On Mon, 2008-10-27 at 18:02 -0700, Mirco Tischler wrote:
> Hi
>
> I have problems connecting to wlan with my iwl4965 chip and 2.6.28-rc2 (didn't
> test -rc1). Networkmanager just times out every time.
> Only after playing around with rmmod/modprobe and killswitch I am sometimes
> able to connect. But I cannot reproduce this reliably yet.
>
> This is what I get in dmesg (from modprobe iwlagn to connection failure):
>
> [ 1881.884690] iwlagn: Intel(R) Wireless WiFi Link AGN driver for
> Linux,1.3.27kds
> [ 1881.884692] iwlagn: Copyright(c) 2003-2008 Intel Corporation
> [ 1881.884809] iwlagn 0000:06:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
> [ 1881.884836] iwlagn 0000:06:00.0: setting latency timer to 64
> [ 1881.884952] iwlagn: Detected Intel Wireless WiFi Link 4965AGN REV=0x4
> [ 1881.935149] iwlagn: Tunable channels: 13 802.11bg, 19 802.11a channels
> [ 1881.936327] phy4: Selected rate control algorithm 'iwl-agn-rs'
> [ 1885.965591] iwlagn 0000:06:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
> [ 1885.965874] iwlagn 0000:06:00.0: irq 41 for MSI/MSI-X
> [ 1885.965974] iwlagn 0000:06:00.0: firmware: requesting iwlwifi-4965-2.ucode
> [ 1886.181600] Registered led device: iwl-phy4:radio
> [ 1886.181618] Registered led device: iwl-phy4:assoc
> [ 1886.181635] Registered led device: iwl-phy4:RX
> [ 1886.181654] Registered led device: iwl-phy4:TX
> [ 1886.213631] ADDRCONF(NETDEV_UP): wlan0: link is not ready
> [ 1887.572768] iwlagn: TX Power requested while scanning!
Looks good so far. But I failed to see what are you trying to connect?
Can you try "iwconfig wlan0 channnel <> essid <>"?
> What might be related is this early on boot:
>
> [ 0.163009] cfg80211: Calling CRDA to update world regulatory domain
> [ 0.163012] cfg80211: Failed calling CRDA
> [ 0.163024] cfg80211: calling CRDA failed - unable to update world
> regulatory domain, using static definition
Does your AP work on A band?
Thanks,
-yi
On Wed, 29 Oct 2008 14:18:04 +0800, Zhu Yi wrote:
> Can you try attached patch?
>
> Thanks,
> -yi
I did. Didn't solve the problems but here's the output:
[ 880.195147] iwlagn: Intel(R) Wireless WiFi Link AGN driver for Linux,1.3.27kds
[ 880.195150] iwlagn: Copyright(c) 2003-2008 Intel Corporation
[ 880.195300] iwlagn 0000:06:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
[ 880.195327] iwlagn 0000:06:00.0: setting latency timer to 64
[ 880.195444] iwlagn: Detected Intel Wireless WiFi Link 4965AGN REV=0x4
[ 880.245264] iwlagn: Tunable channels: 13 802.11bg, 19 802.11a channels
[ 880.247969] phy3: Selected rate control algorithm 'iwl-agn-rs'
[ 884.274477] iwlagn 0000:06:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
[ 884.275247] iwlagn 0000:06:00.0: irq 41 for MSI/MSI-X
[ 884.275361] iwlagn 0000:06:00.0: firmware: requesting iwlwifi-4965-2.ucode
[ 884.489886] Registered led device: iwl-phy3:radio
[ 884.489933] Registered led device: iwl-phy3:assoc
[ 884.489973] Registered led device: iwl-phy3:RX
[ 884.490012] Registered led device: iwl-phy3:TX
[ 884.526082] ADDRCONF(NETDEV_UP): wlan0: link is not ready
[ 884.784601] iwlagn: TX Power requested while scanning!
[ 884.784611] Pid: 9839, comm: iwlagn/1 Not tainted 2.6.28-rc2-oldreg-debug #27
[ 884.784617] Call Trace:
[ 884.784632] [<ffffffff8101a374>] ? native_smp_send_reschedule+0x4d/0x4f
[ 884.784658] [<ffffffffa01c1082>] iwl4965_send_tx_power+0x5a/0xf2f [iwlagn]
[ 884.784668] [<ffffffff813c065c>] ? _spin_unlock_irqrestore+0x31/0x3c
[ 884.784677] [<ffffffff8103341f>] ? try_to_wake_up+0x222/0x234
[ 884.784685] [<ffffffff8103343e>] ? default_wake_function+0xd/0xf
[ 884.784693] [<ffffffff8104ace2>] ? autoremove_wake_function+0x11/0x38
[ 884.784703] [<ffffffff8102aee1>] ? __wake_up_common+0x46/0x76
[ 884.784710] [<ffffffff813c065c>] ? _spin_unlock_irqrestore+0x31/0x3c
[ 884.784717] [<ffffffff8102c04f>] ? __wake_up+0x43/0x4f
[ 884.784725] [<ffffffff81047dff>] ? queue_work+0x59/0x68
[ 884.784735] [<ffffffff81395eab>] ?ieee80211_mlme_notify_scan_completed+0xc7/0xd9
[ 884.784743] [<ffffffff81393708>] ? ieee80211_scan_completed+0x2e8/0x32c
[ 884.784770] [<ffffffffa00d5425>] ? iwl_rx_queue_restock+0xc6/0x143 [iwlcore]
[ 884.784790] [<ffffffffa00cf290>] iwl_set_tx_power+0x6b/0x71 [iwlcore]
[ 884.784812] [<ffffffffa00de3f6>] iwl_bg_scan_completed+0xbc/0xcd [iwlcore]
[ 884.784834] [<ffffffffa00de33a>] ? iwl_bg_scan_completed+0x0/0xcd [iwlcore]
[ 884.784842] [<ffffffff81047298>] run_workqueue+0x95/0x138
[ 884.784849] [<ffffffff81047430>] worker_thread+0xf5/0x109
[ 884.784856] [<ffffffff8104acd1>] ? autoremove_wake_function+0x0/0x38
[ 884.784864] [<ffffffff8104733b>] ? worker_thread+0x0/0x109
[ 884.784871] [<ffffffff8104a956>] kthread+0x49/0x76
[ 884.784879] [<ffffffff8100d129>] child_rip+0xa/0x11
[ 884.784886] [<ffffffff8104a90d>] ? kthread+0x0/0x76
[ 884.784892] [<ffffffff8100d11f>] ? child_rip+0x0/0x11
On Tue, 28 Oct 2008 19:36:34 +0100, Johannes Berg wrote:
> Can we have a dmesg output with the "TX power requested" line posted
> here?
>
> johannes
You mean the hole output? Cause I posted the wireless part in my initial mail.
Anyway, I'll have to recreate it cause I unfortunately didn't store it and I
can't find any saved dmesg in /var/log that still contains that line. So be
patient please ;)
Mirco
On Tue, 28 Oct 2008 10:19:50 +0800, Zhu Yi wrote:
> You don't need this command if wpa_supplicant works correctly. Is it a
> hidden SSID AP?
No it isn't. Did I mention it worked well with 2.6.27?
> Thanks,
> -yi
On Tue, 2008-10-28 at 17:24 +0100, Mirco Tischler wrote:
> > This is a different bug. He has the wireless worked sometimes.
> >
> > I think your problem relates to 5GHz channel usage. We have already a
> > patch queued to resolve the problem. We will have regulatory framework
> > reworked to better solve the problem. But I guess it will be in .29 time
> > frame.
> >
> > Thanks,
> > -yi
> I tested Arjan's solution nevertheless and it seems to do the job. I can't find
> the "TX Power requested" line anymore on dmesg output with that config option
> enabled anymore and it connects without problems.
Can we have a dmesg output with the "TX power requested" line posted
here?
johannes
On Wed, 2008-10-29 at 00:24 +0800, Mirco Tischler wrote:
>
> I tested it quickly and not very thoroughly on the wpa-eap encrypted
> network at
> university and it showed the same behaviour: no connection without
> CONFIG_WIRELESS_OLD_REGULATORY and no problems with it enabled.
Interesting. The only difference between without and with
CONFIG_WIRELESS_OLD_REGULATORY is the txpower is lower and you can only
do passive scan. Can you try attached patch?
Thanks,
-yi
On Thu, 30 Oct 2008 10:55:38 -0700, reinette chatre wrote:
>
> That helped us to obtain more information. Could you please try commit
> 9685e33a2c919c1e95cb7072988fc1e0f0bad3d9 ?
>
Solves the problem for me.
Thanks
Mirco
On Tue, 28 Oct 2008 09:33:37 +0800
Zhu Yi <[email protected]> wrote:
> On Mon, 2008-10-27 at 18:02 -0700, Mirco Tischler wrote:
> > Hi
> >
> > I have problems connecting to wlan with my iwl4965 chip and
> > 2.6.28-rc2 (didn't test -rc1). Networkmanager just times out every
> > time. Only after playing around with rmmod/modprobe and killswitch
> > I am sometimes able to connect. But I cannot reproduce this
> > reliably yet.
> >
> > This is what I get in dmesg (from modprobe iwlagn to connection
> > failure):
> >
> > [ 1881.884690] iwlagn: Intel(R) Wireless WiFi Link AGN driver for
> > Linux,1.3.27kds
> > [ 1881.884692] iwlagn: Copyright(c) 2003-2008 Intel Corporation
> > [ 1881.884809] iwlagn 0000:06:00.0: PCI INT A -> GSI 18 (level,
> > low) -> IRQ 18 [ 1881.884836] iwlagn 0000:06:00.0: setting latency
> > timer to 64 [ 1881.884952] iwlagn: Detected Intel Wireless WiFi
> > Link 4965AGN REV=0x4 [ 1881.935149] iwlagn: Tunable channels: 13
> > 802.11bg, 19 802.11a channels [ 1881.936327] phy4: Selected rate
> > control algorithm 'iwl-agn-rs' [ 1885.965591] iwlagn 0000:06:00.0:
> > PCI INT A -> GSI 18 (level, low) -> IRQ 18 [ 1885.965874] iwlagn
> > 0000:06:00.0: irq 41 for MSI/MSI-X [ 1885.965974] iwlagn
> > 0000:06:00.0: firmware: requesting iwlwifi-4965-2.ucode
> > [ 1886.181600] Registered led device: iwl-phy4:radio [ 1886.181618]
> > Registered led device: iwl-phy4:assoc [ 1886.181635] Registered led
> > device: iwl-phy4:RX [ 1886.181654] Registered led device:
> > iwl-phy4:TX [ 1886.213631] ADDRCONF(NETDEV_UP): wlan0: link is not
> > ready [ 1887.572768] iwlagn: TX Power requested while scanning!
>
> Looks good so far. But I failed to see what are you trying to connect?
> Can you try "iwconfig wlan0 channnel <> essid <>"?
>
I had this same issue
turning on CONFIG_WIRELESS_OLD_REGULATORY fixed it for me....
--
Arjan van de Ven Intel Open Source Technology Centre
For development, discussion and tips for power savings,
visit http://www.lesswatts.org
On Tue, 28 Oct 2008 09:33:37 +0800, Zhu Yi wrote:
>
> Looks good so far. But I failed to see what are you trying to connect?
> Can you try "iwconfig wlan0 channnel <> essid <>"?
Sorry. I should have supplied that information on the first try. The network is
a standard WPA-protected G-band network at 2.437Ghz (channel 6).
That command seems to help a bit. At least I was able to connect more often
after executing it. I will observe that in detail tomorrow though.
>
> Does your AP work on A band?
>
> Thanks,
> -yi
On Tue, 28 Oct 2008 13:23:34 +0800, Zhu Yi wrote:
> On Mon, 2008-10-27 at 22:32 -0600, Arjan van de Ven wrote:
> > I had this same issue
> >
> > turning on CONFIG_WIRELESS_OLD_REGULATORY fixed it for me....
>
> Arjan,
>
> This is a different bug. He has the wireless worked sometimes.
>
> I think your problem relates to 5GHz channel usage. We have already a
> patch queued to resolve the problem. We will have regulatory framework
> reworked to better solve the problem. But I guess it will be in .29 time
> frame.
>
> Thanks,
> -yi
I tested Arjan's solution nevertheless and it seems to do the job. I can't find
the "TX Power requested" line anymore on dmesg output with that config option
enabled anymore and it connects without problems.
This is how it looks now:
[ 777.680750] iwlagn: Intel(R) Wireless WiFi Link AGN driver for Linux,1.3.27kds
[ 777.680753] iwlagn: Copyright(c) 2003-2008 Intel Corporation
[ 777.680878] iwlagn 0000:06:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
[ 777.680908] iwlagn 0000:06:00.0: setting latency timer to 64
[ 777.680977] iwlagn: Detected Intel Wireless WiFi Link 4965AGN REV=0x4
[ 777.731255] iwlagn: Tunable channels: 13 802.11bg, 19 802.11a channels
[ 777.732432] phy1: Selected rate control algorithm 'iwl-agn-rs'
[ 781.764539] iwlagn 0000:06:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 18
[ 781.764821] iwlagn 0000:06:00.0: irq 41 for MSI/MSI-X
[ 781.764919] iwlagn 0000:06:00.0: firmware: requesting iwlwifi-4965-2.ucode
[ 781.980719] Registered led device: iwl-phy1:radio
[ 781.980764] Registered led device: iwl-phy1:assoc
[ 781.980808] Registered led device: iwl-phy1:RX
[ 781.980856] Registered led device: iwl-phy1:TX
[ 782.014001] ADDRCONF(NETDEV_UP): wlan0: link is not ready
[ 783.633713] wlan0: authenticate with AP 00:c0:a8:e5:e5:a5
[ 783.635637] wlan0: authenticated
[ 783.635644] wlan0: associate with AP 00:c0:a8:e5:e5:a5
[ 783.642770] wlan0: RX AssocResp from 00:c0:a8:e5:e5:a5 (capab=0x431 status=0
aid=1)
[ 783.642776] wlan0: associated [ 783.669684] ADDRCONF(NETDEV_CHANGE): wlan0:
link becomes ready
I tested it quickly and not very thoroughly on the wpa-eap encrypted network at
university and it showed the same behaviour: no connection without
CONFIG_WIRELESS_OLD_REGULATORY and no problems with it enabled.
Thanks
Mirco
In between I found out that suspend and hibernation are broken too. I would
open a new thread but in the "Suspend to RAM regression in 2.6.28-rc2
(bisected)" thread that handled a similar problem iwlwifi was blamed, at least
in the mails I read. So maybe this is related here too.
The machine suspends and hibernates fine, but doesn't wake up anymore correctly.
On hibernation I end up with a black screen and only the mouse which I can
still move. But anything else is locked up. I can't switch to virtual console
with ctrl+alt+f? nor can I restart the xserver. I can't even connect on
network. The only thing I can do is move the mouse on a black screen or push
the power button.
On suspend it is a little better. After wakeup I see the frames of the windows
but not the content, and the gnome panel. And again I can move the mouse but I
can't interact with anything. At least I can connect over network using ssh.
But processes like compiz don't respond to kill -9 PID and when I try to kill
Xorg the machine completely locks up.
Another thing that I just noticed is very slow boot. From startup of udev until
short before X starts the boot lasts minutes. What I noticed is that there are
long phases of no hard disk activity. And I don't really have the heart to
write this but I have the impression that keyboard activity (i.e. pressing
random keys) causes the harddisk to read again and the hole boot to go on.
Sorry for this long subjective description of symptoms only without any
kernel messages or anything but that's all I can deliver at the moment. Maybe
tomorrow I find more than only half an hour to dig into this.
Thanks again
Mirco
On Mon, 2008-10-27 at 18:59 -0700, Mirco Tischler wrote:
> > Looks good so far. But I failed to see what are you trying to
> connect?
> > Can you try "iwconfig wlan0 channnel <> essid <>"?
> Sorry. I should have supplied that information on the first try. The
> network is
> a standard WPA-protected G-band network at 2.437Ghz (channel 6).
> That command seems to help a bit. At least I was able to connect more
> often after executing it.
You don't need this command if wpa_supplicant works correctly. Is it a
hidden SSID AP?
Thanks,
-yi
On Mon, 2008-10-27 at 22:32 -0600, Arjan van de Ven wrote:
> I had this same issue
>
> turning on CONFIG_WIRELESS_OLD_REGULATORY fixed it for me....
Arjan,
This is a different bug. He has the wireless worked sometimes.
I think your problem relates to 5GHz channel usage. We have already a
patch queued to resolve the problem. We will have regulatory framework
reworked to better solve the problem. But I guess it will be in .29 time
frame.
Thanks,
-yi
On Tue, 2008-10-28 at 23:58 +0100, Mirco Tischler wrote:
> On Tue, 28 Oct 2008 19:36:34 +0100, Johannes Berg wrote:
> > Can we have a dmesg output with the "TX power requested" line posted
> > here?
> You mean the hole output? Cause I posted the wireless part in my initial mail.
Sorry, I got confused, the initial mail is fine.
johannes
On Thu, 2008-10-30 at 10:24 -0700, Mirco Tischler wrote:
> On Wed, 29 Oct 2008 14:18:04 +0800, Zhu Yi wrote:
> > Can you try attached patch?
> >
> > Thanks,
> > -yi
> I did. Didn't solve the problems but here's the output:
That helped us to obtain more information. Could you please try commit
9685e33a2c919c1e95cb7072988fc1e0f0bad3d9 ?
Reproduced here for convenience:
commit 9685e33a2c919c1e95cb7072988fc1e0f0bad3d9
Author: Mohamed Abbas <[email protected]>
Date: Thu Oct 23 23:48:54 2008 -0700
iwlwifi: clear scanning bits upon failure
In iwl_bg_request_scan function, if we could not send a
scan command it will go to done.
In done it does the right thing to call mac80211 with
scan complete, but the problem is STATUS_SCAN_HW is still
set causing any future scan to fail. Fix by clearing the scanning status
bits if scan fails.
Signed-off-by: Mohamed Abbas <[email protected]>
Signed-off-by: Reinette Chatre <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
diff --git a/drivers/net/wireless/iwlwifi/iwl-scan.c b/drivers/net/wireless/iwlwifi/iwl-scan.c
index 3379b41..b0abf89 100644
--- a/drivers/net/wireless/iwlwifi/iwl-scan.c
+++ b/drivers/net/wireless/iwlwifi/iwl-scan.c
@@ -834,6 +834,13 @@ static void iwl_bg_request_scan(struct work_struct *data)
return;
done:
+ /* Cannot perform scan. Make sure we clear scanning
+ * bits from status so next scan request can be performed.
+ * If we don't clear scanning status bit here all next scan
+ * will fail
+ */
+ clear_bit(STATUS_SCAN_HW, &priv->status);
+ clear_bit(STATUS_SCANNING, &priv->status);
/* inform mac80211 scan aborted */
queue_work(priv->workqueue, &priv->scan_completed);
mutex_unlock(&priv->mutex);