linux-kernel - Re: [PATCH 1/7] input: alps: Set correct name of psmouse device in alps

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: <201412200953.38259@pali>
Date:	Sat, 20 Dec 2014 09:53:38 +0100
From:	Pali Rohár <pali.rohar@...il.com>
To:	Dmitry Torokhov <dmitry.torokhov@...il.com>
Cc:	Hans de Goede <hdegoede@...hat.com>,
	Vadim Klishko <vadim@...que.com>, linux-input@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/7] input: alps: Set correct name of psmouse device in alps_init()

On Tuesday 16 December 2014 12:58:20 Pali Rohár wrote:
> On Tuesday 16 December 2014 06:02:34 Dmitry Torokhov wrote:
> > Hi Pali,
> > 
> > On Fri, Nov 14, 2014 at 08:38:20PM +0100, Pali Rohár wrote:
> > > On some laptops after starting them from off state (not
> > > after reboot), function alps_probe_trackstick_v3() (called
> > > from function alps_identify()) does not detect trackstick.
> > > To fix this problem we need to reset device. But function
> > > alps_identify() is called also from alps_detect() and we
> > > do not want to reset device in detect function because it
> > > will slow down initialization of all other non alps
> > > devices.
> > > 
> > > This patch moves code for setting correct device name &
> > > protocol from function alps_detect() to alps_init() which
> > > already doing full device reset.
> > > 
> > > So this patch removes need to do trackstick detection in
> > > alps_detect() function.
> > > 
> > > Signed-off-by: Pali Rohár <pali.rohar@...il.com>
> > > ---
> > > 
> > >  drivers/input/mouse/alps.c |   17 ++++++++++++++---
> > >  1 file changed, 14 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/input/mouse/alps.c
> > > b/drivers/input/mouse/alps.c index 8d85c79..9ffa98d 100644
> > > --- a/drivers/input/mouse/alps.c
> > > +++ b/drivers/input/mouse/alps.c
> > > @@ -2392,6 +2392,10 @@ int alps_init(struct psmouse
> > > *psmouse)
> > > 
> > >  	if (input_register_device(priv->dev2))
> > >  	
> > >  		goto init_fail;
> > > 
> > > +	if (!(priv->flags & ALPS_DUALPOINT))
> > > +		psmouse->name = "GlidePoint TouchPad";
> > > +	psmouse->model = priv->proto_version;
> > > +
> > > 
> > >  	psmouse->protocol_handler = alps_process_byte;
> > >  	psmouse->poll = alps_poll;
> > >  	psmouse->disconnect = alps_disconnect;
> > > 
> > > @@ -2422,11 +2426,18 @@ int alps_detect(struct psmouse
> > > *psmouse, bool set_properties)
> > > 
> > >  		return -1;
> > >  	
> > >  	if (set_properties) {
> > > 
> > > +		/*
> > > +		 * NOTE: To detect model and trackstick presence we
> 
> need
> 
> > > to do +		 *       full device reset. To speed up
> 
> detection
> 
> > > and prevent +		 *       calling duplicate 
initialization
> > > sequence (both in +		 *       alps_detect() and
> > > alps_init()) we set model/protocol +		 *       version 
and
> > > correct name in alps_init() (which will +		 *       do
> 
> full
> 
> > > device reset). For now set name to DualPoint. +		 */
> > > 
> > >  		psmouse->vendor = "ALPS";
> > > 
> > > -		psmouse->name = dummy.flags & ALPS_DUALPOINT ?
> > > -				"DualPoint TouchPad" : "GlidePoint";
> > > -		psmouse->model = dummy.proto_version << 8;
> > > +		psmouse->name = "DualPoint TouchPad";
> > > 
> > >  	}
> > > 
> > > +
> > 
> > I do not quite like the way we change the device description
> > back and forth. Do you think we could allocate the "real"
> > priv structure in alps_detect() and have alps_init() expect
> > to find it (and free it if set_properties is false). This
> > way we'd go through initialization once in detect, it will
> > be authoritative, and we would set the name of the device
> > properly from the beginning.
> > 
> > Thanks.
> 
> No without introducing another psmouse_reset call. I want to
> reduce time of loading driver, so I think this is better.

Also psmouse-base.c call psmouse_reset between alps_detect() and 
alps_init(). We must do trackstick detection after psmouse_reset, 
but intorducing another psmouse_reset in alps_detect() will slow 
down initialization for ALPS devices. Also do not remember that 
psmouse_reset will lock both keyboard & mouse for one second 
(sometimes for two) on my laptop. I played with this and I think 
my patch is good approach how to do trackstick detection without 
slowdown or other side effects.

Also we do not need to know if ALPS touchpad has trackstick 
device in alps_detect() phase. In alps_detect() we just need to 
know if PS/2 device is ALPS or not. From psmouse/serio point of 
view there is only one ALPS device on PS/2/i8042 bus. Just 
alps.ko driver parse PS/2 packets and detect which packet has 
touchpad header and which trackpoint. So we really do not need to 
know if trackstick is there or not in alps_detect() function.

-- 
Pali Rohár
pali.rohar@...il.com

Download attachment "signature.asc " of type "application/pgp-signature" (199 bytes)