Re: [PATCH] PM / Domains: add debugfs listing of struct generic_pm_domain-s

Re: [PATCH] PM / Domains: add debugfs listing of struct generic_pm_domain-s

[Geert Uytterhoeven]

Hi Maciej,

On Wed, Sep 10, 2014 at 1:28 PM, Maciej Matraszek
<m.matraszek@samsung.com> wrote:
> Add /sys/kernel/debug/pm_genpd/pm_genpd_summary file, which
> lists power domains in the system, their statuses and attached devices,
> resembling /sys/kernel/debug/clk/clk_summary.
>
> Currently it is impossible to inspect (from userland) whether
> a power domain is on or off. And, if it is on, which device blocks it
> from powering down. This change allows developers working on
> embedded devices power efficiency to list all necessary information
> about generic power domains in one place.

Thanks, this is _very_ useful!

> The content of pm_genpd/pm_genpd_summary file is generated by iterating
> over all generic power domain in the system, and, for each,
> over registered devices.

... and over the subdomains, if present.

I can confirm that part is working, too, cfr. the slaves column on r8a7740

    domain                      status         slaves
           /device                                      runtime status
----------------------------------------------------------------------
a4su                            off
a3sg                            off
a3sm                            on
a3sp                            on
    /devices/e6600000.pwm                               suspended
    /devices/e6c50000.serial                            active
    /devices/e6850000.sd                                suspended
    /devices/e6bd0000.mmc                               active
a4s                             on               a3sp, a3sm, a3sg
    /devices/e6900000.irqpin                            unsupported
    /devices/e6900004.irqpin                            unsupported
    /devices/e6900008.irqpin                            unsupported
    /devices/e690000c.irqpin                            unsupported
    /devices/e9a00000.ethernet                          active
a3rv                            off
a4r                             off              a3rv
    /devices/fff20000.i2c                               suspended
a4lc                            off
c5                              on               a4lc, a4r, a4s, a4su
    /devices/e6050000.pfc                               unsupported
    /devices/e6138000.timer                             active

> Signed-off-by: Maciej Matraszek <m.matraszek@samsung.com>

Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] PM / Domains: add debugfs listing of struct generic_pm_domain-s

[Maciej Matraszek]

Hi Geert!

On Thu, 2014-09-11 at 10:51 +0200, Geert Uytterhoeven wrote:
> Hi Maciej,
> 
> On Wed, Sep 10, 2014 at 1:28 PM, Maciej Matraszek
> <m.matraszek@samsung.com> wrote:
> > Add /sys/kernel/debug/pm_genpd/pm_genpd_summary file, which
> > lists power domains in the system, their statuses and attached devices,
> > resembling /sys/kernel/debug/clk/clk_summary.
> >
> > Currently it is impossible to inspect (from userland) whether
> > a power domain is on or off. And, if it is on, which device blocks it
> > from powering down. This change allows developers working on
> > embedded devices power efficiency to list all necessary information
> > about generic power domains in one place.
> 
> Thanks, this is _very_ useful!
> 
> > The content of pm_genpd/pm_genpd_summary file is generated by iterating
> > over all generic power domain in the system, and, for each,
> > over registered devices.
> 
> ... and over the subdomains, if present.
> 
> I can confirm that part is working, too, cfr. the slaves column on r8a7740
> 
>     domain                      status         slaves
>            /device                                      runtime status
> ----------------------------------------------------------------------
> a4su                            off
> a3sg                            off
> a3sm                            on
> a3sp                            on
>     /devices/e6600000.pwm                               suspended
>     /devices/e6c50000.serial                            active
>     /devices/e6850000.sd                                suspended
>     /devices/e6bd0000.mmc                               active
> a4s                             on               a3sp, a3sm, a3sg
>     /devices/e6900000.irqpin                            unsupported
>     /devices/e6900004.irqpin                            unsupported
>     /devices/e6900008.irqpin                            unsupported
>     /devices/e690000c.irqpin                            unsupported
>     /devices/e9a00000.ethernet                          active
> a3rv                            off
> a4r                             off              a3rv
>     /devices/fff20000.i2c                               suspended
> a4lc                            off
> c5                              on               a4lc, a4r, a4s, a4su
>     /devices/e6050000.pfc                               unsupported
>     /devices/e6138000.timer                             active

> > Signed-off-by: Maciej Matraszek <m.matraszek@samsung.com>
> 
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>

Thanks for testing, Geert, I'm glad you like it :)!
BTW Can I use your output as example in the commit message for v2?
It's much more informative (though 'runtime status' alignment seems
odd).

> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
> 
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds

-- 
Maciej Matraszek
Samsung R&D Institute Poland
Samsung Electronics

Re: [PATCH] PM / Domains: add debugfs listing of struct generic_pm_domain-s

[Geert Uytterhoeven]

Hi Maciej,

On Thu, Sep 11, 2014 at 6:45 PM, Maciej Matraszek
<m.matraszek@samsung.com> wrote:
> BTW Can I use your output as example in the commit message for v2?
> It's much more informative

Sure, no problem.

> (though 'runtime status' alignment seems odd).

What's wrong with it?

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] PM / Domains: add debugfs listing of struct generic_pm_domain-s

[Maciej Matraszek]

On czw, 2014-09-11 at 19:27 +0200, Geert Uytterhoeven wrote:
> Hi Maciej,
> 
> On Thu, Sep 11, 2014 at 6:45 PM, Maciej Matraszek
> <m.matraszek@samsung.com> wrote:
> > BTW Can I use your output as example in the commit message for v2?
> > It's much more informative
> 
> Sure, no problem.
> 
> > (though 'runtime status' alignment seems odd).
> 
> What's wrong with it?

Nothing big, just for such short device names it doesn't look aesthetic to me :).

Regards,
-- 
Maciej Matraszek
Samsung R&D Institute Poland
Samsung Electronics

Re: [PATCH] PM / Domains: add debugfs listing of struct generic_pm_domain-s

[Geert Uytterhoeven]

Hi Maciej,

On Fri, Sep 12, 2014 at 6:19 PM, Maciej Matraszek
<m.matraszek@samsung.com> wrote:
>> > (though 'runtime status' alignment seems odd).
>>
>> What's wrong with it?
>
> Nothing big, just for such short device names it doesn't look aesthetic to me :).

In legacy mode, using platform code (without DT), the device names are
longer :-)

BTW, the list of slaves is indented by two more spaces than the column title:

    domain                      status         slaves
a4s                             on               a3sp, a3sm, a3sg

This is also the case in V2.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds