_guest()" later, after dmi_scan(), to decide whether to
issue a warning or not.


-- 
Li, Yan

"Everything that is really great and inspiring is created by the
individual who can labor in freedom."
              - Albert Einstein, in Out of My Later Years (1950)