From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Mon, 6 Sep 2010 21:34:31 -0700 From: Arjan van de Ven To: "Dong, Chuanxiao" Subject: Re: [RFC][PATCH 1/1]nand/denali: Add runtime pm support for denali controller driver Message-ID: <20100906213431.3758f14f@infradead.org> In-Reply-To: <5D8008F58939784290FAB48F54975198278A3799CB@shsmsx502.ccr.corp.intel.com> References: <20100906080838.GA6528@intel.com> <20100906125327.036fbcaf@infradead.org> <5D8008F58939784290FAB48F54975198278A3798F3@shsmsx502.ccr.corp.intel.com> <20100906203453.453b1da2@infradead.org> <5D8008F58939784290FAB48F54975198278A3799CB@shsmsx502.ccr.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: "linux-mtd@lists.infradead.org" , "dwmw2@infradead.org" , "Gao, Yunpeng" , "Yuan, Hang" List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Tue, 7 Sep 2010 12:03:40 +0800 "Dong, Chuanxiao" wrote: > > > if you take the refcount every time you get a command, and drop a > > count each time a command completes.... this is all not too > > important.. it'll just work > If I take the refcount every time I got a command and drop a count > each time a command completes, the controller device will be resumed > every time when it gets a command and suspended after it finishes a > command. For a 4k page NAND, processing just 1M data will cause > device resume/suspend 256 times....This will influence the > performance... so this is where you can tell the runtime PM system to use a timer, and only turn on power saving after some idle time. you don't need to do that yourself... the runtime PM system has this capability built in! -- Arjan van de Ven Intel Open Source Technology Centre For development, discussion and tips for power savings, visit http://www.lesswatts.org