From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mailrelay005.isp.belgacom.be (mailrelay005.isp.belgacom.be [195.238.6.171]) by ozlabs.org (Postfix) with ESMTP id 20650DE5C1 for ; Fri, 18 Apr 2008 21:34:39 +1000 (EST) From: Laurent Pinchart To: linuxppc-dev@ozlabs.org Subject: Re: [RFC fs_enet: Convert MII bitbang driver to use GPIO lib Date: Fri, 18 Apr 2008 13:34:29 +0200 References: <200804161640.42693.laurentp@cse-semaphore.com> <20080416160540.GB442@polina.dev.rtsoft.ru> <200804161809.40805.laurentp@cse-semaphore.com> In-Reply-To: <200804161809.40805.laurentp@cse-semaphore.com> MIME-Version: 1.0 Content-Type: multipart/signed; boundary="nextPart3467580.lIyNInjBNs"; protocol="application/pgp-signature"; micalg=pgp-sha1 Message-Id: <200804181334.34229.laurentp@cse-semaphore.com> Cc: Scott Wood , netdev@vger.kernel.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , --nextPart3467580.lIyNInjBNs Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline On Wednesday 16 April 2008 18:09, Laurent Pinchart wrote: > On Wednesday 16 April 2008 18:05, Anton Vorontsov wrote: > > On Wed, Apr 16, 2008 at 04:40:42PM +0200, Laurent Pinchart wrote: > > > This patch converts the MII bitband driver to use GPIO lib for GPIO > > > access. The driver can now handle MDC and MDIO on different GPIO bank= s. > > >=20 > > > The patch depends on Anton Vorontsov GPIO lib support scheduled for > > > 2.6.26. It is by no means complete, I just would like to get some > > > feedback on the approach. I'll resubmit it when the CPM2 GPIO support > > > patches will be available in the powerpc git tree. > >=20 > > Cool! By the way, maybe it is worth splitting it into completely separa= te > > driver, e.g. net/mdio_gpio.c? >=20 > Splitting it into a completely separate driver makes sense. > > > Plus, keep in mind that somebody will eventually want this cool stuff w= ith > > platform_device bindings in addition. :-)=20 >=20 > I'm sure that person will be happy to implement platform_device bindings = :-) I had a first try at moving mdio gpio code into a separate driver. Very little code is OF-independant, so the driver should probably be called= =20 mdio_of_gpio.c or mdio_ofgpio.c. Scott Wood was concerned in=20 http://patchwork.ozlabs.org/linuxppc/patch?id=3D17490 that the gpio lib mig= ht=20 be an unnecessary burden for memory-constraint platforms. Should we keep tw= o=20 mdio bitbang drivers, one with direct access to the ports and one using gpi= o=20 lib ? The later solves the concurrent access issues present in the current= =20 fs_enet mdio bitbang driber. I'll submit a patch for whichever solution gets selected (modifying the=20 current fs enet mdio bitbang driver to use the gpio lib, or creating a new= =20 driver). =2D-=20 Laurent Pinchart CSE Semaphore Belgium Chaussee de Bruxelles, 732A B-1410 Waterloo Belgium T +32 (2) 387 42 59 =46 +32 (2) 387 42 75 --nextPart3467580.lIyNInjBNs Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) iD8DBQBICIdK8y9gWxC9vpcRAr7PAKCIt21AmihWsIvGxygiqEfJMkDGCQCghddh uM+IR/Bns8ILK7Inc/RBgzc= =7Vam -----END PGP SIGNATURE----- --nextPart3467580.lIyNInjBNs--