[PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[Stefan Lippers-Hollmann]

Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
Cc: stable <stable@kernel.org> [2.6.39+]
---
 drivers/net/wireless/rtlwifi/rtl8192cu/sw.c |    3 +++
 1 files changed, 5 insertions(+), 0 deletions(-)

--- a/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
+++ b/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
@@ -322,6 +322,9 @@ static struct usb_device_id rtl8192c_usb
 	{RTL_USB_DEVICE(0x2001, 0x330a, rtl92cu_hal_cfg)}, /*D-Link-Alpha*/
 	{RTL_USB_DEVICE(0x2019, 0xab2b, rtl92cu_hal_cfg)}, /*Planex -Abocom*/
 	{RTL_USB_DEVICE(0x7392, 0x7822, rtl92cu_hal_cfg)}, /*Edimax -Edimax*/
+
+	/****** 8192CUS Dongle ********/
+	{RTL_USB_DEVICE(0x050D, 0x2103, rtl92cu_hal_cfg)}, //Belkin - Edimax
 	{}
 };
 

Re: [PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[Stefan Lippers-Hollmann]

Hi

On Tuesday 02 August 2011, Stefan Lippers-Hollmann wrote:
> Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
> Cc: stable <stable@kernel.org> [2.6.39+]
> ---
>  drivers/net/wireless/rtlwifi/rtl8192cu/sw.c |    3 +++
>  1 files changed, 5 insertions(+), 0 deletions(-)
> 
> --- a/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
> +++ b/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
> @@ -322,6 +322,9 @@ static struct usb_device_id rtl8192c_usb
>  	{RTL_USB_DEVICE(0x2001, 0x330a, rtl92cu_hal_cfg)}, /*D-Link-Alpha*/
>  	{RTL_USB_DEVICE(0x2019, 0xab2b, rtl92cu_hal_cfg)}, /*Planex -Abocom*/
>  	{RTL_USB_DEVICE(0x7392, 0x7822, rtl92cu_hal_cfg)}, /*Edimax -Edimax*/
> +
> +	/****** 8192CUS Dongle ********/
> +	{RTL_USB_DEVICE(0x050D, 0x2103, rtl92cu_hal_cfg)}, //Belkin - Edimax
>  	{}
>  };

Looking at the current(?) vendor driver
4309c61f45497de0241781a507df53e4 *RTL8192CU_linux_v3.0.2164.20110715.zip
there appear to be several more USB IDs which might be worth adding; so
far I only submitted a patch for a tested USB ID.

I've dropped IDs already present in rtl8192cu, the two lines starting 
with '!' appear to be typos in either of the drivers:

#ifdef CONFIG_RTL8192C
        /*=== Realtek demoboard ===*/

        /****** 8188CUS ********/
        {USB_DEVICE(USB_VENDER_ID_REALTEK, 0x817F)},//8188RU

        /****** 8192CUS ********/
        {USB_DEVICE(USB_VENDER_ID_REALTEK, 0x8178)},//8192cu 2*2
        {USB_DEVICE(USB_VENDER_ID_REALTEK, 0x8186)},//8192CE-VAU USB minCard

        /*=== Customer ID ===*/
        /****** 8188CUS Dongle ********/
        {USB_DEVICE(0x4855, 0x0090)},// - Feixun
        {USB_DEVICE(0x13D3, 0x3357)},// - AzureWave
        {USB_DEVICE(0x0DF6, 0x005C)},//Sitecom - Edimax
        {USB_DEVICE(0x0BDA, 0x5088)},//Thinkware - CC&C
        {USB_DEVICE(0x4856, 0x0091)},//NetweeN - Feixun
        {USB_DEVICE(0x9846, 0x9041)},//Netgear - Cameo
        {USB_DEVICE(0x2019, 0x4902)},//Planex - Etop
        {USB_DEVICE(0x2019, 0xAB2E)},//SW-WF02-AD15 -Abocom

        /****** 8188CE-VAU ********/
!       {USB_DEVICE(0x13D3, 0x3359)},// - Azwave
!       {USB_DEVICE(0x13D3, 0x3358)},// - Azwave

        /****** 8188CUS Slim Solo********/
        {USB_DEVICE(0x04F2, 0xAFF7)},//XAVI - XAVI
        {USB_DEVICE(0x04F2, 0xAFF9)},//XAVI - XAVI
        {USB_DEVICE(0x04F2, 0xAFFA)},//XAVI - XAVI

        /****** 8188CUS Slim Combo ********/
        {USB_DEVICE(0x04F2, 0xAFF8)},//XAVI - XAVI
        {USB_DEVICE(0x04F2, 0xAFFB)},//XAVI - XAVI
        {USB_DEVICE(0x04F2, 0xAFFC)},//XAVI - XAVI
        {USB_DEVICE(0x2019, 0x1201)},//Planex - Vencer

        /****** 8192CUS Dongle ********/
        {USB_DEVICE(0x4855, 0x0091)},// - Feixun
        {USB_DEVICE(0x050D, 0x2102)},//Belkin - Sercomm
        {USB_DEVICE(0x050D, 0x2103)},//Belkin - Edimax
        {USB_DEVICE(0x20F4, 0x624D)},//TRENDnet
        {USB_DEVICE(0x0DF6, 0x0061)},//Sitecom - Edimax
        {USB_DEVICE(0x0B05, 0x17AB)},//ASUS - Edimax
#endif
        {}      /* Terminating entry */
};

While I can't confirm it myself, as I don't have any rtl8192cu hardware
locally, I've gotten feedback[1] that temporarily adding new IDs via

# modprobe rtl8192cu 
# echo -n "050d 2103" > /sys/bus/usb/drivers/rtl8192cu/new_id

appears not be successful and might Oops the kernel[2]. Given that I
can't test it myself and considering that the reporter's kernel is
tainted, I can't claim this for sure - but I could ask the original 
reporter to test it again.

Regards
	Stefan Lippers-Hollmann

[1]	http://aptosid.com/index.php?name=PNphpBB2&file=viewtopic&t=1550&start=0
[2]	http://pastebin.com/ppZJ0GmU

Re: [PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[Larry Finger]

On 08/02/2011 03:43 PM, Stefan Lippers-Hollmann wrote:
> Hi
>
> On Tuesday 02 August 2011, Stefan Lippers-Hollmann wrote:
>> Signed-off-by: Stefan Lippers-Hollmann<s.l-h@gmx.de>
>> Cc: stable<stable@kernel.org>  [2.6.39+]
>> ---
>>   drivers/net/wireless/rtlwifi/rtl8192cu/sw.c |    3 +++
>>   1 files changed, 5 insertions(+), 0 deletions(-)
>>
>> --- a/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
>> +++ b/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
>> @@ -322,6 +322,9 @@ static struct usb_device_id rtl8192c_usb
>>   	{RTL_USB_DEVICE(0x2001, 0x330a, rtl92cu_hal_cfg)}, /*D-Link-Alpha*/
>>   	{RTL_USB_DEVICE(0x2019, 0xab2b, rtl92cu_hal_cfg)}, /*Planex -Abocom*/
>>   	{RTL_USB_DEVICE(0x7392, 0x7822, rtl92cu_hal_cfg)}, /*Edimax -Edimax*/
>> +
>> +	/****** 8192CUS Dongle ********/
>> +	{RTL_USB_DEVICE(0x050D, 0x2103, rtl92cu_hal_cfg)}, //Belkin - Edimax
>>   	{}
>>   };
>
> Looking at the current(?) vendor driver
> 4309c61f45497de0241781a507df53e4 *RTL8192CU_linux_v3.0.2164.20110715.zip
> there appear to be several more USB IDs which might be worth adding; so
> far I only submitted a patch for a tested USB ID.
>
> I've dropped IDs already present in rtl8192cu, the two lines starting
> with '!' appear to be typos in either of the drivers:
>
> #ifdef CONFIG_RTL8192C
>          /*=== Realtek demoboard ===*/
>
>          /****** 8188CUS ********/
>          {USB_DEVICE(USB_VENDER_ID_REALTEK, 0x817F)},//8188RU
>
>          /****** 8192CUS ********/
>          {USB_DEVICE(USB_VENDER_ID_REALTEK, 0x8178)},//8192cu 2*2
>          {USB_DEVICE(USB_VENDER_ID_REALTEK, 0x8186)},//8192CE-VAU USB minCard
>
>          /*=== Customer ID ===*/
>          /****** 8188CUS Dongle ********/
>          {USB_DEVICE(0x4855, 0x0090)},// - Feixun
>          {USB_DEVICE(0x13D3, 0x3357)},// - AzureWave
>          {USB_DEVICE(0x0DF6, 0x005C)},//Sitecom - Edimax
>          {USB_DEVICE(0x0BDA, 0x5088)},//Thinkware - CC&C
>          {USB_DEVICE(0x4856, 0x0091)},//NetweeN - Feixun
>          {USB_DEVICE(0x9846, 0x9041)},//Netgear - Cameo
>          {USB_DEVICE(0x2019, 0x4902)},//Planex - Etop
>          {USB_DEVICE(0x2019, 0xAB2E)},//SW-WF02-AD15 -Abocom
>
>          /****** 8188CE-VAU ********/
> !       {USB_DEVICE(0x13D3, 0x3359)},// - Azwave
> !       {USB_DEVICE(0x13D3, 0x3358)},// - Azwave
>
>          /****** 8188CUS Slim Solo********/
>          {USB_DEVICE(0x04F2, 0xAFF7)},//XAVI - XAVI
>          {USB_DEVICE(0x04F2, 0xAFF9)},//XAVI - XAVI
>          {USB_DEVICE(0x04F2, 0xAFFA)},//XAVI - XAVI
>
>          /****** 8188CUS Slim Combo ********/
>          {USB_DEVICE(0x04F2, 0xAFF8)},//XAVI - XAVI
>          {USB_DEVICE(0x04F2, 0xAFFB)},//XAVI - XAVI
>          {USB_DEVICE(0x04F2, 0xAFFC)},//XAVI - XAVI
>          {USB_DEVICE(0x2019, 0x1201)},//Planex - Vencer
>
>          /****** 8192CUS Dongle ********/
>          {USB_DEVICE(0x4855, 0x0091)},// - Feixun
>          {USB_DEVICE(0x050D, 0x2102)},//Belkin - Sercomm
>          {USB_DEVICE(0x050D, 0x2103)},//Belkin - Edimax
>          {USB_DEVICE(0x20F4, 0x624D)},//TRENDnet
>          {USB_DEVICE(0x0DF6, 0x0061)},//Sitecom - Edimax
>          {USB_DEVICE(0x0B05, 0x17AB)},//ASUS - Edimax
> #endif
>          {}      /* Terminating entry */
> };
>
> While I can't confirm it myself, as I don't have any rtl8192cu hardware
> locally, I've gotten feedback[1] that temporarily adding new IDs via
>
> # modprobe rtl8192cu
> # echo -n "050d 2103">  /sys/bus/usb/drivers/rtl8192cu/new_id
>
> appears not be successful and might Oops the kernel[2]. Given that I
> can't test it myself and considering that the reporter's kernel is
> tainted, I can't claim this for sure - but I could ask the original
> reporter to test it again.
>
> Regards
> 	Stefan Lippers-Hollmann
>
> [1]	http://aptosid.com/index.php?name=PNphpBB2&file=viewtopic&t=1550&start=0
> [2]	http://pastebin.com/ppZJ0GmU

I have been holding a few new ID's that are in the vendor driver awaiting a 
test. These include the following:

+       {RTL_USB_DEVICE(USB_VENDER_ID_REALTEK, 0x817f, rtl92cu_hal_cfg)},
+       {RTL_USB_DEVICE(0x13d3, 0x3357, rtl92cu_hal_cfg)}, /* AzureWave */
+       {RTL_USB_DEVICE(0x13d3, 0x3358, rtl92cu_hal_cfg)}, /*Azwave 8188CE-VAU*/
-       {RTL_USB_DEVICE(0x3359, 0x13d3, rtl92cu_hal_cfg)},
+       {RTL_USB_DEVICE(0x13d3, 0x3359, rtl92cu_hal_cfg)},
+       {RTL_USB_DEVICE(0x4855, 0x0090, rtl92cu_hal_cfg)}, /* Feixun */
+       {RTL_USB_DEVICE(0x4855, 0x0091, rtl92cu_hal_cfg)}, /* NetweeN-Feixun */
+       {RTL_USB_DEVICE(0x9846, 0x9041, rtl92cu_hal_cfg)}, /* Netgear Cameo */
+       {RTL_USB_DEVICE(0x050d, 0x2102, rtl92cu_hal_cfg)}, /*Belkin-Sercomm*/
+       {RTL_USB_DEVICE(0x050d, 0x2103, rtl92cu_hal_cfg)}, /*Belkin-Edimax*/

The deleted entry has the vendor and product reversed. The first one in the list 
has been tested and works.

Based on your report, I will drop the 050d entries. If you get a confirmation 
that 050d,2103 works, then submit that patch.

Larry

Re: [PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[Stefan Lippers-Hollmann]

Hi

On Wednesday 03 August 2011, Larry Finger wrote:
> On 08/02/2011 03:43 PM, Stefan Lippers-Hollmann wrote:
> > On Tuesday 02 August 2011, Stefan Lippers-Hollmann wrote:
[...]
> >> +	{RTL_USB_DEVICE(0x050D, 0x2103, rtl92cu_hal_cfg)}, //Belkin - Edimax

Sorry about the C++ style comments, I meant to fix it, but forgot about
it in the end.

[...]
> > [1]	http://aptosid.com/index.php?name=PNphpBB2&file=viewtopic&t=1550&start=0
> > [2]	http://pastebin.com/ppZJ0GmU
> 
> I have been holding a few new ID's that are in the vendor driver awaiting a 
> test. These include the following:
> 
> +       {RTL_USB_DEVICE(USB_VENDER_ID_REALTEK, 0x817f, rtl92cu_hal_cfg)},
> +       {RTL_USB_DEVICE(0x13d3, 0x3357, rtl92cu_hal_cfg)}, /* AzureWave */
> +       {RTL_USB_DEVICE(0x13d3, 0x3358, rtl92cu_hal_cfg)}, /*Azwave 8188CE-VAU*/
> -       {RTL_USB_DEVICE(0x3359, 0x13d3, rtl92cu_hal_cfg)},
> +       {RTL_USB_DEVICE(0x13d3, 0x3359, rtl92cu_hal_cfg)},
> +       {RTL_USB_DEVICE(0x4855, 0x0090, rtl92cu_hal_cfg)}, /* Feixun */
> +       {RTL_USB_DEVICE(0x4855, 0x0091, rtl92cu_hal_cfg)}, /* NetweeN-Feixun */
> +       {RTL_USB_DEVICE(0x9846, 0x9041, rtl92cu_hal_cfg)}, /* Netgear Cameo */
> +       {RTL_USB_DEVICE(0x050d, 0x2102, rtl92cu_hal_cfg)}, /*Belkin-Sercomm*/
> +       {RTL_USB_DEVICE(0x050d, 0x2103, rtl92cu_hal_cfg)}, /*Belkin-Edimax*/
> 
> The deleted entry has the vendor and product reversed. The first one in the list 
> has been tested and works.
> 
> Based on your report, I will drop the 050d entries. If you get a confirmation 
> that 050d,2103 works, then submit that patch.

I have now asked the reporter to check the state of rtl8192cu again, 
current state is (with a kernel including my original patch):
- functionality confirmed on an unencrypted network
- WPA2-psk not conclusively tested, it may be out of range for his device.
("EDIT: With my netbook I can get an IP address and use the neighbor's
  unsecured network. Also, the Belkin device reports it at -47 dBm 
  signal strength, but my Toshiba netbook gives it -80 to -83 (which 
  may explain why the Belkin couldn't get an IP address), so there are 
  still some issues to sort.")
- no Oops if his USB ID is compiled into the module, that only happens
  if he tries to inject it into an unpatched kernel with:
  echo -n "050d 2103">  /sys/bus/usb/drivers/rtl8192cu/new_id
  (and I'm reluctant to raise this as potential issue completely, due
  to the tainted kernel condition on his system).

Hopefully he'll answer my further inquiries about the stability of his 
Belkin n300 Micro USB soon.

Regards
	Stefan Lippers-Hollmann

Re: [PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[Larry Finger]

On 08/02/2011 05:35 PM, Stefan Lippers-Hollmann wrote:
> Hi
>
> On Wednesday 03 August 2011, Larry Finger wrote:
>> On 08/02/2011 03:43 PM, Stefan Lippers-Hollmann wrote:
>>> On Tuesday 02 August 2011, Stefan Lippers-Hollmann wrote:
> [...]
>>>> +	{RTL_USB_DEVICE(0x050D, 0x2103, rtl92cu_hal_cfg)}, //Belkin - Edimax
>
> Sorry about the C++ style comments, I meant to fix it, but forgot about
> it in the end.
>
> [...]
>>> [1]	http://aptosid.com/index.php?name=PNphpBB2&file=viewtopic&t=1550&start=0
>>> [2]	http://pastebin.com/ppZJ0GmU
>>
>> I have been holding a few new ID's that are in the vendor driver awaiting a
>> test. These include the following:
>>
>> +       {RTL_USB_DEVICE(USB_VENDER_ID_REALTEK, 0x817f, rtl92cu_hal_cfg)},
>> +       {RTL_USB_DEVICE(0x13d3, 0x3357, rtl92cu_hal_cfg)}, /* AzureWave */
>> +       {RTL_USB_DEVICE(0x13d3, 0x3358, rtl92cu_hal_cfg)}, /*Azwave 8188CE-VAU*/
>> -       {RTL_USB_DEVICE(0x3359, 0x13d3, rtl92cu_hal_cfg)},
>> +       {RTL_USB_DEVICE(0x13d3, 0x3359, rtl92cu_hal_cfg)},
>> +       {RTL_USB_DEVICE(0x4855, 0x0090, rtl92cu_hal_cfg)}, /* Feixun */
>> +       {RTL_USB_DEVICE(0x4855, 0x0091, rtl92cu_hal_cfg)}, /* NetweeN-Feixun */
>> +       {RTL_USB_DEVICE(0x9846, 0x9041, rtl92cu_hal_cfg)}, /* Netgear Cameo */
>> +       {RTL_USB_DEVICE(0x050d, 0x2102, rtl92cu_hal_cfg)}, /*Belkin-Sercomm*/
>> +       {RTL_USB_DEVICE(0x050d, 0x2103, rtl92cu_hal_cfg)}, /*Belkin-Edimax*/
>>
>> The deleted entry has the vendor and product reversed. The first one in the list
>> has been tested and works.
>>
>> Based on your report, I will drop the 050d entries. If you get a confirmation
>> that 050d,2103 works, then submit that patch.
>
> I have now asked the reporter to check the state of rtl8192cu again,
> current state is (with a kernel including my original patch):
> - functionality confirmed on an unencrypted network
> - WPA2-psk not conclusively tested, it may be out of range for his device.
> ("EDIT: With my netbook I can get an IP address and use the neighbor's
>    unsecured network. Also, the Belkin device reports it at -47 dBm
>    signal strength, but my Toshiba netbook gives it -80 to -83 (which
>    may explain why the Belkin couldn't get an IP address), so there are
>    still some issues to sort.")
> - no Oops if his USB ID is compiled into the module, that only happens
>    if he tries to inject it into an unpatched kernel with:
>    echo -n "050d 2103">   /sys/bus/usb/drivers/rtl8192cu/new_id
>    (and I'm reluctant to raise this as potential issue completely, due
>    to the tainted kernel condition on his system).
>
> Hopefully he'll answer my further inquiries about the stability of his
> Belkin n300 Micro USB soon.

I think that ID and 050d:2102 should be added to the kernel. AFAIK, all 
RTL8192CU and all RTL8188CU chips are alike, at least as far as the programming 
is concerned. At least if the device works with the new IDs, it likely works as 
well as any other device. There are some reports of low performance from some 
models, but none of them crash Linux. In addition, there is no diagnostic info 
that would allow the debugging of the problem. My specific device does not show 
this problem.

As you have seen, I sent the patch for the others.

Larry

Re: [PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[Stefan Lippers-Hollmann]

Hi

On Wednesday 03 August 2011, Stefan Lippers-Hollmann wrote:
> On Wednesday 03 August 2011, Larry Finger wrote:
> > On 08/02/2011 03:43 PM, Stefan Lippers-Hollmann wrote:
> > > On Tuesday 02 August 2011, Stefan Lippers-Hollmann wrote:
> > > [1]	http://aptosid.com/index.php?name=PNphpBB2&file=viewtopic&t=1550&start=0
> > > [2]	http://pastebin.com/ppZJ0GmU
[...]
> - no Oops if his USB ID is compiled into the module, that only happens
>   if he tries to inject it into an unpatched kernel with:
>   echo -n "050d 2103">  /sys/bus/usb/drivers/rtl8192cu/new_id
>   (and I'm reluctant to raise this as potential issue completely, due
>   to the tainted kernel condition on his system).

I can reproduce the reported Oops by injecting a (wrong) USB ID into 
rtl8192cu myself on an untainted system. For reproducing the Oops, I 
inject an unknown (wrong, the device I'm abusing for this test is a 
Netgear WG111T, Atheros AR5523 based - no linux driver exists for this 
device and no module claims its device IDs 0x1385 0x4251) USB ID 
through sysfs facilities into rtl8192cu; all tests with kernel 
3.0 + current -stable queue:

# cat /proc/sys/kernel/tainted
0
# modprobe rtl8192cu
# echo -n "1385 4251" > /sys/bus/usb/drivers/rtl8192cu/new_id
[hotplug device]
# dmesg
[...]
[   86.420998] usbcore: registered new interface driver rtl8192cu
[  110.778029] usb 1-7: new high speed USB device number 5 using ehci_hcd
[  110.893824] usb 1-7: New USB device found, idVendor=1385, idProduct=4251
[  110.893830] usb 1-7: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[  110.893836] usb 1-7: Product: WG111T
[  110.893841] usb 1-7: Manufacturer: Atheros Communications Inc
[  110.893845] usb 1-7: SerialNumber: 1.0
[  110.894760] BUG: unable to handle kernel NULL pointer dereference at 0000000000000018
[  110.894903] IP: [<ffffffffa0604747>] rtl_usb_probe+0x19b/0x74d [rtlwifi]
[  110.895012] PGD 0 
[  110.895012] Oops: 0000 [#1] PREEMPT SMP 
[  110.895012] CPU 0 
[  110.895012] Modules linked in: rtl8192cu rtl8192c_common rtlwifi mac80211 cfg80211 rfkill cpufreq_stats cpufreq_ondemand cpufreq_powersave cpufreq_conservative ppdev lp cpufreq_performance af_packet fuse nls_utf8 ntfs powernow_k8 freq_table mperf tda18218 af9013 ir_lirc_codec lirc_dev ir_sony_decoder ir_jvc_decoder ir_rc6_decoder ir_rc5_decoder ir_nec_decoder dvb_usb_af9015 dvb_usb dvb_core rc_core edac_core snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm snd_seq snd_timer snd_seq_device radeon parport_pc ttm drm_kms_helper drm i2c_algo_bit k8temp pcspkr evdev psmouse edac_mce_amd serio_raw power_supply parport snd i2c_nforce2 i2c_core button soundcore processor snd_page_alloc ext4 mbcache jbd2 crc16 dm_mod btrfs zlib_deflate crc32c libcrc32c sg sr_mod sd_mod cdrom crc_t10dif ata_generic usbhid hid floppy thermal fan firewire_ohci pata_acpi firewire_core crc_itu_t pata_amd sata_nv e1000 libata ohci_hcd forcedeth scsi_mod ehci_hcd usbcore ssb mmc_core pcmcia pcmcia_core [last unloaded: scsi_wait_scan]
[  110.895012] 
[  110.895012] Pid: 577, comm: khubd Not tainted 3.0-0.slh.6-aptosid-amd64 #1 MICRO-STAR INTERNATIONAL CO., LTD MS-7185/MS-7185
[  110.895012] RIP: 0010:[<ffffffffa0604747>]  [<ffffffffa0604747>] rtl_usb_probe+0x19b/0x74d [rtlwifi]
[  110.895012] RSP: 0018:ffff88011da77a60  EFLAGS: 00010246
[  110.895012] RAX: 0000000000000000 RBX: ffff88011fc09ca0 RCX: 0000000000000000
[  110.895012] RDX: ffffffffa0608618 RSI: ffffffffa0607819 RDI: ffff88011fc08520
[  110.895012] RBP: ffff88011f0a7088 R08: 000000000000ffff R09: 000000000000ffff
[  110.895012] R10: 0000000000000000 R11: 000000000000fffb R12: ffff88011d5a6c30
[  110.895012] R13: ffff88011fd54090 R14: ffff88011fc08520 R15: ffff88011fc09ca0
[  110.895012] FS:  00007f6e9e1d37a0(0000) GS:ffff880127c00000(0000) knlGS:0000000000000000
[  110.895012] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[  110.895012] CR2: 0000000000000018 CR3: 000000011fe0f000 CR4: 00000000000006f0
[  110.895012] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  110.895012] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[  110.895012] Process khubd (pid: 577, threadinfo ffff88011da76000, task ffff88011dc5b700)
[  110.895012] Stack:
[  110.895012]  0000000000000246 ffffffff8129f3d2 ffff88011fd54090 ffff88011f0a7000
[  110.895012]  ffff88011f0a7088 ffff88011d5a6c30 ffff88011f0a7000 ffffffffa0639528
[  110.895012]  ffff88011fd54090 ffff88011d5a6c00 0000000000000000 ffffffffa0051f05
[  110.895012] Call Trace:
[  110.895012]  [<ffffffff8129f3d2>] ? __pm_runtime_set_status+0x132/0x1e0
[  110.895012]  [<ffffffffa0051f05>] ? usb_probe_interface+0xe5/0x1e0 [usbcore]
[  110.895012]  [<ffffffff8129657f>] ? driver_probe_device+0x6f/0x190
[  110.895012]  [<ffffffff81296740>] ? __driver_attach+0xa0/0xa0
[  110.895012]  [<ffffffff812952d3>] ? bus_for_each_drv+0x53/0x80
[  110.895012]  [<ffffffff812964cf>] ? device_attach+0x9f/0xc0
[  110.895012]  [<ffffffff81295c7d>] ? bus_probe_device+0x2d/0x50
[  110.895012]  [<ffffffff81293f8d>] ? device_add+0x5ad/0x670
[  110.895012]  [<ffffffff812927b2>] ? dev_set_name+0x42/0x50
[  110.895012]  [<ffffffffa0050577>] ? usb_set_configuration+0x5d7/0x6d0 [usbcore]
[  110.895012]  [<ffffffffa005904b>] ? generic_probe+0x3b/0xa0 [usbcore]
[  110.895012]  [<ffffffff8129657f>] ? driver_probe_device+0x6f/0x190
[  110.895012]  [<ffffffff81296740>] ? __driver_attach+0xa0/0xa0
[  110.895012]  [<ffffffff812952d3>] ? bus_for_each_drv+0x53/0x80
[  110.895012]  [<ffffffff812964cf>] ? device_attach+0x9f/0xc0
[  110.895012]  [<ffffffff81295c7d>] ? bus_probe_device+0x2d/0x50
[  110.895012]  [<ffffffff81293f8d>] ? device_add+0x5ad/0x670
[  110.895012]  [<ffffffffa0048cd2>] ? usb_new_device+0x122/0x1b0 [usbcore]
[  110.895012]  [<ffffffffa0049cb0>] ? hub_thread+0x7c0/0x1310 [usbcore]
[  110.895012]  [<ffffffff813c8ed7>] ? schedule+0x2b7/0x860
[  110.895012]  [<ffffffff81044616>] ? try_to_wake_up+0x1b6/0x270
[  110.895012]  [<ffffffff8106a020>] ? abort_exclusive_wait+0xb0/0xb0
[  110.895012]  [<ffffffffa00494f0>] ? usb_remote_wakeup+0x40/0x40 [usbcore]
[  110.895012]  [<ffffffff8106985e>] ? kthread+0x7e/0x90
[  110.895012]  [<ffffffff813cd1a4>] ? kernel_thread_helper+0x4/0x10
[  110.895012]  [<ffffffff810697e0>] ? kthread_worker_fn+0x180/0x180
[  110.895012]  [<ffffffff813cd1a0>] ? gs_change+0x13/0x13
[  110.895012] Code: 29 60 a0 48 c7 83 60 06 00 00 70 29 60 a0 48 c7 83 68 06 00 00 50 29 60 a0 48 c7 83 70 06 00 00 f0 2b 60 a0 49 8b 87 38 27 00 00 
[  110.895012]  8b 40 18 ff 50 10 49 8b 87 38 27 00 00 4c 89 f7 48 8b 40 18 
[  110.895012] RIP  [<ffffffffa0604747>] rtl_usb_probe+0x19b/0x74d [rtlwifi]
[  110.895012]  RSP <ffff88011da77a60>
[  110.895012] CR2: 0000000000000018
[  110.944291] ---[ end trace aa8fb4280a86af53 ]---
# cat /proc/sys/kernel/tainted
128

The same procedure (after rebooting) doesn't result in an Oops with 
other modules, although they can't initialize this 'unknown' device 
either, using rt2500usb as example:
# cat /proc/sys/kernel/tainted
0
# modprobe rt2500usb
# echo -n "1385 4251" > /sys/bus/usb/drivers/rt2500usb/new_id
[hotplug device]
# dmesg
[...]
[  112.676033] usb 1-7: new high speed USB device number 5 using ehci_hcd
[  112.791945] usb 1-7: New USB device found, idVendor=1385, idProduct=4251
[  112.791951] usb 1-7: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[  112.791957] usb 1-7: Product: WG111T
[  112.791961] usb 1-7: Manufacturer: Atheros Communications Inc
[  112.791965] usb 1-7: SerialNumber: 1.0
[  112.895037] usb 1-7: reset high speed USB device number 5 using ehci_hcd
[  113.108950] phy0 -> rt2x00usb_vendor_request: Error - Vendor Request 0x09 failed for offset 0x0000 with error -32.
[  113.208947] phy0 -> rt2x00usb_vendor_request: Error - Vendor Request 0x07 failed for offset 0x04d0 with error -32.
[  113.308946] phy0 -> rt2x00usb_vendor_request: Error - Vendor Request 0x06 failed for offset 0x04ce with error -32.
[  113.408950] phy0 -> rt2x00usb_vendor_request: Error - Vendor Request 0x07 failed for offset 0x04d0 with error -32.
[  113.508946] phy0 -> rt2x00usb_vendor_request: Error - Vendor Request 0x07 failed for offset 0x04ce with error -32.
[  113.608450] phy0 -> rt2x00usb_vendor_request: Error - Vendor Request 0x07 failed for offset 0x0400 with error -32.
[  113.608457] phy0 -> rt2500usb_init_eeprom: Error - Invalid RT chipset detected.
[  113.608463] phy0 -> rt2x00lib_probe_dev: Error - Failed to allocate device.
[everything fine, as expected no issues - the wrong device just can't 
 get initialized]
# cat /proc/sys/kernel/tainted
0
# reboot

So at least the reported Oops is not specific to 050d:2103, but happens
for any USB ID injected at runtime through sysfs and therefore appears
to be a general issue with rtl8192cu.

Regards
	Stefan Lippers-Hollmann

Re: [PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[Larry Finger]

On 08/02/2011 06:20 PM, Stefan Lippers-Hollmann wrote:
>
> So at least the reported Oops is not specific to 050d:2103, but happens
> for any USB ID injected at runtime through sysfs and therefore appears
> to be a general issue with rtl8192cu.

As your system and mine are both x86_64, I had hoped to use your dump to find 
the offending statement. Unfortunately, the codes don't exactly match. I'll 
duplicate your experiment here and send you a trial patch.

Thanks for the report.

Larry

Re: [PATCH] rtl8192cu: add USB ID for Belkin n300 Micro USB wireless adapter

[Larry Finger]

On 08/02/2011 03:29 PM, Stefan Lippers-Hollmann wrote:
> Signed-off-by: Stefan Lippers-Hollmann<s.l-h@gmx.de>
> Cc: stable<stable@kernel.org>  [2.6.39+]
> ---
>   drivers/net/wireless/rtlwifi/rtl8192cu/sw.c |    3 +++
>   1 files changed, 5 insertions(+), 0 deletions(-)
>
> --- a/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
> +++ b/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
> @@ -322,6 +322,9 @@ static struct usb_device_id rtl8192c_usb
>   	{RTL_USB_DEVICE(0x2001, 0x330a, rtl92cu_hal_cfg)}, /*D-Link-Alpha*/
>   	{RTL_USB_DEVICE(0x2019, 0xab2b, rtl92cu_hal_cfg)}, /*Planex -Abocom*/
>   	{RTL_USB_DEVICE(0x7392, 0x7822, rtl92cu_hal_cfg)}, /*Edimax -Edimax*/
> +
> +	/****** 8192CUS Dongle ********/
> +	{RTL_USB_DEVICE(0x050D, 0x2103, rtl92cu_hal_cfg)}, //Belkin - Edimax
>   	{}
>   };

NACK. That C++ comment should be fixed.

See my response to the second Email for more discussion.

Larry