From: Takashi Iwai Subject: Re: [PATCH v3 3/3] conf: topology: Generate Private data binary blobs Date: Thu, 21 Apr 2016 13:00:26 +0200 Message-ID: References: <1461220058-1378-1-git-send-email-vinod.koul@intel.com> <1461220058-1378-4-git-send-email-vinod.koul@intel.com> <20160421105654.GP2274@localhost> Mime-Version: 1.0 (generated by SEMI 1.14.6 - "Maruoka") Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mx2.suse.de (mx2.suse.de [195.135.220.15]) by alsa0.perex.cz (Postfix) with ESMTP id 9A53C265173 for ; Thu, 21 Apr 2016 13:00:29 +0200 (CEST) In-Reply-To: <20160421105654.GP2274@localhost> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: alsa-devel-bounces@alsa-project.org Sender: alsa-devel-bounces@alsa-project.org To: Vinod Koul Cc: liam.r.girdwood@linux.intel.com, patches.audio@intel.com, alsa-devel@alsa-project.org, broonie@kernel.org, Shreyas NC List-Id: alsa-devel@alsa-project.org On Thu, 21 Apr 2016 12:56:54 +0200, Vinod Koul wrote: > > On Thu, Apr 21, 2016 at 09:23:25AM +0200, Takashi Iwai wrote: > > On Thu, 21 Apr 2016 08:27:38 +0200, > > Vinod Koul wrote: > > > > > > From: Shreyas NC > > > > > > The DSP modules need private data and that is provided as binary > > > blob. These blobs are compiled from C structures which specify module > > > configuration. > > > > > > Signed-off-by: Shreyas NC > > > Signed-off-by: Vinod Koul > > > > What if this program is executed on big endian system? > > So we made the data structure m,emebers as __lexx shouldn't that take care > of this, or we can say, we don't support BE systems. > > Am not sure if someone running BE system will be intrested in Intel aDSP :D It's not about whether it runs on a BE system. The most likely problem is about the package build. The firmware sub-package is usually handled as a noarch one that is used commonly among all architectures. If you generate a different data per architecture and write to the same file, it results in the inconsistent package, since the firmware generated by a BE host may be used for LE machine. So, this needs a bit more consideration. If we don't provide the firmware binary but only rely on the file creation at build time, it won't get the complete set depending on the running machine. Maybe we should add the binary files in addition to the f/w generation code in case anyone needs to modify the data. Takashi