From mboxrd@z Thu Jan 1 00:00:00 1970 From: gregory.clement@free-electrons.com (Gregory CLEMENT) Date: Wed, 21 Nov 2012 09:17:26 +0100 Subject: [PATCH V3 1/3] arm: dma mapping: Export dma ops functions In-Reply-To: <50AC8D09.6050104@samsung.com> References: <1353448587-2937-1-git-send-email-gregory.clement@free-electrons.com> <1353448587-2937-2-git-send-email-gregory.clement@free-electrons.com> <50AC7D59.1020000@samsung.com> <50AC8B41.7080009@free-electrons.com> <50AC8D09.6050104@samsung.com> Message-ID: <50AC8E16.3020006@free-electrons.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 11/21/2012 09:12 AM, Marek Szyprowski wrote: > Hello, > > On 11/21/2012 9:05 AM, Gregory CLEMENT wrote: >> On 11/21/2012 08:06 AM, Marek Szyprowski wrote: >>> On 11/20/2012 10:56 PM, Gregory CLEMENT wrote: >>>> Expose the DMA operations functions. Until now only the dma_ops >>>> structs in a whole or some dma operation were exposed. This patch >>>> exposes all the dma coherents operations. They can be reused when an >>>> architecture or a driver need to create its own set of dma_operation. >>>> >>>> Signed-off-by: Gregory CLEMENT >>> >>> Besides arm_dma_set_mask() function I see no reason to export the other >>> arm dma related functions. >> >> The idea was to let other people use the arm dma related functions, >> for their own dma ops. But for the mvebu machines we only need >> arm_dma_set_mask() indeed. >> >> So you prefer that I only expose arm_dma_set_mask() and let future >> user expose other function if they need it, right? > > I would prefer to avoid exporting functions which are not used anywhere > else. This improves readability of the code and simply forces others to > think twice before they use some static function and check if their use > case is really correct. OK so I will send a new version in a few minutes. Do you think these last changes will be enough for getting your Acked-by? Thanks, Greogry