From mboxrd@z Thu Jan 1 00:00:00 1970 From: broonie@opensource.wolfsonmicro.com (Mark Brown) Date: Wed, 24 Nov 2010 15:43:46 +0000 Subject: [PATCH 1/9] spi/pxa2xx: don't use subys initcall for driver init In-Reply-To: References: <1290597207-29838-1-git-send-email-bigeasy@linutronix.de> <1290597207-29838-2-git-send-email-bigeasy@linutronix.de> <4CED1C95.8070300@linutronix.de> <20101124141623.GH24970@rakim.wolfsonmicro.main> Message-ID: <20101124154346.GJ24970@rakim.wolfsonmicro.main> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Nov 24, 2010 at 08:30:37AM -0700, Grant Likely wrote: > It would be easy enough to defer cpufreq initialization based on some > platform defined criteria. It's just an initcall. However, I don't Though it'd be much better to fix the thing where cpufreq doesn't have a device anyway. > know what the impact area is for subsystems that depend on cpufreq. > Are there drivers that depend on cpufreq before initialized? OMAP? It'd be fairly surprising if something were to have a strict dependency, though - more likely to be notifiers I'd expect. Usually cpufreq comes up fairly late.