linux-kernel - Re: [PATCH 4/5] memstick: rtsx_usb_ms: Support runtime power management

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Pine.LNX.4.44L0.1807251357000.1467-100000@iolanthe.rowland.org>
Date:   Wed, 25 Jul 2018 13:59:58 -0400 (EDT)
From:   Alan Stern <stern@...land.harvard.edu>
To:     Kai-Heng Feng <kai.heng.feng@...onical.com>
cc:     arnd@...db.de, <gregkh@...uxfoundation.org>,
        <ulf.hansson@...aro.org>, <bauer.chen@...ltek.com>,
        <ricky_wu@...ltek.com>, <linux-kernel@...r.kernel.org>,
        <linux-usb@...r.kernel.org>
Subject: Re: [PATCH 4/5] memstick: rtsx_usb_ms: Support runtime power management

On Wed, 25 Jul 2018, Kai-Heng Feng wrote:

> In order to let host's parent device, rtsx_usb, to use USB remote wake
> up signaling to do card detection, it needs to be suspended. Hence it's
> necessary to add runtime PM support for the memstick host.
> 
> To keep memstick host stays suspended when it's not in use, convert the
> card detection function from kthread to delayed_work, which can be
> scheduled when the host is resumed and can be canceled when the host is
> suspended.
> 
> Use an idle function check if there's no card and the power mode is
> MEMSTICK_POWER_OFF. If both criteria are met, put the device to suspend.
> 
> Signed-off-by: Kai-Heng Feng <kai.heng.feng@...onical.com>

I'm not familiar enough with this driver to comment on specific
instances, but you should carefully go through the code and make sure
that you use pm_runtime_get_noresume() and pm_runtime_put_noidle()
correctly.

The basic idea is to use these when you know beforehand that the device
is already at full power (for _get_noresume) and the usage count will
not go to 0 (for _put_noidle).  If you aren't certain of these
requirements then you should call pm_runtime_get_sync() and
pm_runtime_put().

Alan Stern