From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Brown Subject: Re: SoC codec read callback definition Date: Thu, 23 Jul 2009 16:46:04 +0100 Message-ID: <20090723154603.GA30390@sirena.org.uk> References: <67059DBF19D7214F9C66BB0EA91BA90E59BA9D93@dlee04.ent.ti.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from cassiel.sirena.org.uk (cassiel.sirena.org.uk [80.68.93.111]) by alsa0.perex.cz (Postfix) with ESMTP id C030E2453E for ; Thu, 23 Jul 2009 17:46:04 +0200 (CEST) Content-Disposition: inline In-Reply-To: <67059DBF19D7214F9C66BB0EA91BA90E59BA9D93@dlee04.ent.ti.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: alsa-devel-bounces@alsa-project.org Errors-To: alsa-devel-bounces@alsa-project.org To: "Lopez Cruz, Misael" Cc: "alsa-devel@alsa-project.org" List-Id: alsa-devel@alsa-project.org On Thu, Jul 23, 2009 at 02:50:53AM -0500, Lopez Cruz, Misael wrote: > However, actual implementation of that callback in many CODEC drivers > may return a negative error value (i.e. when a register is out of cache > reg size). Should the callback definition be changed to return _int_ > instead of _unsigned int_? Some drivers (like wm8900, wm9081) handle > invalid registers with BUG_ON, in that case it's not necessary to > return the negative error code. There's two classes of error. Reads from the hardware might possibly fail at runtime, while reads from cache should never fail - an attempt to read from an uncached register is a clear bug in the driver. When the code was originally written there were essentially no devices with readback so the cache was the only option. > What should be the best way to handle this? Or is it ok as currently > implemented? Well, it was just a thought. Ideally we'd have error codes but that would loose us a bit of data and IIRC there's at least one codec in there with 32 bit registers which would need to be handled.