[PATCH] ARM: OMAP2+: hwmod: Rearm wake-up interrupts for DT when MUSB is idled

[PATCH] ARM: OMAP2+: hwmod: Rearm wake-up interrupts for DT when MUSB is idled

[Tony Lindgren]

Looks like MUSB cable removal can cause wake-up interrupts to
stop working for device tree based booting at least for UART3
even as nothing is dynamically remuxed. This can be fixed by
calling reconfigure_io_chain() for device tree based booting
in hwmod code. Note that we already do that for legacy booting
if the legacy mux is configured.

My guess is that this is related to UART3 and MUSB ULPI
hsusb0_data0 and hsusb0_data1 support for Carkit mode that
somehow affect the configured IO chain for UART3 and require
rearming the wake-up interrupts.

In general, for device tree based booting, pinctrl-single
calls the rearm hook that in turn calls reconfigure_io_chain
so calling reconfigure_io_chain should not be needed from the
hwmod code for other events.

So let's limit the hwmod rearming of iochain only to
HWMOD_FORCE_MSTANDBY where MUSB is currently the only user
of it. If we see other devices needing similar changes we can
add more checks for it.

Cc: Paul Walmsley <paul@pwsan.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>

--- a/arch/arm/mach-omap2/omap_hwmod.c
+++ b/arch/arm/mach-omap2/omap_hwmod.c
@@ -2185,6 +2185,8 @@ static int _enable(struct omap_hwmod *oh)
 			 oh->mux->pads_dynamic))) {
 		omap_hwmod_mux(oh->mux, _HWMOD_STATE_ENABLED);
 		_reconfigure_io_chain();
+	} else if (oh->flags & HWMOD_FORCE_MSTANDBY) {
+		_reconfigure_io_chain();
 	}
 
 	_add_initiator_dep(oh, mpu_oh);
@@ -2291,6 +2293,8 @@ static int _idle(struct omap_hwmod *oh)
 	if (oh->mux && oh->mux->pads_dynamic) {
 		omap_hwmod_mux(oh->mux, _HWMOD_STATE_IDLE);
 		_reconfigure_io_chain();
+	} else if (oh->flags & HWMOD_FORCE_MSTANDBY) {
+		_reconfigure_io_chain();
 	}
 
 	oh->_state = _HWMOD_STATE_IDLE;

[PATCH] ARM: OMAP2+: hwmod: Rearm wake-up interrupts for DT when MUSB is idled

[Paul Walmsley]

On Mon, 25 Aug 2014, Tony Lindgren wrote:

> Looks like MUSB cable removal can cause wake-up interrupts to
> stop working for device tree based booting at least for UART3
> even as nothing is dynamically remuxed. This can be fixed by
> calling reconfigure_io_chain() for device tree based booting
> in hwmod code. 

Weird, nice find.  Do you know if this applies to all OMAPs, or just 
OMAP3?

> Note that we already do that for legacy booting
> if the legacy mux is configured.
> 
> My guess is that this is related to UART3 and MUSB ULPI
> hsusb0_data0 and hsusb0_data1 support for Carkit mode that
> somehow affect the configured IO chain for UART3 and require
> rearming the wake-up interrupts.
> 
> In general, for device tree based booting, pinctrl-single
> calls the rearm hook that in turn calls reconfigure_io_chain
> so calling reconfigure_io_chain should not be needed from the
> hwmod code for other events.
> 
> So let's limit the hwmod rearming of iochain only to
> HWMOD_FORCE_MSTANDBY where MUSB is currently the only user
> of it. If we see other devices needing similar changes we can
> add more checks for it.

Could you please add a new hwmod flag for this case?  Maybe something like 
HWMOD_RECONFIG_IO_CHAIN?  I think that would make the code easier to read 
and more maintainable, since the workaround would then be explicitly 
connected with that particular workaround's flag.  Looks like we have 
several flag bits left.  Plus if we wind up having to set 
HWMOD_FORCE_MSTANDBY for other devices that might not need the I/O chain 
reconfiguration, we won't have to reimplement this flag workaround.


regards


- Paul


> 
> Cc: Paul Walmsley <paul@pwsan.com>
> Signed-off-by: Tony Lindgren <tony@atomide.com>
> 
> --- a/arch/arm/mach-omap2/omap_hwmod.c
> +++ b/arch/arm/mach-omap2/omap_hwmod.c
> @@ -2185,6 +2185,8 @@ static int _enable(struct omap_hwmod *oh)
>  			 oh->mux->pads_dynamic))) {
>  		omap_hwmod_mux(oh->mux, _HWMOD_STATE_ENABLED);
>  		_reconfigure_io_chain();
> +	} else if (oh->flags & HWMOD_FORCE_MSTANDBY) {
> +		_reconfigure_io_chain();
>  	}
>  
>  	_add_initiator_dep(oh, mpu_oh);
> @@ -2291,6 +2293,8 @@ static int _idle(struct omap_hwmod *oh)
>  	if (oh->mux && oh->mux->pads_dynamic) {
>  		omap_hwmod_mux(oh->mux, _HWMOD_STATE_IDLE);
>  		_reconfigure_io_chain();
> +	} else if (oh->flags & HWMOD_FORCE_MSTANDBY) {
> +		_reconfigure_io_chain();
>  	}
>  
>  	oh->_state = _HWMOD_STATE_IDLE;
> 


- Paul

[PATCH] ARM: OMAP2+: hwmod: Rearm wake-up interrupts for DT when MUSB is idled

[Tony Lindgren]

* Paul Walmsley <paul@pwsan.com> [140827 09:38]:
> On Mon, 25 Aug 2014, Tony Lindgren wrote:
> 
> > Looks like MUSB cable removal can cause wake-up interrupts to
> > stop working for device tree based booting at least for UART3
> > even as nothing is dynamically remuxed. This can be fixed by
> > calling reconfigure_io_chain() for device tree based booting
> > in hwmod code. 
> 
> Weird, nice find.  Do you know if this applies to all OMAPs, or just 
> OMAP3?

I guess it may also apply to others too, but so far I've only
seen this on omap3 occasionally.

> > Note that we already do that for legacy booting
> > if the legacy mux is configured.
> > 
> > My guess is that this is related to UART3 and MUSB ULPI
> > hsusb0_data0 and hsusb0_data1 support for Carkit mode that
> > somehow affect the configured IO chain for UART3 and require
> > rearming the wake-up interrupts.
> > 
> > In general, for device tree based booting, pinctrl-single
> > calls the rearm hook that in turn calls reconfigure_io_chain
> > so calling reconfigure_io_chain should not be needed from the
> > hwmod code for other events.
> > 
> > So let's limit the hwmod rearming of iochain only to
> > HWMOD_FORCE_MSTANDBY where MUSB is currently the only user
> > of it. If we see other devices needing similar changes we can
> > add more checks for it.
> 
> Could you please add a new hwmod flag for this case?  Maybe something like 
> HWMOD_RECONFIG_IO_CHAIN?  I think that would make the code easier to read 
> and more maintainable, since the workaround would then be explicitly 
> connected with that particular workaround's flag.  Looks like we have 
> several flag bits left.  Plus if we wind up having to set 
> HWMOD_FORCE_MSTANDBY for other devices that might not need the I/O chain 
> reconfiguration, we won't have to reimplement this flag workaround.

OK that's a good idea as this may not be always HWMOD_FORCE_MSTANDBY.
I guess I initially assumed that this was always related to
HWMOD_FORCE_MSTANDBY as that's what seemed to hang the system until
I figured it was just the wake-up events that stopped working :)

I'll do that as a follow-up patch as I already sent the pull request
for the $subject one as it was hanging my tests occasionally.

Regards,

Tony