From mboxrd@z Thu Jan 1 00:00:00 1970 From: Felipe Balbi Subject: Re: [net-next PATCH 1/1] drivers: net: cpsw: Add support for new CPSW IP version Date: Wed, 31 Jul 2013 23:26:17 +0300 Message-ID: <20130731202617.GB3737@radagast> References: <20130731144957.GC4904@netboy> <20130731152827.GB25618@radagast> <20130731163845.GB4234@netboy> <20130731184525.GA629@radagast> <20130731192229.GB8027@netboy> <20130731194332.GA900@radagast> <20130731194523.GB900@radagast> <20130731200428.GD8027@netboy> <20130731200756.GA3737@radagast> <20130731202006.GE8027@netboy> Reply-To: Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="vGgW1X5XWziG23Ko" Return-path: Content-Disposition: inline In-Reply-To: <20130731202006.GE8027@netboy> Sender: netdev-owner@vger.kernel.org To: Richard Cochran Cc: Felipe Balbi , Mugunthan V N , netdev@vger.kernel.org, davem@davemloft.net, linux-omap@vger.kernel.org List-Id: linux-omap@vger.kernel.org --vGgW1X5XWziG23Ko Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hi, On Wed, Jul 31, 2013 at 10:20:07PM +0200, Richard Cochran wrote: > On Wed, Jul 31, 2013 at 11:07:56PM +0300, Felipe Balbi wrote: > >=20 > > what I'm saying is that we can give new IP revision a chance to work if > > they have no programming model differences (except for, perhaps, new > > features and different erratas). >=20 > But it also has a chance to fail when there are differences. > Comparing CPSW V1 with V2, it appears that TI likes to move the > registers around between versions. To me, this is reason enough to > make the driver defensive. oh well, we can go on and on with this. Unfortunately we (SW team) don't have control over the HW folks. We strongly suggest that they don't break SW compatibility, and that's starting to become true. You can very well expect next version of CPSW to be SW compatible. If it isn't, then TI will send patches to add a new revision check and treat it well. We are the first ones to have access to new versions of all our IPs anyway. And, IMHO, even if HW engineers decides to move registers around in CPSW v3, that still doesn't chage the fact that defaulting to highest known revision is a good practice. Bailing out just because the revision check isn't what you expect it to be is a very poor practice and leads to periodic patches updating 'switch' statements all over the place. --=20 balbi --vGgW1X5XWziG23Ko Content-Type: application/pgp-signature; name="signature.asc" Content-Description: Digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.12 (GNU/Linux) iQIcBAEBAgAGBQJR+XLpAAoJEIaOsuA1yqREyZoQAK3VNCGpmQC9iDx5NAn+az3a JapntkM8HLZeMZZgjWP74fazKyqy01DUcT1dja6aVvoHNGFuOeUtisvR4ke9o/R0 97PJEliZpODdfMEX10C5NxRlEQfjflmdL6pSjFvIldB6AlBjV9Ng+MgVH35WqnMf /96WkIvSg8gyNZY6czEx900gPVZ8yE1dSI4WTFQ3pvLMe+ruhijRJSZyDFgetrvc elG/qOChw+e3Rbvsw3HtGQGJpA0k3UcH5UDqdwLv3IsrKhAB2qPY4U1PlAWdkCxo A13xB4WhrXCFgnhMrCbhdQaXgdCdeCrKBCk4+VQ9Y0FmC0RRXg/3tSUVqk4BT6V1 s+mt2SjuGOlB1IOeopI+RdYVLmzeLsxdF19YdVuLvWMawsJAVzstq+P0rjv62HQX CtSMzF4DssIIiMhN23Lp7NUjbcfcniVzwUArJERnAIKtdjHAMErQXZWJPIS1KXRc fMNcU2EQ2EQM4Hlc2KYEGlMd6TdC6Bm2EyD0ubuRQNf8oT4LvBAre1qjLkiClzo1 xkhhK12E0e0sWxp0A4clr9sBd8DsV1PgaqneHUKxuPzQJvHXj//budJXJMzCGDK2 fflOORs9VWY6rWiKykJ8Nww+kjMYA+M7DbYM0WAVTzHE7+UvZ+uFpkLOsgDklVbM sPgmpdgYEFaoPdldGuHF =oeHk -----END PGP SIGNATURE----- --vGgW1X5XWziG23Ko--