RFC: Possible fix for wireless issue / Power consumption?

[Michael Trimarchi]

Nicola Mfb wrote:
> On Wed, Jul 1, 2009 at 11:32 PM, Paul Fertser<fercerpav at gmail.com> wrote:
> [...]
>   
>> consumption in some weird way, as suggested by that log. :-/ Also i'm
>> not sure why this fix is needed as there was working kernel revision
>> with the same driver and it looks like nobody's analised what change
>> actually introduced the regression.
>>     
>
> Just to remember that 2.6.28 kernels have wifi issue too, I just got
> that fso disabling the wifi interface:
>
> [ 2952.960000] ar6000_ioctl_siwauth(): Unknown flag 0xc
> [ 2952.960000] channel hint set to 2437
> [ 2952.970000] AR6000 disconnected
> [ 2960.315000] debug_hdr_ptr: 0x800018b0
> [ 2961.130000] ar6000_wow interrupt
> [ 2962.985000] Unable to handle kernel NULL pointer dereference at
> virtual address 0000002c
> [ 2962.985000] pgd = c20b0000
> [ 2962.990000] [0000002c] *pgd=3202b031, *pte=00000000, *ppte=00000000
> [ 2963.000000] Internal error: Oops: 17 [#1] PREEMPT
> [ 2963.000000] Modules linked in: ipv6 uinput
> snd_soc_neo1973_gta02_wm8753 snd_soc_s3c24xx_i2s snd_soc_s3c24xx
> snd_soc_wm8753 snd_soc_core snd_pcm snd_timer snd_page_alloc snd
> g_ether btusb bluetooth ohci_hcd ar6000
> [ 2963.000000] CPU: 0    Not tainted  (2.6.28-20090105.git69b2aa26 #1)
> [ 2963.000000] PC is at wmi_cmd_send+0x8c/0xb4 [ar6000]
> [ 2963.000000] LR is at wmi_cmd_send+0x54/0xb4 [ar6000]
> [ 2963.000000] pc : [<bf00f694>]    lr : [<bf00f65c>]    psr: 80000013
> [ 2963.000000] sp : c2011d28  ip : c2011d28  fp : c2011d44
> [ 2963.000000] r10: 00008b18  r9 : c6cb6000  r8 : 00000000
> [ 2963.000000] r7 : 00000000  r6 : 00000000  r5 : c6ca5320  r4 : 00000009
> [ 2963.000000] r3 : 00000000  r2 : 00000000  r1 : c6ca5320  r0 : c6cdb882
> [ 2963.000000] Flags: Nzcv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
> [ 2963.000000] Control: c000717f  Table: 320b0000  DAC: 00000015
> [ 2963.000000] Process wpa_supplicant (pid: 2156, stack limit = 0xc2010268)
> [ 2963.000000] Stack: (0xc2011d28 to 0xc2012000)
> [...]
> [ 2963.000000] Backtrace:
> [ 2963.000000] [<bf00f608>] (wmi_cmd_send+0x0/0xb4 [ar6000]) from
> [<bf0118ac>] (wmi_bssfilter_cmd+0x8c/0x90 [ar6000])
> [ 2963.000000]  r7:00000000 r6:c6ca5320 r5:00000000 r4:c6cdb884
> [ 2963.000000] [<bf011820>] (wmi_bssfilter_cmd+0x0/0x90 [ar6000]) from
> [<bf009938>] (ar6000_ioctl_siwscan+0x11c/0x148 [ar6000])
> [ 2963.000000]  r8:00000000 r7:c2010000 r6:c6cb64e0 r5:c2011d7c r4:00000000
> [ 2963.000000] [<bf00981c>] (ar6000_ioctl_siwscan+0x0/0x148 [ar6000])
> from [<c02c4494>] (ioctl_standard_call+0x288/0x3c4)
> [ 2963.000000]  r8:c2011e88 r7:00000000 r6:c02ebd1c r5:c2011e78 r4:0000013c
> [ 2963.000000] [<c02c420c>] (ioctl_standard_call+0x0/0x3c4) from
> [<c02c3dd8>] (wext_handle_ioctl+0xbc/0x224)
> [ 2963.000000] [<c02c3d1c>] (wext_handle_ioctl+0x0/0x224) from
> [<c0253d64>] (dev_ioctl+0x6b4/0x71c)
> [ 2963.000000]  r8:00000000 r7:c08f3d8c r6:c2011e78 r5:00008b18 r4:00000000
> [ 2963.000000] [<c02536b0>] (dev_ioctl+0x0/0x71c) from [<c02435fc>]
> (sock_ioctl+0x220/0x260)
> [ 2963.000000] [<c02433dc>] (sock_ioctl+0x0/0x260) from [<c00b13d4>]
> (vfs_ioctl+0x3c/0x9c)
> [ 2963.000000]  r6:00008b18 r5:03f88bd0 r4:c2368080
> [ 2963.000000] [<c00b1398>] (vfs_ioctl+0x0/0x9c) from [<c00b19c8>]
> (do_vfs_ioctl+0x48c/0x4e4)
> [ 2963.000000]  r6:00008b18 r5:c2368080 r4:c76f8088
> [ 2963.000000] [<c00b153c>] (do_vfs_ioctl+0x0/0x4e4) from [<c00b1a60>]
> (sys_ioctl+0x40/0x60)
> [ 2963.000000] [<c00b1a20>] (sys_ioctl+0x0/0x60) from [<c002fea0>]
> (ret_fast_syscall+0x0/0x2c)
> [ 2963.000000]  r7:00000036 r6:0008bc60 r5:00000000 r4:00000002
> [ 2963.000000] Code: e1a01005 ebffff5b e3a02001 e1a01005 (e597002c)
> [ 2963.025000] ---[ end trace eb7a4de88470bd43 ]---
>
> uname -a
>
> Linux debian-gta02 2.6.28-20090105.git69b2aa26 #1 PREEMPT Tue Mar 3
> 20:22:28 UTC 2009 armv4tl GNU/Linux
>
> The 99% of time I cannot reboot the device and have to remove batteries.
> In my 1 year freerunner experience I may say that WiFi never worked
> fine, of course I did not tryied every single kernel but all the
> default shipped with shr/om/asu/fso and their feed updates.
> Some considerations: the hardware is broken? there was no big effort
> by OM to patch the driver?
>   
This bugs is fixed in the latest kernel, but I still prensent in the
2.6.28

Michael
> Regards
>
>    Nicola
>
>
>