[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-id: <1443120167.38671418974875751.JavaMail.weblogic@epmlwas08d>
Date: Fri, 19 Dec 2014 07:41:15 +0000 (GMT)
From: MyungJoo Ham <myungjoo.ham@...sung.com>
To: 이종화 <jonghwa3.lee@...sung.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Cc: "linux-pm@...r.kernel.org" <linux-pm@...r.kernel.org>,
"sre@...nel.org" <sre@...nel.org>,
"dbaryshkov@...il.com" <dbaryshkov@...il.com>,
"dwmw2@...radead.org" <dwmw2@...radead.org>,
"anton@...msg.org" <anton@...msg.org>,
"pavel@....cz" <pavel@....cz>,
최찬우 <cw00.choi@...sung.com>
Subject: Re: [PATCH RESEND v2 02/10] power: charger-manager: Use
power_supply_changed() not private uevent.
>
> Whenever battery status is changed, charger manager tries to trigger uevent
> through private interface. This patch modifies it to use power_supply_changed()
> since it belongs to power supply subsystem.
>
> Signed-off-by: Jonghwa Lee <jonghwa3.lee@...sung.com>
The original uevent_notify() has two additional mechanisms:
C1. Save events in suspend-again context and show them up at wakeup.
C2. If the new event is a duplicated event, ignore it.
Questions:
Q1. Have you checked if C1 is met with the modification? Besides, have
you made it sure that the modification won't change the behavior of
suspend-again context? (whether "theoretical" or "experimental")
Q2. Do you still support C2?
For example, if we have notifited the user that we are charging
30 seconds ago, we should never bother the user with another message
that declares that it is charging unless we have notified that
we are not charging since then.
Cheers,
MyungJoo.
> ---
> drivers/power/charger-manager.c | 91 +++++----------------------------------
> 1 file changed, 11 insertions(+), 80 deletions(-)
>
Powered by blists - more mailing lists