From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andy Gross Subject: Re: [PATCH 2/4] firmware: qcom: scm: add iommu scm calls for pg table Date: Tue, 30 Aug 2016 16:15:32 -0500 Message-ID: <20160830211532.GE24683@hector.attlocal.net> References: <1471622000-1906-1-git-send-email-stanimir.varbanov@linaro.org> <1471622000-1906-3-git-send-email-stanimir.varbanov@linaro.org> <14e4872a-7775-4221-140a-e08d28ae9098@codeaurora.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <14e4872a-7775-4221-140a-e08d28ae9098@codeaurora.org> Sender: linux-arm-msm-owner@vger.kernel.org To: "Gupta, Puja" Cc: Stanimir Varbanov , Ohad Ben-Cohen , Bjorn Andersson , Stephen Boyd , Rob Herring , Mark Rutland , Rob Clark , Srinivas Kandagatla , linux-kernel@vger.kernel.org, linux-arm-msm@vger.kernel.org, linux-remoteproc@vger.kernel.org, devicetree@vger.kernel.org List-Id: devicetree@vger.kernel.org On Wed, Aug 24, 2016 at 11:35:29AM -0700, Gupta, Puja wrote: > On 8/19/2016 8:53 AM, Stanimir Varbanov wrote: > >Those two scm calls are used to get the size of secure iommu > >page table and to pass physical memory address for this page > >table. The calls are used by remoteproc venus driver to load > >the firmware. > Do we really need these additional scm calls for venus? why can't we just > reuse existing __qcom_scm_pas_mem_setup() call? Keep in mind this is with the smmu-v1 on 8916. It seems like the downstream kernel has to do extra stuff when dealing with the older smmu. The newer platforms handle this quasi-transparently. Regards, Andy