Re: [PATCH v3 1/5] PM: sleep: Resume children after resuming the parent

[Tudor Ambarus <tudor.ambarus@linaro.org>]

On 7/14/25 11:35 AM, Tudor Ambarus wrote:
> 
> 
> On 7/14/25 8:29 AM, Rafael J. Wysocki wrote:
>>> diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c
>>> index d9d4fc58bc5a..0e186bc38a00 100644
>>> --- a/drivers/base/power/main.c
>>> +++ b/drivers/base/power/main.c
>>> @@ -1281,6 +1281,27 @@ static void dpm_async_suspend_parent(struct device *dev, async_func_t func)
>>>                 dpm_async_with_cleanup(dev->parent, func);
>>>  }
>>>
>>> +static void dpm_async_suspend_complete_all(struct list_head *device_list)
>>> +{
>>> +       struct device *dev;
>>> +
>>> +
>>> +       pr_err("tudor: %s: enter\n", __func__);
>>> +       guard(mutex)(&async_wip_mtx);
>>> +
>>> +       list_for_each_entry_reverse(dev, device_list, power.entry) {
>>> +               /*
>>> +                * In case the device is being waited for and async processing
>>> +                * has not started for it yet, let the waiters make progress.
>>> +                */
>>> +               pr_err("tudor: %s: in device list\n", __func__);
>>> +               if (!dev->power.work_in_progress) {
>>> +                       pr_err("tudor: %s: call complete_all\n", __func__);
>>> +                       complete_all(&dev->power.completion);
>>> +               }
>>> +       }
>>> +}
>>> +
>>>  /**
>>>   * resume_event - Return a "resume" message for given "suspend" sleep state.
>>>   * @sleep_state: PM message representing a sleep state.
>>> @@ -1459,6 +1480,7 @@ static int dpm_noirq_suspend_devices(pm_message_t state)
>>>                 mutex_lock(&dpm_list_mtx);
>>>
>>>                 if (error || async_error) {
>>> +                       dpm_async_suspend_complete_all(&dpm_late_early_list);
>>>                         /*
>>>                          * Move all devices to the target list to resume them
>>>                          * properly.
>>> @@ -1663,6 +1685,7 @@ int dpm_suspend_late(pm_message_t state)
>>>                 mutex_lock(&dpm_list_mtx);
>>>
>>>                 if (error || async_error) {
>>> +                       dpm_async_suspend_complete_all(&dpm_late_early_list);
>>>                         /*
>>>                          * Move all devices to the target list to resume them
>>>                          * properly.
>>> @@ -1959,6 +1982,7 @@ int dpm_suspend(pm_message_t state)
>>>                 mutex_lock(&dpm_list_mtx);
>>>
>>>                 if (error || async_error) {
>>> +                       dpm_async_suspend_complete_all(&dpm_late_early_list);
>> -> There is a bug here which is not present in the patch I've sent.
> 
> My bad, I edited by hand, sorry.
> 
>>
>> It should be
>>
>>         dpm_async_suspend_complete_all(&dpm_prepared_list);
> 
> 
> Wonderful, it seems this makes suspend happy on downstream pixel6!
> I'm running some more tests and get back to you in a few hours.
> 

Solves failures on pixel6 downstream:
Reported-and-tested-by: Tudor Ambarus <tudor.ambarus@linaro.org>

Thanks!