From: Subbaraman Narayanamurthy <quic_subbaram@quicinc.com>
To: Sebastian Reichel <sebastian.reichel@collabora.com>,
"Manaf Meethalavalappu Pallikunhi" <quic_manafm@quicinc.com>
Cc: <linux-pm@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<gregkh@google.com>
Subject: Re: [RESEND 1/2] power_supply: Register cooling device outside of probe
Date: Mon, 27 Feb 2023 13:46:52 -0800 [thread overview]
Message-ID: <fd372789-d39e-08f9-ae44-7702733155ae@quicinc.com> (raw)
In-Reply-To: <20220609221224.t5k7i4w4dfjza5xc@mercury.elektranox.org>
On 6/9/22 3:12 PM, Sebastian Reichel wrote:
> Hi,
>
> On Wed, Jun 01, 2022 at 12:00:53AM +0530, Manaf Meethalavalappu Pallikunhi wrote:
>> Registering the cooling device from the probe can result in the
>> execution of get_property() function before it gets initialized.
>>
>> To avoid this, register the cooling device from a workqueue
>> instead of registering in the probe.
>>
>> Signed-off-by: Manaf Meethalavalappu Pallikunhi <quic_manafm@quicinc.com>
>> ---
> This removes error handling from the psy_register_cooler() call, so
> it introduces a new potential problem. If power_supply_get_property()
> is called to early -EAGAIN is returned. So can you elaborate the problem
> that you are seeing with the current code?
>
> -- Sebastian
When the device boots up with all the vendor modules getting loaded,
here is what we're seeing when booting up with 6.1.11 recently. First
log is printed with adding a pr_err() in __power_supply_register().
[ 7.008938][ T682] power_supply battery: psy_register_cooler failed, rc=-11
[ 7.030941][ T682] qti_battery_charger: probe of qcom,battery_charger failed with error -11
Here, our downstream qti_battery_charger driver exposes the following
power supply properties POWER_SUPPLY_PROP_CHARGE_CONTROL_LIMIT and
POWER_SUPPLY_PROP_CHARGE_CONTROL_LIMIT_MAX under a power supply device.
This is happening because of the following call sequence,
battery_chg_probe() ->
power_supply_register() ->
psy_register_cooler() ->
thermal_cooling_device_register() ->
cdev->ops->get_max_state() ->
ps_get_max_charge_cntl_limit() ->
power_supply_get_property()
ends up calling power_supply_get_property() to read CHARGE_CONTROL_LIMIT
property.
However, it returns -EAGAIN because psy->initialized is set to true
later after psy_register_cooler() succeeds. So, this ends up in a
driver probe failure forever.
-Subbaraman
>
>> drivers/power/supply/power_supply_core.c | 10 ++++------
>> 1 file changed, 4 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/power/supply/power_supply_core.c b/drivers/power/supply/power_supply_core.c
>> index 385814a14a0a..74623c4977db 100644
>> --- a/drivers/power/supply/power_supply_core.c
>> +++ b/drivers/power/supply/power_supply_core.c
>> @@ -132,6 +132,7 @@ void power_supply_changed(struct power_supply *psy)
>> }
>> EXPORT_SYMBOL_GPL(power_supply_changed);
>>
>> +static int psy_register_cooler(struct power_supply *psy);
>> /*
>> * Notify that power supply was registered after parent finished the probing.
>> *
>> @@ -139,6 +140,8 @@ EXPORT_SYMBOL_GPL(power_supply_changed);
>> * calling power_supply_changed() directly from power_supply_register()
>> * would lead to execution of get_property() function provided by the driver
>> * too early - before the probe ends.
>> + * Also, registering cooling device from the probe will execute the
>> + * get_property() function. So register the cooling device after the probe.
>> *
>> * Avoid that by waiting on parent's mutex.
>> */
>> @@ -156,6 +159,7 @@ static void power_supply_deferred_register_work(struct work_struct *work)
>> }
>>
>> power_supply_changed(psy);
>> + psy_register_cooler(psy);
>>
>> if (psy->dev.parent)
>> mutex_unlock(&psy->dev.parent->mutex);
>> @@ -1261,10 +1265,6 @@ __power_supply_register(struct device *parent,
>> if (rc)
>> goto register_thermal_failed;
>>
>> - rc = psy_register_cooler(psy);
>> - if (rc)
>> - goto register_cooler_failed;
>> -
>> rc = power_supply_create_triggers(psy);
>> if (rc)
>> goto create_triggers_failed;
>> @@ -1294,8 +1294,6 @@ __power_supply_register(struct device *parent,
>> add_hwmon_sysfs_failed:
>> power_supply_remove_triggers(psy);
>> create_triggers_failed:
>> - psy_unregister_cooler(psy);
>> -register_cooler_failed:
>> psy_unregister_thermal(psy);
>> register_thermal_failed:
>> device_del(dev);
next prev parent reply other threads:[~2023-02-27 21:47 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-31 18:30 [RESEND 1/2] power_supply: Register cooling device outside of probe Manaf Meethalavalappu Pallikunhi
2022-05-31 18:30 ` [RESEND 2/2] power_supply: Use of-thermal cdev registration API Manaf Meethalavalappu Pallikunhi
2022-06-09 22:17 ` Sebastian Reichel
2022-06-09 22:12 ` [RESEND 1/2] power_supply: Register cooling device outside of probe Sebastian Reichel
2023-02-27 21:46 ` Subbaraman Narayanamurthy [this message]
2023-02-28 0:23 ` Sebastian Reichel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=fd372789-d39e-08f9-ae44-7702733155ae@quicinc.com \
--to=quic_subbaram@quicinc.com \
--cc=gregkh@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=quic_manafm@quicinc.com \
--cc=sebastian.reichel@collabora.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).