iwl3945 regression: WARNING: at net/wireless/core.c:291

iwl3945 regression: WARNING: at net/wireless/core.c:291

[Kalle Valo]

Hi Reinette,

I updated to latest wireless-testing and iwl3945 didn't load. I'm using
Lenovo x60s with 32 bit kernel and debian unstable.

The error message:

[   17.484147] ------------[ cut here ]------------
[   17.489314] WARNING: at net/wireless/core.c:291 wiphy_register+0xa7/0x1bf [cfg80211]()
[   17.494687] Hardware name: 1703Y1F
[   17.499921] Modules linked in: snd_seq snd_timer iwl3945(+) snd_seq_device iwlcore pcmcia(+) thinkpad_acpi snd rfkill mac80211 btusb soundcore video backlight yenta_socket rsrc_nonstatic lib80211 bluetooth snd_page_alloc processor battery ac output pcmcia_core rng_core psmouse pcspkr cfg80211 evdev i2c_i801 nvram button ext3 jbd mbcache sha256_generic aes_i586 aes_generic cbc dm_crypt dm_mirror dm_region_hash dm_log dm_snapshot dm_mod sd_mod ata_generic ata_piix uhci_hcd libata scsi_mod sdhci_pci sdhci mmc_core led_class e1000e ide_pci_generic ide_core ehci_hcd usbcore thermal fan
[   17.529567] Pid: 1924, comm: modprobe Not tainted 2.6.30-rc1 #152
[   17.535677] Call Trace:
[   17.541699]  [<c0123415>] warn_slowpath+0x71/0xa8
[   17.547679]  [<c011beb8>] ? enqueue_task_fair+0x131/0x17f
[   17.553539]  [<c01100e7>] ? native_smp_send_reschedule+0x45/0x47
[   17.559499]  [<c011a114>] ? resched_task+0x5d/0x60
[   17.565318]  [<f898438e>] wiphy_register+0xa7/0x1bf [cfg80211]
[   17.571152]  [<f85c6271>] ieee80211_register_hw+0xd8/0x354 [mac80211]
[   17.577000]  [<f871018f>] ? iwl_enable_interrupts+0xca/0xd2 [iwlcore]
[   17.582789]  [<f876b4ac>] iwl3945_pci_probe+0xa9b/0xbb7 [iwl3945]
[   17.588529]  [<c01e70b4>] local_pci_probe+0xe/0x10
[   17.594313]  [<c01e7ad1>] pci_device_probe+0x43/0x66
[   17.600163]  [<c0246e45>] driver_probe_device+0x80/0x10c
[   17.606094]  [<c0246f14>] __driver_attach+0x43/0x5f
[   17.612075]  [<c0246884>] bus_for_each_dev+0x3d/0x67
[   17.617988]  [<c0246d17>] driver_attach+0x14/0x16
[   17.623902]  [<c0246ed1>] ? __driver_attach+0x0/0x5f
[   17.629781]  [<c02462d2>] bus_add_driver+0xd2/0x1e9
[   17.635605]  [<c0247131>] driver_register+0x8b/0xe8
[   17.641438]  [<c01e7df3>] __pci_register_driver+0x38/0x91
[   17.647263]  [<f86c9048>] iwl3945_init+0x48/0x65 [iwl3945]
[   17.653060]  [<c0101137>] do_one_initcall+0x4a/0x10c
[   17.658825]  [<f86c9000>] ? iwl3945_init+0x0/0x65 [iwl3945]
[   17.664628]  [<c0136a52>] ? __blocking_notifier_call_chain+0x40/0x4c
[   17.670478]  [<c0143946>] sys_init_module+0x87/0x186
[   17.676382]  [<c0102948>] sysenter_do_call+0x12/0x26
[   17.682219] ---[ end trace 6325ce0ce8671ee8 ]---
[   17.687950] iwl3945 0000:03:00.0: Failed to register hw (error -22)
[   17.693817] iwl3945 0000:03:00.0: PCI INT A disabled
[   17.699462] iwl3945: probe of 0000:03:00.0 failed with error -22

And the warning from code:

	if (WARN_ON(wiphy->max_scan_ssids < 1))
		return -EINVAL;

It should be easy to fix, I just don't have time to send a patch right
now.

-- 
Kalle Valo

[PATCH] fix iwl3945 registration regression

[Johannes Berg]

I forgot that iwl3945 registration is separate from iwlagn.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
---
Hi Kalle, Reinette,

This looks like my mistake, sorry!

> [   17.489314] WARNING: at net/wireless/core.c:291 wiphy_register+0xa7/0x1bf [cfg80211]()

> And the warning from code:
> 
> 	if (WARN_ON(wiphy->max_scan_ssids < 1))
> 		return -EINVAL;

I thought 3945 were already unified so that the setup code was common,
but it turns out not to be. This patch will fix it.

Alternatively the scanning rework patches you (Reinette) have will also
fix it -- do you just want to send those to John instead?

johannes

 drivers/net/wireless/iwlwifi/iwl3945-base.c |    2 ++
 1 file changed, 2 insertions(+)

--- wireless-testing.orig/drivers/net/wireless/iwlwifi/iwl3945-base.c	2009-04-09 09:53:35.000000000 +0200
+++ wireless-testing/drivers/net/wireless/iwlwifi/iwl3945-base.c	2009-04-09 09:53:57.000000000 +0200
@@ -4913,6 +4913,8 @@ static int iwl3945_setup_mac(struct iwl_
 
 	hw->wiphy->custom_regulatory = true;
 
+	hw->wiphy->max_scan_ssids = 1; /* WILL FIX */
+
 	/* Default value; 4 EDCA QOS priorities */
 	hw->queues = 4;
 

Re: [PATCH] fix iwl3945 registration regression

[Kalle Valo]

On Thu, Apr 9, 2009 at 10:56 AM, Johannes Berg
<johannes@sipsolutions.net> wrote:
> I forgot that iwl3945 registration is separate from iwlagn.
>
> Signed-off-by: Johannes Berg <johannes@sipsolutions.net>

Thanks, this fixed the problem for me.

Tested-by: Kalle Valo <kalle.valo@iki.fi>

Although I wasn't able to get iwl3945 working yet:

Apr  9 17:06:57 tikku kernel: [   16.586612] iwl3945: Intel(R)
PRO/Wireless 3945ABG/BG Network Connection driver for Linux, 1.2.26kd
Apr  9 17:06:57 tikku kernel: [   16.591406] iwl3945: Copyright(c)
2003-2009 Intel Corporation
Apr  9 17:06:57 tikku kernel: [   16.596203] iwl3945 0000:03:00.0: PCI
INT A -> GSI 17 (level, low) -> IRQ 17
Apr  9 17:06:57 tikku kernel: [   16.600996] iwl3945 0000:03:00.0:
setting latency timer to 64
Apr  9 17:06:57 tikku kernel: [   16.660038] iwl3945 0000:03:00.0:
Tunable channels: 13 802.11bg, 23 802.11a channels
Apr  9 17:06:57 tikku kernel: [   16.664799] iwl3945 0000:03:00.0:
Detected Intel Wireless WiFi Link 3945ABG
Apr  9 17:06:57 tikku kernel: [   16.669786] iwl3945 0000:03:00.0: irq
29 for MSI/MSI-X
Apr  9 17:06:57 tikku kernel: [   16.675157] phy0: Selected rate
control algorithm 'iwl-3945-rs'
Apr  9 17:06:57 tikku kernel: [   16.681674] phy0: Failed to initialize wep: -2
Apr  9 17:06:57 tikku kernel: [   16.695543] iwl3945 0000:03:00.0:
Failed to register hw (error -2)
Apr  9 17:06:57 tikku kernel: [   16.700576] iwl3945 0000:03:00.0: PCI
INT A disabled
Apr  9 17:06:57 tikku kernel: [   16.705359] iwl3945: probe of
0000:03:00.0 failed with error -2

Maybe this is the infamous module loading bug I see people talking
about? I'll investigate it later.

> This looks like my mistake, sorry!

No problem! That's why we are here, testing latest and greatest wireless code :)

Kalle

Re: [PATCH] fix iwl3945 registration regression

[Larry Finger]

Kalle Valo wrote:
> On Thu, Apr 9, 2009 at 10:56 AM, Johannes Berg
> <johannes@sipsolutions.net> wrote:
>> I forgot that iwl3945 registration is separate from iwlagn.
>>
>> Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
> 
> Thanks, this fixed the problem for me.
> 
> Tested-by: Kalle Valo <kalle.valo@iki.fi>
> 
> Although I wasn't able to get iwl3945 working yet:
> 
> Apr  9 17:06:57 tikku kernel: [   16.586612] iwl3945: Intel(R)
> PRO/Wireless 3945ABG/BG Network Connection driver for Linux, 1.2.26kd
> Apr  9 17:06:57 tikku kernel: [   16.591406] iwl3945: Copyright(c)
> 2003-2009 Intel Corporation
> Apr  9 17:06:57 tikku kernel: [   16.596203] iwl3945 0000:03:00.0: PCI
> INT A -> GSI 17 (level, low) -> IRQ 17
> Apr  9 17:06:57 tikku kernel: [   16.600996] iwl3945 0000:03:00.0:
> setting latency timer to 64
> Apr  9 17:06:57 tikku kernel: [   16.660038] iwl3945 0000:03:00.0:
> Tunable channels: 13 802.11bg, 23 802.11a channels
> Apr  9 17:06:57 tikku kernel: [   16.664799] iwl3945 0000:03:00.0:
> Detected Intel Wireless WiFi Link 3945ABG
> Apr  9 17:06:57 tikku kernel: [   16.669786] iwl3945 0000:03:00.0: irq
> 29 for MSI/MSI-X
> Apr  9 17:06:57 tikku kernel: [   16.675157] phy0: Selected rate
> control algorithm 'iwl-3945-rs'
> Apr  9 17:06:57 tikku kernel: [   16.681674] phy0: Failed to initialize wep: -2
> Apr  9 17:06:57 tikku kernel: [   16.695543] iwl3945 0000:03:00.0:
> Failed to register hw (error -2)
> Apr  9 17:06:57 tikku kernel: [   16.700576] iwl3945 0000:03:00.0: PCI
> INT A disabled
> Apr  9 17:06:57 tikku kernel: [   16.705359] iwl3945: probe of
> 0000:03:00.0 failed with error -2
> 
> Maybe this is the infamous module loading bug I see people talking
> about? I'll investigate it later.

Yes it is. The key is the "Failed to initialize wep", which is caused by a
failure to load modules ecb and arc4. Apply that patch and all will be well.

Larry

Re: [PATCH] fix iwl3945 registration regression

[John W. Linville]

On Thu, Apr 09, 2009 at 09:46:37AM -0500, Larry Finger wrote:
> Kalle Valo wrote:

> > Maybe this is the infamous module loading bug I see people talking
> > about? I'll investigate it later.
> 
> Yes it is. The key is the "Failed to initialize wep", which is caused by a
> failure to load modules ecb and arc4. Apply that patch and all will be well.

FWIW, I reverted the problem causing commit in the master branch
of wireless-testing.  Hopefully the fix will be in -rc2 when
wireless-testing gets there.

John
-- 
John W. Linville		Someday the world will need a hero, and you
linville@tuxdriver.com			might be all we have.  Be ready.

Re: [PATCH] fix iwl3945 registration regression

[Larry Finger]

John W. Linville wrote:
> 
> FWIW, I reverted the problem causing commit in the master branch
> of wireless-testing.  Hopefully the fix will be in -rc2 when
> wireless-testing gets there.

It should be. Rusty Russell has it in his tree already and the regression is
well known on LKML and to Rafael Wysocki.

Larry

Re: [PATCH] fix iwl3945 registration regression

[Kalle Valo]

"John W. Linville" <linville@tuxdriver.com> writes:

> On Thu, Apr 09, 2009 at 09:46:37AM -0500, Larry Finger wrote:
>> Kalle Valo wrote:
>
>> > Maybe this is the infamous module loading bug I see people talking
>> > about? I'll investigate it later.
>> 
>> Yes it is. The key is the "Failed to initialize wep", which is caused by a
>> failure to load modules ecb and arc4. Apply that patch and all will be well.
>
> FWIW, I reverted the problem causing commit in the master branch
> of wireless-testing.  

Nice, thanks. I have Johannes patch applied on top of latest
wireless-testing and iwl3945 works again.

-- 
Kalle Valo

Re: [PATCH] fix iwl3945 registration regression

[reinette chatre]

Johannes,

On Thu, 2009-04-09 at 00:56 -0700, Johannes Berg wrote:

> Alternatively the scanning rework patches you (Reinette) have will also
> fix it -- do you just want to send those to John instead?

I surely will send those patches. I was just waiting for
wireless-testing to contain the patches on which they depend. 

Reinette