Date: Mon, 23 Jul 2007 16:20:08 +1000
From: David Gibson <hermes@gibson.dropbear.id.au>
To: Pavel Roskin <proski@gnu.org>
Cc: Faidon Liambotis <paravoid@debian.org>,
	linux-wireless@vger.kernel.org
Subject: Re: [PATCH 2.6.23 3/3] [wireless] orinoco: create a Kconfig option for Prism2
Message-ID: <20070723062008.GG3272@localhost.localdomain>
References: <20070722131751.GA3009@void.cube.gr> <20070723014210.GB3272@localhost.localdomain> <1185171040.2108.46.camel@dv>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <1185171040.2108.46.camel@dv>
Sender: linux-wireless-owner@vger.kernel.org

On Mon, Jul 23, 2007 at 02:10:40AM -0400, Pavel Roskin wrote:
> On Mon, 2007-07-23 at 11:42 +1000, David Gibson wrote:
> > > Create an option to disable support for Prism devices in orinoco driver which
> > > disabled the IDs for the PCMCIA module and creates a Kconfig dependency for
> > > the Prism PCI modules.
> > 
> > I very much dislike this being a compile time only CONFIG option.  A
> > module parameter which enables or disables the device IDs would be
> > much nicer, if possible.  It can default to off (orinoco will not
> > claim Prism cards).
> 
> I'm afraid it's not possible.  There is only one device map in the
> driver, and it's read by depmod to generate module map tables.

Ah, yes.  Bugger :(

> 
> If the run-time device table is different from the compile-time table,
> this would be really confusing.  Either the module would be loaded to
> support the device but would refuse to do so, or the module would need
> to be loaded manually for some devices, even if no other module is
> available to support the device.
> 
> Besides, orinoco_pci would become useless by default.
> 
> Please see my message.  For TMD bridges, we don't know the firmware
> flavor before we know the firmware version.  For PLX bridges, we can
> read the CIS first and match the PCMCIA table, but it would need to be
> done manually, without the PCMCIA code.
> 
> I don't see any clear solution without some kind of the device handover
> from Orinoco to hostap or vice versa.
> 
> I would probably start with cutting Agere and Symbol devices from
> hostap, which is a much more clear-cut case than cutting working Prism
> support from Orinoco.

Hrm.  Eck.

I remember once in the distant past Jouni was talking about a possible
merge of hostap and orinoco, starting with the low-level device
manipulation stuff (hermes.c, basically).  But nothing ever came of it.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson