Re: [PATCHv2] i2c: omap: Use noirq system sleep pm ops to idle device for suspend

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Tony,

On Thu, Jan 10, 2019 at 07:59:16AM -0800, Tony Lindgren wrote:
> We currently get the following error with pixcir_ts driver during a
> suspend resume cycle:
> 
> omap_i2c 4802a000.i2c: controller timed out
> pixcir_ts 1-005c: pixcir_int_enable: can't read reg 0x34 : -110
> pixcir_ts 1-005c: Failed to disable interrupt generation: -110
> pixcir_ts 1-005c: Failed to stop
> dpm_run_callback(): pixcir_i2c_ts_resume+0x0/0x98
> [pixcir_i2c_ts] returns -110
> PM: Device 1-005c failed to resume: error -110
> 
> And at least am437x based devices with pixcir_ts will fail to resume
> to a touchscreen that is configured as the wakeup-source in device
> tree for these devices.
> 
> This is because pixcir_ts tries to reconfigure it's registers for
> noirq suspend which fails. This also leaves i2c-omap in enabled state
> for suspend.

I didn't fully get this. Does it reconfigure the registers for noirq or
during noirq? If the former, why exactly does it fail?

> Let's fix the pixcir_ts issue and make sure i2c-omap is suspended by
> adding SET_NOIRQ_SYSTEM_SLEEP_PM_OPS. In the long run the solution
> seems to be to move the handling of suspended adapters to the i2c
> core, but that still needs some more work.

I don't know if you followed this, but we failed with a generic handling
within the core. We have helpers now which drivers can apply
individually.

> Let's also get rid of some ifdefs while at it and replace them with
> __maybe_unused as SET_RUNTIME_PM_OPS and SET_NOIRQ_SYSTEM_SLEEP_PM_OPS
> already deal with the various PM Kconfig options.
> 
> Cc: Dave Gerlach <d-gerlach@xxxxxx>
> Cc: Grygorii Strashko <grygorii.strashko@xxxxxx>
> Cc: Keerthy <j-keerthy@xxxxxx>
> Cc: Tero Kristo <t-kristo@xxxxxx>
> Reported-by: Keerthy <j-keerthy@xxxxxx>
> Signed-off-by: Tony Lindgren <tony@xxxxxxxxxxx>
> ---
> 
> Wolfram, this is still pending, can you please apply this a fix for
> the -rc series if no objections?

Yes, I can do that (when I have no further questions ;))

Regards,

   Wolfram

Attachment: signature.asc
Description: PGP signature


[Index of Archives]     [Linux Arm (vger)]     [ARM Kernel]     [ARM MSM]     [Linux Tegra]     [Linux WPAN Networking]     [Linux Wireless Networking]     [Maemo Users]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux