From mboxrd@z Thu Jan  1 00:00:00 1970
From: peter.griffin@linaro.org (Peter Griffin)
Date: Wed, 26 Oct 2016 09:56:20 +0100
Subject: [PATCH] ARM: sti: stih410-clocks: Add PROC_STFE as a critical
 clock
In-Reply-To: <20161026085141.GJ8574@dell>
References: <1476804957-24000-1-git-send-email-peter.griffin@linaro.org>
 <20161024085304.GE14477@dell>
 <20161024134416.GB10440@griffinp-ThinkPad-X1-Carbon-2nd>
 <20161025074257.GF8574@dell>
 <20161025093926.GA3206@griffinp-ThinkPad-X1-Carbon-2nd>
 <20161025100141.GH8574@dell>
 <20161025134940.GA3993@griffinp-ThinkPad-X1-Carbon-2nd>
 <20161026085141.GJ8574@dell>
Message-ID: <20161026085620.GA5476@griffinp-ThinkPad-X1-Carbon-2nd>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Hi Lee,

On Wed, 26 Oct 2016, Lee Jones wrote:

> > > > If the clock is enabled when Linux boots, the Linux clock framework *needs*
> > > > to assume the hardware may have been used in previous boot stages, and it should
> > > > not attempt to disable the clock.
> > > 
> > > None of the boot loaders we use do this.
> > 
> > But the Linux kernel isn't just used by us. It is not uncommon for STB
> > bootloaders to get information from the frontend as part of the boot process. 
> 
> Okay, this is the clincher.  Since we need to support non-standard
> bootloaders, it's difficult to guarantee that the clock will be
> disabled at boot.  For this reason, I believe that we can call this a
> critical clock.
> 
That's good news as the STi maintainer already acked and applied the patch.

Peter.