From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vince Hsu Subject: Re: [PATCH 1/11] ARM: tegra: add function to control the GPU rail clamp Date: Mon, 5 Jan 2015 14:55:38 +0800 Message-ID: <54AA356A.8080309@nvidia.com> References: <1419331204-26679-1-git-send-email-vinceh@nvidia.com> <1419331204-26679-2-git-send-email-vinceh@nvidia.com> <1419426990.2179.7.camel@lynxeye.de> <549B7638.2010405@nvidia.com> <1419539683.2165.6.camel@lynxeye.de> <54A0C148.6030400@nvidia.com> <1419957752.4082.2.camel@lynxeye.de> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8"; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1419957752.4082.2.camel-8ppwABl0HbeELgA04lAiVw@public.gmane.org> Sender: linux-tegra-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Lucas Stach Cc: thierry.reding-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org, swarren-3lzwWm7+Weoh9ZMKESR00Q@public.gmane.org, gnurou-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org, bskeggs-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org, martin.peres-GANU6spQydw@public.gmane.org, seven-FA6nBp6kBxZzu6KWmfFNGwC/G2K4zDHf@public.gmane.org, samuel.pitoiset-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org, nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org, linux-tegra-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-tegra@vger.kernel.org On 12/31/2014 12:42 AM, Lucas Stach wrote: > Am Montag, den 29.12.2014, 10:49 +0800 schrieb Vince Hsu: > > [...] > >>>> That's a read fence to assure the post of the previous writes through >>>> Tegra interconnect. (copy-paster from >>>> https://android.googlesource.com/kernel/tegra.git/+/28b107dcb3aa122de8e94e48af548140d519298f) >>> I see what it does, the question is more about why this is needed. >>> What is the Tegra interconnect? According to the TRM the Tegra contains >>> some standard AXI <-> AHB <-> APB bridges. That a read is needed to >>> assure the write is posted to the APB bus seems to imply that there is >>> some write buffering in one of those bridges. Can we get this documented >>> somewhere? >> The TRM does mention a read after the write. Check the section 32.2.2.3. >> > Unfortunately this doesn't seem to be included in the public TRM. It > would be nice if this could be documented either in the next version of > the TRM or as a public Appnote. > It should be in the latest public TRM. Could you check again? Thanks, Vince