From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755753Ab1KAAhe (ORCPT ); Mon, 31 Oct 2011 20:37:34 -0400 Received: from mail-gx0-f174.google.com ([209.85.161.174]:45713 "EHLO mail-gx0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753611Ab1KAAhc (ORCPT ); Mon, 31 Oct 2011 20:37:32 -0400 Date: Mon, 31 Oct 2011 17:37:26 -0700 From: Tejun Heo To: "Rafael J. Wysocki" , linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org, linux-usb@vger.kernel.org, Seth Forshee , Oleg Nesterov , Alan Stern , Greg Kroah-Hartman Subject: [PATCH pm 1/2] usb_storage: don't use set_freezable_with_signal() Message-ID: <20111101003726.GM18855@google.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org The current implementation of set_freezable_with_signal() is buggy and tricky to get right. usb-storage is the only user and its use can be avoided trivially. All usb-storage wants is to be able to sleep with timeout and get woken up if freezing() becomes true. This can be trivially implemented by doing interruptible wait w/ freezing() included in the wait condition. There's no reason to use set_freezable_with_signal(). Perform interruptible wait on freezing() instead of using set_freezable_with_signal(), which is scheduled for removal. Signed-off-by: Tejun Heo Cc: Oleg Nesterov Cc: "Rafael J. Wysocki" Cc: Seth Forshee Cc: Alan Stern Cc: Greg Kroah-Hartman --- These two patches are on top of "freezer: fix various bugs and simplify implementation, take#2" patchset[1] and are also available in the following git branch. git://git.kernel.org/pub/scm/linux/kernel/git/tj/misc.git pm-kill-freezable_with_signal If usb-storage ppl are okay with it, I think routing this through pm would be the easiest. Oh, and this definitely is for the next merge window. Thank you. [1] http://thread.gmane.org/gmane.linux.kernel/1209247 drivers/usb/storage/usb.c | 13 +++++++------ 1 files changed, 7 insertions(+), 6 deletions(-) diff --git a/drivers/usb/storage/usb.c b/drivers/usb/storage/usb.c index c325e69..aa84b3d 100644 --- a/drivers/usb/storage/usb.c +++ b/drivers/usb/storage/usb.c @@ -831,7 +831,8 @@ static int usb_stor_scan_thread(void * __us) dev_dbg(dev, "device found\n"); - set_freezable_with_signal(); + set_freezable(); + /* * Wait for the timeout to expire or for a disconnect * @@ -839,16 +840,16 @@ static int usb_stor_scan_thread(void * __us) * fail to freeze, but we can't be non-freezable either. Nor can * khubd freeze while waiting for scanning to complete as it may * hold the device lock, causing a hang when suspending devices. - * So we request a fake signal when freezing and use - * interruptible sleep to kick us out of our wait early when - * freezing happens. + * So instead of using wait_event_freezable(), explicitly test + * for (DONT_SCAN || freezing) in interruptible wait and proceed + * if any of DONT_SCAN, freezing or timeout has happened. */ if (delay_use > 0) { dev_dbg(dev, "waiting for device to settle " "before scanning\n"); wait_event_interruptible_timeout(us->delay_wait, - test_bit(US_FLIDX_DONT_SCAN, &us->dflags), - delay_use * HZ); + test_bit(US_FLIDX_DONT_SCAN, &us->dflags) || + freezing(current), delay_use * HZ); } /* If the device is still connected, perform the scanning */ -- 1.7.3.1