Re: [PATCH 1/2] cs5535-gpio: add AMD CS5535/CS5536 GPIO driver support

From: "Tobias Müller" <Tobias_Mueller@twam.info>
To: David Brownell <david-b@pacbell.net>
Cc: Andres Salomon <dilinger@collabora.co.uk>,
	akpm@linux-foundation.org, Randy Dunlap <randy.dunlap@oracle.com>,
	deepak@laptop.org, Takashi Iwai <tiwai@suse.de>,
	linux-kernel@vger.kernel.org, linux-geode@lists.infradead.org,
	jordan@cosmicpenguin.net, cjb@laptop.org
Subject: Re: [PATCH 1/2] cs5535-gpio: add AMD CS5535/CS5536 GPIO driver  support
Date: Sat, 15 Aug 2009 12:59:48 +0200	[thread overview]
Message-ID: <17be05570908150359t2cfafac3if43a2dc55a1a7195@mail.gmail.com> (raw)
In-Reply-To: <200907011532.34677.david-b@pacbell.net>

2009/7/2 David Brownell <david-b@pacbell.net>:
> Correct-but-annoying ... and maybe worth changing.  The
> direction *displayed* may not reflect the actual hardware
> until after that GPIO signal is initialized.  Boot firmware
> may well have set the direction; Linux shouldn't change it
> without explicit instructions to do so.
OK. If this is a know problem with all GPIO drivers, we can
skip these to two lines to set in pin a defined state.

> So it would be nice to remove the heuristic.  The best
> way would be to add a new method to query gpio direction.
> Then that when it's available, instead of the heuristic.
This can be implemented when the GPIO interface supports it.

The problem with the mask still occurs. I changed to default
mask to 0x0F7FFFFF, so that all pins except the reserved
ones and the power-pin (which needs special handling) are
enabled. Shall we keep it this way (which I prefer) or delete
it entirely?

Sorry, for the late reply, holidays were nice. :)
  Tobias