linux-kernel - Re: [RFC 1/2] USB: OTG: Hold wakeupsource when VBUS present

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]

Date:	Fri, 29 Aug 2014 15:50:15 -0500
From:	Felipe Balbi <balbi@...com>
To:	Todd Poynor <toddpoynor@...gle.com>
CC:	<balbi@...com>, Kiran Kumar Raparthy <kiran.kumar@...aro.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	<linux-usb@...r.kernel.org>,
	Android Kernel Team <kernel-team@...roid.com>,
	John Stultz <john.stultz@...aro.org>,
	Arve Hj�nnev�g <arve@...roid.com>,
	Benoit Goby <benoit@...roid.com>,
	Ruchi Kandoi <kandoiruchi@...gle.com>,
	Badhri Jagan Sridharan <badhri@...gle.com>
Subject: Re: [RFC 1/2] USB: OTG: Hold wakeupsource when VBUS present

Hi,

On Wed, Aug 27, 2014 at 02:58:30PM -0700, Todd Poynor wrote:
> On Fri, Aug 22, 2014 at 10:12 AM, Felipe Balbi <balbi@...com> wrote:
> ...
> > you never explain why this is needed and you have also added some
> > information to commit log which shouldn't be here.
> 
> Android uses this to prevent suspend from interfering with USB
> peripheral traffic, notably adb.

aaaaah, that's because Android freezes userspace right ?

> The wakeup source is held only when USB is connected and enumerated
> for a host session (I might be using wrong terminology here).  It may
> not be necessary on a platform that implements wakeup on incoming USB
> traffic, although it is likely adb and other protocols would need to
> hold wakeup sources at certain times.
> 
> ...
> >> +static struct otgws_lock vbus_lock;
> >
> > should be per-PHY
> 
> One of the reasons this was done as a separate driver and via
> notifiers was to keep the (original Android wakelock) logic out of the
> USB code.  If the general idea is something that finds favor with the
> USB and PM folks then perhaps adding a wakeup source per PHY in the
> PHY driver would be better.

it's best to do it per-PHY, but it doesn't have to be written into each
and every PHY driver, do it at the core and abstract that away from PHY
drivers.

-- 
balbi

Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)