[PATCH V2] b43legacy: Don't cancel the restart workqueue in wireless_core_exit

[PATCH V2] b43legacy: Don't cancel the restart workqueue in wireless_core_exit

[Larry Finger]

From: Michael Buesch <mb@bu3sch.de>

The wq must be canceled later on rmmod. It's nonfatal, if
the wq runs on a device that's not started or down. It will
handle these cases.
But syncing in wireless_core_exit() will cause a deadlock with
the restart_work. (restart work cancels itself)

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: Larry Finger <larry.finger@lwfinger.net>
---

John,

Sorry, but I sent the bare patch in the first version.

Larry

 drivers/net/wireless/b43legacy/main.c |    2 ++
 1 file changed, 2 insertions(+)

Index: wireless-dev/drivers/net/wireless/b43legacy/main.c
===================================================================
--- wireless-dev.orig/drivers/net/wireless/b43legacy/main.c
+++ wireless-dev/drivers/net/wireless/b43legacy/main.c
@@ -3021,6 +3021,7 @@ static void b43legacy_wireless_core_exit
 	B43legacy_WARN_ON(b43legacy_status(dev) > B43legacy_STAT_INITIALIZED);
 	if (b43legacy_status(dev) != B43legacy_STAT_INITIALIZED)
 		return;
+	b43legacy_set_status(dev, B43legacy_STAT_UNINIT);
 
 	b43legacy_rng_exit(dev->wl);
 	b43legacy_pio_free(dev);
@@ -3520,6 +3521,7 @@ static void b43legacy_one_core_detach(st
 
 	wldev = ssb_get_drvdata(dev);
 	wl = wldev->wl;
+	cancel_work_sync(&wldev->restart_work);
 	b43legacy_debugfs_remove_device(wldev);
 	b43legacy_wireless_core_detach(wldev);
 	list_del(&wldev->list);

Re: [PATCH V2] b43legacy: Don't cancel the restart workqueue in wireless_core_exit

[John W. Linville]

Is this the correct patch?  The first hunk conflicts with an earlier
patch ("b43legacy: Fix cancellation of work queues").

John

On Mon, Sep 10, 2007 at 11:59:20AM -0500, Larry Finger wrote:
> From: Michael Buesch <mb@bu3sch.de>
> 
> The wq must be canceled later on rmmod. It's nonfatal, if
> the wq runs on a device that's not started or down. It will
> handle these cases.
> But syncing in wireless_core_exit() will cause a deadlock with
> the restart_work. (restart work cancels itself)
> 
> Signed-off-by: Michael Buesch <mb@bu3sch.de>
> Signed-off-by: Larry Finger <larry.finger@lwfinger.net>
> ---
> 
> John,
> 
> Sorry, but I sent the bare patch in the first version.
> 
> Larry
> 
>  drivers/net/wireless/b43legacy/main.c |    2 ++
>  1 file changed, 2 insertions(+)
> 
> Index: wireless-dev/drivers/net/wireless/b43legacy/main.c
> ===================================================================
> --- wireless-dev.orig/drivers/net/wireless/b43legacy/main.c
> +++ wireless-dev/drivers/net/wireless/b43legacy/main.c
> @@ -3021,6 +3021,7 @@ static void b43legacy_wireless_core_exit
>  	B43legacy_WARN_ON(b43legacy_status(dev) > B43legacy_STAT_INITIALIZED);
>  	if (b43legacy_status(dev) != B43legacy_STAT_INITIALIZED)
>  		return;
> +	b43legacy_set_status(dev, B43legacy_STAT_UNINIT);
>  
>  	b43legacy_rng_exit(dev->wl);
>  	b43legacy_pio_free(dev);
> @@ -3520,6 +3521,7 @@ static void b43legacy_one_core_detach(st
>  
>  	wldev = ssb_get_drvdata(dev);
>  	wl = wldev->wl;
> +	cancel_work_sync(&wldev->restart_work);
>  	b43legacy_debugfs_remove_device(wldev);
>  	b43legacy_wireless_core_detach(wldev);
>  	list_del(&wldev->list);
> -
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
John W. Linville
linville@tuxdriver.com

Re: [PATCH V2] b43legacy: Don't cancel the restart workqueue in wireless_core_exit

[Larry Finger]

John W. Linville wrote:
> Is this the correct patch?  The first hunk conflicts with an earlier
> patch ("b43legacy: Fix cancellation of work queues").
> 
> John

Yes, this one is correct. The first one was without a commit message, and had an error.

Larry

Re: [PATCH V2] b43legacy: Don't cancel the restart workqueue in wireless_core_exit

[John W. Linville]

On Mon, Sep 17, 2007 at 12:05:36PM -0500, Larry Finger wrote:
> John W. Linville wrote:
> > Is this the correct patch?  The first hunk conflicts with an earlier
> > patch ("b43legacy: Fix cancellation of work queues").
> > 
> > John
> 
> Yes, this one is correct. The first one was without a commit message, and had an error.

Hmmmm, not that one.  There was another patch that (looking closer)
you sent me and cc'ed bcm43xx-dev on or about 30 August.  I guess
I'll drop that one and use this one instead.

Thanks,

John
-- 
John W. Linville
linville@tuxdriver.com

Re: [PATCH V2] b43legacy: Don't cancel the restart workqueue in wireless_core_exit

[Larry Finger]

John W. Linville wrote:
> On Mon, Sep 17, 2007 at 12:05:36PM -0500, Larry Finger wrote:
>> John W. Linville wrote:
>>> Is this the correct patch?  The first hunk conflicts with an earlier
>>> patch ("b43legacy: Fix cancellation of work queues").
>>>
>>> John
>> Yes, this one is correct. The first one was without a commit message, and had an error.
> 
> Hmmmm, not that one.  There was another patch that (looking closer)
> you sent me and cc'ed bcm43xx-dev on or about 30 August.  I guess
> I'll drop that one and use this one instead.

No, the Aug 30 one is needed. Somehow my patch sets got confused. Please apply the August 30 patch
and drop the Sept. 10 one. I'll fix it after your tree gets updated. I don't think it will cause any
problems. As you said, the first hunk is already in the Aug. 30 patch.

Larry