Re: port-i386/38935: x86 bus_dmamap_sync() should use memory barrier

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Tue, Jun 10, 2008 at 09:30:02PM +0000, Andrew Doran wrote:
> The following reply was made to PR port-i386/38935; it has been noted by 
> GNATS.
> 
> From: Andrew Doran <ad%netbsd.org@localhost>
> To: gnats-bugs%netbsd.org@localhost
> Cc: 
> Subject: Re: port-i386/38935: x86 bus_dmamap_sync() should use memory barrier
> Date: Tue, 10 Jun 2008 22:25:12 +0100
> 
>  On Tue, Jun 10, 2008 at 09:20:02PM +0000, Andrew Doran wrote:
>  
>  >  I think x86_lfence() without any if() would be enough. Does it solve the
>  >  problem for you?
>  
>  Bah, that will always evaluate to the equivalent of x86_mfence(). For this,

What would always evaluate to x86_mfence ? x86_lfence(), or my code ?

>  membar_consumer() would a be better test since it will be patched to 'lfence'
>  during boot.

where is it defined ? I'd like to know what it would look like on
xen (xen doesn't use x86/patch.c for now).

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--