From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755926AbZKFPlg (ORCPT ); Fri, 6 Nov 2009 10:41:36 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753717AbZKFPlf (ORCPT ); Fri, 6 Nov 2009 10:41:35 -0500 Received: from cantor2.suse.de ([195.135.220.15]:34502 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753175AbZKFPle (ORCPT ); Fri, 6 Nov 2009 10:41:34 -0500 Date: Fri, 06 Nov 2009 16:41:38 +0100 Message-ID: From: Takashi Iwai To: "Jan Beulich" Cc: , , , , Subject: Re: [PATCH] x86: adjust GFP mask handling for coherent allocations (take 2) In-Reply-To: <4AF18187020000780001D8AA@vpn.id2.novell.com> References: <4AF18187020000780001D8AA@vpn.id2.novell.com> User-Agent: Wanderlust/2.15.6 (Almost Unreal) SEMI/1.14.6 (Maruoka) FLIM/1.14.9 (=?UTF-8?B?R29qxY0=?=) APEL/10.7 Emacs/23.1 (x86_64-suse-linux-gnu) MULE/6.0 (HANACHIRUSATO) MIME-Version: 1.0 (generated by SEMI 1.14.6 - "Maruoka") Content-Type: text/plain; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org At Wed, 04 Nov 2009 12:28:39 +0000, Jan Beulich wrote: > > Rather than forcing GFP flags and DMA mask to be inconsistent, GFP > flags should be determined even for the fallback device through > dma_alloc_coherent_mask()/dma_alloc_coherent_gfp_flags(). This > restores 64-bit behavior as it was prior to commits > 8965eb19386fdf5ccd0ef8b02593eb8560aa3416 and > 4a367f3a9dbf2e7ffcee4702203479809236ee6e (not sure why there are two of > them), where GFP_DMA was forced on for 32-bit, but not for 64-bit, with > the slight adjustment that afaict even 32-bit doesn't need this without > CONFIG_ISA. > > Signed-off-by: Jan Beulich > Cc: Takashi Iwai Acked-by: Takashi Iwai thanks, Takashi > Cc: Jesse Barnes > > --- > arch/x86/include/asm/dma-mapping.h | 10 +++++++--- > arch/x86/kernel/pci-dma.c | 6 ++---- > 2 files changed, 9 insertions(+), 7 deletions(-) > > --- linux-2.6.32-rc6/arch/x86/include/asm/dma-mapping.h 2009-11-04 13:23:36.000000000 +0100 > +++ 2.6.32-rc6-x86-dma-mask-gfp/arch/x86/include/asm/dma-mapping.h 2009-10-27 11:10:28.000000000 +0100 > @@ -14,6 +14,12 @@ > #include > #include > > +#ifdef CONFIG_ISA > +#define ISA_DMA_BIT_MASK DMA_BIT_MASK(24) > +#else > +#define ISA_DMA_BIT_MASK DMA_BIT_MASK(32) > +#endif > + > extern dma_addr_t bad_dma_address; > extern int iommu_merge; > extern struct device x86_dma_fallback_dev; > @@ -124,10 +130,8 @@ dma_alloc_coherent(struct device *dev, s > if (dma_alloc_from_coherent(dev, size, dma_handle, &memory)) > return memory; > > - if (!dev) { > + if (!dev) > dev = &x86_dma_fallback_dev; > - gfp |= GFP_DMA; > - } > > if (!is_device_dma_capable(dev)) > return NULL; > --- linux-2.6.32-rc6/arch/x86/kernel/pci-dma.c 2009-11-04 13:23:36.000000000 +0100 > +++ 2.6.32-rc6-x86-dma-mask-gfp/arch/x86/kernel/pci-dma.c 2009-10-27 11:14:11.000000000 +0100 > @@ -45,12 +45,10 @@ int iommu_pass_through __read_mostly; > dma_addr_t bad_dma_address __read_mostly = 0; > EXPORT_SYMBOL(bad_dma_address); > > -/* Dummy device used for NULL arguments (normally ISA). Better would > - be probably a smaller DMA mask, but this is bug-to-bug compatible > - to older i386. */ > +/* Dummy device used for NULL arguments (normally ISA). */ > struct device x86_dma_fallback_dev = { > .init_name = "fallback device", > - .coherent_dma_mask = DMA_BIT_MASK(32), > + .coherent_dma_mask = ISA_DMA_BIT_MASK, > .dma_mask = &x86_dma_fallback_dev.coherent_dma_mask, > }; > EXPORT_SYMBOL(x86_dma_fallback_dev); > > >