Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 31 Aug 2010 15:39:56 -0400
From:      John Baldwin <jhb@freebsd.org>
To:        Matthew Jacob <mj@feral.com>
Cc:        svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org, Pyun YongHyeon <yongari@freebsd.org>
Subject:   Re: svn commit: r212061 - head/sys/dev/bge
Message-ID:  <201008311539.57185.jhb@freebsd.org>
In-Reply-To: <4C7D41E9.7060907@feral.com>
References:  <201008311733.o7VHXmxX037013@svn.freebsd.org> <201008311350.17175.jhb@freebsd.org> <4C7D41E9.7060907@feral.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tuesday, August 31, 2010 1:54:49 pm Matthew Jacob wrote:
>   But not amd64 please.
> 
> > Keep in mind the PAE case where you cannot effectively specify a 4GB
> > boundary.  I used a 2GB boundary for twa(4) in the PAE case to deal
> > with the boundary issue.  Probably though, bus_dma should just always
> > enforce a 4GB boundary, at least on x86.

Yes, thinking about this more, only i386 + PAE is special.  All other cases 
could represent the 4GB boundary restriction in a bus dma tag for the PCI bus 
(or in the platform-specific Host-PCI bridge drivers).  For i386 + PAE it 
might make sense to always enforce a 4GB boundary in the bus_dma code itself.

-- 
John Baldwin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201008311539.57185.jhb>