From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: pvops dom0: no sound after boot; possibly causedby swiotlb Date: Fri, 5 Feb 2010 11:24:29 -0500 Message-ID: <20100205162429.GA7887@phenom.dumpdata.com> References: <17472149.3960821265244940127.JavaMail.servlet@kundenserver> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <17472149.3960821265244940127.JavaMail.servlet@kundenserver> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Ronny.Hegewald@online.de Cc: xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org > No, your first statement was correct. It would always return the same mask for 32-bit and 64-bit, because of the 32-bitness of the return-variable 'dma_mask'. > > So the following patch maybe would be more appropriate? I think it isn't neccessary. The 'dmal_alloc_coherent_mask' returns an unsigned long. On 32-bit 'unsigned long' is a 4-byte value, and on 64-bit that is 8-byte value. So I think your previous patch: http://lists.xensource.com/archives/html/xen-devel/2010-02/msg00038.html is correct as when it is compiled under 64-bit it would return a 64-bit value and when compiled under 32-bit, it would return a 32-bit value. This would mimick the behavior of upstream logic.