From mboxrd@z Thu Jan 1 00:00:00 1970 From: Richard Cochran Subject: Re: [PATCH 2/6] drivers: net: ethernet: davinci_cpdma: add clear api for statistics interrupt Date: Mon, 22 Oct 2012 13:31:04 +0200 Message-ID: <20121022113103.GC2285@netboy.at.omicron.at> References: <1350427518-7230-1-git-send-email-mugunthanvnm@ti.com> <1350427518-7230-3-git-send-email-mugunthanvnm@ti.com> <20121018024808.GB2867@netboy.at.omicron.at> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "netdev@vger.kernel.org" , "davem@davemloft.net" To: "N, Mugunthan V" Return-path: Received: from mail-wi0-f178.google.com ([209.85.212.178]:63526 "EHLO mail-wi0-f178.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751375Ab2JVLbO (ORCPT ); Mon, 22 Oct 2012 07:31:14 -0400 Received: by mail-wi0-f178.google.com with SMTP id hr7so2349736wib.1 for ; Mon, 22 Oct 2012 04:31:13 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Mon, Oct 22, 2012 at 10:42:26AM +0000, N, Mugunthan V wrote: > > What happens when Hardware timestamps are used? They are not being used. > It will be good to > notify the application immediately using the interrupts instead of > using workqueue which can delay the event in hardware for max about > HZ*8. Yes, it would make sense to use the interrupt in that case, but I did not see any code supporting external events. In any case, I still want to handle the overflow in the same way as in the other PHC drivers, for the sake of consistency, and to avoid wasting CPU time "in interrupt" with silly counter overflow code. Thanks, Richard