From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Brownell Subject: Re: Wakeup-events implementation Date: Tue, 17 Aug 2010 11:37:22 -0700 (PDT) Message-ID: <324856.49167.qm@web180313.mail.gq1.yahoo.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: linux-pm-bounces@lists.linux-foundation.org Errors-To: linux-pm-bounces@lists.linux-foundation.org To: =?iso-8859-1?Q?Arve_Hj=F8nnev=E5g?= , Alan Stern Cc: Linux-pm mailing list List-Id: linux-pm@vger.kernel.org --- On Tue, 8/17/10, Alan Stern wrote: I wasn't > aware that memory cards could generate wakeup events. Almost all MMC/SD hardware can generate reliable insert/remove events, in my experience. So MMC/SD drivers I've modified tend to treat them as wake events ... userspace can respond without polling constantly "did it change yet, huh, did it???" Unfortunately Pierre spent a lot of time hard-wiring UNSAFE_RESUME assumptions into the MMC core, so lots of that stuff is currently (and needlessly) broken. And of more experimental interest, something I don't recall being discussed here: I/O completions as wake events. Start an I/O, enter low power state, and wake when it completes. Easily done on lots of embedded hardware..