From mboxrd@z Thu Jan 1 00:00:00 1970 From: Adrian Hunter Subject: Re: [PATCH 2/2] mmc: sdhci-pci: add runtime pm support Date: Mon, 17 Oct 2011 11:19:37 +0300 Message-ID: <4E9BE519.2020908@intel.com> References: <1317645214-12316-1-git-send-email-adrian.hunter@intel.com> <1317645214-12316-3-git-send-email-adrian.hunter@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from mga02.intel.com ([134.134.136.20]:18181 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750992Ab1JQITj (ORCPT ); Mon, 17 Oct 2011 04:19:39 -0400 In-Reply-To: Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Chris Ball Cc: linux-mmc On 16/10/11 05:26, Chris Ball wrote: > Hi, > > On Sat, Oct 15 2011, Chris Ball wrote: >> Card insertion detection after boot no longer works (nothing new appears >> in dmesg), so I'm unable to use the reader. Are you able to try to >> figure this out? I'm happy to try any patches you suggest. > > If I boot with a card inserted, I still see repeated "irq while runtime > suspended" messages for 80 seconds after boot, but the card *does* work. > I don't see further "irq while runtime suspended" messages during or > after I/O to the card at that point. > > So, we have a problem of unexpected irqs while runtime suspended, and a > problem of broken card detection when booted without a card inserted > (on my laptop). Unfortunately I do not have any information about the unexpected irqs. Card detect will not work because the interrupt is masked while the host controller is runtime suspended, on the grounds that the host controller may be powered off and so cannot handle interrupts. An alternate mechanism is needed to enable the host controller to wake-up from a card detect signal. So Runtime PM needs to be opt-in. I have sent a patch.