linux-kernel - Re: [Bug #13337] [post 2.6.29 regression] hang during suspend of b44/b43 modules

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <1247334464.30647.109.camel@johannes.local>
Date:	Sat, 11 Jul 2009 19:47:44 +0200
From:	Johannes Berg <johannes@...solutions.net>
To:	Jan Scholz <scholz@...s.uni-frankfurt.de>
Cc:	"Rafael J. Wysocki" <rjw@...k.pl>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Kernel Testers List <kernel-testers@...r.kernel.org>,
	Tomas Janousek <tomi@...i.cz>,
	"John W. Linville" <linville@...driver.com>
Subject: Re: [Bug #13337] [post 2.6.29 regression] hang during suspend of
 b44/b43 modules

On Sat, 2009-07-11 at 19:20 +0200, Jan Scholz wrote:
> Johannes Berg <johannes@...solutions.net> writes:
> 
> > On Sat, 2009-07-11 at 16:07 +0200, Jan Scholz wrote:
> >
> >> > Jan, can you run with console_suspend=0 (or whatever it is now, Rafael
> >> > will correct me if I'm wrong) that should work on your platform, and put
> >> > a bunch of printk calls into __ieee80211_suspend in net/mac80211/pm.c to
> >> > see where in there it's hanging?
> >> 
> >> Did that, seems like it hangs in
> >>   drv_remove_interface(local, &conf);
> >
> > Ok, thanks! I was going to suggest doing the same in
> > b43_op_remove_interface in drivers/net/wireless/b43/main.c, but I think
> > the problem is just that we have an ordering problem and call drv_stop
> > before drv_remove_interface. I'll try to come up with a fix soon --
> > might not be able to today though, in the meantime I guess you can just
> > move the drv_stop call to the end of the function and tell me if that
> > works -- that would be racy but should be ok most of the time.
> 
> I moved drv_stop to the end of the function (see patch below) and now
> suspend works fine.

Thanks, that's what I suspected -- I'll take a closer look into what is
required to make it race-free against RX.

johannes

> Jan
> 
> diff --git a/net/mac80211/pm.c b/net/mac80211/pm.c
> index 7a549f9..0ac15fb 100644
> --- a/net/mac80211/pm.c
> +++ b/net/mac80211/pm.c
> @@ -58,12 +58,6 @@ int __ieee80211_suspend(struct ieee80211_hw *hw)
>  	/* flush again, in case driver queued work */
>  	flush_workqueue(local->hw.workqueue);
>  
> -	/* stop hardware - this must stop RX */
> -	if (local->open_count) {
> -		ieee80211_led_radio(local, false);
> -		drv_stop(local);
> -	}
> -
>  	/* remove STAs */
>  	spin_lock_irqsave(&local->sta_lock, flags);
>  	list_for_each_entry(sta, &local->sta_list, list) {
> @@ -111,6 +105,12 @@ int __ieee80211_suspend(struct ieee80211_hw *hw)
>  		drv_remove_interface(local, &conf);
>  	}
>  
> +	/* stop hardware - this must stop RX */
> +	if (local->open_count) {
> +		ieee80211_led_radio(local, false);
> +		drv_stop(local);
> +	}
> +
>  	local->suspended = true;
>  	local->quiescing = false;
> 

Download attachment "signature.asc" of type "application/pgp-signature" (802 bytes)