VirtualBox

Opened 13 years ago

Closed 4 years ago

#9511 closed defect (obsolete)

USB WiFi ath9k_htc device not initialized in Linux guest

Reported by: solamour Owned by:
Component: USB Version: VirtualBox 4.1.2
Keywords: Cc:
Guest type: Linux Host type: Linux

Description (last modified by Frank Mehnert)

I'm running VirtualBox 4.1.2 in Ubuntu 10.04 host with Gentoo Linux as the guest. I've been trying to use a USB WiFi device (Atheros chipset) in the guest OS, but so far no luck yet.

  • When I put everything in the guest OS in a real machine (i.e. no VirtualBox guest), the USB WiFi device works OK.
  • If the guest OS is Windows XP, I end up with an yellow exclamation in the Device Manager.
  • If the guest OS is Linux, I'm getting the following kernel message.

[ 3.282211] ath9k_htc 1-1:1.0: usb_probe_interface [ 3.282222] ath9k_htc 1-1:1.0: usb_probe_interface - got id [ 3.772725] usb 1-1: ath9k_htc: Transferred FW: ar9271.fw, size: 51312 [ 4.772194] ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive [ 4.772287] Failed to initialize the device [ 4.774250] ath9k_htc: probe of 1-1:1.0 failed with error -22 [ 4.775452] usbcore: registered new interface driver ath9k_htc

# lsusb | grep -i ath Bus 001 Device 003: ID 0cf3:9271 Atheros Communications, Inc. AR9271 802.11n

My guess is that the USB WiFi device itself and the setup in the Linux guest might not be an issue, because when I use a real machine, everything works perfectly. sol

Attachments (1)

dmesg.jpg (37.2 KB ) - added by NetSec144 9 years ago.
The ath9k_htc dmesg results

Download all attachments as: .zip

Change History (59)

comment:1 by tdrd, 12 years ago

Confirmed this issue using Mac OS X Lion host with BT5R1 guest. USB device was same as above, 0cf3:9271.

Exact same host/guest/hardware worked with Parallels Desktop for Mac 7.

comment:2 by haozhun, 12 years ago

Confirmed this issue using Windows 7 x64 host with Debian Squeeze (backports linux-image-686 3.2 and backports firmware-atheros 0.35). Virtualbox version is 4.1.8

Device: TP-Link WN821N v3 (using Atheros AR7010+AR9287)

Tested the device on Debian without using virtualbox, and it works fine. Tested the device on Debian using virtualbox, got the following in /var/log/syslog

kernel: [   98.212097] usb 2-2: USB disconnect, device number 3
kernel: [  102.654780] usb 1-1: new high-speed USB device number 2 using ehci_hcd
kernel: [  103.279004] usb 1-1: New USB device found, idVendor=0cf3, idProduct=7015
kernel: [  103.279014] usb 1-1: New USB device strings: Mfr=16, Product=32, SerialNumber=48
kernel: [  103.279020] usb 1-1: Product: UB95
kernel: [  103.279025] usb 1-1: Manufacturer: ATHEROS
kernel: [  103.279030] usb 1-1: SerialNumber: 12345
kernel: [  103.597849] usb 1-1: ath9k_htc: Transferred FW: htc_7010.fw, size: 72992
kernel: [  104.596310] ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive
kernel: [  104.596328] Failed to initialize the device
kernel: [  104.605694] ath9k_htc: probe of 1-1:1.0 failed with error -22

Does any one know how this can be solved? Or any work around? Is this a bug in Virtualbox USB support code or in ath9k_htc?

Last edited 12 years ago by haozhun (previous) (diff)

comment:3 by Zullwick, 12 years ago

Same issue using Virtualbox 4.1.10 r76836 with Host system Windows 7 x64 Ultimate and guest system Backtrack 5. Using TP-Link TL-WN722N USB adapter.

USB adapter works fine in Backtrack 5 when booted up from a USB stick but not in a virtual machine. Has same issue in Mint 12 x64.

Any fix on the way?

[   50.397906] usb 1-2: ath9k_htc: Transferred FW: ar9271.fw, size: 51312
[   51.390246] ath9k_htc 1-2:1.0: ath9k_htc: Target is unresponsive
[   51.390251] Failed to initialize the device
[   51.401861] ath9k_htc: probe of 1-2:1.0 failed with error -22
[   51.401876] usbcore: registered new interface driver ath9k_htc

comment:4 by Frank Mehnert, 12 years ago

Description: modified (diff)

USB is a very complex interface and there are always devices which will not properly work in VirtualBox guests. If you can, please don't pass that USB device into the guest but let it drive by the host and use the normal VirtualBox networking functions to access the network over the USB network card.

comment:5 by Digininja, 12 years ago

Any movement on fixing this? I can't just passed through networking as I want access to the actual wifi card for monitor and master modes.

I'm seeing the error with Debian guest on a Windows 7 host.

comment:6 by Mr-Protocol, 12 years ago

I (and I am sure many others) would be appreciative for a fix for this. Direct access to the card expands so many more possibilities.

If any info is needed to help resolve this, please let us know.

My setup: Win7 x64 Host BackTrack 5r3 Guest ALFA AWUS036NHA - USB WiFi Card

Version 0, edited 12 years ago by Mr-Protocol (next)

comment:7 by AlexBotezatu, 11 years ago

I can confirm this same behaviour on vmware. Had spent hours in trying to figure it out on both virtualbox and vmware. However, there is a workaround in vmware. Using auto-connect feature, physically extract the device from the usb slot of the host machine and re-insert it back, and it magically works. Every single time. If you disconnect the device via the menu, it behaves exactly like above (target is unresponsive, etc) in VirtualBox and Vmware guests alike.

Confirmed it on two different guest distro's (gentoo and mint) with a mint 12 host. Hope this helps.

Last edited 11 years ago by AlexBotezatu (previous) (diff)

comment:8 by rmfalco89, 11 years ago

Same issue using Virtual Box 4.2.4 on Ubuntu 10.10 host. Trying to make my netgear n150 working on Backtrack guest.

[ 1144.636298] usb 1-1: new high-speed USB device number 4 using ehci_hcd
[ 1145.086750] usb 1-1: ath9k_htc: Firmware htc_9271.fw requested
[ 1145.612418] usb 1-1: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272
[ 1146.612298] ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive
[ 1146.612331] ath9k_htc: Failed to initialize the device
[ 1146.621503] usb 1-1: ath9k_htc: USB layer deinitialized
[ 1160.682062] usb 1-1: USB disconnect, device number 4

Tried in all the ways I can think of, but nothing, it seems that virtual box doesn't make possible to view the adapter as really connected to the usb port of the guest system. sad :(

Last edited 11 years ago by rmfalco89 (previous) (diff)

comment:9 by expert_vision, 11 years ago

Same issue using VirtualBox 4.2.6 on Windows 7 64-bit host with Ubuntu 10.10 guest. It also doesn't work in Windows XP guest, after the driver have been successfully installed, with the following error:

This device cannot start. (Code 10)

USB capture method in VirtualBox should probably be changed completely, as there are a lot of issues currently regarding USB connectivity, especially with USB 2.0 devices; and USB connection is vital for many developing work-sets.

comment:10 by Alessander Botti Benevides, 11 years ago

I confirm this issue for the following configuration:

VirtualBox version:4.2.6
USB WiFi device:TP-LINK TL-WN722N
Host:Mac OS X Version 10.7.5
Guest:Kubuntu 12.10 64 bit, kernel 3.5.0-23-generic

comment:11 by iriki, 11 years ago

Same issue (they just keep coming!!)

VirtualBox version: 4.2.6; USB WiFi device: TP-LINK TL-WL722N; Host: Windows 7 Ultimate 32 bits; Guest: Ubuntu 32 bits

comment:12 by JayZeeGP, 11 years ago

Same Issue :( VirtualBox version: 4.1.18 Guest: BackTrack Host: Kubuntu USB WiFi device:TP-LINK TL-WL722N

comment:13 by FabiettoAOL, 11 years ago

Same Issue :( VirtualBox version: 4.1.18 Guest: BackTrack Host: Windows 7 x64 -- USB WiFi device:TP-LINK TL-WL722N. Virtualbox sees it like eth2 and it0s impossible to connect through it. It's passed 1 year and still the problem exists.... maybe it's better to start to use VMplayer?!

comment:14 by ferama, 11 years ago

Same problem here. VirtualBox 4.2.10

comment:15 by Shane, 11 years ago

Oracle / Virtual Box Developers,

I have been following this ticket for a while now. I always update Virtual Box to the latest patches. I am now at 4.2.10 r84104. I am also experiencing the same USB errors. My setup is as follows:

HOST OS is Windows 7 64bit Ultimate
Guest OS's are Linux Mint 14 Mate and Ubuntu 12.04 LTS
I get the same error messages in syslog from both Guest OS's. See Below:

ath9k_htc 1-2:1.0: >ath9k_htc: Target is unresponsive
ath9k_htc: Failed to initialize the device
usb 1-2: >ath9k_htc: USB layer deinitialized

I have a TL-WN722N USB WiFi Adapter that I am trying to use. This is a known compatible card with both Linux versions listed above. I have tested it with Linux installed directly on physical hardware and it works perfectly. Just not in the Virtual Machine.

Oracle,
Please look into this and provide an update as soon as possible.
Thank You.

Last edited 11 years ago by Shane (previous) (diff)

comment:16 by Manozco, 11 years ago

I have the same problem

Host: Win 7 x64

Guest: Kali Linux 1.0 (Backtrack 6)

VBox version: 4.2.10 r84104

Wifi Chip: TP Link WN722N (Atheros AR9271)

comment:17 by narusuke, 11 years ago

me too

Host : Linux Mint 14 MATE (Kernel Linux 3.5.0-22-generic)

Guest : Linux Kali / Backtrack 6 (Kernel Linux 3.7-trunk-amd64)

VBox Version : 4.2.10 r84104 (Latest Version)

Wireless Adaptor : Alfa AWUS036NHA

Chipset : Atheros AR9271 / Atheros UB91C

On the host my alfa work perfectly, but on my guest is like this :

[ 2023.183715] usb 1-1: Product: UB91C
[ 2023.183723] usb 1-1: Manufacturer: ATHEROS
[ 2023.183730] usb 1-1: SerialNumber: 12345
[ 2023.185712] usb 1-1: ath9k_htc: Firmware htc_9271.fw requested
[ 2023.623632] usb 1-1: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272
[ 2024.624291] ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive
[ 2024.624344] ath9k_htc: Failed to initialize the device
[ 2024.646098] usb 1-1: ath9k_htc: USB layer deinitialized

To VBox programmers/developers, please solve it, thanks

Last edited 11 years ago by narusuke (previous) (diff)

comment:18 by bonzz, 11 years ago

confirmed

Host: Debian Wheezy 3.2.0-4-amd64 #1 SMP Debian 3.2.41-2 x86_64 GNU/Linux
vBox: virtualbox-4.2_4.2.12-84980~Debian~wheezy_amd64.deb from virtualbox apt repo
Guest: Kali Linux 1.0
Wifi: Alfa AWUS036NHA
Chip: Atheros AR9271 / UB91C

I am able to provide a capture (wireshark + usbmon) upon request from virtualbox.

comment:19 by narusuke, 11 years ago

still same with latest vbox virtualbox-4.2_4.2.12-84980~Ubuntu~precise_amd64

i still unable to see wlan

i use compat-drivers-3.9-rc4-2 (newest) so what s wrong with my laptop? i ve followed http://forums.hak5.org/index.php?/topic/27344-alfa-and-wmware/page-2#entry218754 but still same bonzz can you share your experience?

comment:20 by Frank Mehnert, 11 years ago

The preferred way to get Internet over wlan into a VM is to use the WLAN adapter on the host and using normal NAT for the VM. Passing USB WLAN adapters to the guest is almost untested.

in reply to:  20 comment:21 by Alessander Botti Benevides, 11 years ago

Replying to frank:

The preferred way to get Internet over wlan into a VM is to use the WLAN adapter on the host and using normal NAT for the VM. Passing USB WLAN adapters to the guest is almost untested.

In my case, the host (Mac OS X) does not support the device TP-LINK TL-WN722N. Therefore, NAT is not an option. That's why I'm trying to pass it to the guest. If anyone knows how to use TP-LINK TL-WN722N in Mac, please drop a comment.

comment:22 by IvanX, 11 years ago

And, in my case, I want to be able to pass AppleTalk packets over Wi-Fi. These don't survive the virtual network adapter in Virtualbox, whether or not the host is in NAT or bridged mode. Natively recognizing a Wi-Fi USB adapter would solve the problem (and Parallels Desktop actually does solve it; I can use a TP-Link TL-WN721N on the emulated USB port in Linux without difficulty there).

in reply to:  20 comment:23 by niadev67, 11 years ago

Replying to frank:

The preferred way to get Internet over wlan into a VM is to use the WLAN adapter on the host and using normal NAT for the VM. Passing USB WLAN adapters to the guest is almost untested.

I too require this feature. My VBox host is Solaris. I need to pass-through my TP-LinK Wireless N USB Adapter TL-WN722N to a Linux VM to create a Wireless Access Point. I have been waiting for a fix for a long time and still hope for a resolution soon so I can retire my old and slow current Wireless AP hardware.

comment:24 by patroashish, 11 years ago

I am having same issue (ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive) with a TP Link wireless card while using Ubuntu 12.04 as the Guest OS. An urgent resolution of the issue would be really helpful!

comment:25 by richardebreyer, 11 years ago

I would like to control in some detail the behavior of a PCI attached wireless device, the Atheros 9285, from within a virtual box – a virtual Ubuntu server operating as a Fedora VM, so I need a direct link to the PCI device. So far, no luck. Does anyone out there have a suggestion?

comment:26 by Fran Simó, 11 years ago

The same with TP-Link TL-WN721N
Host: Win 7 Enterprise SP1
Guest: Ubuntu 12.04 (Linux version 3.2.0-23-generic)

Last edited 11 years ago by Fran Simó (previous) (diff)

comment:27 by Zare, 11 years ago

Hello Oracle Team,

Firstly I'd like to say that I'm very appreciative of VirtualBox and all the work you guys have put into this awesome piece of free software over the past years. Although I've seen a steady flow of great improvements, there are some issues, such as the one listed in this ticket, which have been open for a while.

As I'm sure you're aware by know, there are many good and robust wifi dongles/cards which are utilizing AR9271 Atheros chipset, but most of them unfortunately don't have native Mac Os X support, which is why people like me have to rely your USB pass-through support. In my case, the guest system is Debian, which gives me a "Target is unresponsive" error.

Atheros have recently published open firmware for this chipset (https://github.com/qca/open-ath9k-htc-firmware), so I was wondering if you might consider trying to fix this issue in one of your next releases? If so, it'd be great if we could get an ETA on it?

Regards,

Zarko

comment:28 by Zare, 11 years ago

Hi All,

For anyone with the same problem (Host: OS 10.6.8, Guest: Kali Linux), I can confirm that my Atheros AR9271 card works perfectly in VMWare Fusion 5.0, I've tested it this morning.

Regards,

Zarko

comment:29 by hcelizondo, 11 years ago

Same issue with VB 4.3.0 Beta2 r88827. Host win7 ultimate, guest: Kali-Linux. Sad but true.


Sep 19 10:00:45 kali kernel: [ 157.484456] usb 1-1: new high-speed USB device number 2 using ehci_hcd

Sep 19 10:00:45 kali kernel: [ 157.898431] usb 1-1: New USB device found, idVendor=0cf3, idProduct=9271

Sep 19 10:00:45 kali kernel: [ 157.898435] usb 1-1: New USB device strings: Mfr=16, Product=32, SerialNumber=48

Sep 19 10:00:45 kali kernel: [ 157.898436] usb 1-1: Product: USB2.0 WLAN

Sep 19 10:00:45 kali kernel: [ 157.898438] usb 1-1: Manufacturer: ATHEROS

Sep 19 10:00:45 kali kernel: [ 157.898439] usb 1-1: SerialNumber: 12345

Sep 19 10:00:46 kali kernel: [ 158.018680] cfg80211: Calling CRDA to update world regulatory domain

Sep 19 10:00:46 kali kernel: [ 158.075984] cfg80211: World regulatory domain updated:

Sep 19 10:00:46 kali kernel: [ 158.075988] cfg80211: (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp)

Sep 19 10:00:46 kali kernel: [ 158.075990] cfg80211: (2402000 KHz - 2472000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)

Sep 19 10:00:46 kali kernel: [ 158.075992] cfg80211: (2457000 KHz - 2482000 KHz @ 20000 KHz), (300 mBi, 2000 mBm)

Sep 19 10:00:46 kali kernel: [ 158.075993] cfg80211: (2474000 KHz - 2494000 KHz @ 20000 KHz), (300 mBi, 2000 mBm)

Sep 19 10:00:46 kali kernel: [ 158.075994] cfg80211: (5170000 KHz - 5250000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)

Sep 19 10:00:46 kali kernel: [ 158.075996] cfg80211: (5735000 KHz - 5835000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)

Sep 19 10:00:46 kali kernel: [ 158.105961] usb 1-1: ath9k_htc: Firmware htc_9271.fw requested

Sep 19 10:00:46 kali kernel: [ 158.106439] usbcore: registered new interface driver ath9k_htc

Sep 19 10:00:46 kali kernel: [ 158.590527] usb 1-1: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272

Sep 19 10:00:47 kali kernel: [ 159.588108] ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive

Sep 19 10:00:47 kali kernel: [ 159.588118] ath9k_htc: Failed to initialize the device

Sep 19 10:00:47 kali kernel: [ 159.590952] usb 1-1: ath9k_htc: USB layer deinitialized


Edit. Same Host and same guest working perfectly on VMware workstation 10 trial version

Last edited 11 years ago by hcelizondo (previous) (diff)

comment:30 by Thierry M, 10 years ago

FYI

Same issue with latest virtual box 4.3 with Netgear WNA1100. Host/Guest: Windows 7 64 bits/Debian Wheezy

Works with VM Player 5.X.

Last edited 10 years ago by Thierry M (previous) (diff)

comment:31 by hany3, 10 years ago

Unfortunately the issue still present with latest version of VirtualBox 4.3.2.90405 which FAILS to connect to my usb wireless adapter TL-WN722N which uses Atheros AR9271 chipset using any driver whether provided by tp-link or atheros in both scenarios virtualbox fails to connect to them and sadly in both scenarios VMware efficiently connects to them

it's shocking to see a bug reports started 2 years ago and the virtualbox still unable to fix this bug until now

how many complaints should be made in this thread to make the virtualbox development team give some attention to this silly bug in virtual box , this the main reason why i'm switched to vmware detaching virtualbox for the lack support and attention we are getting from the development team

comment:32 by Klaus Espenlaub, 10 years ago

Sorry to state the obvious - but we don't have the necessary hardware. I only see complaints, and no one offers any kind of help. Remember, you're taking the free ride so it's stretching things to expect that we can justify investing time and money into something which for some funny reason no customer has ever mentioned.

We do what's possible, but as long as we don't have the hardware we have to rely on the community. Bugs don't fix themselves magically if many people in a bug ticket say that they're affected, too. I know that it's cheap gear, but in a big corporation every cent needs to be justified, which is only easily done if there's a customer with a support contract anywhere in sight.

in reply to:  32 comment:33 by hany3, 10 years ago

Replying to klaus:

if making virtualbox shareware for more resources will make it a better product with some support so let it be about this issue , it's not logic to implement a usb support while excluding one of the most common usb wireless adapters currently , the 10 Dollars priced tp-link TL-WN722N and its equivalent Atheros AR9271 usb wirless adapters

Last edited 10 years ago by hany3 (previous) (diff)

comment:34 by Klaus Espenlaub, 10 years ago

Making VirtualBox shareware, just for some USB WiFi adapter support which as Frank stated above should be handled by the host OS? I can't even imagine what the reaction from the open source community would be. Don't think this gets us closer to a solution.

in reply to:  34 comment:35 by hany3, 10 years ago

Replying to klaus: i didn't suggest that until u indirectly referred to by asking for some resources to solve that issue , actually there are many open source freewares out there provide excellent support than for ex. micro$oft do

in reply to:  34 comment:36 by keremer, 10 years ago

Replying to klaus:

Making VirtualBox shareware, just for some USB WiFi adapter support which as Frank stated above should be handled by the host OS? I can't even imagine what the reaction from the open source community would be. Don't think this gets us closer to a solution.

Dear Klaus,

I am having difficulty to understand your position. All these users are talking about some problem about the way VirtualBox handles the USB allocation to guests. But it somehow comes down to whether the resource allocation for finding a solution is justifiable or not. This is not just a problem limited to the Atheros USB Driver. When it comes to making USB hardware available to VirtualBox guests, it is always a gamble whether it would work or not. There are many loopholes VirtualBox USB Capture for guests. I myself having the same issue exactly as described by many people above. Also I have returned another USB Wi-Fi adapter which was not even posible to attach to VirtualBox guest except for only some occassions.

I can use the same Atheros hardware without any problem with VMWare Workstation or when I boot my system directly with BTR/Kali/Ubuntu. So shall we continue on having this discussion whether it is worth to fix the problem or get down to scrutinize the USB capture code? This is the real question. You say nobody gives you a hand what is that you need to start on working over the USB Capture issue?

comment:37 by Frank Mehnert, 10 years ago

Reading complaints like it's shocking to see a bug reports started 2 years ago and the virtualbox still unable to fix this bug until now doesn't improve our willingness to invest time.

And as we mentioned several times in other tickets: We have to prioritize our work as we have limited resources. Even if a bug is open for several years it does not mean that the priority of such a bug increases. You might be aware that there are about 4,000 tickets open in the public bug tracker. And of course, every reporter thinks that his issue is the most important one. Sorry to say, but we have a different view.

in reply to:  37 comment:38 by keremer, 10 years ago

Replying to frank:

Reading complaints like it's shocking to see a bug reports started 2 years ago and the virtualbox still unable to fix this bug until now doesn't improve our willingness to invest time.

And as we mentioned several times in other tickets: We have to prioritize our work as we have limited resources. Even if a bug is open for several years it does not mean that the priority of such a bug increases. You might be aware that there are about 4,000 tickets open in the public bug tracker. And of course, every reporter thinks that his issue is the most important one. Sorry to say, but we have a different view.

OK Thank you for the information. Trying to attach USB devices to a Virtual system is a low priority issue with almost no interest to anyone.

comment:39 by Klaus Espenlaub, 10 years ago

keremer, you're misinterpreting our replies. It is NOT low priority to have working USB passthrough, but in this particular case there are two prerequisites for making progress, and that's what's causing the complete stand still:

  1. Having an affected device (we don't care if it's ath9k or not)
  2. Having free developer resources

At the moment condition 1 isn't met (we have no USB devices available which can be made working), this won't make progress. If the USB passthrough code would have trouble with a wide variety of devices then I'd agree that starting the investigation from the source code side is a good approach, but this isn't the case. The vast majority of devices does work very well, just extremely few don't, and then staring at the sources is a waste of time.

We'd have USB debugging tools available, but without an affected device they don't help.

I think I repeated "affected device" by now enough to make it obvious that this is the key to make progress, and in my previous message I already mentioned that even getting such cheap device needs a lot of effort, which won't happen if everyone is busy and budgets are tight.

That's what Frank summarized in a somewhat frustrated way as "prioritize work".

As all my previous subtle hints were only abused as starting points for complete misinterpretations of the situation I'll spell it out now: if we miraculously find such a device (which is working in general) in our mail then condition 1 will be out of the way, and while that doesn't guarantee immediate developer availability (can still take months) it's resolving the so far blocking prerequisite. We promise to send it back if the owner wants that, and in any case this will be acknowledged in the documentation as a contribution if desired.

in reply to:  39 ; comment:40 by keremer, 10 years ago

Replying to klaus:

keremer, you're misinterpreting our replies. It is NOT low priority to have working USB passthrough, but in this particular case there are two prerequisites for making progress, and that's what's causing the complete stand still:

  1. Having an affected device (we don't care if it's ath9k or not)
  2. Having free developer resources

This is what I am trying to tell.. I don't know *any* USB Wi-Fi adapter that works with your USB passthrough. Currently I have tested 3 different adapters with different chipsets. To get them workign was always hectic. Sometimes it worked to boot the system and then attaching the adapter (BCM, ATH and RealTek) solved the issue temporarily until you remove/attach and sometimes did not. Sometimes VirtualBox could isolate the adapter successfully from the hypervisor and attach it to the virtual gust. Sometimes it did not.

In fact it seems that there's a problem with USB passthrough not only a single chipset. Please use whatever USB Wireless adapter at your hand and see if it is removed from the system and attached to the guest every time you plug it in and out.

At the moment condition 1 isn't met (we have no USB devices available which can be made working), this won't make progress. If the USB passthrough code would have trouble with a wide variety of devices then I'd agree that starting the investigation from the source code side is a good approach, but this isn't the case. The vast majority of devices does work very well, just extremely few don't, and then staring at the sources is a waste of time.

We'd have USB debugging tools available, but without an affected device they don't help.

I think I repeated "affected device" by now enough to make it obvious that this is the key to make progress, and in my previous message I already mentioned that even getting such cheap device needs a lot of effort, which won't happen if everyone is busy and budgets are tight.

That's what Frank summarized in a somewhat frustrated way as "prioritize work".

As all my previous subtle hints were only abused as starting points for complete misinterpretations of the situation I'll spell it out now: if we miraculously find such a device (which is working in general) in our mail then condition 1 will be out of the way, and while that doesn't guarantee immediate developer availability (can still take months) it's resolving the so far blocking prerequisite. We promise to send it back if the owner wants that, and in any case this will be acknowledged in the documentation as a contribution if desired.

I guess you don't need a miracle to correct this situation because it is about USB passthrough and not specific devices. Besides if you please indicate the correct channels I am sure someone could donate an adapter for you. If it were only a single chip issue there were not so many complaints about USB Passtrhrough in this or different treads. It is the real miracle how could you evade this for such a long time though.

In fact nobody expects immediate developer dispatch but there has been several major releases during the last two years and there was no improvement in USB passthrough about the attachment of any wireless USB devices.

comment:41 by keremer, 10 years ago

The interesting thing is USB VID is always different than it were in a Physical PC with each adapter (not only Atheros). As other people indicated it is detected by the Guest but it is unable to communicate further with the device after it detects. This is why I have to use VMWare Workstaion although all my other Virtual Guests are in VirtualBox and happy to use VB for virtualization.

Last edited 10 years ago by keremer (previous) (diff)

in reply to:  40 comment:42 by IvanX, 10 years ago

Replying to keremer:

I don't know *any* USB Wi-Fi adapter that works with your USB passthrough.

I've been able to successfully and predictably use a Realtek RTL8188CUS and a Ralink RA5370 in Virtualbox (Mac) hosting Debian 7.3.0 i386 guest.

Chipsets I haven't been able to be accepted by Virtualbox are the Atheros 9271 WiFi dongle and the ASIX 88178 and 88772 USB-to-Ethernet dongles.

I think it's just hit and miss, in my experience.

Last edited 10 years ago by IvanX (previous) (diff)

comment:43 by nashcontrol, 10 years ago

Working in Kali guest over Virtual Box 4.3.6!! (Thanks to my buddies at Hacktics ASC)

Steps:

  1. apt-get update
  2. apt-get upgrade
  3. apt-get dist-upgrade
  4. apt-get install linux-headers-$(uname -r)
  5. Shutdown Linux guest OS
  6. Download VirtualBox Extension Pack (https://www.virtualbox.org/wiki/Downloads)
  7. Apply extension at: File -> Preferences -> Extensions
  8. Configure the Linux guest OS and enable USB 2.0 Controller
  9. Connect and enable device
  10. Enjoy

comment:44 by chrismsnz, 10 years ago

Hi,

I am currently experiencing this bug and am willing to put some time in to help get it fixed. Please let me know how I can help.

Here are the circumstances in which I have experienced this issue:

  • Virtualbox Version: 4.3.6, 4.1.12 (ubuntu repos)
  • USB Extension Pack: Installed
  • Host OS: Ubuntu 12.04 64-bit and Debian 7.0 Wheezy 64-bit
  • Host Kernel: Default Ubuntu (3.5.0-45) and Wheezy kernels
  • Guest: Kali Linux 32-bit
  • Guest Kernel: 3.12-pae, 3.7-pae
  • Guest Additions: Installed

Device is an Alfa AWUS036NHA, chipset AR9271.

The following snippet is the kernel log when the device is inserted and passed through to the guest:

[   73.410377] usb 1-1: ath9k_htc: Firmware htc_9271.fw requested
[   73.410582] usbcore: registered new interface driver ath9k_htc
[   73.418759] usb 1-1: firmware: direct-loading firmware htc_9271.fw
[   73.869611] usb 1-1: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272
[   74.868328] ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive
[   74.868347] ath9k_htc: Failed to initialize the device
[   75.026956] usb 1-1: ath9k_htc: USB layer deinitialized
[   95.252377] usb 1-1: USB disconnect, device number 2

I currently have spare hardware and am willing to test combinations of host/guests as required, just need some direction on how to get started.

Cheers,

Chris.

chris at nevermind dot co dot nz

comment:45 by sneak, 10 years ago

I've also been bitten by this bug. Running latest Ubuntu 13.10 stable in latest stable virtualbox (4.3.6) with latest stable virtualbox matching extensions (also 4.3.6) with latest stable vagrant setting the whole thing up. Using standard off the shelf atheros ath9k-based usb wifi.

[  295.873968] usb 1-1: USB disconnect, device number 2
[  298.576324] usb 1-1: new high-speed USB device number 3 using ehci-pci
[  298.848768] usb 1-1: New USB device found, idVendor=0cf3, idProduct=9271
[  298.848775] usb 1-1: New USB device strings: Mfr=16, Product=32, SerialNumber=48
[  298.848779] usb 1-1: Product: UB91C
[  298.848782] usb 1-1: Manufacturer: ATHEROS
[  298.848786] usb 1-1: SerialNumber: 12345
[  298.872690] cfg80211: Calling CRDA to update world regulatory domain
[  298.878435] cfg80211: World regulatory domain updated:
[  298.878440] cfg80211:   (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp)
[  298.878443] cfg80211:   (2402000 KHz - 2472000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[  298.878446] cfg80211:   (2457000 KHz - 2482000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[  298.878449] cfg80211:   (2474000 KHz - 2494000 KHz @ 20000 KHz), (300 mBi, 2000 mBm)
[  298.878451] cfg80211:   (5170000 KHz - 5250000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[  298.878453] cfg80211:   (5735000 KHz - 5835000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[  298.901818] usb 1-1: ath9k_htc: Firmware htc_9271.fw requested
[  298.902484] usbcore: registered new interface driver ath9k_htc
[  299.346601] usb 1-1: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272
[  300.344196] ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive
[  300.344655] ath9k_htc: Failed to initialize the device
[  300.357050] usb 1-1: ath9k_htc: USB layer deinitialized
root@vagrant-ubuntu-saucy-64:~# lsusb
Bus 001 Device 003: ID 0cf3:9271 Atheros Communications, Inc. AR9271 802.11n
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub

Happy to help troubleshoot in any way required.

Jeffrey Paul

sneak at datavibe dot net

Last edited 10 years ago by sneak (previous) (diff)

comment:46 by kas21, 10 years ago

I am also experiencing this issue using VirtualBox 4.3.12 on Arch Linux. I have found that installing the Expansion Pack and creating a USB Filter for the WiFi card will allow the guest to start the card successfully. This doesn't fix everything as the card will stop (and restart) while using it. Please, please fix this issue. I love using VirtualBox but this issue is forcing me over to VMware in order to use this hardware. Thanks for all time and hard work.

comment:47 by nnikolic, 9 years ago

Got a workaround for me :

Hypervisor : Ubuntu Desktop 14.04
VirtualBox : 4.3.18 r96516
VirtualGuesAdditions : yes
Guests : Ubuntu 12.04, Ubuntu 14.04
Wifi Dongle : ID 0846:9030 NetGear, Inc. WNA1100 Wireless-N 150 [Atheros AR9271]
Driver : ath9k_htc

  1. Get Vendor & Product ID from wifi dongle ( with lsusb)
  2. check if the module ath9k_htc on your hypervisor is loaded
  3. if yes, unload module ( modprobe -r ath9k_htc)
  4. create a blacklist file in /etc/modprobe.d/ for the module : vim /etc/modprobe.d/blacklist-ath9k.conf type in the new file : blacklist ath9k_htc save file ...
  5. repluge the wifi dongle
  6. create manualy for your VM guest a USB filter (settings -> usb -> usb connector with blue circle) and add the Vendor & Product ID
  7. start your VM
  8. with dmesg | grep htc you shoud get :

[ 2.981893] usb 1-1: ath9k_htc: Firmware htc_9271.fw requested
[ 2.982436] usbcore: registered new interface driver ath9k_htc
[ 3.329829] usb 1-1: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272
[ 3.580004] ath9k_htc 1-1:1.0: ath9k_htc: HTC initialized with 33 credits
[ 4.263725] ath9k_htc 1-1:1.0: ath9k_htc: FW Version: 1.3

Thats it ! I hope I could help ....

comment:48 by zebul666, 9 years ago

this is working here with virtualbox 4.3.20 and a kali linux 1.0.9 vm and TP-LINK TL-WN722N But the network freezes after some random time like one hour.

I have to unplug and replug the usb dongle (if the VM does not crash when I unplug it)

See https://forums.virtualbox.org/viewtopic.php?f=3&t=64884

comment:49 by takersdown, 9 years ago

Im using centos 6.6 on VirtualBox 4.3.26

And my TP-LINK TL-WN722N didnt work. What should i have to do?

comment:50 by W0xing, 9 years ago

Hi I have a VBOX 4.3.28 in a debian jessie host. Im trying to use this TP-LINK TL-WN722N in a Kali linux vm, I follow the steps above, but I don't have any positive results, below I post the exit of the commands.

kali vm

# dmesg | grep htc

[ 104.615589] usbcore: registered new interface driver ath9k_htc

debian host

# modprobe -r ath9k_htc

# dmesg | grep htc

[11548.082446] usb 3-3: ath9k_htc: Firmware htc_9271.fw requested

[11548.082564] usb 3-3: firmware: direct-loading firmware htc_9271.fw

[11548.082624] usbcore: registered new interface driver ath9k_htc

[11548.366012] usb 3-3: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272

[11548.603648] ath9k_htc 3-3:1.0: ath9k_htc: HTC initialized with 33 credits

[11548.872194] ath9k_htc 3-3:1.0: ath9k_htc: FW Version: 1.3

[11597.119334] usbcore: deregistering interface driver ath9k_htc

[11597.211629] usb 3-3: ath9k_htc: USB layer deinitialized

[11597.211649] ath9k_htc: Driver unloaded

# modprobe ath9k_htc

# dmesg | grep htc

[18429.471157] usbcore: deregistering interface driver ath9k_htc

[18429.585548] usb 3-3: ath9k_htc: USB layer deinitialized

[18429.585582] ath9k_htc: Driver unloaded

[19192.948825] usb 3-3: ath9k_htc: Firmware htc_9271.fw requested

[19192.948862] usb 3-3: firmware: direct-loading firmware htc_9271.fw

[19192.948895] usbcore: registered new interface driver ath9k_htc

[19193.241513] usb 3-3: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272

[19193.479179] ath9k_htc 3-3:1.0: ath9k_htc: HTC initialized with 33 credits

[19193.745977] ath9k_htc 3-3:1.0: ath9k_htc: FW Version: 1.3

anyone can help?

in reply to:  39 comment:51 by dylan88, 9 years ago

Replying to klaus:

keremer, you're misinterpreting our replies. It is NOT low priority to have working USB passthrough, but in this particular case there are two prerequisites for making progress, and that's what's causing the complete stand still:

  1. Having an affected device (we don't care if it's ath9k or not)
  2. Having free developer resources

At the moment condition 1 isn't met (we have no USB devices available which can be made working), this won't make progress. If the USB passthrough code would have trouble with a wide variety of devices then I'd agree that starting the investigation from the source code side is a good approach, but this isn't the case. The vast majority of devices does work very well, just extremely few don't, and then staring at the sources is a waste of time.

Is this still an issue? I'd really like to have USB passthrough work properly. If you post the mailing address with Attn To: line so it gets to the right place I will have an atheros wifi dongle in the mail to you tomorrow.

by NetSec144, 9 years ago

Attachment: dmesg.jpg added

The ath9k_htc dmesg results

comment:52 by NetSec144, 9 years ago

I am having this same issue in vialab CE lunbuntu guest on an Ubuntu 14.04 LTS host. The error is the same as everyone else who post here. The "Taret is unresponsive" with an error -22.

I've tried the linux-firmware update as well as atheros-firmware and both have the same result. Is there another driver to try?

comment:53 by conixa, 9 years ago

Everyone please stop panicking.

To the devs: it's obvious there is a problem here, what do you need to get it fixed? Money, a wireless card, a Ferrari?

Everyone else: if you want to get it working while the devs are fixing it, just follow nnikolic comment: https://www.virtualbox.org/ticket/9511#comment:47 If you do as he describes it will work'''

comment:54 by Klaus Espenlaub, 9 years ago

I said several times (at least once in a PM to a forum user) that we need the hardware. The forum user offered sending one to us, but so far nothing happened.

It's a sad truth that buying a USB device with a specific chip is almost impossible. Every vendor sells whatever they can get for the lowest price, which means that there's no guarantee we'd get what we're after. We already have USB WiFi cards, but none of them have an Atheros chip.

in reply to:  50 comment:55 by SamuelBR, 9 years ago

Replying to W0xing:

Hi I have a VBOX 4.3.28 in a debian jessie host. Im trying to use this TP-LINK TL-WN722N in a Kali linux vm, I follow the steps above, but I don't have any positive results, below I post the exit of the commands.

kali vm

# dmesg | grep htc

[ 104.615589] usbcore: registered new interface driver ath9k_htc

debian host

# modprobe -r ath9k_htc

# dmesg | grep htc

[11548.082446] usb 3-3: ath9k_htc: Firmware htc_9271.fw requested

[11548.082564] usb 3-3: firmware: direct-loading firmware htc_9271.fw

[11548.082624] usbcore: registered new interface driver ath9k_htc

[11548.366012] usb 3-3: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272

[11548.603648] ath9k_htc 3-3:1.0: ath9k_htc: HTC initialized with 33 credits

[11548.872194] ath9k_htc 3-3:1.0: ath9k_htc: FW Version: 1.3

[11597.119334] usbcore: deregistering interface driver ath9k_htc

[11597.211629] usb 3-3: ath9k_htc: USB layer deinitialized

[11597.211649] ath9k_htc: Driver unloaded

# modprobe ath9k_htc

# dmesg | grep htc

[18429.471157] usbcore: deregistering interface driver ath9k_htc

[18429.585548] usb 3-3: ath9k_htc: USB layer deinitialized

[18429.585582] ath9k_htc: Driver unloaded

[19192.948825] usb 3-3: ath9k_htc: Firmware htc_9271.fw requested

[19192.948862] usb 3-3: firmware: direct-loading firmware htc_9271.fw

[19192.948895] usbcore: registered new interface driver ath9k_htc

[19193.241513] usb 3-3: ath9k_htc: Transferred FW: htc_9271.fw, size: 51272

[19193.479179] ath9k_htc 3-3:1.0: ath9k_htc: HTC initialized with 33 credits

[19193.745977] ath9k_htc 3-3:1.0: ath9k_htc: FW Version: 1.3

anyone can help?

Hi W0xing, I had the same issue with my VM (Kali 2.0) running as a guest on the same version of yours VBox (4.3.28), however my host is openSUSE 12.2 instead of Debian. I followed the same steps from nnikolic's tutorial above, and it didn't work too. Well, I discovered that the real problem wasn't about USB passthrough, and was about firmware that came with Kali. I finally got the device working after I followed the steps proposed by nnikolic and installed the package firmware-atheros from jessie-backports repository. Unfortunatelly, the version of firmware-atheros, that came with Kali 2.0 (0.44kali), doesn't work with TP-LINK TL-WN722N.

All you have to do is:

Remove the old package (firmware-atheros 0.44kali):

apt-get remove --purge firmware-atheros

Add jessie-backports apt repository:

echo “deb http://ftp.br.debian.org/debian jessie-backports main contrib non-free” >> /etc/apt/sources.list
apt-get update

And install the new firmware-atheros (0.44~bpo8+1):

apt-get install -t jessie-backports firmware-atheros

Remember, follow the steps proposed by nnikolic and install the new firmware-atheros from jessie-backports. Certainly it will work!

Best Regards

Last edited 9 years ago by SamuelBR (previous) (diff)

comment:56 by Thierry M, 8 years ago

Issue is still present in 5.0.8.

Fix proposed by SamuelRR does not works for Debian 7.2 (wheezy) with backported firmware, and Netgear N150: target is still unresponsive:

[67859.299277] usb 1-1: firmware: agent loaded htc_9271.fw into memory
[67859.663508] usb 1-1: ath9k_htc: Transferred FW: htc_9271.fw, size: 50980
[67860.661840] ath9k_htc 1-1:1.0: ath9k_htc: Target is unresponsive
[67860.664805] ath9k_htc: probe of 1-1:1.0 failed with error -22
[67860.665215] usbcore: registered new interface driver ath9k_htc

(50980-> size of new firmware).

comment:57 by michaln, 8 years ago

The AR9271 works with newer Linux guest kernels, starting around 3.13 or so. It indeed fails with older Linux kernels when using EHCI emulation due to a VirtualBox bug.

The AR9271 also fails with older Linux guest kernels when using xHCI emulation. Again newer guest kernels work starting around 3.5. This can't be fixed in VirtualBox.

The ath9k_htc driver is poorly written and schedules bulk transfers on interrupt endpoints. That causes both the EHCI and xHCI problems. The EHCI one we should be able to work around, the xHCI one not so.

There are separate problems with Windows guests.

comment:58 by aeichner, 4 years ago

Resolution: obsolete
Status: newclosed
Note: See TracTickets for help on using tickets.

© 2023 Oracle
ContactPrivacy policyTerms of Use