From: "Rafael J. Wysocki" <rjw@sisk.pl> To: "Ohad Ben-Cohen" <ohad@wizery.com> Cc: Alan Stern <stern@rowland.harvard.edu>, linux-pm@lists.linux-foundation.org, Johannes Berg <johannes@sipsolutions.net>, linux-wireless@vger.kernel.org, linux-mmc@vger.kernel.org, Ido Yariv <ido@wizery.com>, Kevin Hilman <khilman@deeprootsystems.com> Subject: Re: [linux-pm] subtle pm_runtime_put_sync race and sdio functions Date: Tue, 28 Dec 2010 21:04:06 +0100 [thread overview] Message-ID: <201012282104.06232.rjw@sisk.pl> (raw) In-Reply-To: <AANLkTinewad6YxXvn5gnJQ+HJn41Sg96319GQn9KRKWj@mail.gmail.com> On Tuesday, December 28, 2010, Ohad Ben-Cohen wrote: > On Sun, Dec 26, 2010 at 8:37 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote: > > So, it only happens during asynchronous suspend? In other words, if suspend > > is synchronous, everything should be fine, right? > > Not necessarily. So it's not a race after all, is it? > Consider this simple scenario, where a device was added after the mmc > host controller, but before mac80211. In this case its suspend handler > will have the chance to abort system suspend after mac80211 already > told our driver to power down the device (but the device wasn't > powered down yet, because the driver used pm_runtime_put_sync() which > is disabled). Well, first, you shouldn't rely on pm_runtime_put_sync() to actually _suspend_ the device at any point. What it does is to call pm_runtime_idle() for the device, which isn't guaranteed to suspend it. If you want the device to be suspended, you should use pm_runtime_put_noidle(device); pm_runtime_suspend(device); or, alternatively pm_runtime_put_sync_suspend(device); (which equivalent to the above pair of callbacks, but is not available in kernels prior to 2.6.37-rc1). Second, what you'd really want to do (I guess) is: pm_runtime_put_noidle(device); device->bus->pm->runtime_suspend(device); (I have omitted all of the usual checks for simplicity), because that would _unconditionally_ put your device into a low-power state. No? The problem is at this point the PM core will think the device is still RPM_ACTIVE, so it will be necessary to additionally do something like: pm_runtime_disable(device); pm_runtime_set_suspended(device); pm_runtime_enable(device); Of course, you'll need to ensure there are no races between that and any other code path that may want to resume the device simultaneously. And here it backfires, because you have to synchronize not only with runtime resume, but also with system suspend and possibly resume. Rafael
WARNING: multiple messages have this Message-ID (diff)
From: "Rafael J. Wysocki" <rjw-KKrjLPT3xs0@public.gmane.org> To: Ohad Ben-Cohen <ohad-Ix1uc/W3ht7QT0dZR+AlfA@public.gmane.org> Cc: Alan Stern <stern-nwvwT67g6+6dFdvTe/nMLpVzexx5G7lz@public.gmane.org>, linux-pm-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org, Johannes Berg <johannes-cdvu00un1VgdHxzADdlk8Q@public.gmane.org>, linux-wireless-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-mmc-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Ido Yariv <ido-Ix1uc/W3ht7QT0dZR+AlfA@public.gmane.org>, Kevin Hilman <khilman-1D3HCaltpLuhEniVeURVKkEOCMrvLtNR@public.gmane.org> Subject: Re: [linux-pm] subtle pm_runtime_put_sync race and sdio functions Date: Tue, 28 Dec 2010 21:04:06 +0100 [thread overview] Message-ID: <201012282104.06232.rjw@sisk.pl> (raw) In-Reply-To: <AANLkTinewad6YxXvn5gnJQ+HJn41Sg96319GQn9KRKWj-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org> On Tuesday, December 28, 2010, Ohad Ben-Cohen wrote: > On Sun, Dec 26, 2010 at 8:37 PM, Rafael J. Wysocki <rjw-KKrjLPT3xs0@public.gmane.org> wrote: > > So, it only happens during asynchronous suspend? In other words, if suspend > > is synchronous, everything should be fine, right? > > Not necessarily. So it's not a race after all, is it? > Consider this simple scenario, where a device was added after the mmc > host controller, but before mac80211. In this case its suspend handler > will have the chance to abort system suspend after mac80211 already > told our driver to power down the device (but the device wasn't > powered down yet, because the driver used pm_runtime_put_sync() which > is disabled). Well, first, you shouldn't rely on pm_runtime_put_sync() to actually _suspend_ the device at any point. What it does is to call pm_runtime_idle() for the device, which isn't guaranteed to suspend it. If you want the device to be suspended, you should use pm_runtime_put_noidle(device); pm_runtime_suspend(device); or, alternatively pm_runtime_put_sync_suspend(device); (which equivalent to the above pair of callbacks, but is not available in kernels prior to 2.6.37-rc1). Second, what you'd really want to do (I guess) is: pm_runtime_put_noidle(device); device->bus->pm->runtime_suspend(device); (I have omitted all of the usual checks for simplicity), because that would _unconditionally_ put your device into a low-power state. No? The problem is at this point the PM core will think the device is still RPM_ACTIVE, so it will be necessary to additionally do something like: pm_runtime_disable(device); pm_runtime_set_suspended(device); pm_runtime_enable(device); Of course, you'll need to ensure there are no races between that and any other code path that may want to resume the device simultaneously. And here it backfires, because you have to synchronize not only with runtime resume, but also with system suspend and possibly resume. Rafael -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2010-12-28 20:04 UTC|newest] Thread overview: 179+ messages / expand[flat|nested] mbox.gz Atom feed top 2010-12-09 23:37 subtle pm_runtime_put_sync race and sdio functions Ohad Ben-Cohen 2010-12-10 0:00 ` Rafael J. Wysocki 2010-12-10 0:00 ` [linux-pm] " Rafael J. Wysocki 2010-12-10 17:24 ` Alan Stern 2010-12-10 17:24 ` [linux-pm] " Alan Stern 2010-12-10 22:01 ` Rafael J. Wysocki 2010-12-10 22:01 ` [linux-pm] " Rafael J. Wysocki 2010-12-10 23:02 ` Ohad Ben-Cohen 2010-12-10 23:02 ` [linux-pm] " Ohad Ben-Cohen 2010-12-10 22:59 ` Ohad Ben-Cohen 2010-12-10 22:59 ` [linux-pm] " Ohad Ben-Cohen 2010-12-11 1:17 ` Ohad Ben-Cohen 2010-12-11 14:53 ` Alan Stern 2010-12-11 1:17 ` Ohad Ben-Cohen 2010-12-11 14:50 ` Alan Stern 2010-12-18 13:29 ` Ohad Ben-Cohen 2010-12-18 13:29 ` [linux-pm] " Ohad Ben-Cohen 2010-12-18 14:16 ` David Vrabel 2010-12-18 15:12 ` Ohad Ben-Cohen 2010-12-18 15:12 ` Ohad Ben-Cohen 2010-12-18 14:16 ` David Vrabel 2010-12-18 15:07 ` [linux-pm] " Rafael J. Wysocki 2010-12-18 16:00 ` Ohad Ben-Cohen 2010-12-18 16:00 ` [linux-pm] " Ohad Ben-Cohen 2010-12-18 16:40 ` Johannes Berg 2010-12-18 19:08 ` Ohad Ben-Cohen 2010-12-18 19:08 ` Ohad Ben-Cohen 2010-12-18 21:30 ` Alan Stern 2010-12-18 21:30 ` Alan Stern 2010-12-18 22:57 ` Rafael J. Wysocki 2010-12-18 22:57 ` [linux-pm] " Rafael J. Wysocki 2010-12-18 21:30 ` Alan Stern 2010-12-18 22:52 ` Rafael J. Wysocki 2010-12-18 22:52 ` [linux-pm] " Rafael J. Wysocki 2010-12-18 19:08 ` Ohad Ben-Cohen 2010-12-18 21:29 ` [linux-pm] " Alan Stern 2010-12-18 21:29 ` Alan Stern 2010-12-18 21:29 ` Alan Stern 2010-12-18 22:50 ` Rafael J. Wysocki 2010-12-18 22:50 ` [linux-pm] " Rafael J. Wysocki 2010-12-18 16:40 ` Johannes Berg 2010-12-18 22:47 ` [linux-pm] " Rafael J. Wysocki 2010-12-18 22:47 ` Rafael J. Wysocki 2010-12-19 7:48 ` Ohad Ben-Cohen 2010-12-19 7:48 ` Ohad Ben-Cohen 2010-12-19 7:48 ` Ohad Ben-Cohen 2010-12-19 10:22 ` Rafael J. Wysocki 2010-12-19 10:22 ` [linux-pm] " Rafael J. Wysocki 2010-12-20 3:37 ` Alan Stern 2010-12-20 3:37 ` Alan Stern 2010-12-20 21:17 ` Rafael J. Wysocki 2010-12-21 0:57 ` Alan Stern 2010-12-21 0:57 ` Alan Stern 2010-12-21 21:31 ` Rafael J. Wysocki 2010-12-22 1:42 ` Alan Stern 2010-12-22 1:42 ` Alan Stern 2010-12-22 12:29 ` Rafael J. Wysocki 2010-12-22 12:29 ` Rafael J. Wysocki 2011-01-26 23:28 ` Kevin Hilman 2011-01-26 23:28 ` Kevin Hilman 2011-01-27 18:13 ` Alan Stern 2011-01-27 18:13 ` [linux-pm] " Alan Stern 2011-01-27 18:13 ` Alan Stern 2011-01-27 19:22 ` Kevin Hilman 2011-01-27 19:22 ` [linux-pm] " Kevin Hilman 2011-01-27 19:22 ` Kevin Hilman 2011-01-27 19:49 ` Alan Stern 2011-01-27 19:49 ` [linux-pm] " Alan Stern 2011-01-27 19:49 ` Alan Stern 2011-01-27 20:15 ` Kevin Hilman 2011-01-27 20:15 ` [linux-pm] " Kevin Hilman 2011-01-27 20:15 ` Kevin Hilman 2011-01-27 22:18 ` Vitaly Wool 2011-01-27 22:18 ` Vitaly Wool 2011-01-27 22:18 ` Vitaly Wool 2011-01-27 23:21 ` Rafael J. Wysocki 2011-01-27 23:21 ` [linux-pm] " Rafael J. Wysocki 2011-01-27 23:49 ` Kevin Hilman 2011-01-27 23:49 ` [linux-pm] " Kevin Hilman 2011-01-27 23:11 ` Rafael J. Wysocki 2011-01-27 23:11 ` [linux-pm] " Rafael J. Wysocki 2011-01-27 18:20 ` Vitaly Wool 2011-01-27 18:20 ` [linux-pm] " Vitaly Wool 2011-01-27 18:20 ` Vitaly Wool 2011-01-27 18:54 ` Kevin Hilman 2011-01-27 18:54 ` [linux-pm] " Kevin Hilman 2010-12-22 12:29 ` Rafael J. Wysocki 2010-12-22 1:42 ` Alan Stern 2010-12-21 21:31 ` Rafael J. Wysocki 2010-12-21 0:57 ` Alan Stern 2010-12-20 21:17 ` Rafael J. Wysocki 2010-12-20 3:37 ` Alan Stern 2010-12-21 22:23 ` Kevin Hilman 2010-12-21 22:23 ` [linux-pm] " Kevin Hilman 2010-12-22 1:48 ` Alan Stern 2010-12-22 1:48 ` Alan Stern 2010-12-22 1:48 ` Alan Stern 2010-12-23 7:51 ` Ohad Ben-Cohen 2010-12-23 7:51 ` [linux-pm] " Ohad Ben-Cohen 2010-12-23 16:03 ` Alan Stern 2010-12-23 16:03 ` [linux-pm] " Alan Stern 2010-12-23 16:03 ` Alan Stern 2010-12-25 7:34 ` Ohad Ben-Cohen 2010-12-25 7:34 ` Ohad Ben-Cohen 2010-12-25 16:21 ` Alan Stern 2010-12-25 16:21 ` Alan Stern 2010-12-25 20:58 ` Ohad Ben-Cohen 2010-12-25 20:58 ` [linux-pm] " Ohad Ben-Cohen 2010-12-25 20:58 ` Ohad Ben-Cohen 2010-12-25 21:50 ` Vitaly Wool 2010-12-25 21:50 ` [linux-pm] " Vitaly Wool 2010-12-26 5:27 ` Ohad Ben-Cohen 2010-12-26 5:27 ` [linux-pm] " Ohad Ben-Cohen 2010-12-25 21:54 ` Vitaly Wool 2010-12-25 21:54 ` Vitaly Wool 2010-12-25 21:54 ` Vitaly Wool 2010-12-26 2:48 ` [linux-pm] " Alan Stern 2010-12-26 2:48 ` Alan Stern 2010-12-26 5:55 ` Ohad Ben-Cohen 2010-12-26 5:55 ` [linux-pm] " Ohad Ben-Cohen 2010-12-26 5:55 ` Ohad Ben-Cohen 2010-12-26 11:45 ` Rafael J. Wysocki 2010-12-26 12:43 ` Ohad Ben-Cohen 2010-12-26 12:43 ` [linux-pm] " Ohad Ben-Cohen 2010-12-26 12:43 ` Ohad Ben-Cohen 2010-12-26 18:35 ` Rafael J. Wysocki 2010-12-28 19:11 ` Ohad Ben-Cohen 2010-12-28 19:11 ` [linux-pm] " Ohad Ben-Cohen 2010-12-28 19:21 ` Rafael J. Wysocki 2010-12-28 19:21 ` [linux-pm] " Rafael J. Wysocki 2010-12-28 19:21 ` Rafael J. Wysocki 2010-12-28 19:34 ` Ohad Ben-Cohen 2010-12-28 20:36 ` Rafael J. Wysocki 2010-12-28 20:36 ` [linux-pm] " Rafael J. Wysocki 2010-12-28 19:34 ` Ohad Ben-Cohen 2010-12-26 18:35 ` Rafael J. Wysocki 2010-12-26 14:53 ` [linux-pm] " Ohad Ben-Cohen 2010-12-26 18:37 ` Rafael J. Wysocki 2010-12-26 18:37 ` [linux-pm] " Rafael J. Wysocki 2010-12-28 19:15 ` Ohad Ben-Cohen 2010-12-28 20:04 ` Rafael J. Wysocki 2010-12-28 20:04 ` Rafael J. Wysocki [this message] 2010-12-28 20:04 ` [linux-pm] " Rafael J. Wysocki 2010-12-28 20:41 ` Ohad Ben-Cohen 2010-12-28 20:41 ` [linux-pm] " Ohad Ben-Cohen 2010-12-28 20:41 ` Ohad Ben-Cohen 2010-12-28 19:15 ` Ohad Ben-Cohen 2010-12-26 14:53 ` Ohad Ben-Cohen 2010-12-26 11:45 ` Rafael J. Wysocki 2010-12-26 17:00 ` [linux-pm] " Alan Stern 2010-12-26 17:00 ` Alan Stern 2010-12-28 19:04 ` Ohad Ben-Cohen 2010-12-28 19:04 ` [linux-pm] " Ohad Ben-Cohen 2010-12-28 19:04 ` Ohad Ben-Cohen 2010-12-28 21:46 ` Alan Stern 2010-12-28 21:46 ` [linux-pm] " Alan Stern 2010-12-28 21:46 ` Alan Stern 2010-12-29 6:34 ` Ohad Ben-Cohen 2010-12-30 4:25 ` Alan Stern 2010-12-30 4:25 ` Alan Stern 2010-12-30 4:25 ` Alan Stern 2010-12-29 6:34 ` Ohad Ben-Cohen 2010-12-29 8:01 ` Ohad Ben-Cohen 2010-12-29 8:01 ` [linux-pm] " Ohad Ben-Cohen 2010-12-30 4:30 ` Alan Stern 2010-12-30 4:30 ` [linux-pm] " Alan Stern 2010-12-30 4:30 ` Alan Stern 2010-12-26 17:00 ` Alan Stern 2010-12-26 2:48 ` Alan Stern 2010-12-25 16:21 ` Alan Stern 2010-12-25 7:34 ` Ohad Ben-Cohen 2010-12-18 22:47 ` Rafael J. Wysocki 2010-12-18 15:07 ` Rafael J. Wysocki 2010-12-18 21:20 ` [linux-pm] " Alan Stern 2010-12-18 23:03 ` Rafael J. Wysocki 2010-12-18 23:03 ` Rafael J. Wysocki 2010-12-19 10:00 ` [linux-pm] " Ohad Ben-Cohen 2010-12-19 10:00 ` Ohad Ben-Cohen 2010-12-18 21:20 ` Alan Stern
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=201012282104.06232.rjw@sisk.pl \ --to=rjw@sisk.pl \ --cc=ido@wizery.com \ --cc=johannes@sipsolutions.net \ --cc=khilman@deeprootsystems.com \ --cc=linux-mmc@vger.kernel.org \ --cc=linux-pm@lists.linux-foundation.org \ --cc=linux-wireless@vger.kernel.org \ --cc=ohad@wizery.com \ --cc=stern@rowland.harvard.edu \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.