linux-kernel - Re: [PATCH 1/5] i8042: intel-8042 DT documentation

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: <20150211000125.5d66bfb5@v1ron-s7>
Date:	Wed, 11 Feb 2015 00:01:25 +0300
From:	Roman Volkov <v1ron@...os.org>
To:	Dmitry Torokhov <dmitry.torokhov@...il.com>
Cc:	Mark Rutland <mark.rutland@....com>,
	Rob Herring <robh+dt@...nel.org>,
	Pawel Moll <Pawel.Moll@....com>,
	Ian Campbell <ijc+devicetree@...lion.org.uk>,
	Kumar Gala <galak@...eaurora.org>,
	"grant.likely@...aro.org" <grant.likely@...aro.org>,
	Hans de Goede <hdegoede@...hat.com>,
	Jiri Kosina <jkosina@...e.cz>,
	Wolfram Sang <wsa@...-dreams.de>,
	"linux-input@...r.kernel.org" <linux-input@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"devicetree@...r.kernel.org" <devicetree@...r.kernel.org>,
	Roman Volkov <v1ron@...l.ru>,
	Tony Prisk <linux@...sktech.co.nz>
Subject: Re: [PATCH 1/5] i8042: intel-8042 DT documentation

В Tue, 3 Feb 2015 11:32:02 -0800
Dmitry Torokhov <dmitry.torokhov@...il.com> пишет:

> On Tue, Feb 03, 2015 at 11:38:16AM +0000, Mark Rutland wrote:
> > On Mon, Feb 02, 2015 at 09:48:46PM +0000, Roman Volkov wrote:
> > > Documentation for 'intel,8042' DT compatible node.
> > > 
> > > Signed-off-by: Tony Prisk <linux@...sktech.co.nz>
> > > Signed-off-by: Roman Volkov <v1ron@...os.org>
> > > ---
> > >  .../devicetree/bindings/input/intel-8042.txt       | 29
> > > ++++++++++++++++++++++ 1 file changed, 29 insertions(+)
> > >  create mode 100644
> > > Documentation/devicetree/bindings/input/intel-8042.txt
> > > 
> > > diff --git
> > > a/Documentation/devicetree/bindings/input/intel-8042.txt
> > > b/Documentation/devicetree/bindings/input/intel-8042.txt new file
> > > mode 100644 index 0000000..2aea7ec --- /dev/null
> > > +++ b/Documentation/devicetree/bindings/input/intel-8042.txt
> > > @@ -0,0 +1,29 @@
> > > +* Intel 8042 Keyboard Controller
> > > +
> > > +Required properties:
> > > +- compatible: should be "intel,8042"
> > > +- regs: memory for keyboard controller
> > > +- interrupts: two interrupts should be specified (keyboard and
> > > aux)
> > 
> > Is it possible only one of these is wired up?
> 
> Yes, and we should support this case. The core of i8042 does.
> 

Do we need to just read these IRQ numbers and leave them negative if
absent? Will it be acceptable? This would look like:

i8042_kbd_irq = platform_get_irq_byname(pdev, "kbd");

Testing shows it prints "Invalid argument" error -22 when an IRQ is
absent and we are not using nokbd/noaux module options.

> 
> > 
> > It might be worth using interrupt-names.
> > 
> > > +- command-reg: offset in memory for command register
> > > +- status-reg: offset in memory for status register
> > > +- data-reg: offset in memory for data register
> > > +
> > > +Optional properties:
> > > +- init-reset: Controller should be reset on init and cleanup
> > 
> > Why is this necessary? Can't we just always reset it?
> 
> We do not reset by default on x86 because BIOS takes care of this for
> us and quite often firmware that emulates i8042 gets confused if we
> try to reset it too. Non non-x86 we reset by default. I think we
> should do the same for OF case  (reset) and not use this property.
> 
> > 
> > > +
> > > +Optional Linux-specific properties:
> > > +- linux,kbd_phys_desc: defaults to i8042/serio0
> > > +- linux,aux_phys_desc: defaults to i8042/serio1
> > > +- linux,mux_phys_desc: defaults to i8042/serio%d
> > 
> > As a general note, s/_/-/ in property names please.
> > 
> > That said, I don't follow why we should have these at all. I don't
> > understand what the description is intended to mean.
> > 
> > In general we want to avoid Linux-specific properties. If a DTB
> > needs to know about the inernals of an OS it's likely to be fragile
> > and broken over time.
> 
> Right, the desc were carried over from older days to keep dmesg
> familiar. With OF it is new platforms so just settle on a generic
> description and use it instead of allowing to specify through DT.
> 
> > 
> > > +
> > > +
> > > +Example:
> > > +	keyboard@...08800 {
> > > +		compatible = "intel,8042";
> > > +		reg = <0xd8008800 0x100>;
> > > +		interrupts = <23 4>;
> > 
> > If this is intended to be two interrupts, please bracket them
> > individually, e.g.
> > 
> > 	interrupts = <23>, <4>;
> > 
> > > +		command-reg = <0x04>;
> > > +		status-reg = <0x04>;
> > 
> > Same address?
> > 
> > > +		data-reg = <0x00>;
> > > +		mux-ports = <2>;
> > 
> > This wasn't documented above.
> 
> I think active MUX is purely x86 concept, I have never heard of it
> being used anywhere else.
> 
> Thanks.
> 

--
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/