linux-kernel - Re: [RFT/PATCH] i8042: remove polling timer (v6)

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: <20061030090851.GA2687@suse.cz>
Date:	Mon, 30 Oct 2006 10:08:51 +0100
From:	Vojtech Pavlik <vojtech@...e.cz>
To:	Dmitry Torokhov <dtor@...ightbb.com>
Cc:	Dave Neuer <mr.fred.smoothie@...ox.com>,
	LKML <linux-kernel@...r.kernel.org>,
	Andrew Morton <akpm@...l.org>
Subject: Re: [RFT/PATCH] i8042: remove polling timer (v6)

On Sun, Oct 29, 2006 at 10:34:00PM -0500, Dmitry Torokhov wrote:
> On Sunday 29 October 2006 18:20, Dave Neuer wrote:
> > On 8/23/06, Dmitry Torokhov <dtor@...ightbb.com> wrote:
> > > Hi everyone,
> > >
> > > Here is another version of the patch removing polling timer from i8042
> > > which is needed if we want tickless kernel. Keyboards should now work
> > > on boxes that do not have mouse plugged in. PLease give it a test.
> > 
> >  What's the intent of this; just to allow tickless? Or is it also to
> > make the i8042 driver less racy? I ask because I've applied this over
> > (a modified) 2.6.18 on my Compaq Presario X1010us laptop which has
> > been driving me crazy w/ Synaptics problems and keyboard problems
> > (intermittent, but   frequent enough lately that I finally figured I
> > needed to do something about it).
> > 
> > If removing raciness is part of the goal, isn't the window in
> > i8042_aux_write still a problem?
> > 
> > 	if (port->mux == -1)
> > 		retval = i8042_command(&c, I8042_CMD_AUX_SEND);
> > 	else
> > 		retval = i8042_command(&c, I8042_CMD_MUX_SEND + port->mux);
> > 
> >         /* i8042_command has re-enabled interrupts;
> >            what happens if real interrupt happens here, before we call
> > the ISR ourselves? */
> > 
> > 	i8042_interrupt(0, NULL, NULL);
> > 	return retval;
> > }
> 
> Hi Dave,
> 
> i8042_interrupt() uses spinlock to serialize access to the KBC so if real
> interrupt happens before we call i8042_interrupt() manually (and it should
> normally happen) it will just process the response and second i8042_interrupt()
> will be just a no-op.

This would, however, create two reads of the i8042 controller
back-to-back, which has been a problem on old i8042's: IIRC IBM
documentation states that between the reads there should be a delay.

-- 
Vojtech Pavlik
Director SuSE Labs
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/