From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Johannes Berg <johannes@sipsolutions.net>,
linux-pm <linux-pm@lists.linux-foundation.org>,
Ralf Baechle <ralf@linux-mips.org>,
Jamey Hicks <jamey@crl.dec.com>
Subject: Re: [RFC] apm-emulation: implement notify/ack for /sys/power/state events
Date: Mon, 17 Dec 2007 07:03:00 +1100 [thread overview]
Message-ID: <1197835380.21886.0.camel@pasglop> (raw)
In-Reply-To: <200712162015.59963.rjw@sisk.pl>
On Sun, 2007-12-16 at 20:15 +0100, Rafael J. Wysocki wrote:
> > + *
> > + * Unfortunately we cannot do a timeout because then we'd
> > + * suspend again right away if the process that had apm_bios
> > + * open and that we timed out waiting for "acknowledges" the
> > + * event after we have resumed. If suspend doesn't work because
> > + * of a rogue process, just kill that process.
> > + *
> > + * FIXME: is the suspends_pending == 1 test racy?
I think we can do the timeout thing easily... Just mark the fd's that
haven't ack'ed with a flag that makes us ignore the next ack...
Ben.
> Make suspends_pending atomic? Then, you'd not need the locking around it.
>
> > + */
> > + err = wait_event_interruptible(apm_suspend_waitqueue,
> > + suspends_pending == 1);
> > +
> > + mutex_lock(&state_lock);
> > + suspends_pending--;
> > + mutex_unlock(&state_lock);
> > +
> > + if (!err)
> > + return NOTIFY_OK;
> > +
> > + /* interrupted by signal */
> > + return NOTIFY_BAD;
> > +
> > + case PM_POST_SUSPEND:
> > + /* TODO: maybe grab error code, needs core changes */
> > + apm_after_resume(0);
> > + return NOTIFY_OK;
> > +
> > + default:
> > + return NOTIFY_DONE;
> > + }
> > +}
> > +
> > +struct notifier_block apm_notif_block = {
> > + .notifier_call = apm_suspend_notifier,
> > +};
> > +
> > static int __init apm_init(void)
> > {
> > int ret;
> > @@ -588,7 +671,7 @@ static int __init apm_init(void)
> > if (IS_ERR(kapmd_tsk)) {
> > ret = PTR_ERR(kapmd_tsk);
> > kapmd_tsk = NULL;
> > - return ret;
> > + goto out;
> > }
> > wake_up_process(kapmd_tsk);
> >
> > @@ -597,16 +680,27 @@ static int __init apm_init(void)
> > #endif
> >
> > ret = misc_register(&apm_device);
> > - if (ret != 0) {
> > - remove_proc_entry("apm", NULL);
> > - kthread_stop(kapmd_tsk);
> > - }
> > + if (ret)
> > + goto out_stop;
> > +
> > + ret = register_pm_notifier(&apm_notif_block);
> > + if (ret)
> > + goto out_unregister;
> >
> > + return 0;
> > +
> > + out_unregister:
> > + misc_deregister(&apm_device);
> > + out_stop:
> > + remove_proc_entry("apm", NULL);
> > + kthread_stop(kapmd_tsk);
> > + out:
> > return ret;
> > }
> >
> > static void __exit apm_exit(void)
> > {
> > + unregister_pm_notifier(&apm_notif_block);
> > misc_deregister(&apm_device);
> > remove_proc_entry("apm", NULL);
> >
> > --- everything.orig/kernel/power/main.c 2007-12-16
> 15:23:18.689514811 +0100
> > +++ everything/kernel/power/main.c 2007-12-16 15:31:06.179512207
> +0100
> > @@ -25,6 +25,7 @@
> > #include "power.h"
> >
> > BLOCKING_NOTIFIER_HEAD(pm_chain_head);
> > +EXPORT_SYMBOL_GPL(pm_chain_head);
> >
> > DEFINE_MUTEX(pm_mutex);
>
next prev parent reply other threads:[~2007-12-16 20:03 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-12-16 17:58 [RFC] apm-emulation: implement notify/ack for /sys/power/state events Johannes Berg
2007-12-16 19:15 ` Rafael J. Wysocki
2007-12-16 18:59 ` Johannes Berg
2007-12-16 19:34 ` Johannes Berg
2007-12-16 20:56 ` Rafael J. Wysocki
2007-12-16 20:49 ` Johannes Berg
2007-12-16 21:03 ` Johannes Berg
2007-12-16 20:03 ` Benjamin Herrenschmidt [this message]
2007-12-16 20:09 ` Johannes Berg
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=1197835380.21886.0.camel@pasglop \
--to=benh@kernel.crashing.org \
--cc=jamey@crl.dec.com \
--cc=johannes@sipsolutions.net \
--cc=linux-pm@lists.linux-foundation.org \
--cc=ralf@linux-mips.org \
--cc=rjw@sisk.pl \
/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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox