[PATCH] ARM: OMAP2+: hwmod_core: enable optional clocks before main clock

[Tony Lindgren]

* Keerthy <j-keerthy at ti.com> [171222 09:43]:
> 
> 
> On Friday 22 December 2017 02:56 PM, Tero Kristo wrote:
> > The optional clocks must be enabled before the main clock after the
> > transition to clkctrl controlled clocks is done. Otherwise the module
> > we attempt to enable might be stuck in transition.
> > 
> > Reported-by: Keerthy <j-keerthy at ti.com>
> > Signed-off-by: Tero Kristo <t-kristo at ti.com>
> 
> Tested the patch on both DRA7-EVM and DRA72-EVM. I no longer see the issue:
> 
> Tested-by: Keerthy <j-keerthy at ti.com>
> 
> > ---
> > Hi Tony,
> > 
> > This patch fixes a regression seen in linux-next, where certain peripherals
> > fail to enable after the clkctrl changes are in. The case seen has been
> > with mcasp3, where it fails to transition to enabled during the audio
> > driver probe. Not sure where you want to pick this up, maybe as early
> > rc fixes if its too late to push this to linux-next?

Thanks applying into omap-for-v4.16/soc, yeah we should
get it merged during the merge window as a fix.

Regards,

Tony



> >  arch/arm/mach-omap2/omap_hwmod.c | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> > 
> > diff --git a/arch/arm/mach-omap2/omap_hwmod.c b/arch/arm/mach-omap2/omap_hwmod.c
> > index 7324048..340d05c 100644
> > --- a/arch/arm/mach-omap2/omap_hwmod.c
> > +++ b/arch/arm/mach-omap2/omap_hwmod.c
> > @@ -976,6 +976,9 @@ static int _enable_clocks(struct omap_hwmod *oh)
> >  
> >  	pr_debug("omap_hwmod: %s: enabling clocks\n", oh->name);
> >  
> > +	if (oh->flags & HWMOD_OPT_CLKS_NEEDED)
> > +		_enable_optional_clocks(oh);
> > +
> >  	if (oh->_clk)
> >  		clk_enable(oh->_clk);
> >  
> > @@ -984,9 +987,6 @@ static int _enable_clocks(struct omap_hwmod *oh)
> >  			clk_enable(os->_clk);
> >  	}
> >  
> > -	if (oh->flags & HWMOD_OPT_CLKS_NEEDED)
> > -		_enable_optional_clocks(oh);
> > -
> >  	/* The opt clocks are controlled by the device driver. */
> >  
> >  	return 0;
> >