From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paul Mundt Subject: Re: [PATCH RFC 0/8] Dynamic clock devices Date: Mon, 15 Nov 2010 19:01:50 +0900 Message-ID: <20101115100150.GA25973@linux-sh.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-api-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Richard Cochran Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-api-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Alan Cox , Arnd Bergmann , Christoph Lameter , John Stultz , Peter Zijlstra , Thomas Gleixner List-Id: linux-api@vger.kernel.org On Thu, Nov 04, 2010 at 08:26:34PM +0100, Richard Cochran wrote: > Okay, here is a work in progress, not well tested, but I would like to > get some feedback whether the direction is good or not. > > The first patch introduces clock devices which can appear and > disappear like usb devices (and, I suppose, hot plug PCI but I am not > too sure that what is offered here would really work in that case). > The subsequent patches convert the clock_ and timer_ system calls, one > by one. > Sorry to jump in late, but this only just caught my attention. While I have no strong feelings on this series one way or the other, the naming is a bit unfortunate. The clock device / clkdev naming is already in use as an extension to the clock framework and is used by a wide variety of embedded platforms already, with a pending patch to move it in to the generic namespace (grepping for clkdev will give you an idea). The idea behind that interface is similar in that it deals with the dynamic creation and teardown of clocks, but is decoupled from timekeeping. It's also reasonable to assume that devices with dynamic clocks tracked through clkdev will wish to also use this interface in the timekeeping case, so it would be good to settle on something less ambiguous in advance.