From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vasco Steinmetz Subject: Re: Slow performance of Odroid XU4 - cpufreq-dt patch needed? Date: Thu, 14 Jan 2016 18:30:29 +0100 Message-ID: <5697DB35.5000601@kyberraum.net> References: <56967861.10305@kyberraum.net> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Return-path: Received: from nexus.metro.kyberraum.net ([195.47.195.230]:42566 "EHLO mx1.kyberraum.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756281AbcANRac (ORCPT ); Thu, 14 Jan 2016 12:30:32 -0500 In-Reply-To: Sender: linux-samsung-soc-owner@vger.kernel.org List-Id: linux-samsung-soc@vger.kernel.org To: Krzysztof Kozlowski Cc: linux-samsung-soc@vger.kernel.org Hi Krzysztof, > Where there is no cpufreq driver, everything depends on the > bootloader. Usually the bootloader sets the initial value of clocks > and regulators. It shouldn't be the lowest value, though. Right. The u-bloat in use is the pretty old one provided by hk (modified u-boot-2012 IIRC), so that may explain unsuitable initial clock values. Compiled a recent u-boot-2015.10 and also one from trunk, but it's image size exceeds the hardcoded offsets within the signed firmware bootloader from hk required to load u-boot. Had also no success with another hk firmware bootloader I found which could boot >500K u-boot images. It's a mess on this board (compared to panda/beagle[bone]s). Since I haven't got a 1.8v ttl uart cable with that rare connector, debugging/analyzing bootloader or early kernel issues such as cpu init or modifying the kconfig is pretty hard. > There was a patch from Anand Moon ("ARM: dts: use vmmc-supply of > emmc/sd for exynos5422-odroidxu3") but it did not reach the level of > acceptance for merging. It would be nice to continue that work. I saw that and hoped he'd repost an improved version based on your comments. I'm new to this list so I will not try to bother you by submitting patches (yet ;-). Cheers, Vasco