mbox series

[PATCH/RFC,0/3] pmdomain: renesas: rmobile-sysc: Remove serial console handling

Message ID cover.1716811405.git.geert+renesas@glider.be
Headers show
Series pmdomain: renesas: rmobile-sysc: Remove serial console handling | expand

Message

Geert Uytterhoeven May 27, 2024, 12:41 p.m. UTC
Hi all,

Since commit a47cf07f60dcb02d ("serial: core: Call
device_set_awake_path() for console port"), the serial driver properly
handles the case where the serial console is part of the awake path, and
it looked like we could start removing special serial console handling
from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
Unfortunately the devil is in the details, as usual...

Earlycon relies on the serial port to be initialized by the firmware
and/or bootloader.  Linux is not aware of any hardware dependencies that
must be met to keep the port working, and thus cannot guarantee they
stay met, until the full serial driver takes over.

E.g. all unused clocks and unused PM Domains are disabled in a late
initcall.  As this happens after the full serial driver has taken over,
the serial port's clock and/or PM Domain are no longer deemed unused,
and this is typically not a problem.

However, if the serial port's clock or PM Domain is shared with another
device, and that other device is runtime-suspended before the full
serial driver has probed, the serial port's clock and/or PM Domain will
be disabled inadvertently.  Any subsequent serial console output will
cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
ports share their PM Domain with several other I/O devices.  After the
use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
before the full serial driver takes over, the PM Domain containing the
early serial port is powered down, causing a lock-up when booted with
"earlycon".

This RFC patch series aims to provide a mechanism for handling this, and
to fix it for the PM Domain case:
  1. The first patch provides a mechanism to let the clock and/or PM
     Domain subsystem or drivers handle this, by exporting the clock and
     PM Domain dependencies for the serial port, as available in the
     system's device tree,
  2. The second patch introduces a new flag to handle a PM domain that
     must be kept powered-on during early boot, and by setting this flag
     if the PM Domain contains the serial console (originally I handled
     this inside rmobile-sysc, but it turned out to be easy to
     generalize this to other platforms in the core PM Domain code).
  3. The third patch removes the no longer needed special console
     handling from the R-Mobile SYSC PM Domain driver.

I did not fix the similar clock issue, as it is more complex (there can
be multiple clocks, and each clock provider can have its own value of
#clock-cells), and I do not need it for Renesas ARM platforms.

This has been tested on the APE6-EVM, Armadillo-800-EVA, and KZM-A9-GT
development boards, with and without earlycon, including s2ram with and
without no_console_suspend.

Notes:
  - This should not be needed on RZ/G3S, where each serial port device
    has its own PM Domain,
  - drivers/clk/imx/clk.c and drivers/pmdomain/imx/scu-pd.c have special
    handling for the of_stdout device, but is probably not affected, as
    each serial port seems to share its PM Domain only with the serial
    port's clock controller.

Thanks for your comments!

Geert Uytterhoeven (3):
  earlycon: Export clock and PM Domain info from FDT
  pmdomain: core: Avoid earlycon power-down
  pmdomain: renesas: rmobile-sysc: Remove serial console handling

 drivers/pmdomain/core.c                 | 24 ++++++++++++++++--
 drivers/pmdomain/renesas/rmobile-sysc.c | 33 +------------------------
 drivers/tty/serial/earlycon.c           | 14 ++++++++++-
 include/linux/pm_domain.h               |  4 +++
 include/linux/serial_core.h             | 10 ++++++++
 5 files changed, 50 insertions(+), 35 deletions(-)

Comments

Ulf Hansson June 5, 2024, 9:34 a.m. UTC | #1
+ Tomi

On Mon, 27 May 2024 at 14:41, Geert Uytterhoeven
<geert+renesas@glider.be> wrote:
>
>         Hi all,
>
> Since commit a47cf07f60dcb02d ("serial: core: Call
> device_set_awake_path() for console port"), the serial driver properly
> handles the case where the serial console is part of the awake path, and
> it looked like we could start removing special serial console handling
> from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
> Unfortunately the devil is in the details, as usual...
>
> Earlycon relies on the serial port to be initialized by the firmware
> and/or bootloader.  Linux is not aware of any hardware dependencies that
> must be met to keep the port working, and thus cannot guarantee they
> stay met, until the full serial driver takes over.
>
> E.g. all unused clocks and unused PM Domains are disabled in a late
> initcall.  As this happens after the full serial driver has taken over,
> the serial port's clock and/or PM Domain are no longer deemed unused,
> and this is typically not a problem.
>
> However, if the serial port's clock or PM Domain is shared with another
> device, and that other device is runtime-suspended before the full
> serial driver has probed, the serial port's clock and/or PM Domain will
> be disabled inadvertently.  Any subsequent serial console output will
> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
> ports share their PM Domain with several other I/O devices.  After the
> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
> before the full serial driver takes over, the PM Domain containing the
> early serial port is powered down, causing a lock-up when booted with
> "earlycon".

Hi Geert,

Thanks for the detailed description of the problem! As pointed out in
regards to another similar recent patch [1], this is indeed a generic
problem, not limited to the serial console handling.

At Linaro Connect a few weeks ago I followed up with Saravana from the
earlier discussions at LPC last fall. We now have a generic solution
for genpd drafted on plain paper, based on fw_devlink and the
->sync_state() callback. I am currently working on the genpd series,
while Saravana will re-spin the series (can't find the link to the
last version) for the clock framework. Ideally, we want these things
to work in a very similar way.

That said, allow me to post the series for genpd in a week or two to
see if it can solve your problem too, for the serial console.

Kind regards
Uffe

[1]
https://lore.kernel.org/linux-arm-kernel/CAPDyKFqShuq98qV5nSPzSqwLLUZ7LxLvp1eihGRBkU4qUKdWwQ@mail.gmail.com/

>
> This RFC patch series aims to provide a mechanism for handling this, and
> to fix it for the PM Domain case:
>   1. The first patch provides a mechanism to let the clock and/or PM
>      Domain subsystem or drivers handle this, by exporting the clock and
>      PM Domain dependencies for the serial port, as available in the
>      system's device tree,
>   2. The second patch introduces a new flag to handle a PM domain that
>      must be kept powered-on during early boot, and by setting this flag
>      if the PM Domain contains the serial console (originally I handled
>      this inside rmobile-sysc, but it turned out to be easy to
>      generalize this to other platforms in the core PM Domain code).
>   3. The third patch removes the no longer needed special console
>      handling from the R-Mobile SYSC PM Domain driver.
>
> I did not fix the similar clock issue, as it is more complex (there can
> be multiple clocks, and each clock provider can have its own value of
> #clock-cells), and I do not need it for Renesas ARM platforms.

I will defer to Sarvana here, but ideally his series for the clock
framework should solve this case too.

>
> This has been tested on the APE6-EVM, Armadillo-800-EVA, and KZM-A9-GT
> development boards, with and without earlycon, including s2ram with and
> without no_console_suspend.
>
> Notes:
>   - This should not be needed on RZ/G3S, where each serial port device
>     has its own PM Domain,
>   - drivers/clk/imx/clk.c and drivers/pmdomain/imx/scu-pd.c have special
>     handling for the of_stdout device, but is probably not affected, as
>     each serial port seems to share its PM Domain only with the serial
>     port's clock controller.
>
> Thanks for your comments!
>
> Geert Uytterhoeven (3):
>   earlycon: Export clock and PM Domain info from FDT
>   pmdomain: core: Avoid earlycon power-down
>   pmdomain: renesas: rmobile-sysc: Remove serial console handling
>
>  drivers/pmdomain/core.c                 | 24 ++++++++++++++++--
>  drivers/pmdomain/renesas/rmobile-sysc.c | 33 +------------------------
>  drivers/tty/serial/earlycon.c           | 14 ++++++++++-
>  include/linux/pm_domain.h               |  4 +++
>  include/linux/serial_core.h             | 10 ++++++++
>  5 files changed, 50 insertions(+), 35 deletions(-)
>
> --

Kind regards
Uffe
Tomi Valkeinen June 5, 2024, 10:41 a.m. UTC | #2
Hi Ulf,

On 05/06/2024 12:34, Ulf Hansson wrote:
> + Tomi
> 
> On Mon, 27 May 2024 at 14:41, Geert Uytterhoeven
> <geert+renesas@glider.be> wrote:
>>
>>          Hi all,
>>
>> Since commit a47cf07f60dcb02d ("serial: core: Call
>> device_set_awake_path() for console port"), the serial driver properly
>> handles the case where the serial console is part of the awake path, and
>> it looked like we could start removing special serial console handling
>> from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
>> Unfortunately the devil is in the details, as usual...
>>
>> Earlycon relies on the serial port to be initialized by the firmware
>> and/or bootloader.  Linux is not aware of any hardware dependencies that
>> must be met to keep the port working, and thus cannot guarantee they
>> stay met, until the full serial driver takes over.
>>
>> E.g. all unused clocks and unused PM Domains are disabled in a late
>> initcall.  As this happens after the full serial driver has taken over,
>> the serial port's clock and/or PM Domain are no longer deemed unused,
>> and this is typically not a problem.
>>
>> However, if the serial port's clock or PM Domain is shared with another
>> device, and that other device is runtime-suspended before the full
>> serial driver has probed, the serial port's clock and/or PM Domain will
>> be disabled inadvertently.  Any subsequent serial console output will
>> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
>> ports share their PM Domain with several other I/O devices.  After the
>> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
>> before the full serial driver takes over, the PM Domain containing the
>> early serial port is powered down, causing a lock-up when booted with
>> "earlycon".
> 
> Hi Geert,
> 
> Thanks for the detailed description of the problem! As pointed out in
> regards to another similar recent patch [1], this is indeed a generic
> problem, not limited to the serial console handling.
> 
> At Linaro Connect a few weeks ago I followed up with Saravana from the
> earlier discussions at LPC last fall. We now have a generic solution
> for genpd drafted on plain paper, based on fw_devlink and the
> ->sync_state() callback. I am currently working on the genpd series,
> while Saravana will re-spin the series (can't find the link to the
> last version) for the clock framework. Ideally, we want these things
> to work in a very similar way.
> 
> That said, allow me to post the series for genpd in a week or two to
> see if it can solve your problem too, for the serial console.

Both the genpd and the clock solutions will make suppliers depend on all 
their consumers to be probed, right?

I think it is a solution, and should be worked on, but it has the 
drawback that suppliers that have consumers that will possibly never be 
probed, will also never be able to turn off unused resources.

This was specifically the case with the TI ti-sci pmdomain case I was 
looking at: the genpd driver (ti_sci_pm_domains.c) provides a lot of 
genpds for totally unrelated devices, and so if, e.g., you don't have or 
don't want to load a driver for the GPU, all PDs are affected.

Even here the solutions you mention will help: instead of things getting 
broken because genpds get turned off while they are actually in use, the 
genpds will be kept enabled, thus fixing the breakage. Unfortunately, 
they'll be kept enabled forever.

I've been ill for quite a while so I haven't had the chance to look at 
this more, but before that I was hacking around a bit with something I 
named .partial_sync_state(). .sync_state() gets called when all the 
consumers have probed, but .partial_sync_state() gets called when _a_ 
consumer has been probed.

For the .sync_state() things are easy for the driver, as it knows 
everything related has been probed, but for .partial_sync_state() the 
driver needs to track resources internally. .partial_sync_state() will 
tell the driver that a consumer device has probed, the driver can then 
find out which specific resources (genpds in my case) that consumer 
refers to, and then... Well, that's how far I got with my hacks =).

So, I don't know if this .partial_sync_state() can even work, but I 
think we do need something more on top of the .sync_state().

  Tomi

> 
> Kind regards
> Uffe
> 
> [1]
> https://lore.kernel.org/linux-arm-kernel/CAPDyKFqShuq98qV5nSPzSqwLLUZ7LxLvp1eihGRBkU4qUKdWwQ@mail.gmail.com/
> 
>>
>> This RFC patch series aims to provide a mechanism for handling this, and
>> to fix it for the PM Domain case:
>>    1. The first patch provides a mechanism to let the clock and/or PM
>>       Domain subsystem or drivers handle this, by exporting the clock and
>>       PM Domain dependencies for the serial port, as available in the
>>       system's device tree,
>>    2. The second patch introduces a new flag to handle a PM domain that
>>       must be kept powered-on during early boot, and by setting this flag
>>       if the PM Domain contains the serial console (originally I handled
>>       this inside rmobile-sysc, but it turned out to be easy to
>>       generalize this to other platforms in the core PM Domain code).
>>    3. The third patch removes the no longer needed special console
>>       handling from the R-Mobile SYSC PM Domain driver.
>>
>> I did not fix the similar clock issue, as it is more complex (there can
>> be multiple clocks, and each clock provider can have its own value of
>> #clock-cells), and I do not need it for Renesas ARM platforms.
> 
> I will defer to Sarvana here, but ideally his series for the clock
> framework should solve this case too.
> 
>>
>> This has been tested on the APE6-EVM, Armadillo-800-EVA, and KZM-A9-GT
>> development boards, with and without earlycon, including s2ram with and
>> without no_console_suspend.
>>
>> Notes:
>>    - This should not be needed on RZ/G3S, where each serial port device
>>      has its own PM Domain,
>>    - drivers/clk/imx/clk.c and drivers/pmdomain/imx/scu-pd.c have special
>>      handling for the of_stdout device, but is probably not affected, as
>>      each serial port seems to share its PM Domain only with the serial
>>      port's clock controller.
>>
>> Thanks for your comments!
>>
>> Geert Uytterhoeven (3):
>>    earlycon: Export clock and PM Domain info from FDT
>>    pmdomain: core: Avoid earlycon power-down
>>    pmdomain: renesas: rmobile-sysc: Remove serial console handling
>>
>>   drivers/pmdomain/core.c                 | 24 ++++++++++++++++--
>>   drivers/pmdomain/renesas/rmobile-sysc.c | 33 +------------------------
>>   drivers/tty/serial/earlycon.c           | 14 ++++++++++-
>>   include/linux/pm_domain.h               |  4 +++
>>   include/linux/serial_core.h             | 10 ++++++++
>>   5 files changed, 50 insertions(+), 35 deletions(-)
>>
>> --
> 
> Kind regards
> Uffe
Ulf Hansson June 5, 2024, 10:53 a.m. UTC | #3
On Wed, 5 Jun 2024 at 12:41, Tomi Valkeinen
<tomi.valkeinen@ideasonboard.com> wrote:
>
> Hi Ulf,
>
> On 05/06/2024 12:34, Ulf Hansson wrote:
> > + Tomi
> >
> > On Mon, 27 May 2024 at 14:41, Geert Uytterhoeven
> > <geert+renesas@glider.be> wrote:
> >>
> >>          Hi all,
> >>
> >> Since commit a47cf07f60dcb02d ("serial: core: Call
> >> device_set_awake_path() for console port"), the serial driver properly
> >> handles the case where the serial console is part of the awake path, and
> >> it looked like we could start removing special serial console handling
> >> from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
> >> Unfortunately the devil is in the details, as usual...
> >>
> >> Earlycon relies on the serial port to be initialized by the firmware
> >> and/or bootloader.  Linux is not aware of any hardware dependencies that
> >> must be met to keep the port working, and thus cannot guarantee they
> >> stay met, until the full serial driver takes over.
> >>
> >> E.g. all unused clocks and unused PM Domains are disabled in a late
> >> initcall.  As this happens after the full serial driver has taken over,
> >> the serial port's clock and/or PM Domain are no longer deemed unused,
> >> and this is typically not a problem.
> >>
> >> However, if the serial port's clock or PM Domain is shared with another
> >> device, and that other device is runtime-suspended before the full
> >> serial driver has probed, the serial port's clock and/or PM Domain will
> >> be disabled inadvertently.  Any subsequent serial console output will
> >> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
> >> ports share their PM Domain with several other I/O devices.  After the
> >> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
> >> before the full serial driver takes over, the PM Domain containing the
> >> early serial port is powered down, causing a lock-up when booted with
> >> "earlycon".
> >
> > Hi Geert,
> >
> > Thanks for the detailed description of the problem! As pointed out in
> > regards to another similar recent patch [1], this is indeed a generic
> > problem, not limited to the serial console handling.
> >
> > At Linaro Connect a few weeks ago I followed up with Saravana from the
> > earlier discussions at LPC last fall. We now have a generic solution
> > for genpd drafted on plain paper, based on fw_devlink and the
> > ->sync_state() callback. I am currently working on the genpd series,
> > while Saravana will re-spin the series (can't find the link to the
> > last version) for the clock framework. Ideally, we want these things
> > to work in a very similar way.
> >
> > That said, allow me to post the series for genpd in a week or two to
> > see if it can solve your problem too, for the serial console.
>
> Both the genpd and the clock solutions will make suppliers depend on all
> their consumers to be probed, right?
>
> I think it is a solution, and should be worked on, but it has the
> drawback that suppliers that have consumers that will possibly never be
> probed, will also never be able to turn off unused resources.
>
> This was specifically the case with the TI ti-sci pmdomain case I was
> looking at: the genpd driver (ti_sci_pm_domains.c) provides a lot of
> genpds for totally unrelated devices, and so if, e.g., you don't have or
> don't want to load a driver for the GPU, all PDs are affected.
>
> Even here the solutions you mention will help: instead of things getting
> broken because genpds get turned off while they are actually in use, the
> genpds will be kept enabled, thus fixing the breakage. Unfortunately,
> they'll be kept enabled forever.
>
> I've been ill for quite a while so I haven't had the chance to look at
> this more, but before that I was hacking around a bit with something I
> named .partial_sync_state(). .sync_state() gets called when all the
> consumers have probed, but .partial_sync_state() gets called when _a_
> consumer has been probed.
>
> For the .sync_state() things are easy for the driver, as it knows
> everything related has been probed, but for .partial_sync_state() the
> driver needs to track resources internally. .partial_sync_state() will
> tell the driver that a consumer device has probed, the driver can then
> find out which specific resources (genpds in my case) that consumer
> refers to, and then... Well, that's how far I got with my hacks =).
>
> So, I don't know if this .partial_sync_state() can even work, but I
> think we do need something more on top of the .sync_state().

Thanks for the update!

You certainly have a point, but rather than implementing some platform
specific method, I think we should be able enforce the call to
->sync_state(), based upon some condition/timeout - and even if all
consumers haven't been probed.

[...]

Kind regards
Uffe
Tomi Valkeinen June 5, 2024, 11:16 a.m. UTC | #4
On 05/06/2024 13:53, Ulf Hansson wrote:
> On Wed, 5 Jun 2024 at 12:41, Tomi Valkeinen
> <tomi.valkeinen@ideasonboard.com> wrote:
>>
>> Hi Ulf,
>>
>> On 05/06/2024 12:34, Ulf Hansson wrote:
>>> + Tomi
>>>
>>> On Mon, 27 May 2024 at 14:41, Geert Uytterhoeven
>>> <geert+renesas@glider.be> wrote:
>>>>
>>>>           Hi all,
>>>>
>>>> Since commit a47cf07f60dcb02d ("serial: core: Call
>>>> device_set_awake_path() for console port"), the serial driver properly
>>>> handles the case where the serial console is part of the awake path, and
>>>> it looked like we could start removing special serial console handling
>>>> from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
>>>> Unfortunately the devil is in the details, as usual...
>>>>
>>>> Earlycon relies on the serial port to be initialized by the firmware
>>>> and/or bootloader.  Linux is not aware of any hardware dependencies that
>>>> must be met to keep the port working, and thus cannot guarantee they
>>>> stay met, until the full serial driver takes over.
>>>>
>>>> E.g. all unused clocks and unused PM Domains are disabled in a late
>>>> initcall.  As this happens after the full serial driver has taken over,
>>>> the serial port's clock and/or PM Domain are no longer deemed unused,
>>>> and this is typically not a problem.
>>>>
>>>> However, if the serial port's clock or PM Domain is shared with another
>>>> device, and that other device is runtime-suspended before the full
>>>> serial driver has probed, the serial port's clock and/or PM Domain will
>>>> be disabled inadvertently.  Any subsequent serial console output will
>>>> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
>>>> ports share their PM Domain with several other I/O devices.  After the
>>>> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
>>>> before the full serial driver takes over, the PM Domain containing the
>>>> early serial port is powered down, causing a lock-up when booted with
>>>> "earlycon".
>>>
>>> Hi Geert,
>>>
>>> Thanks for the detailed description of the problem! As pointed out in
>>> regards to another similar recent patch [1], this is indeed a generic
>>> problem, not limited to the serial console handling.
>>>
>>> At Linaro Connect a few weeks ago I followed up with Saravana from the
>>> earlier discussions at LPC last fall. We now have a generic solution
>>> for genpd drafted on plain paper, based on fw_devlink and the
>>> ->sync_state() callback. I am currently working on the genpd series,
>>> while Saravana will re-spin the series (can't find the link to the
>>> last version) for the clock framework. Ideally, we want these things
>>> to work in a very similar way.
>>>
>>> That said, allow me to post the series for genpd in a week or two to
>>> see if it can solve your problem too, for the serial console.
>>
>> Both the genpd and the clock solutions will make suppliers depend on all
>> their consumers to be probed, right?
>>
>> I think it is a solution, and should be worked on, but it has the
>> drawback that suppliers that have consumers that will possibly never be
>> probed, will also never be able to turn off unused resources.
>>
>> This was specifically the case with the TI ti-sci pmdomain case I was
>> looking at: the genpd driver (ti_sci_pm_domains.c) provides a lot of
>> genpds for totally unrelated devices, and so if, e.g., you don't have or
>> don't want to load a driver for the GPU, all PDs are affected.
>>
>> Even here the solutions you mention will help: instead of things getting
>> broken because genpds get turned off while they are actually in use, the
>> genpds will be kept enabled, thus fixing the breakage. Unfortunately,
>> they'll be kept enabled forever.
>>
>> I've been ill for quite a while so I haven't had the chance to look at
>> this more, but before that I was hacking around a bit with something I
>> named .partial_sync_state(). .sync_state() gets called when all the
>> consumers have probed, but .partial_sync_state() gets called when _a_
>> consumer has been probed.
>>
>> For the .sync_state() things are easy for the driver, as it knows
>> everything related has been probed, but for .partial_sync_state() the
>> driver needs to track resources internally. .partial_sync_state() will
>> tell the driver that a consumer device has probed, the driver can then
>> find out which specific resources (genpds in my case) that consumer
>> refers to, and then... Well, that's how far I got with my hacks =).
>>
>> So, I don't know if this .partial_sync_state() can even work, but I
>> think we do need something more on top of the .sync_state().
> 
> Thanks for the update!
> 
> You certainly have a point, but rather than implementing some platform
> specific method, I think we should be able enforce the call to
> ->sync_state(), based upon some condition/timeout - and even if all
> consumers haven't been probed.

Hmm, I think that was already implemented in some of the serieses out 
there (or even in mainline already?), as I remember doing some 
experiments with it. I don't like it much, though.

With a simple timeout, it'll always be just a bit too early for some 
user (nfs mount took a bit more time than expected -> board frozen).

The only condition I can see that would somewhat work is a manual 
trigger from the userspace. The boot scripts could then signal the 
kernel when all the modules have been loaded and probably a suitable, 
platform/use case specific amount of time has passed to allow the 
drivers to probe.

It just feels a bit too much of a "let's hope this work" approach.

That said, the timeout/condition is probably acceptable for many cases, 
where turning off a resource forcefully will just result in, say, a 
temporarily blanked display, or something else that gets fixed if and 
when the proper driver is probed.

Unfortunately, here with the case I have, the whole board gets halted if 
the display subsystem genpd is turned off and the display driver is 
loaded after that.

  Tomi
Claudiu Beznea June 11, 2024, 4:52 a.m. UTC | #5
Hi, Geert,

On 27.05.2024 15:41, Geert Uytterhoeven wrote:
> 	Hi all,
> 
> Since commit a47cf07f60dcb02d ("serial: core: Call
> device_set_awake_path() for console port"), the serial driver properly
> handles the case where the serial console is part of the awake path, and
> it looked like we could start removing special serial console handling
> from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
> Unfortunately the devil is in the details, as usual...
> 
> Earlycon relies on the serial port to be initialized by the firmware
> and/or bootloader.  Linux is not aware of any hardware dependencies that
> must be met to keep the port working, and thus cannot guarantee they
> stay met, until the full serial driver takes over.
> 
> E.g. all unused clocks and unused PM Domains are disabled in a late
> initcall.  As this happens after the full serial driver has taken over,
> the serial port's clock and/or PM Domain are no longer deemed unused,
> and this is typically not a problem.
> 
> However, if the serial port's clock or PM Domain is shared with another
> device, and that other device is runtime-suspended before the full
> serial driver has probed, the serial port's clock and/or PM Domain will
> be disabled inadvertently.  Any subsequent serial console output will
> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
> ports share their PM Domain with several other I/O devices.  After the
> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
> before the full serial driver takes over, the PM Domain containing the
> early serial port is powered down, causing a lock-up when booted with
> "earlycon".
> 
> This RFC patch series aims to provide a mechanism for handling this, and
> to fix it for the PM Domain case:
>   1. The first patch provides a mechanism to let the clock and/or PM
>      Domain subsystem or drivers handle this, by exporting the clock and
>      PM Domain dependencies for the serial port, as available in the
>      system's device tree,
>   2. The second patch introduces a new flag to handle a PM domain that
>      must be kept powered-on during early boot, and by setting this flag
>      if the PM Domain contains the serial console (originally I handled
>      this inside rmobile-sysc, but it turned out to be easy to
>      generalize this to other platforms in the core PM Domain code).
>   3. The third patch removes the no longer needed special console
>      handling from the R-Mobile SYSC PM Domain driver.
> 
> I did not fix the similar clock issue, as it is more complex (there can
> be multiple clocks, and each clock provider can have its own value of
> #clock-cells), and I do not need it for Renesas ARM platforms.
> 
> This has been tested on the APE6-EVM, Armadillo-800-EVA, and KZM-A9-GT
> development boards, with and without earlycon, including s2ram with and
> without no_console_suspend.
> 
> Notes:
>   - This should not be needed on RZ/G3S, where each serial port device
>     has its own PM Domain,

For the record, I've tested this series on RZ/G3S. All good with it.
If any, you can add my:

Tested-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

Thank you,
Claudiu Beznea

>   - drivers/clk/imx/clk.c and drivers/pmdomain/imx/scu-pd.c have special
>     handling for the of_stdout device, but is probably not affected, as
>     each serial port seems to share its PM Domain only with the serial
>     port's clock controller.
> 
> Thanks for your comments!
> 
> Geert Uytterhoeven (3):
>   earlycon: Export clock and PM Domain info from FDT
>   pmdomain: core: Avoid earlycon power-down
>   pmdomain: renesas: rmobile-sysc: Remove serial console handling
> 
>  drivers/pmdomain/core.c                 | 24 ++++++++++++++++--
>  drivers/pmdomain/renesas/rmobile-sysc.c | 33 +------------------------
>  drivers/tty/serial/earlycon.c           | 14 ++++++++++-
>  include/linux/pm_domain.h               |  4 +++
>  include/linux/serial_core.h             | 10 ++++++++
>  5 files changed, 50 insertions(+), 35 deletions(-)
>
Saravana Kannan June 21, 2024, 1:07 a.m. UTC | #6
On Wed, Jun 5, 2024 at 4:16 AM Tomi Valkeinen
<tomi.valkeinen@ideasonboard.com> wrote:
>
> On 05/06/2024 13:53, Ulf Hansson wrote:
> > On Wed, 5 Jun 2024 at 12:41, Tomi Valkeinen
> > <tomi.valkeinen@ideasonboard.com> wrote:
> >>
> >> Hi Ulf,
> >>
> >> On 05/06/2024 12:34, Ulf Hansson wrote:
> >>> + Tomi
> >>>
> >>> On Mon, 27 May 2024 at 14:41, Geert Uytterhoeven
> >>> <geert+renesas@glider.be> wrote:
> >>>>
> >>>>           Hi all,
> >>>>
> >>>> Since commit a47cf07f60dcb02d ("serial: core: Call
> >>>> device_set_awake_path() for console port"), the serial driver properly
> >>>> handles the case where the serial console is part of the awake path, and
> >>>> it looked like we could start removing special serial console handling
> >>>> from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
> >>>> Unfortunately the devil is in the details, as usual...
> >>>>
> >>>> Earlycon relies on the serial port to be initialized by the firmware
> >>>> and/or bootloader.  Linux is not aware of any hardware dependencies that
> >>>> must be met to keep the port working, and thus cannot guarantee they
> >>>> stay met, until the full serial driver takes over.
> >>>>
> >>>> E.g. all unused clocks and unused PM Domains are disabled in a late
> >>>> initcall.  As this happens after the full serial driver has taken over,
> >>>> the serial port's clock and/or PM Domain are no longer deemed unused,
> >>>> and this is typically not a problem.
> >>>>
> >>>> However, if the serial port's clock or PM Domain is shared with another
> >>>> device, and that other device is runtime-suspended before the full
> >>>> serial driver has probed, the serial port's clock and/or PM Domain will
> >>>> be disabled inadvertently.  Any subsequent serial console output will
> >>>> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
> >>>> ports share their PM Domain with several other I/O devices.  After the
> >>>> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
> >>>> before the full serial driver takes over, the PM Domain containing the
> >>>> early serial port is powered down, causing a lock-up when booted with
> >>>> "earlycon".
> >>>
> >>> Hi Geert,
> >>>
> >>> Thanks for the detailed description of the problem! As pointed out in
> >>> regards to another similar recent patch [1], this is indeed a generic
> >>> problem, not limited to the serial console handling.
> >>>
> >>> At Linaro Connect a few weeks ago I followed up with Saravana from the
> >>> earlier discussions at LPC last fall. We now have a generic solution
> >>> for genpd drafted on plain paper, based on fw_devlink and the
> >>> ->sync_state() callback. I am currently working on the genpd series,
> >>> while Saravana will re-spin the series (can't find the link to the
> >>> last version) for the clock framework. Ideally, we want these things
> >>> to work in a very similar way.
> >>>
> >>> That said, allow me to post the series for genpd in a week or two to
> >>> see if it can solve your problem too, for the serial console.
> >>
> >> Both the genpd and the clock solutions will make suppliers depend on all
> >> their consumers to be probed, right?
> >>
> >> I think it is a solution, and should be worked on, but it has the
> >> drawback that suppliers that have consumers that will possibly never be
> >> probed, will also never be able to turn off unused resources.
> >>
> >> This was specifically the case with the TI ti-sci pmdomain case I was
> >> looking at: the genpd driver (ti_sci_pm_domains.c) provides a lot of
> >> genpds for totally unrelated devices, and so if, e.g., you don't have or
> >> don't want to load a driver for the GPU, all PDs are affected.
> >>
> >> Even here the solutions you mention will help: instead of things getting
> >> broken because genpds get turned off while they are actually in use, the
> >> genpds will be kept enabled, thus fixing the breakage. Unfortunately,
> >> they'll be kept enabled forever.
> >>
> >> I've been ill for quite a while so I haven't had the chance to look at
> >> this more, but before that I was hacking around a bit with something I
> >> named .partial_sync_state(). .sync_state() gets called when all the
> >> consumers have probed, but .partial_sync_state() gets called when _a_
> >> consumer has been probed.
> >>
> >> For the .sync_state() things are easy for the driver, as it knows
> >> everything related has been probed, but for .partial_sync_state() the
> >> driver needs to track resources internally. .partial_sync_state() will
> >> tell the driver that a consumer device has probed, the driver can then
> >> find out which specific resources (genpds in my case) that consumer
> >> refers to, and then... Well, that's how far I got with my hacks =).
> >>
> >> So, I don't know if this .partial_sync_state() can even work, but I
> >> think we do need something more on top of the .sync_state().
> >
> > Thanks for the update!
> >
> > You certainly have a point, but rather than implementing some platform
> > specific method, I think we should be able enforce the call to
> > ->sync_state(), based upon some condition/timeout - and even if all
> > consumers haven't been probed.
>
> Hmm, I think that was already implemented in some of the serieses out
> there (or even in mainline already?), as I remember doing some
> experiments with it. I don't like it much, though.
>
> With a simple timeout, it'll always be just a bit too early for some
> user (nfs mount took a bit more time than expected -> board frozen).
>
> The only condition I can see that would somewhat work is a manual
> trigger from the userspace. The boot scripts could then signal the
> kernel when all the modules have been loaded and probably a suitable,
> platform/use case specific amount of time has passed to allow the
> drivers to probe.

This is also already supported in mainline.

Devices with sync_state() implementations (once Ulf adds it) will have
a state_synced file in sysfs. It shows where it has been called yet or
not. But you can also echo 1 into it to force the sync_state()
callback (only if it hasn't been called already). So, yeah, all
methods of handling this are available if you implement the
sync_state() callback.

By default it's all strict (wait till all consumers probe
successfully). But you can set it to timeout (fw_devlink.sync_state).
And you also have the option I mentioned above that you can use with
both cases.

-Saravana

>
> It just feels a bit too much of a "let's hope this work" approach.
>
> That said, the timeout/condition is probably acceptable for many cases,
> where turning off a resource forcefully will just result in, say, a
> temporarily blanked display, or something else that gets fixed if and
> when the proper driver is probed.
>
> Unfortunately, here with the case I have, the whole board gets halted if
> the display subsystem genpd is turned off and the display driver is
> loaded after that.
>
>   Tomi
>
Geert Uytterhoeven June 21, 2024, 7:07 a.m. UTC | #7
Hi Saravana,

On Fri, Jun 21, 2024 at 3:08 AM Saravana Kannan <saravanak@google.com> wrote:
> On Wed, Jun 5, 2024 at 4:16 AM Tomi Valkeinen
> <tomi.valkeinen@ideasonboard.com> wrote:
> > On 05/06/2024 13:53, Ulf Hansson wrote:
> > > On Wed, 5 Jun 2024 at 12:41, Tomi Valkeinen
> > > <tomi.valkeinen@ideasonboard.com> wrote:
> > >> On 05/06/2024 12:34, Ulf Hansson wrote:
> > >>> On Mon, 27 May 2024 at 14:41, Geert Uytterhoeven
> > >>> <geert+renesas@glider.be> wrote:
> > >>>> Since commit a47cf07f60dcb02d ("serial: core: Call
> > >>>> device_set_awake_path() for console port"), the serial driver properly
> > >>>> handles the case where the serial console is part of the awake path, and
> > >>>> it looked like we could start removing special serial console handling
> > >>>> from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
> > >>>> Unfortunately the devil is in the details, as usual...
> > >>>>
> > >>>> Earlycon relies on the serial port to be initialized by the firmware
> > >>>> and/or bootloader.  Linux is not aware of any hardware dependencies that
> > >>>> must be met to keep the port working, and thus cannot guarantee they
> > >>>> stay met, until the full serial driver takes over.
> > >>>>
> > >>>> E.g. all unused clocks and unused PM Domains are disabled in a late
> > >>>> initcall.  As this happens after the full serial driver has taken over,
> > >>>> the serial port's clock and/or PM Domain are no longer deemed unused,
> > >>>> and this is typically not a problem.

Let's call this "Case A".

> > >>>>
> > >>>> However, if the serial port's clock or PM Domain is shared with another
> > >>>> device, and that other device is runtime-suspended before the full
> > >>>> serial driver has probed, the serial port's clock and/or PM Domain will
> > >>>> be disabled inadvertently.  Any subsequent serial console output will
> > >>>> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
> > >>>> ports share their PM Domain with several other I/O devices.  After the
> > >>>> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
> > >>>> before the full serial driver takes over, the PM Domain containing the
> > >>>> early serial port is powered down, causing a lock-up when booted with
> > >>>> "earlycon".

Let's call this "Case B".

> > >>>
> > >>> Thanks for the detailed description of the problem! As pointed out in
> > >>> regards to another similar recent patch [1], this is indeed a generic
> > >>> problem, not limited to the serial console handling.
> > >>>
> > >>> At Linaro Connect a few weeks ago I followed up with Saravana from the
> > >>> earlier discussions at LPC last fall. We now have a generic solution
> > >>> for genpd drafted on plain paper, based on fw_devlink and the
> > >>> ->sync_state() callback. I am currently working on the genpd series,
> > >>> while Saravana will re-spin the series (can't find the link to the
> > >>> last version) for the clock framework. Ideally, we want these things
> > >>> to work in a very similar way.
> > >>>
> > >>> That said, allow me to post the series for genpd in a week or two to
> > >>> see if it can solve your problem too, for the serial console.
> > >>
> > >> Both the genpd and the clock solutions will make suppliers depend on all
> > >> their consumers to be probed, right?
> > >>
> > >> I think it is a solution, and should be worked on, but it has the
> > >> drawback that suppliers that have consumers that will possibly never be
> > >> probed, will also never be able to turn off unused resources.
> > >>
> > >> This was specifically the case with the TI ti-sci pmdomain case I was
> > >> looking at: the genpd driver (ti_sci_pm_domains.c) provides a lot of
> > >> genpds for totally unrelated devices, and so if, e.g., you don't have or
> > >> don't want to load a driver for the GPU, all PDs are affected.
> > >>
> > >> Even here the solutions you mention will help: instead of things getting
> > >> broken because genpds get turned off while they are actually in use, the
> > >> genpds will be kept enabled, thus fixing the breakage. Unfortunately,
> > >> they'll be kept enabled forever.
> > >>
> > >> I've been ill for quite a while so I haven't had the chance to look at
> > >> this more, but before that I was hacking around a bit with something I
> > >> named .partial_sync_state(). .sync_state() gets called when all the
> > >> consumers have probed, but .partial_sync_state() gets called when _a_
> > >> consumer has been probed.
> > >>
> > >> For the .sync_state() things are easy for the driver, as it knows
> > >> everything related has been probed, but for .partial_sync_state() the
> > >> driver needs to track resources internally. .partial_sync_state() will
> > >> tell the driver that a consumer device has probed, the driver can then
> > >> find out which specific resources (genpds in my case) that consumer
> > >> refers to, and then... Well, that's how far I got with my hacks =).
> > >>
> > >> So, I don't know if this .partial_sync_state() can even work, but I
> > >> think we do need something more on top of the .sync_state().
> > >
> > > Thanks for the update!
> > >
> > > You certainly have a point, but rather than implementing some platform
> > > specific method, I think we should be able enforce the call to
> > > ->sync_state(), based upon some condition/timeout - and even if all
> > > consumers haven't been probed.
> >
> > Hmm, I think that was already implemented in some of the serieses out
> > there (or even in mainline already?), as I remember doing some
> > experiments with it. I don't like it much, though.
> >
> > With a simple timeout, it'll always be just a bit too early for some
> > user (nfs mount took a bit more time than expected -> board frozen).
> >
> > The only condition I can see that would somewhat work is a manual
> > trigger from the userspace. The boot scripts could then signal the
> > kernel when all the modules have been loaded and probably a suitable,
> > platform/use case specific amount of time has passed to allow the
> > drivers to probe.
>
> This is also already supported in mainline.
>
> Devices with sync_state() implementations (once Ulf adds it) will have
> a state_synced file in sysfs. It shows where it has been called yet or
> not. But you can also echo 1 into it to force the sync_state()
> callback (only if it hasn't been called already). So, yeah, all
> methods of handling this are available if you implement the
> sync_state() callback.
>
> By default it's all strict (wait till all consumers probe
> successfully). But you can set it to timeout (fw_devlink.sync_state).
> And you also have the option I mentioned above that you can use with
> both cases.

So the idea is to disable unused genpds and clocks from the genpd
resp. clock's driver .sync_state() callback, instead of from a late
initcall?  That would indeed solve issues related to "Case A".

However, how to solve "Case B"? Ignore disabling genpds or clocks
before .sync_state() callback() has been called?
That would cause issues for cases where the clock must be disabled,
cfr.
    "[PATCH RFC 0/3] Add clk_disable_unprepare_sync()"
    https://lore.kernel.org/all/20240131160947.96171-1-biju.das.jz@bp.renesas.com/
    "[PATCH v3 0/3] Add clk_poll_disable_unprepare()"
    https://lore.kernel.org/linux-renesas-soc/20240318110842.41956-1-biju.das.jz@bp.renesas.com/

> > It just feels a bit too much of a "let's hope this work" approach.
> >
> > That said, the timeout/condition is probably acceptable for many cases,
> > where turning off a resource forcefully will just result in, say, a
> > temporarily blanked display, or something else that gets fixed if and
> > when the proper driver is probed.
> >
> > Unfortunately, here with the case I have, the whole board gets halted if
> > the display subsystem genpd is turned off and the display driver is
> > loaded after that.

Tomi: Do you have more details? The genpd must be controlling something
critical that must never be turned off, or perhaps the display driver
lacks some initialization?

Thanks!

Gr{oetje,eeting}s,

                        Geert
Tomi Valkeinen June 21, 2024, 8:49 a.m. UTC | #8
On 21/06/2024 10:07, Geert Uytterhoeven wrote:
> Hi Saravana,
> 
> On Fri, Jun 21, 2024 at 3:08 AM Saravana Kannan <saravanak@google.com> wrote:
>> On Wed, Jun 5, 2024 at 4:16 AM Tomi Valkeinen
>> <tomi.valkeinen@ideasonboard.com> wrote:
>>> On 05/06/2024 13:53, Ulf Hansson wrote:
>>>> On Wed, 5 Jun 2024 at 12:41, Tomi Valkeinen
>>>> <tomi.valkeinen@ideasonboard.com> wrote:
>>>>> On 05/06/2024 12:34, Ulf Hansson wrote:
>>>>>> On Mon, 27 May 2024 at 14:41, Geert Uytterhoeven
>>>>>> <geert+renesas@glider.be> wrote:
>>>>>>> Since commit a47cf07f60dcb02d ("serial: core: Call
>>>>>>> device_set_awake_path() for console port"), the serial driver properly
>>>>>>> handles the case where the serial console is part of the awake path, and
>>>>>>> it looked like we could start removing special serial console handling
>>>>>>> from PM Domain drivers like the R-Mobile SYSC PM Domain driver.
>>>>>>> Unfortunately the devil is in the details, as usual...
>>>>>>>
>>>>>>> Earlycon relies on the serial port to be initialized by the firmware
>>>>>>> and/or bootloader.  Linux is not aware of any hardware dependencies that
>>>>>>> must be met to keep the port working, and thus cannot guarantee they
>>>>>>> stay met, until the full serial driver takes over.
>>>>>>>
>>>>>>> E.g. all unused clocks and unused PM Domains are disabled in a late
>>>>>>> initcall.  As this happens after the full serial driver has taken over,
>>>>>>> the serial port's clock and/or PM Domain are no longer deemed unused,
>>>>>>> and this is typically not a problem.
> 
> Let's call this "Case A".
> 
>>>>>>>
>>>>>>> However, if the serial port's clock or PM Domain is shared with another
>>>>>>> device, and that other device is runtime-suspended before the full
>>>>>>> serial driver has probed, the serial port's clock and/or PM Domain will
>>>>>>> be disabled inadvertently.  Any subsequent serial console output will
>>>>>>> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
>>>>>>> ports share their PM Domain with several other I/O devices.  After the
>>>>>>> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
>>>>>>> before the full serial driver takes over, the PM Domain containing the
>>>>>>> early serial port is powered down, causing a lock-up when booted with
>>>>>>> "earlycon".
> 
> Let's call this "Case B".
> 
>>>>>>
>>>>>> Thanks for the detailed description of the problem! As pointed out in
>>>>>> regards to another similar recent patch [1], this is indeed a generic
>>>>>> problem, not limited to the serial console handling.
>>>>>>
>>>>>> At Linaro Connect a few weeks ago I followed up with Saravana from the
>>>>>> earlier discussions at LPC last fall. We now have a generic solution
>>>>>> for genpd drafted on plain paper, based on fw_devlink and the
>>>>>> ->sync_state() callback. I am currently working on the genpd series,
>>>>>> while Saravana will re-spin the series (can't find the link to the
>>>>>> last version) for the clock framework. Ideally, we want these things
>>>>>> to work in a very similar way.
>>>>>>
>>>>>> That said, allow me to post the series for genpd in a week or two to
>>>>>> see if it can solve your problem too, for the serial console.
>>>>>
>>>>> Both the genpd and the clock solutions will make suppliers depend on all
>>>>> their consumers to be probed, right?
>>>>>
>>>>> I think it is a solution, and should be worked on, but it has the
>>>>> drawback that suppliers that have consumers that will possibly never be
>>>>> probed, will also never be able to turn off unused resources.
>>>>>
>>>>> This was specifically the case with the TI ti-sci pmdomain case I was
>>>>> looking at: the genpd driver (ti_sci_pm_domains.c) provides a lot of
>>>>> genpds for totally unrelated devices, and so if, e.g., you don't have or
>>>>> don't want to load a driver for the GPU, all PDs are affected.
>>>>>
>>>>> Even here the solutions you mention will help: instead of things getting
>>>>> broken because genpds get turned off while they are actually in use, the
>>>>> genpds will be kept enabled, thus fixing the breakage. Unfortunately,
>>>>> they'll be kept enabled forever.
>>>>>
>>>>> I've been ill for quite a while so I haven't had the chance to look at
>>>>> this more, but before that I was hacking around a bit with something I
>>>>> named .partial_sync_state(). .sync_state() gets called when all the
>>>>> consumers have probed, but .partial_sync_state() gets called when _a_
>>>>> consumer has been probed.
>>>>>
>>>>> For the .sync_state() things are easy for the driver, as it knows
>>>>> everything related has been probed, but for .partial_sync_state() the
>>>>> driver needs to track resources internally. .partial_sync_state() will
>>>>> tell the driver that a consumer device has probed, the driver can then
>>>>> find out which specific resources (genpds in my case) that consumer
>>>>> refers to, and then... Well, that's how far I got with my hacks =).
>>>>>
>>>>> So, I don't know if this .partial_sync_state() can even work, but I
>>>>> think we do need something more on top of the .sync_state().
>>>>
>>>> Thanks for the update!
>>>>
>>>> You certainly have a point, but rather than implementing some platform
>>>> specific method, I think we should be able enforce the call to
>>>> ->sync_state(), based upon some condition/timeout - and even if all
>>>> consumers haven't been probed.
>>>
>>> Hmm, I think that was already implemented in some of the serieses out
>>> there (or even in mainline already?), as I remember doing some
>>> experiments with it. I don't like it much, though.
>>>
>>> With a simple timeout, it'll always be just a bit too early for some
>>> user (nfs mount took a bit more time than expected -> board frozen).
>>>
>>> The only condition I can see that would somewhat work is a manual
>>> trigger from the userspace. The boot scripts could then signal the
>>> kernel when all the modules have been loaded and probably a suitable,
>>> platform/use case specific amount of time has passed to allow the
>>> drivers to probe.
>>
>> This is also already supported in mainline.
>>
>> Devices with sync_state() implementations (once Ulf adds it) will have
>> a state_synced file in sysfs. It shows where it has been called yet or
>> not. But you can also echo 1 into it to force the sync_state()
>> callback (only if it hasn't been called already). So, yeah, all
>> methods of handling this are available if you implement the
>> sync_state() callback.
>>
>> By default it's all strict (wait till all consumers probe
>> successfully). But you can set it to timeout (fw_devlink.sync_state).
>> And you also have the option I mentioned above that you can use with
>> both cases.
> 
> So the idea is to disable unused genpds and clocks from the genpd
> resp. clock's driver .sync_state() callback, instead of from a late
> initcall?  That would indeed solve issues related to "Case A".
> 
> However, how to solve "Case B"? Ignore disabling genpds or clocks
> before .sync_state() callback() has been called?
> That would cause issues for cases where the clock must be disabled,
> cfr.
>      "[PATCH RFC 0/3] Add clk_disable_unprepare_sync()"
>      https://lore.kernel.org/all/20240131160947.96171-1-biju.das.jz@bp.renesas.com/
>      "[PATCH v3 0/3] Add clk_poll_disable_unprepare()"
>      https://lore.kernel.org/linux-renesas-soc/20240318110842.41956-1-biju.das.jz@bp.renesas.com/
> 
>>> It just feels a bit too much of a "let's hope this work" approach.
>>>
>>> That said, the timeout/condition is probably acceptable for many cases,
>>> where turning off a resource forcefully will just result in, say, a
>>> temporarily blanked display, or something else that gets fixed if and
>>> when the proper driver is probed.
>>>
>>> Unfortunately, here with the case I have, the whole board gets halted if
>>> the display subsystem genpd is turned off and the display driver is
>>> loaded after that.
> 
> Tomi: Do you have more details? The genpd must be controlling something
> critical that must never be turned off, or perhaps the display driver
> lacks some initialization?

I don't know the exact HW level details. It may be a HW bug or possibly 
a firmware issue. But what I see is simple:

If the display subsystem is powered and a video output is enabled, 
turning off the PD causes the display subsystem to go to a bad state, 
after which the next register access will halt the cpu.

This happens quite easily if the bootloader has enabled a display: when 
the kernel's tidss driver probes, the device framework will make sure 
the PD is enabled (which is fine, it's basically a no-op). But if the 
tidss returns an error, like EPROBE_DEFER, the device framework will 
disable the PD. When the tidss driver probes again later, it will cause 
a halt at a register access.

  Tomi
Ulf Hansson July 9, 2024, 3:20 p.m. UTC | #9
[...]

> > > >>>>
> > > >>>> However, if the serial port's clock or PM Domain is shared with another
> > > >>>> device, and that other device is runtime-suspended before the full
> > > >>>> serial driver has probed, the serial port's clock and/or PM Domain will
> > > >>>> be disabled inadvertently.  Any subsequent serial console output will
> > > >>>> cause a crash or system lock-up.  E.g. on R/SH-Mobile SoCs, the serial
> > > >>>> ports share their PM Domain with several other I/O devices.  After the
> > > >>>> use of pwm (Armadillo-800-EVA) or i2c (KZM-A9-GT) during early boot,
> > > >>>> before the full serial driver takes over, the PM Domain containing the
> > > >>>> early serial port is powered down, causing a lock-up when booted with
> > > >>>> "earlycon".
>
> Let's call this "Case B".
>
> > > >>>
> > > >>> Thanks for the detailed description of the problem! As pointed out in
> > > >>> regards to another similar recent patch [1], this is indeed a generic
> > > >>> problem, not limited to the serial console handling.
> > > >>>
> > > >>> At Linaro Connect a few weeks ago I followed up with Saravana from the
> > > >>> earlier discussions at LPC last fall. We now have a generic solution
> > > >>> for genpd drafted on plain paper, based on fw_devlink and the
> > > >>> ->sync_state() callback. I am currently working on the genpd series,
> > > >>> while Saravana will re-spin the series (can't find the link to the
> > > >>> last version) for the clock framework. Ideally, we want these things
> > > >>> to work in a very similar way.
> > > >>>
> > > >>> That said, allow me to post the series for genpd in a week or two to
> > > >>> see if it can solve your problem too, for the serial console.

I managed to hit the vacation period before I was able to post the
series. I will pick it up this week and hopefully should be able to
post something next week.

> > > >>
> > > >> Both the genpd and the clock solutions will make suppliers depend on all
> > > >> their consumers to be probed, right?
> > > >>
> > > >> I think it is a solution, and should be worked on, but it has the
> > > >> drawback that suppliers that have consumers that will possibly never be
> > > >> probed, will also never be able to turn off unused resources.
> > > >>
> > > >> This was specifically the case with the TI ti-sci pmdomain case I was
> > > >> looking at: the genpd driver (ti_sci_pm_domains.c) provides a lot of
> > > >> genpds for totally unrelated devices, and so if, e.g., you don't have or
> > > >> don't want to load a driver for the GPU, all PDs are affected.
> > > >>
> > > >> Even here the solutions you mention will help: instead of things getting
> > > >> broken because genpds get turned off while they are actually in use, the
> > > >> genpds will be kept enabled, thus fixing the breakage. Unfortunately,
> > > >> they'll be kept enabled forever.
> > > >>
> > > >> I've been ill for quite a while so I haven't had the chance to look at
> > > >> this more, but before that I was hacking around a bit with something I
> > > >> named .partial_sync_state(). .sync_state() gets called when all the
> > > >> consumers have probed, but .partial_sync_state() gets called when _a_
> > > >> consumer has been probed.
> > > >>
> > > >> For the .sync_state() things are easy for the driver, as it knows
> > > >> everything related has been probed, but for .partial_sync_state() the
> > > >> driver needs to track resources internally. .partial_sync_state() will
> > > >> tell the driver that a consumer device has probed, the driver can then
> > > >> find out which specific resources (genpds in my case) that consumer
> > > >> refers to, and then... Well, that's how far I got with my hacks =).
> > > >>
> > > >> So, I don't know if this .partial_sync_state() can even work, but I
> > > >> think we do need something more on top of the .sync_state().
> > > >
> > > > Thanks for the update!
> > > >
> > > > You certainly have a point, but rather than implementing some platform
> > > > specific method, I think we should be able enforce the call to
> > > > ->sync_state(), based upon some condition/timeout - and even if all
> > > > consumers haven't been probed.
> > >
> > > Hmm, I think that was already implemented in some of the serieses out
> > > there (or even in mainline already?), as I remember doing some
> > > experiments with it. I don't like it much, though.
> > >
> > > With a simple timeout, it'll always be just a bit too early for some
> > > user (nfs mount took a bit more time than expected -> board frozen).
> > >
> > > The only condition I can see that would somewhat work is a manual
> > > trigger from the userspace. The boot scripts could then signal the
> > > kernel when all the modules have been loaded and probably a suitable,
> > > platform/use case specific amount of time has passed to allow the
> > > drivers to probe.
> >
> > This is also already supported in mainline.
> >
> > Devices with sync_state() implementations (once Ulf adds it) will have
> > a state_synced file in sysfs. It shows where it has been called yet or
> > not. But you can also echo 1 into it to force the sync_state()
> > callback (only if it hasn't been called already). So, yeah, all
> > methods of handling this are available if you implement the
> > sync_state() callback.
> >
> > By default it's all strict (wait till all consumers probe
> > successfully). But you can set it to timeout (fw_devlink.sync_state).
> > And you also have the option I mentioned above that you can use with
> > both cases.
>
> So the idea is to disable unused genpds and clocks from the genpd
> resp. clock's driver .sync_state() callback, instead of from a late
> initcall?  That would indeed solve issues related to "Case A".
>
> However, how to solve "Case B"? Ignore disabling genpds or clocks
> before .sync_state() callback() has been called?
> That would cause issues for cases where the clock must be disabled,
> cfr.
>     "[PATCH RFC 0/3] Add clk_disable_unprepare_sync()"
>     https://lore.kernel.org/all/20240131160947.96171-1-biju.das.jz@bp.renesas.com/
>     "[PATCH v3 0/3] Add clk_poll_disable_unprepare()"
>     https://lore.kernel.org/linux-renesas-soc/20240318110842.41956-1-biju.das.jz@bp.renesas.com/
>

For genpd, the plan is to check the initial state of the PM domain. It
can be powered-on or powered-off and if it's powered-on, we should not
allow it to be powered-off until after ->sync_state() have been
called.

The similar approach is what Saravanna is trying to implement for
clocks, I think.

In the end, we simply need to try out these approaches to see if they
really work. Although, based on previous discussions (LKML +
F2F-conferences), I think there should be a good chance for us.

[...]

Kind regards
Uffe