From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ezequiel Garcia Subject: Re: [PATCH 2/3] nand: omap2: Remove horrible ifdefs to fix module probe Date: Sun, 7 Sep 2014 12:16:56 -0300 Message-ID: <20140907151656.GA1048@arch.cereza> References: <1410033389-32357-1-git-send-email-ezequiel@vanguardiasur.com.ar> <1410033389-32357-3-git-send-email-ezequiel@vanguardiasur.com.ar> <540B785C.3070400@pek-sem.com> <20140906231728.GA4663@arch.cereza> <540C26F7.7020005@pek-sem.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mail-qg0-f46.google.com ([209.85.192.46]:53440 "EHLO mail-qg0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752375AbaIGPSI (ORCPT ); Sun, 7 Sep 2014 11:18:08 -0400 Received: by mail-qg0-f46.google.com with SMTP id q107so4434044qgd.19 for ; Sun, 07 Sep 2014 08:18:07 -0700 (PDT) Content-Disposition: inline In-Reply-To: <540C26F7.7020005@pek-sem.com> Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: pekon Cc: Roger Quadros , Brian Norris , Tony Lindgren , linux-mtd@lists.infradead.org, Guido =?iso-8859-1?Q?Mart=EDnez?= , linux-omap On 07 Sep 03:05 PM, pekon wrote: [..] > As ECC-scheme is selected in GPMC driver based on DTS settings, so > any mis-match is easily handled there. > Moreover, error will occur when we change ECC-scheme on-the-fly, > which is still not supported by framework, and require many other > updates if we plan to support that in near future. > So considering ECC-scheme as static configuration is a safe assumption. > > But surely you can drop new check in omap2_nand_ecc_check(), which > is already covered in @@gpmc_probe_nand_child(...) > > However, I leave it to you and rogerq@ti.com (as he currently > maintains OMAP NAND driver from TI side) to decide how to go about it. > As far as I can see, your proposal only cover the devicetree-probed case and it won't work for legacy boards. Given it seems we still support legacy, I'd say your proposal would break things. Not to mention that I still prefer to fail at probe time if the driver cannot deal with the selected ECC, and that's exactly what omap2_nand_ecc_check() does. -- Ezequiel Garcia, VanguardiaSur www.vanguardiasur.com.ar