Re: [RFC][PATCH] misc: Introduce reboot_reason driver

[Bjorn Andersson <bjorn.andersson@sonymobile.com>]

On Wed 09 Dec 17:32 PST 2015, John Stultz wrote:

> On Tue, Dec 8, 2015 at 2:07 PM, Bjorn Andersson
> <bjorn.andersson@sonymobile.com> wrote:
> > On Tue 08 Dec 13:29 PST 2015, John Stultz wrote:
> >> diff --git a/arch/arm/boot/dts/qcom-apq8064-nexus7-flo.dts b/arch/arm/boot/dts/qcom-apq8064-nexus7-flo.dts
> >> index 5183d18..ee5dcb7 100644
> >> --- a/arch/arm/boot/dts/qcom-apq8064-nexus7-flo.dts
> >> +++ b/arch/arm/boot/dts/qcom-apq8064-nexus7-flo.dts
> >> @@ -282,6 +282,15 @@
> >>                       };
> >>               };
> >>
> >> +             reboot_reason: reboot_reason@2a03f65c {
> >> +                     compatible              = "reboot_reason";
> >> +                     reg                     = <0x2A03F65C 0x4>;
> >> +                     reason,none             = <0x77665501>;
> >> +                     reason,bootloader       = <0x77665500>;
> >> +                     reason,recovery         = <0x77665502>;
> >> +                     reason,oem              = <0x6f656d00>;
> >> +             };
> >> +
> >
> > This address refers to IMEM, which is shared with a number of other
> > uses. So I think we should have a simple-mfd (and syscon) with this
> > within.
> 
> So talking with Arnd some more it looked like IMEM was really just
> SRAM. Is that not the case, or is there something else special about
> it?  Does it really need simple-mfd and syscon? I'm still fuzzy on how
> to use those for this.
> 

I'm pretty sure it's just SRAM, but I hadn't looked at that binding
before, sounds like a conceptually better fit.

The part that I was looking for was the convenience of having a regmap
available for the uses that we will find later on, but I guess sram
provides similar means of accessing various pieces of the memory.

> >> +     /* initialize specified reasons from DT */
> >> +     if (!of_property_read_u32(pdev->dev.of_node, "reason,none", &val))
> >> +             reasons[NONE] = val;
> >> +     if (!of_property_read_u32(pdev->dev.of_node, "reason,bootloader", &val))
> >> +             reasons[BOOTLOADER] = val;
> >> +     if (!of_property_read_u32(pdev->dev.of_node, "reason,recovery", &val))
> >> +             reasons[RECOVERY] = val;
> >> +     if (!of_property_read_u32(pdev->dev.of_node, "reason,oem", &val))
> >> +             reasons[OEM] = val;
> >
> > I would like for this to be less hard coded.
> 
> So thinking of this more. Is having something like:
> 
> cmds = "default", "bootloader", "recovery";
> vals  = <0xmagic1>, <0xmagic2>, <0xmagic3>;
> 
> what you're thinking about?

As these are normally just ascii strings I was thinking we could have
them as individual properties and then use for_each_property_of_node()
on the implementation side. But it doesn't really matter.

> 
> This wouldn't quite handle the "oem-N" options as simply, but they
> could define each oem- case explicitly in the DT to support it.
> 

If we have a reasonably dynamic way of defining these there's little to
no reason to treat oem-N specially from the others.

Regards,
Bjorn