From mboxrd@z Thu Jan 1 00:00:00 1970 From: t-kristo@ti.com (Tero Kristo) Date: Tue, 25 Sep 2012 19:33:32 +0300 Subject: [PATCHv7 00/21] ARM: OMAP3+: auto retention support Message-ID: <1348590833-12335-1-git-send-email-t-kristo@ti.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi, This set applies on top of linux-3.6-rc5 + func-pwrst code (from Jean) + omap4 core retention set (from me) + PM usecounting changes set (from me). Changes compared to previous version: - Added proper OMAP4 auto-retention support (now that the PM usecount set below supports OMAP4 also properly) - Merged TPS6236x support to this set (patches 16,18-20), without this there will be an issue with patch #21, as panda board will misbehave without TPS support and auto-ret enabled - Patch #15 fixes the addressing of voltage channels for OMAP4, without this the voltage channels either end up changing wrong voltage channel or not changing anything at all - Patch #16 is new based on the discussion earlier on the separate TPS set * uses proper I2C parameters based on board data, this is still in a data table though, as calculating these runtime would be rather complicated (if even possible, as it requires some higher order mathematics according to my understanding) - Patch #17 enables high speed I2C communication for voltage channel on OMAP4 (this now works thanks to patch #16) - Patch #18 is needed for 4460 boards, otherwise the boot-up OPP detection fails and voltage control doesn't work at all on 4460 boards - Patch #19 was pulled from the TPS support set - Patch #20 was pulled from the TPS support set, with following changes: * board setup for the GPIOs was sanitized a bit * dropped a number of unused defines from the code Working branch available: git://gitorious.org/~kristo/omap-pm/omap-pm-work.git branch: mainline-3.6-rc5-omap-auto-ret-v7 Testing done: - hw used: omap3 beagle rev c4, omap4460 gp panda, omap4430 emu blaze - suspend + cpuidle - measured core + mpu voltage rails on all setups to verify that the voltage transitions work properly during idle -Tero