From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jaroslav Kysela Subject: Re: [PATCH 2/2] add API to allow disabling period interrupt Date: Tue, 2 Nov 2010 09:34:02 +0100 (CET) Message-ID: References: <1288649614-31581-1-git-send-email-pierre-louis.bossart@intel.com> <1288649614-31581-2-git-send-email-pierre-louis.bossart@intel.com> <4CCFCA6A.8010803@ladisch.de> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mail1.perex.cz (unknown [212.20.107.53]) by alsa0.perex.cz (Postfix) with ESMTP id CD4FC2445B for ; Tue, 2 Nov 2010 09:34:03 +0100 (CET) In-Reply-To: <4CCFCA6A.8010803@ladisch.de> 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: Clemens Ladisch Cc: Pierre-Louis Bossart , alsa-devel@alsa-project.org, Pierre-Louis Bossart List-Id: alsa-devel@alsa-project.org On Tue, 2 Nov 2010, Clemens Ladisch wrote: >> + * When period interrupts are disabled, the application must not use poll() or >> + * any functions that could block on this device. > > We should also mention that the application must call some update > function often enough. Reading this - the poll() inside the driver may return POLLERR when no period updates are invoked in the driver. It does not make much sense to block the application when we know that the wake up function is not supported. Also all wait functions in alsa-lib should be reviewed and changed to return an error. Also, the blocking behaviour should not be allowed - it means that apps must use SND_PCM_NONBLOCK or snd_pcm_nonblock(pcm, 1) for this mode. Jaroslav ----- Jaroslav Kysela Linux Kernel Sound Maintainer ALSA Project, Red Hat, Inc.