From mboxrd@z Thu Jan 1 00:00:00 1970 From: bjorn.andersson@sonymobile.com (Bjorn Andersson) Date: Thu, 30 Oct 2014 08:04:31 -0700 Subject: [RFC 5/7] soc: qcom: Add Shared Memory Driver In-Reply-To: References: <1412037291-16880-1-git-send-email-bjorn.andersson@sonymobile.com> <1412037291-16880-6-git-send-email-bjorn.andersson@sonymobile.com> <20141030003806.GU28611@sonymobile.com> Message-ID: <20141030150429.GX28611@sonymobile.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Thu 30 Oct 06:34 PDT 2014, Ohad Ben-Cohen wrote: > Hi Bjorn, > > On Thu, Oct 30, 2014 at 2:38 AM, Bjorn Andersson > wrote: > > My concern is that merging SMD into rpmsg will still give us two different > > solutions, except that they will be deeply tangled. > > I agree. > > It's very clear that SMD and rpmsg are different. But fundamentally, > they do exactly the same: new messages are placed in a ring buffer and > then an interrupt is triggered. > > If Qualcomm wanted to adopt rpmsg I'm confident there would have been > no technical barriers to do so. > Their wish list seems long, but I think you're right :) > But given that SMD is not going anywhere, I personally hope that you > could at least avoid having two different implementations of it merged > upstream. > Looking at the lack of functionality in mach-msm it seems like it's quite a lot of work getting to a point where anyone could actually test smd on those devices. If anyone have any input on how I get a piece of user space that will run with mainline and communicate with the modem to test this I'm happily listening. Regards, Bjorn