From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Boyd Subject: Re: [PATCH v3 3/9] media: venus: adding core part and helper functions Date: Fri, 11 Nov 2016 14:54:52 -0800 Message-ID: <20161111225452.GD5177@codeaurora.org> References: <1478540043-24558-1-git-send-email-stanimir.varbanov@linaro.org> <1478540043-24558-4-git-send-email-stanimir.varbanov@linaro.org> <20161110214339.GF16026@codeaurora.org> <550b86f3-e687-ddcd-2f20-d430bbe06940@linaro.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <550b86f3-e687-ddcd-2f20-d430bbe06940@linaro.org> Sender: linux-kernel-owner@vger.kernel.org To: Stanimir Varbanov Cc: Mauro Carvalho Chehab , Hans Verkuil , Andy Gross , Bjorn Andersson , Srinivas Kandagatla , linux-media@vger.kernel.org, linux-kernel@vger.kernel.org, linux-arm-msm@vger.kernel.org List-Id: linux-arm-msm@vger.kernel.org On 11/11, Stanimir Varbanov wrote: > On 11/10/2016 11:43 PM, Stephen Boyd wrote: > > > > Should this be ret |= ? Only the last time through the loop will > > there be an error. Or perhaps we should be bailing out early from > > this loop? > > I think that even if unset_buffers fail we need to free the memory. In > case of an error in firmware while unset buffers we should mark the > session as invalid and abort the session acordingly. Ok. So perhaps we shouldn't have any return value for this function then? -- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project