This fixes the problem causing the following trace:
------------[ cut here ]------------
WARNING: at linux-2.6.34/net/wireless/core.c:633 wdev_cleanup_work+0xb7/0xe0 [cfg80211]()
Hardware name: Latitude C840
Pid: 707, comm: cfg80211 Not tainted 2.6.34.7-0.5-desktop #1
Call Trace:
[<c02065c3>] try_stack_unwind+0x173/0x190
[<c02051cf>] dump_trace+0x3f/0xe0
[<c020662b>] show_trace_log_lvl+0x4b/0x60
[<c0206658>] show_trace+0x18/0x20
[<c064e0b3>] dump_stack+0x6d/0x72
[<c02443ae>] warn_slowpath_common+0x6e/0xb0
[<c0244403>] warn_slowpath_null+0x13/0x20
[<e2db5497>] wdev_cleanup_work+0xb7/0xe0 [cfg80211]
[<c025cfa9>] run_workqueue+0x79/0x170
[<c025d123>] worker_thread+0x83/0xe0
[<c025fef4>] kthread+0x74/0x80
[<c0203826>] kernel_thread_helper+0x6/0x10
---[ end trace 3f0348b3b0c6f4ff ]---
Reported by: Giacomo Comes <[email protected]>
Signed-off-by: David Kilroy <[email protected]>
---
This patch has only been compile tested. Given the trace, I've
guessed that cfg80211 would prefer that orinoco abort the scan on
interface down.
---
drivers/net/wireless/orinoco/main.c | 12 +++++-------
drivers/net/wireless/orinoco/scan.c | 8 ++++++++
drivers/net/wireless/orinoco/scan.h | 1 +
3 files changed, 14 insertions(+), 7 deletions(-)
diff --git a/drivers/net/wireless/orinoco/main.c b/drivers/net/wireless/orinoco/main.c
index e8e2d0f..fa0cf74 100644
--- a/drivers/net/wireless/orinoco/main.c
+++ b/drivers/net/wireless/orinoco/main.c
@@ -1392,10 +1392,9 @@ static void orinoco_process_scan_results(struct work_struct *work)
orinoco_add_hostscan_results(priv, buf, len);
kfree(buf);
- } else if (priv->scan_request) {
+ } else {
/* Either abort or complete the scan */
- cfg80211_scan_done(priv->scan_request, (len < 0));
- priv->scan_request = NULL;
+ orinoco_scan_done(priv, (len < 0));
}
spin_lock_irqsave(&priv->scan_lock, flags);
@@ -1684,6 +1683,8 @@ static int __orinoco_down(struct orinoco_private *priv)
hermes_write_regn(hw, EVACK, 0xffff);
}
+ orinoco_scan_done(priv, true);
+
/* firmware will have to reassociate */
netif_carrier_off(dev);
priv->last_linkstatus = 0xffff;
@@ -1762,10 +1763,7 @@ void orinoco_reset(struct work_struct *work)
orinoco_unlock(priv, &flags);
/* Scanning support: Notify scan cancellation */
- if (priv->scan_request) {
- cfg80211_scan_done(priv->scan_request, 1);
- priv->scan_request = NULL;
- }
+ orinoco_scan_done(priv, true);
if (priv->hard_reset) {
err = (*priv->hard_reset)(priv);
diff --git a/drivers/net/wireless/orinoco/scan.c b/drivers/net/wireless/orinoco/scan.c
index 4300d9d..86cb54c 100644
--- a/drivers/net/wireless/orinoco/scan.c
+++ b/drivers/net/wireless/orinoco/scan.c
@@ -229,3 +229,11 @@ void orinoco_add_hostscan_results(struct orinoco_private *priv,
priv->scan_request = NULL;
}
}
+
+void orinoco_scan_done(struct orinoco_private *priv, bool abort)
+{
+ if (priv->scan_request) {
+ cfg80211_scan_done(priv->scan_request, abort);
+ priv->scan_request = NULL;
+ }
+}
diff --git a/drivers/net/wireless/orinoco/scan.h b/drivers/net/wireless/orinoco/scan.h
index 2dc4e04..27281fb 100644
--- a/drivers/net/wireless/orinoco/scan.h
+++ b/drivers/net/wireless/orinoco/scan.h
@@ -16,5 +16,6 @@ void orinoco_add_extscan_result(struct orinoco_private *priv,
void orinoco_add_hostscan_results(struct orinoco_private *dev,
unsigned char *buf,
size_t len);
+void orinoco_scan_done(struct orinoco_private *priv, bool abort);
#endif /* _ORINOCO_SCAN_H_ */
--
1.7.2.2
On Mon, Nov 29, 2010 at 8:26 PM, John W. Linville
<[email protected]> wrote:
> On Wed, Nov 24, 2010 at 08:33:02PM +0000, David Kilroy wrote:
>> This fixes the problem causing the following trace:
>>
>> ------------[ cut here ]------------
>> WARNING: at linux-2.6.34/net/wireless/core.c:633 wdev_cleanup_work+0xb7/0xe0 [cfg80211]()
>> Hardware name: Latitude C840
>> Pid: 707, comm: cfg80211 Not tainted 2.6.34.7-0.5-desktop #1
>> Call Trace:
>> [<c02065c3>] try_stack_unwind+0x173/0x190
>> [<c02051cf>] dump_trace+0x3f/0xe0
>> [<c020662b>] show_trace_log_lvl+0x4b/0x60
>> [<c0206658>] show_trace+0x18/0x20
>> [<c064e0b3>] dump_stack+0x6d/0x72
>> [<c02443ae>] warn_slowpath_common+0x6e/0xb0
>> [<c0244403>] warn_slowpath_null+0x13/0x20
>> [<e2db5497>] wdev_cleanup_work+0xb7/0xe0 [cfg80211]
>> [<c025cfa9>] run_workqueue+0x79/0x170
>> [<c025d123>] worker_thread+0x83/0xe0
>> [<c025fef4>] kthread+0x74/0x80
>> [<c0203826>] kernel_thread_helper+0x6/0x10
>> ---[ end trace 3f0348b3b0c6f4ff ]---
>>
>> Reported by: Giacomo Comes <[email protected]>
>> Signed-off-by: David Kilroy <[email protected]>
>> ---
>>
>> This patch has only been compile tested. Given the trace, I've
>> guessed that cfg80211 would prefer that orinoco abort the scan on
>> interface down.
>
> Any chance we can get a Tested-by on this one, so I can feel good
> about sending it for 2.6.37?
I was looking for an old PCMCIA orinoco hermes-I or -II card, but I
found a cheat sheet from my very first Linux book from Michael Kofler.
Which key on a EN/US keyboard is equivalent to the one on a DE
keyboard, nice catch :-).
Will look in the other cartons in the cellar...
- Sedat -
On Wed, Nov 24, 2010 at 08:33:02PM +0000, David Kilroy wrote:
> This fixes the problem causing the following trace:
>
> ------------[ cut here ]------------
> WARNING: at linux-2.6.34/net/wireless/core.c:633 wdev_cleanup_work+0xb7/0xe0 [cfg80211]()
> Hardware name: Latitude C840
> Pid: 707, comm: cfg80211 Not tainted 2.6.34.7-0.5-desktop #1
> Call Trace:
> [<c02065c3>] try_stack_unwind+0x173/0x190
> [<c02051cf>] dump_trace+0x3f/0xe0
> [<c020662b>] show_trace_log_lvl+0x4b/0x60
> [<c0206658>] show_trace+0x18/0x20
> [<c064e0b3>] dump_stack+0x6d/0x72
> [<c02443ae>] warn_slowpath_common+0x6e/0xb0
> [<c0244403>] warn_slowpath_null+0x13/0x20
> [<e2db5497>] wdev_cleanup_work+0xb7/0xe0 [cfg80211]
> [<c025cfa9>] run_workqueue+0x79/0x170
> [<c025d123>] worker_thread+0x83/0xe0
> [<c025fef4>] kthread+0x74/0x80
> [<c0203826>] kernel_thread_helper+0x6/0x10
> ---[ end trace 3f0348b3b0c6f4ff ]---
>
> Reported by: Giacomo Comes <[email protected]>
> Signed-off-by: David Kilroy <[email protected]>
> ---
>
> This patch has only been compile tested. Given the trace, I've
> guessed that cfg80211 would prefer that orinoco abort the scan on
> interface down.
Any chance we can get a Tested-by on this one, so I can feel good
about sending it for 2.6.37?
John
--
John W. Linville Someday the world will need a hero, and you
[email protected] might be all we have. Be ready.
On Mon, Nov 29, 2010 at 09:18:27PM +0100, Sedat Dilek wrote:
> On Mon, Nov 29, 2010 at 8:26 PM, John W. Linville
> <[email protected]> wrote:
> > On Wed, Nov 24, 2010 at 08:33:02PM +0000, David Kilroy wrote:
> >> This fixes the problem causing the following trace:
> >>
> >> ------------[ cut here ]------------
> >> WARNING: at linux-2.6.34/net/wireless/core.c:633 wdev_cleanup_work+0xb7/0xe0 [cfg80211]()
> >> Hardware name: Latitude C840
> >> Pid: 707, comm: cfg80211 Not tainted 2.6.34.7-0.5-desktop #1
> >> Call Trace:
> >> ?[<c02065c3>] try_stack_unwind+0x173/0x190
> >> ?[<c02051cf>] dump_trace+0x3f/0xe0
> >> ?[<c020662b>] show_trace_log_lvl+0x4b/0x60
> >> ?[<c0206658>] show_trace+0x18/0x20
> >> ?[<c064e0b3>] dump_stack+0x6d/0x72
> >> ?[<c02443ae>] warn_slowpath_common+0x6e/0xb0
> >> ?[<c0244403>] warn_slowpath_null+0x13/0x20
> >> ?[<e2db5497>] wdev_cleanup_work+0xb7/0xe0 [cfg80211]
> >> ?[<c025cfa9>] run_workqueue+0x79/0x170
> >> ?[<c025d123>] worker_thread+0x83/0xe0
> >> ?[<c025fef4>] kthread+0x74/0x80
> >> ?[<c0203826>] kernel_thread_helper+0x6/0x10
> >> ---[ end trace 3f0348b3b0c6f4ff ]---
> >>
> >> Reported by: Giacomo Comes <[email protected]>
> >> Signed-off-by: David Kilroy <[email protected]>
> >> ---
> >>
> >> This patch has only been compile tested. Given the trace, I've
> >> guessed that cfg80211 would prefer that orinoco abort the scan on
> >> interface down.
> >
> > Any chance we can get a Tested-by on this one, so I can feel good
> > about sending it for 2.6.37?
>
> I was looking for an old PCMCIA orinoco hermes-I or -II card, but I
> found a cheat sheet from my very first Linux book from Michael Kofler.
> Which key on a EN/US keyboard is equivalent to the one on a DE
> keyboard, nice catch :-).
> Will look in the other cartons in the cellar...
>
As I reported in another e-mail, with the patch, the trace is gone.
Unfortunatly it doesn't help make the wireles card work with WPA.
Giacomo
On Tue, Nov 30, 2010 at 2:23 AM, Giacomo Comes <[email protected]> wrote:
> On Mon, Nov 29, 2010 at 09:18:27PM +0100, Sedat Dilek wrote:
>> On Mon, Nov 29, 2010 at 8:26 PM, John W. Linville
>> <[email protected]> wrote:
>> > On Wed, Nov 24, 2010 at 08:33:02PM +0000, David Kilroy wrote:
>> >> This fixes the problem causing the following trace:
>> >>
>> >> ------------[ cut here ]------------
>> >> WARNING: at linux-2.6.34/net/wireless/core.c:633 wdev_cleanup_work+0xb7/0xe0 [cfg80211]()
>> >> Hardware name: Latitude C840
>> >> Pid: 707, comm: cfg80211 Not tainted 2.6.34.7-0.5-desktop #1
>> >> Call Trace:
>> >> ?[<c02065c3>] try_stack_unwind+0x173/0x190
>> >> ?[<c02051cf>] dump_trace+0x3f/0xe0
>> >> ?[<c020662b>] show_trace_log_lvl+0x4b/0x60
>> >> ?[<c0206658>] show_trace+0x18/0x20
>> >> ?[<c064e0b3>] dump_stack+0x6d/0x72
>> >> ?[<c02443ae>] warn_slowpath_common+0x6e/0xb0
>> >> ?[<c0244403>] warn_slowpath_null+0x13/0x20
>> >> ?[<e2db5497>] wdev_cleanup_work+0xb7/0xe0 [cfg80211]
>> >> ?[<c025cfa9>] run_workqueue+0x79/0x170
>> >> ?[<c025d123>] worker_thread+0x83/0xe0
>> >> ?[<c025fef4>] kthread+0x74/0x80
>> >> ?[<c0203826>] kernel_thread_helper+0x6/0x10
>> >> ---[ end trace 3f0348b3b0c6f4ff ]---
>> >>
>> >> Reported by: Giacomo Comes <[email protected]>
>> >> Signed-off-by: David Kilroy <[email protected]>
>> >> ---
>> >>
>> >> This patch has only been compile tested. Given the trace, I've
>> >> guessed that cfg80211 would prefer that orinoco abort the scan on
>> >> interface down.
>> >
>> > Any chance we can get a Tested-by on this one, so I can feel good
>> > about sending it for 2.6.37?
> As I reported in another e-mail, with the patch, the trace is gone.
> Unfortunatly it doesn't help make the wireles card work with WPA.
Sorry for the delay in getting back to you, John. I see you've already
added the patch to wireless-2.6
FWIW, I've reproduced the original trace with an orinoco_usb device
and verified that this patch fixes it.
Dave.