From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Brown Subject: Re: Regulator support for smsc911x Date: Fri, 3 Feb 2012 11:11:34 +0000 Message-ID: <20120203111133.GD3151@opensource.wolfsonmicro.com> References: <20120203074408.GD1990@pengutronix.de> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="Xm/fll+QQv+hsKip" Cc: Fabio Estevam , robert.marklund@stericsson.com, netdev@vger.kernel.org, Sascha Hauer To: Sascha Hauer Return-path: Received: from opensource.wolfsonmicro.com ([80.75.67.52]:50118 "EHLO opensource.wolfsonmicro.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755690Ab2BCLLg (ORCPT ); Fri, 3 Feb 2012 06:11:36 -0500 Content-Disposition: inline In-Reply-To: <20120203074408.GD1990@pengutronix.de> Sender: netdev-owner@vger.kernel.org List-ID: --Xm/fll+QQv+hsKip Content-Type: text/plain; charset=us-ascii Content-Disposition: inline On Fri, Feb 03, 2012 at 08:44:08AM +0100, Sascha Hauer wrote: > On Fri, Feb 03, 2012 at 01:17:23AM -0200, Fabio Estevam wrote: > > Please advise if I should fix it in Kconfig or defconfig, so that I > > can submit a patch. > If you enable it every device will get a regulator when it requests one. > If for some reason for example the initialization order of your devices > changes and a device gets probed before the corresponding (real) > regulator is registered, then this device will continue with the dummy > regulator and probably won't work. Hrm, yeah - the dummy regulator support will be totally broken by the stuff Grant is working on for fixing the general issues with probe ordering. > We could a) live with this problem and default enable the dummy > regulator. We could also b) apply (a fixed version of) the following > patch which simplifies registration of a dummy regulator. Still it's a There's also options c) set up the regulators for the board and d) don't enable the regulator API if the board doesn't use regulators. > bit annoying to have to fix all users once someone adds regulator > support for a driver. Yeah, it's a shame. I really can't see anything else useful we can do though. --Xm/fll+QQv+hsKip Content-Type: application/pgp-signature; name="signature.asc" Content-Description: Digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iQIcBAEBAgAGBQJPK8DYAAoJEBus8iNuMP3dPlkP/RKrmgvKsRDyp9MORJy09uvY D5t8YL7c2RNYyA+uzkeLSwsoJoX4sZJC+j96euTksIBA3BBFGEXShQoFhujFiAPo kEJ5SHU/2ulQMHKcqyJysGF7U60VPfAHElCyS+HgoB8zu5T4X20WP+qPCf/q3XgO 6lgZVmq1R/ezYBkYRv+L5MCYfL5RRPj3yZXgLCub0JfbR02NafgH71gEX+TO3SOw j1eq0MObs5dTcZS/rVjqSRQs+e0gY2fAJ4wnmhTdqaFO8FjR+/Yga5EjURLENxoP pMWOGH/v4DZviCyRWxdL1f3IUBGVB2FC6+sEhOZ0MbfFCGpcSrjq2udlfp35kD/i JoGnW/FR3bNTqel/JacHCnCe0ZwKX8BHqQnzPLBlIo1bTzEhhrjm8BFSA2gwkZmt /2r1jVJkHkwRu2Frc7jUEDl7wh+3BIyTNZRkmI6BW/rao0fhkSLIjk2emxQmMFUN 4EMRGfFX05fe8OrXe/NOl6dpLUacU9qPOkMNAvGeYkeHadw2wLDC5shFWYcLeBtv uO6tFEWjeW0Pv+tBWBBSEovaH9nltpfsXP+RK67pH0ixIIPTh+Au2ApfFqZjzKQ1 RTlMHR9w7H53S5gvyGfxFBrEv8r+LaaSEhaDqAUGOiFNR/Ak8vKVINDB5dGXIqyj Mnh6bECB7sTF2Ch9w95c =Zn7L -----END PGP SIGNATURE----- --Xm/fll+QQv+hsKip--