From: Marek Vasut <marek.vasut@gmail.com>
To: Oliver Neukum <oliver@neukum.org>
Subject: Re: loading firmware while usermodehelper disabled.
Date: Sun, 1 Jan 2012 18:06:06 +0100
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Dave Jones <davej@redhat.com>,
	Linux Kernel <linux-kernel@vger.kernel.org>,
	Larry Finger <Larry.Finger@lwfinger.net>,
	Chaoming Li <chaoming_li@realsil.com.cn>,
	"John W. Linville" <linville@tuxdriver.com>,
	Matthew Garrett <mjg@redhat.com>,
	"Greg Kroah-Hartman" <gregkh@suse.de>,
	USB list <linux-usb@vger.kernel.org>,
	Linux Wireless List <linux-wireless@vger.kernel.org>
References: <20111230235421.GA6054@redhat.com> <201201011328.41311.oliver@neukum.org> <201201011732.23069.marek.vasut@gmail.com>
In-Reply-To: <201201011732.23069.marek.vasut@gmail.com>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="us-ascii"
Message-Id: <201201011806.07150.marek.vasut@gmail.com> (sfid-20120101_180656_417441_F23BF636)
Sender: linux-wireless-owner@vger.kernel.org

> > Am Sonntag, 1. Januar 2012, 10:54:47 schrieb Marek Vasut:
> > > > Am Samstag, 31. Dezember 2011, 19:39:18 schrieb Marek Vasut:
> > > > > maybe we should implement such thing into the firmware loader
> > > > > itself? Allow it -- for example via some node in /dev -- to force
> > > > > loading firmware into some buffer in kernel just before suspend so
> > > > > it'll certainly be readily available at resume time?
> > > > 
> > > > This is difficult. We don't know at suspend time which firmware we
> > > > will need at resume time.
> > > 
> > > That's not actually true ... you can ask the drivers if they need to
> > > load firmware after resume ... that can be implemented and udev can
> > > handle it. The driver should know if the hardware it's controlling is
> > > stupid or not.
> > 
> > Device IDs can morph due to a power loss. After resumption another driver
> > would bind.
> 
> What do you mean? I don't think I quite follow.

Ah I see ... you mean like you can swap the devices (eg. usb devices) when the 
computer is asleep. Yea, that's an issue.

On the other hand, you'd expect those to keep pestering udev to load firmware 
later in the wakeup process (maybe it can be somehow defered?). But if we 
actually implemented some pre-suspend FW cache, it might help the speed of 
resume a bit, right ?

M
> 
> M
> 
> > 	Regards
> > 	
> > 		Oliver