From mboxrd@z Thu Jan 1 00:00:00 1970 From: laurent.pinchart@ideasonboard.com (Laurent Pinchart) Date: Thu, 26 Jan 2012 10:58:25 +0100 Subject: [Linaro-mm-sig] [PATCH 1/3] dma-buf: Introduce dma buffer sharing mechanism In-Reply-To: References: <1324891397-10877-1-git-send-email-sumit.semwal@ti.com> <201201201423.46858.laurent.pinchart@ideasonboard.com> Message-ID: <201201261058.27098.laurent.pinchart@ideasonboard.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi Sumit, On Wednesday 25 January 2012 14:56:52 Semwal, Sumit wrote: > On Fri, Jan 20, 2012 at 6:53 PM, Laurent Pinchart wrote: > > Hi Summit, > > > > Sorry for the late review. I know that this code is now in mainline, but > > I still have a couple of comments. I'll send patches if you agree with > > them. > > Hi Laurent, > > Thanks for your review; apologies for being late in replying - I was > OoO for last couple of days. No worries. [snip] > Let me know if you'd send patches for these, or should I just go ahead and > correct. I'll send patches. Another small comment. The map_dma_buf operation is defined as struct sg_table * (*map_dma_buf)(struct dma_buf_attachment *, enum dma_data_direction); If we want to let exporters cache the sg_table we should return a const struct sg_table *. unmap_dma_buf will then take a const pointer as well, which would need to be cast to a non-const pointer internally. What's your opinion on that ? -- Regards, Laurent Pinchart