Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Content-Type: text/plain;
        charset=us-ascii;
        delsp=yes;
        format=flowed
Mime-Version: 1.0 (Mac OS X Mail 11.5 \(3445.9.1\))
Subject: Re: [PATCH 1/5] misc: rtsx_usb: Use USB remote wakeup signaling for
 card insertion detection
From:   Kai-Heng Feng <kai.heng.feng@canonical.com>
In-Reply-To: <Pine.LNX.4.44L0.1809130936350.1990-100000@iolanthe.rowland.org>
Date:   Mon, 17 Sep 2018 16:07:27 +0800
Cc:     Ulf Hansson <ulf.hansson@linaro.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Linux USB Mailing List <linux-usb@vger.kernel.org>
Content-Transfer-Encoding: 7bit
Message-Id: <481937DD-FB33-4086-B478-32491A01B51A@canonical.com>
References: <Pine.LNX.4.44L0.1809130936350.1990-100000@iolanthe.rowland.org>
To:     Alan Stern <stern@rowland.harvard.edu>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

at 21:41, Alan Stern <stern@rowland.harvard.edu> wrote:

> On Thu, 13 Sep 2018, Kai-Heng Feng wrote:
>
>> I am working on the next version of this series, and the last missing
>> puzzle is to differentiate system-wide resume from runtime resume in
>> usb_driver's resume() and reset_resume() callback.
>>
>> The parent device, rtsx_usb, has two child devices, rtsx_usb_ms and
>> rtsx_usb_sdmmc.
>> pm_request_resume() is used in rtsx_usb's resume() to facilitate USB  
>> remote
>> wakeup signaling, so we don't need to poll the card slot status.
>> But this has a side effect: during system resume the rtsx_usb calls
>> pm_request_resume() in its resume(), so child devices calls their
>> runtime_resume() instead of resume() callback.
>
> Have you actually observed this?  It shouldn't happen.
> pm_request_resume() schedules a runtime resume on the PM work queue,
> but this work queue is frozen during system sleep.  It doesn't unfreeze
> until after all the devices have been restored to full power.  Thus the
> child device's resume() callback should be invoked before
> runtime_resume().

You are right, I read the trace wrong.

It happens right before system-wide suspension, when the PM core resumes it  
to fully functional state.

>
>> So, is it reasonable to pass pm_message_t to resume() and reset_resume()
>> callbacks and use PMSG_IS_AUTO() to differentiate them?
>
> It should not be necessary to do this.

Understood, thanks.

Kai-Heng

>
> Alan Stern