From mboxrd@z Thu Jan 1 00:00:00 1970 From: Pavel Machek Subject: Re: [PATCH v4 1/2] ARM: socfpga: support suspend to ram Date: Tue, 26 May 2015 23:56:17 +0200 Message-ID: <20150526215617.GA3137@amd> References: <1432317735-29413-1-git-send-email-atull@opensource.altera.com> <1432317735-29413-2-git-send-email-atull@opensource.altera.com> <20150525124133.GF9943@amd> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: devicetree-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: atull Cc: Dinh Nguyen , Rob Herring , Arnd Bergmann , Steffen Trumtrar , Pawel Moll , Mark Rutland , Ian Campbell , Kumar Gala , devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Russell King , linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, Delicious Quinoq List-Id: devicetree@vger.kernel.org Hi! > Thanks for the feedback. > > Yes I agree. > > One possibility is that I could add a suspend function to EDAC that > always fails with a helpful message like "Suspend with EDAC not > supported." If someone wants to suspend, they'll have to take the > EDAC out of the device tree so its probe will bomb out. I'm not > sure how normal it is to have suspend functions in the kernel > that always fail, though. "Interesting" solution, but better than config exclusion, I'd say. > > > + /* > > > + * This code assumes that the bootloader has already > > > + * properly configured the sdram controller: > > > > Which bootloaders can do that? > > Whatever bootloader is configuring the sdram controller for the > board. Ok, so direct question: will mainline u-boot 2015.04 configure the controller correctly? > > > + * sdr.ctrlcfg.lowpwreq.selfrfshmask = 3 > > > + * sdr.ctrlcfg.lowpwrtiming.clkdisablecycles = 8 > > > + * sdr.ctrlcfg.dramtiming4.selfrfshexit = 512 > > > > Would it make sense to configure it in Linux? > > The bootloader is responsible for configuring the sdram > controller with parameters appropriate for whatever vendor's DDR > is on the board, so that's where this must go. > > Now I realize I should make this message more general, since it is > true for the socfpga cyclone5 devkit, but may need to be adjusted > for other boards. Aha, ok, the comment really needs to be fixed. Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html