VirtualBox

Ticket #7462 (closed defect: fixed)

Opened 4 years ago

Last modified 10 months ago

USB device not available in a guest using more than 1 CPU

Reported by: Ludovic Owned by:
Priority: major Component: USB
Version: VirtualBox 4.0.8 Keywords:
Cc: Guest type: Linux
Host type: Mac OS X

Description (last modified by frank) (diff)

If I configure my guest OS to use more than 1 CPU then I can't use my USB device. I have a iMac model iMac11,1 with a Intel Core i7 (4 CPU with 2 threads each)

The USB device is a CCID smart card reader with vID=0x08E6 and pID=0x3437. The device works great if the guest uses only 1 CPU. But I get errors if I configure VirtualBox to use 2 or more CPU for the guest. The errors reported by the kernel are:

[   56.092448] usb 2-1: new full speed USB device using ohci_hcd and address 2
[   62.637182] atkbd.c: Spurious NAK on isa0060/serio0. Some program might be trying access hardware directly.
[   71.249524] usb 2-1: device descriptor read/64, error -110
[   86.512425] usb 2-1: device descriptor read/64, error -110
[   86.809503] usb 2-1: new full speed USB device using ohci_hcd and address 3
[  101.976574] usb 2-1: device descriptor read/64, error -110
[  117.240584] usb 2-1: device descriptor read/64, error -110
[  117.536353] usb 2-1: new full speed USB device using ohci_hcd and address 4
[  122.578952] usb 2-1: device descriptor read/8, error -110
[  127.698748] usb 2-1: device descriptor read/8, error -110
[  127.996433] usb 2-1: new full speed USB device using ohci_hcd and address 5
[  133.042903] usb 2-1: device descriptor read/8, error -110
[  138.161658] usb 2-1: device descriptor read/8, error -110
[  138.264556] hub 2-0:1.0: unable to enumerate USB device on port 1

and the lsusb command line tools is locked until the kernel gives up enumerating the device.

I have the problem with Debian testing/squeeze and also with Ubuntu 10.04. It is quiet easy to reproduce.

I also tried with Windows 7 and the device is visible with 1 CPU but not with a 2 CPU guest. So the problem is not limited to a Linux guest or the Linux kernel.

Since the problem is not linked to the guest OS it must be in VirtualBox itself.

Attachments

VBox.log Download (64.2 KB) - added by Ludovic 4 years ago.
VBox.2.log Download (65.7 KB) - added by lkraav 3 years ago.
USB Bus Probe.txt Download (32.2 KB) - added by iegorman 3 years ago.
Mac Mac OS X "Bus Probe" summary of USB flash drive and USB hard drive attached at the same time
USB System Information.txt Download (1.9 KB) - added by iegorman 3 years ago.
Mac OS X "System Information" summary of USB flash drive and USB hard drive attached at the same time
USB System Information.2.txt Download (1.9 KB) - added by iegorman 3 years ago.
Mac OS X "System Information" summary of USB flash drive and USB hard drive attached at the same time
USB System Information.3.txt Download (1.9 KB) - added by iegorman 3 years ago.
Mac OS X "System Information" summary of USB flash drive and USB hard drive attached at the same time
USB-WinXP-2CPU-2011-09-22-VBox.log Download (92.5 KB) - added by iegorman 3 years ago.
VirtualBox log of run with USB hard drive and USB flash drive
USB Flash Drive Imation.txt Download (1.1 KB) - added by iegorman 3 years ago.
Mac OS X "System Information" summary of Imation USB flash drive
USB Flash Drive SanDisk.txt Download (1.1 KB) - added by iegorman 3 years ago.
Mac OS X "System Information" summary of SanDisk USB flash drive
USB-WinXP-2CPU-2011-09-25-VBox.log Download (84.0 KB) - added by iegorman 3 years ago.
VirtualBox log of run with Imation USB flash drive and Sandisk USB flash drive
VBlogs.7z Download (23.4 KB) - added by blueytoo 12 months ago.
VBLogs.1.7z Download (16.0 KB) - added by blueytoo 12 months ago.
Screen shot 2013-05-08 at 1.04.44 am.png Download (138.7 KB) - added by blueytoo 12 months ago.
VirtualBoxVM_2013-05-08-193910_localhost.crash Download (51.8 KB) - added by blueytoo 12 months ago.
2vCPU crash
VBox.App.logs.7z Download (30.3 KB) - added by blueytoo 12 months ago.
VBox application logs
sandisk-usb-probe Download (3.3 KB) - added by blueytoo 11 months ago.
Sandisk SDAD-111
usb-log.2.txt Download (10.1 KB) - added by blueytoo 11 months ago.
RossTech problems log
usb-log.3.txt Download (4.3 KB) - added by blueytoo 11 months ago.
usb-log-when-Rosstech-malfunctioning
VBCrashlogs20130512.7z Download (13.9 KB) - added by blueytoo 11 months ago.
more crashlogs 2vCPU

Change History

Changed 4 years ago by Ludovic

comment:1 Changed 4 years ago by stepheneb

I have confirmed the same behavior.

I have a host OS of Mac OS X 10.6.4, MacBook Pro, Intel Core i7. The guest OS is Windows 7 Professional.

I have setup a USB filter for my device. It appears properly identified in the VirtualBox Active USB devices pulldown menu in the lower right.

If I enable two CPU's Windows 7 reports a failure trying to load device driver software for the new USB hardware.

When I enabled just one CPU Windows saw the new hardware and reported success installing software and adding the USB device.

The USB device in question is a Vernier GoLINK which is a HID device.

comment:2 Changed 4 years ago by iegorman

I have attached a note to Ticket #7346 about a similar problem with Windows XP and a USB "thumb" that I resolved by using a USB filter with "remote" field set to "No" and other fields blank.

comment:3 Changed 4 years ago by Ludovic

My USB configuration filter is set to remote: no. So that is not the problem (nor the solution).

The problem is still present in VirtualBox 3.2.10 r66523

comment:4 Changed 3 years ago by MacVette

I started a thread on this same problem before I came across this bug ticket:

 My Thread

The problem is still present in VirtualBox 3.2.12

I was attempting to use a Thumb drive and an Aladdin HASP security key. Both devices work when setting CPU's to 1. No USB devices work with CPU's >= 2 (other than the keyboard).

My setup:
Host - iMac Core i7 2.93 GHz (Quad-Core) / 16 GB RAM / Snow Leopard / 64-bit Kernel
Guest - Windows 7 Pro 64-bit / VirtualBox Guest Additions installed

comment:5 Changed 3 years ago by Ludovic

The problem is still present in VirtualBox version 4.0.2 r69518

comment:6 Changed 3 years ago by omascia

The problem is still present in VirtualBox version 4.0.4

comment:7 Changed 3 years ago by diosdog

Problem still present in 4.0.6. Macbook Pro quad core i7; host: Mac OS X; guests: Fedora 14 and Gentoo.

comment:8 Changed 3 years ago by 7xacker

I see the same problem on Mac OS X 10.6.7 (10J869) iMac10,1 VirtualBox 4.0.8 and Ubuntu 10.10 as guest OS

yurys@yurys-ubuntu:~$ dmesg <...> [ 31.537302] ISO 9660 Extensions: RRIP_1991A [ 128.000310] usb 2-2: new full speed USB device using ohci_hcd and address 3 [ 143.156245] usb 2-2: device descriptor read/64, error -110 [ 158.420259] usb 2-2: device descriptor read/64, error -110 [ 158.660263] usb 2-2: new full speed USB device using ohci_hcd and address 4 [ 173.824222] usb 2-2: device descriptor read/64, error -110 [ 189.084308] usb 2-2: device descriptor read/64, error -110 [ 189.324264] usb 2-2: new full speed USB device using ohci_hcd and address 5 [ 194.376463] usb 2-2: device descriptor read/8, error -110 [ 199.497255] usb 2-2: device descriptor read/8, error -110 [ 199.744277] usb 2-2: new full speed USB device using ohci_hcd and address 6 [ 204.782119] usb 2-2: device descriptor read/8, error -110 [ 209.900838] usb 2-2: device descriptor read/8, error -110 [ 210.004206] hub 2-0:1.0: unable to enumerate USB device on port 2 yurys@yurys-ubuntu:~$

comment:9 Changed 3 years ago by 7xacker

One addition, setting number of CPUs to 1 fixes the issue for me as well.

comment:10 Changed 3 years ago by chaumes

I can also confirm that this is still a problem in 4.0.8 on Mac OS X host. 9 months and still no action?

comment:11 Changed 3 years ago by CaptSaltyJack

Agreed. 9 months and it's not fixed?? It doesn't even show this ticket as being assigned to anyone.

comment:12 Changed 3 years ago by frank

You will notice that most tickets are not assigned to anybody for the simple reason that it does not matter who of the Oracle team is fixing the bug. And perhaps you will also realize that there are many many other important bugs to fix. Of course you think that this bug is the most important one but the priority is defined by the VirtualBox development team, mainly driven by customer demands. I agree that this is a nasty bug but the current workaround is to use only one guest CPU if you need to setup a VM with USB support.

comment:13 Changed 3 years ago by frank

  • Version changed from VirtualBox 3.2.8 to VirtualBox 4.0.8

comment:14 Changed 3 years ago by CaptSaltyJack

Ok ok.. I agree it's not mission critical. Hope it's fixed soon. BTW, the other workaround is to temporarily move/rename the file /System/Library/Extensions/IOUSBMassStorageClass.kext, but doing so I believe temporarily breaks USB mass storage devices on the Mac itself.

comment:15 Changed 3 years ago by prairie

I have encountered this problem also on an iMac (mid 2011) with quad cores and using just one cpu instead of two in my Win 7 VM resolves it. It would be good to have a warning note in the release information for VB about this situation until it is resolved so other users don't encounter it and have to hunt for the work around.

comment:16 Changed 3 years ago by prairie

Failed to mention that I am on OS X 10.6.8 using VB 4.1, but saw it on 4.0.12, 4.0.10 also.

comment:17 Changed 3 years ago by CaptSaltyJack

This bug is showing up in VBox 4.1, on MacOS X, even with just 1 CPU. Guest is Ubuntu 11.04. Any ideas? Every time I connect the USB to the guest, I can see it show up as "Kingston DT 101 G2" but it's not usable in any way, not even visible in gparted. Something is wrong.

comment:18 Changed 3 years ago by prairie

I'm not seeing the problem with Ubuntu 11.04 running under VBox 4.1 on my 10.6.8 system when only 1 cpu is configured with VT-X and Nested Paging enabled.

comment:19 Changed 3 years ago by CaptSaltyJack

Strange. I've got those options enabled too. I have PAE enabled and IO APIC but I disabled them and got the same result still. I'm completely unable to access the USB thumb drive.

comment:20 Changed 3 years ago by lkraav

I guess I might as well jump on this one. Running VirtualBox 4.1.2. In some earlier versions, possibly 3.x, I've had USB devices working just fine for sure.

Now unable to use a simple USB thumb drive, Linux 2.6 guest, on OS X 10.6.8 host. Selecting the USB device from the window icon causes it to briefly disconnect from host, then immediately reconnect. AFAICT guest never sees it. After that guest keeps running, but it can no longer be shut down by VirtualBox ("VirtualBox failed to stop <machine>" error dialog). I've had to Force Quit the guest machine process for now.

I will take a look at VBox.log in the morning.

Same thumb drive works fine in Linux host VirtualBox for the same guest.

Changed 3 years ago by lkraav

comment:21 Changed 3 years ago by lkraav

unable to get USB device connected no matter what I try. i'm not seeing any error messages getting logged for some reason.

comment:22 Changed 3 years ago by lkraav

noticed that if i don't install USB 2.0 support with the extension pack, i am at least able to power off the virtual machine normally.

comment:23 Changed 3 years ago by iegorman

I can confirm this with VirtualBox 4.1.2 running in Mac OSX Lion on a mid-2009 Macbook Pro

Running Ubuntu 10.4 (Long term Support), I can access USB flash drives if I turn off APIC and set one CPU.

Running Windows XP on one CPU I can access USB Flash drives. Runing Windows XP with two CPUs, I cannot (unrecognized USB device.

However, the WinXP 2-CPU installation will recognize a USB-connected hard disk, and will occasionally recognize some USB Flash Drives.

I am attaching six files from two separate runs of Windows XP with 2 CPUs

First run - recognize SATA hard drive attached via SATA-to-USB adapter. This was part of a run to fix files on the hard drive.

Summary of USB devices from Mac OS Bus Probe utility

USB Bus Probe.txt

Summary of hard drive and Flash Drive from "System Information"

USB System Information.txt

VirtualBox run log

USB-WinXP-2CPU-2011-09-22-VBox.log

Second run - keep trying Imation Flash drive until recognized, then try a SanDisk drive.

Summary of Imation Flash Drive from "System Information"

USB Flash Drive Imation.txt

Summary of SanDisk (Cruzer) Flash Drive from "System Information"

USB Flash Drive SanDisk.txt

VirtualBox run log

USB-WinXP-2CPU-2011-09-25-VBox.log

To sum up: Windows XP running in VirtualBox with two CPUs can recognize USB hard drives and some, but not all, flash drives. Ubuntu has no trouble when running with one CPU, has trouble recognizing some flash drives when running with two CPUs, and I have not tried other flash drives or hard disks on USB in Ubuntu.

Changed 3 years ago by iegorman

Mac Mac OS X "Bus Probe" summary of USB flash drive and USB hard drive attached at the same time

Changed 3 years ago by iegorman

Mac OS X "System Information" summary of USB flash drive and USB hard drive attached at the same time

Changed 3 years ago by iegorman

Mac OS X "System Information" summary of USB flash drive and USB hard drive attached at the same time

Changed 3 years ago by iegorman

Mac OS X "System Information" summary of USB flash drive and USB hard drive attached at the same time

Changed 3 years ago by iegorman

VirtualBox log of run with USB hard drive and USB flash drive

Changed 3 years ago by iegorman

Mac OS X "System Information" summary of Imation USB flash drive

Changed 3 years ago by iegorman

Mac OS X "System Information" summary of SanDisk USB flash drive

Changed 3 years ago by iegorman

VirtualBox log of run with Imation USB flash drive and Sandisk USB flash drive

comment:24 Changed 3 years ago by iegorman

My apologies for screwing up and adding file "USB System Information.txt"

comment:25 Changed 2 years ago by chrismurphy

Macbook Pro 8,2, Mac OS X 10.6.8, VirtualBox 4.1.4. Problem occurs with
Fedora 16 Linux f16.local 3.1.0-5.fc16.x86_64 #1 SMP Thu Oct 27 03:46:50 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux
CentOS6 Linux C6virt.local 2.6.32-71.29.1.el6.x86_64 #1 SMP Mon Jun 27 19:49:27 BST 2011 x86_64 x86_64 x86_64 GNU/Linux

Bug 9841 is a duplicate of this bug.

comment:26 Changed 2 years ago by kubrick

Hi! Same Here,

VirtualBox 4.1.6 Host : Mac OSX 10.7.2 / intel core i7 Guest : Fedora 16 x86_64

CC ME.

François

comment:27 Changed 2 years ago by sid.shetye

Hi,

I can report this same issue on my end too:

VirtualBox: 4.1.6 r74713 + Extension pack (latest for this version) Host : Mac OS X Lion (10.7.2) Guest: Windows 7 x86 (32bits) SP1

With 2 cores, I was only able to get my USB drive recognized once (the first time) with USB filters. Even then I had to mess with the order of USB filters. It never worked after that (40+ attempts), always came up as unrecognized device. I stepped it down to 2 cores after reading this thread and it worked right off the bat the first time itself!

It would be AWESOME if this old issue is resolved soon!

Thanks for the great product! Sid

comment:28 Changed 2 years ago by sid.shetye

I cannot edit the above post but please do not suggest disabling a core as a long term workaround to this problem! My compiles are ~33% faster with both cores versus just one, so it's actually quite a loss of productivity to work around this bug. Thanks!

comment:29 Changed 2 years ago by SouthTexas

I see that this is still an open issue as of version 4.1.16. I'm guessing this must be much more complicated to correct than it appears it should be?

comment:30 Changed 23 months ago by Anders85

I'd like to request that if there's no intent on solving this bug in near future, the documentation for the processor tab in VM settings be updated to inform of the consequences it has on OSX systems. Judging by this bug report, its affecting a lot of people and its the kind of bug that can be difficult to debug for users. So please, just mention it in the docs.

comment:31 Changed 19 months ago by noah_r

Same experience in VirtualBox 4.1.23. Mac OS X Mountain Lion 10.8.2 on Mac Book Pro Retina 2012. Windows 7 guest OS. I spent several hours troubleshooting both USB smart card reader and standard USB Thumb drive. I had the same experience without and then later with USB 2.0 extension installed. I agree that until the bug can be fixed, selecting 2 or more cores on Mac OS X host and 'Enable USB' should indicate a configuration problem. Hours lost troubleshooting until I found this bug report...

comment:32 follow-up: ↓ 33 Changed 19 months ago by dougman82

I've also been experiencing this issue for the last several months. I am currently running a Macbook Pro with OSX Lion. Took me the better part of two days to finally figure out why my Smart Card reader would not work in the Windows 7 guest OS.

I completely agree with Anders85 (comment 30) above: if there are no plans to resolve this bug, it would be greatly helpful to provide the information in official documentation that USB devices may not work with a guest utilizing multiple CPU cores.

This is the type of problem that is maddeningly difficult to figure out because it is completely non-intuitive. Had I not found such bug reports as this, there is no way that I would ever have thought to reduce the VM's CPU cores to 1 in order to get my USB device to properly operate.

comment:33 in reply to: ↑ 32 Changed 18 months ago by blueytoo

Replying to dougman82:

This is the type of problem that is maddeningly difficult to figure out because it is completely non-intuitive. Had I not found such bug reports as this, there is no way that I would ever have thought to reduce the VM's CPU cores to 1 in order to get my USB device to properly operate.

Same problem. Took ages to find the answer. 100% agree with dougman82. Yes changing CPUs to 1 fixed it.

MBP 17" early 2011 2.3GHz; MacOS 10.6.8 Server (64bit); Windows7 (64bit); VB 4.2.2; BootCamp 3.3

comment:34 Changed 14 months ago by vatjjar

I know this is marked as a Mac OS X bug, but just wanted to inform this issue appears also with other host OSes. I'm running Win7 host, Ubuntu linux 12.10 guest with virtualbox 4.2.6 with latest quest additions. I have been trying to setup android development tools in guest OS, which work, but none of my android devices work via USB. The reason is exactly as described above. Also, trying to limit the number of CPU cores of the guest does not help, not at least on my Win7 host. This comment probably does not help too much in solving the original problem, but I still though to make a note about it here.

comment:35 Changed 13 months ago by 65bruceRF

More feedback for you VirtualBox Dev and PM teams. First let me say AWESOME! I am very happy with this product. I honestly believe this system is best of breed and is superior to all other Virtual Host systems I have used, and I believe I have used all of them now. On a more pointed note... Can you please fix this issue, I also can confirm that on a new generation MBP running the latest OSX (10.8.3) ALL of my guest OS's in X64 mode (I do not have any x86 guests) all failed to load USB devices with multiple CPU's provisioned for the Guest this is all flavors of X64 Windows and several flavors of Linux, I have not tried on a guest OSX system. Once each guest was returned to a single CPU all USB devices were recognized, loaded correctly and worked very well (FOR USB 2.0). That is my last request, which I know you and many others are grappling with, USB 3.0 support.... PLEASE expedite the deployment of same, and I am happy to partake in alpha and beta testing (YES I post feedback and test rigorously) as many of my guest OS's are data/database driven and fast access to disk is paramount to their usability, superspeed USB connectivity is nearly enough to work with development and testing of JDBC databases like H2. Please let me know if I can assist in any way.

comment:36 Changed 12 months ago by frank

  • Description modified (diff)

We think we greatly improved USB for OSX hosts for VMs with more than one CPU. Could you install  this test build to check if it fixes your problems? Note, this build should fix problems you observe only with VMs with multiple guest CPUs. It will not fix problems with USB devices which you also observe with VMs with a single guest CPU.

comment:37 Changed 12 months ago by blueytoo

Can't test USB2 without matching extension pack.

comment:38 Changed 12 months ago by frank

I think the fix should also work with the 4.2.12 extension pack but anyway,  here is a 4.2.13 extension pack. Thanks for testing!

Changed 12 months ago by blueytoo

comment:39 Changed 12 months ago by blueytoo

Tried 4.2.13. (testing 8CPU) Have some problems.

  1. Can't shutdown VM without showing "Aborted".
  2. Clicking on "USB" device (actually a PC card) to disconnect (Sandisk card reader which is empty but ticked as captured) trying to solve #1 crashes VB.
  3. Connected USB device (Rosstech) looks to recognise correctly, but then crashes VB. Persistently crashes VB on login.

Crash logs attached (VBlogs.7z)

Last edited 11 months ago by blueytoo (previous) (diff)

comment:40 follow-up: ↓ 44 Changed 12 months ago by blueytoo

Update. I managed to disconnect the unused Sandisk flash card reader during VM startup. I started without the Rosstech USB device being plugged in and plugged in later. Everything appeared OK. Shutdown and the "aborted" VM stated has cleared too.

If I boot VM with Rosstech USB unplugged and have the empty Sandisk card reader attached, then plugging in the Rosstech USB causes a crash.

The Sandisk flash reader is empty. I can't get it to remove via "safely remove". If I try to unattach from VM, it crashes.

comment:41 Changed 12 months ago by blueytoo

Update. System seems presently unstable without Rosstech USB even plugged in. Sandisk flash reader presumably causing problems. More crash logs attached.

Changed 12 months ago by blueytoo

comment:42 Changed 12 months ago by Sid

I'm getting spammed by this ticket. I posted 17 months back and since then we've dumped VirtualBox for VMWare and have no interest in Oracle software. How can I get myself off this ticket notification/email list?

comment:43 Changed 12 months ago by frank

Allright, I've removed users "Sid" and "sid.shetye".

blueytoo, thanks for testing. We will analyze your logs.

comment:44 in reply to: ↑ 40 ; follow-up: ↓ 45 Changed 12 months ago by Hachiman

Replying to blueytoo:

Update. I managed to disconnect the unused Sandisk flash card reader during VM startup. I started without the Rosstech USB device being plugged in and plugged in later. Everything appeared OK. Shutdown and the "aborted" VM stated has cleared too.

If I boot VM with Rosstech USB unplugged and have the empty Sandisk card reader attached, then plugging in the Rosstech USB causes a crash.

The Sandisk flash reader is empty. I can't get it to remove via "safely remove". If I try to unattach from VM, it crashes.

Does all these crashes happen in UP configuration? Does it depends on number of virtual processors (e.g. with 2 vCPU)?

comment:45 in reply to: ↑ 44 Changed 12 months ago by blueytoo

Replying to Hachiman:

Does all these crashes happen in UP configuration? Does it depends on number of virtual processors (e.g. with 2 vCPU)?

Only started testing 8 vCPU. No problems previously noted with 1 vCPU and same hardware. Config screen attached FYI it that helps.

Back to 1 CPU with 4.2.13 - can happily catpure and uncapture the Sandisk flash reader.

Testing 2vCPU. Seemed ok. Recognised Rosstech, uncaptrued it; uncapturing Sandisk flash reader caused another crash. Log attached.

Last edited 12 months ago by blueytoo (previous) (diff)

Changed 12 months ago by blueytoo

Changed 12 months ago by blueytoo

2vCPU crash

Changed 12 months ago by blueytoo

VBox application logs

comment:46 Changed 11 months ago by Hachiman

I've tried to reproduce this crush locally with only flash card reader available here Transcent RDP8 and wasn't successful. Could you please give a hint what model of Sandisk reader do you use or even better (if you have USB prober installed on your mac) dump out put to defect.

comment:47 follow-up: ↓ 48 Changed 11 months ago by blueytoo

Model is: Sandisk SDAD109A11 Digital Media Memory Card to Express Slot Adapter. As I said earlier, it is not technically a "USB" device.

Can do USB prober output later.

comment:48 in reply to: ↑ 47 Changed 11 months ago by Hachiman

Replying to blueytoo:

Model is: Sandisk SDAD109A11 Digital Media Memory Card to Express Slot Adapter. As I said earlier, it is not technically a "USB" device.

Can do USB prober output later.

If you have got linux at hand lsusb -v is also acceptable.

And back to Rosstech... Do I understand correctly that it works as expected with 2 vSMP and doesn't work with 8. right? btw, what kind of device is it?

Changed 11 months ago by blueytoo

Sandisk SDAD-111

Changed 11 months ago by blueytoo

RossTech problems log

comment:49 follow-up: ↓ 50 Changed 11 months ago by blueytoo

Installed .kext for USB prober logging. Now the Sandisk flash reader will happily uncapture and capture. I presume the kext added is the logging IOUSBFamily.kext (but that wasn't what the auto-install dialog of USB Prober said.

RossTech is automotive diagnostic cable with dongle function. Apparently can be very sensitive to bus timing when talking to the car. I have only used its internal test function in the VBox bug testing. (Not so keen on risking corrupting my car's computer while VBox is unstable.) Has worked fine in 1vCPU mode.

All seemed OK temporarily with 2vCPU - both Sandisk capture/uncapture and Rosstech talking to software. With 8vCPU, Rosstech software was unstable - seeming to be capturing random mouse input - multiple flashing software windows screen changes.

Just retesting 2vCPU - Rosstech talking to its software normally. Detach Sandisk flash reader -> crash. Reboot Win7VM, detach sandisk flash reader -> crash.

Crash logs attached + usb-log.3 covers period of last 2 crashes.

Changed 11 months ago by blueytoo

usb-log-when-Rosstech-malfunctioning

Changed 11 months ago by blueytoo

more crashlogs 2vCPU

comment:50 in reply to: ↑ 49 Changed 11 months ago by Hachiman

Replying to blueytoo:

Thank you for update and provided details, unfortunately they don't give me a glue about the issue with SanDisk :(, so let do other way... Please download & install these builds, they don't fix anything new with your issue, but fresh enough that I can use symbol information in investigating core dump. So I'd appreciate if you consult our wiki about getting core dump and upload result core to our ftp server  ftp://ftp.oracle.com/appsdev/incoming for further investigations.

 http://www.virtualbox.org/download/testcase/VirtualBox-4.2.13-85606-OSX.dmg  http://www.virtualbox.org/download/testcase/Oracle_VM_VirtualBox_Extension_Pack-4.2.13-85606.vbox-extpack

comment:51 follow-up: ↓ 52 Changed 11 months ago by blueytoo

Uploading core dump. (ticket7462.core.44848) taking a while....

For some reason,  ftp://ftp.oracle.com/appsdev/incoming/incoming exists and is not accessible - ?should be removed

Wiki on core dumps is missing some information - the /cores folder needs to be created first + permissions! (See  https://developer.apple.com/library/mac/#technotes/tn2004/tn2124.html)

Last edited 11 months ago by blueytoo (previous) (diff)

comment:52 in reply to: ↑ 51 ; follow-up: ↓ 53 Changed 11 months ago by Hachiman

Replying to blueytoo:

Uploading core dump. (ticket7462.core.44848) taking a while....

Great, please ping me via posting note to defect when uploading will be finished (or by mail vasily _dot_ levchenko _at_ oracle _dot_ com).

For some reason,  ftp://ftp.oracle.com/appsdev/incoming/incoming exists and is not accessible - ?should be removed

perhaps someone uploaded file with such name, this is WO folder for security reasons.

Wiki on core dumps is missing some information - the /cores folder needs to be created first + permissions! (See  https://developer.apple.com/library/mac/#technotes/tn2004/tn2124.html)

thanks for note will update our wiki.

comment:53 in reply to: ↑ 52 ; follow-up: ↓ 55 Changed 11 months ago by blueytoo

Replying to Hachiman:

Replying to blueytoo:

For some reason,  ftp://ftp.oracle.com/appsdev/incoming/incoming exists and is not accessible - ?should be removed

perhaps someone uploaded file with such name, this is WO folder for security reasons.

.../incoming/incoming is not writable at all

comment:55 in reply to: ↑ 53 Changed 11 months ago by frank

Replying to blueytoo:

Replying to Hachiman:

Replying to blueytoo:

For some reason,  ftp://ftp.oracle.com/appsdev/incoming/incoming exists and is not accessible - ?should be removed

perhaps someone uploaded file with such name, this is WO folder for security reasons.

.../incoming/incoming is not writable at all

That directory does not exist. Just upload the files to the  ftp://ftp.oracle.com/appsdev/incoming directory, this should work.

But the latest test build uploaded by Hachiman should really fix the crashes you saw (they were surely regressions), therefore we are very interested in your feedback and in feedback from other users as well. Thanks for testing!

comment:56 Changed 11 months ago by blueytoo

So far so good! Tested 2vCPU and 8vCPU. Sandisk attach/unattach, Rosstech, and Gemalto smart card all appear to be functioning normally.

VMs shut down correctly (rather than leaving "aborted" status).

Thanks for all your work fixing this.

comment:57 follow-up: ↓ 58 Changed 11 months ago by lkj

I don't know if it's a side effect of 4.2.13 but I can't save a machine's state. just noting it here. Host OSX 10.8.3

edit: It seems to work with an openSUSE 12.3 guest but not a Debian 7 guest. The openSUSE guests have their 4.2.6 additions, the Debian 7 guest has their 4.1.18 additions.

Last edited 11 months ago by lkj (previous) (diff)

comment:58 in reply to: ↑ 57 Changed 11 months ago by Hachiman

Replying to lkj:

I don't know if it's a side effect of 4.2.13 but I can't save a machine's state. just noting it here. Host OSX 10.8.3

edit: It seems to work with an openSUSE 12.3 guest but not a Debian 7 guest. The openSUSE guests have their 4.2.6 additions, the Debian 7 guest has their 4.1.18 additions.

Perhaps not related (I don't see any issues with saving/restoring here), but please attach log file.

comment:59 Changed 11 months ago by lkj

Well the whole issue was strange. I actually haven't tried to save a state in over two years but in order to have a guest open to the same screen size I had dragged it to, I thought I would try it. I was running a Debian 7 guest which actually had 4.1.18 guest additions pre-installed. The first time I tried saving the state nothing was in the snapshots folder. The second time I tried it it did have a .sav file. However, because the Debian 7 guest was using Gnome 3, when I restored it I had nothing but a black screen. Perhaps guests using mutter can't be restored properly. Maybe the problems were caused by running a guest with 4.1.18 additions.

I have heard the devs. are rewriting the graphics code for Macs so perhaps I will just wait and see what happens.

comment:60 Changed 11 months ago by STA Admin

We had the same issue, i think since 2 years, but we didn't really need 2 CPU for virtual machine. Now things was changed, and I thought to write for this issue, but found this thread. And I can say that 4.2.13 solved the problem for us. Host MacMini i7 4x2GHz with 8GB ram and Mac OS X server 10.7.5, guest Windows Server 2003

comment:61 follow-up: ↓ 62 Changed 11 months ago by CaptSaltyJack

4.2.13 isn't yet out for us mere mortals. ;) So I'll wait to check that out, and confirm whether the update fixes this issue for me.

comment:62 in reply to: ↑ 61 Changed 11 months ago by STA Admin

Replying to CaptSaltyJack:

4.2.13 isn't yet out for us mere mortals. ;) So I'll wait to check that out, and confirm whether the update fixes this issue for me.

You can get it from the link above... like me ;)

comment:63 follow-up: ↓ 65 Changed 11 months ago by CaptSaltyJack

Got it!

Ah shoot. Doesn't work for me. I have a Kingston USB thumb drive. I eject it from my Mac (host), then in VirtualBox go under Devices->USB and choose the Kingston drive. It has a checkmark next to it, but doesn't show up in Nautilus.

comment:64 Changed 11 months ago by STA Admin

Tried now with 4 CPU, and all works flawless. We have an Eutron SmartKey 3 USB dongle, that before (with more than 1 CPU) was impossible to attach... or better, vbox saw it attached, but inside the VM it wasn't available. With 4.2.13 it's finally OK!

comment:65 in reply to: ↑ 63 Changed 11 months ago by blueytoo

Replying to CaptSaltyJack:

Got it!

Ah shoot. Doesn't work for me. I have a Kingston USB thumb drive. I eject it from my Mac (host), then in VirtualBox go under Devices->USB and choose the Kingston drive. It has a checkmark next to it, but doesn't show up in Nautilus.

Did you set up the USB filter?

comment:66 Changed 11 months ago by CaptSaltyJack

Ah! That worked! I didn't realize I had to add it in the USB filters first. Thanks!

comment:67 Changed 10 months ago by aktaylor

Hi all,

I am having the same issue as outlined in this ticket and the link for the fix is not working. I would like to test it out as well.

Thanks!

comment:68 Changed 10 months ago by frank

  • Status changed from new to closed
  • Resolution set to fixed

Finally fixed in VBox 4.2.14! If you have still problems with USB on Mac OS X, please open separate tickets.

comment:69 Changed 10 months ago by ieg

I have confirmed the fix with Windows XP (updated to 2013-06-22) running as a guest wIth 2 CPUs in Mac OS X 10.8.4 - An IMATION 2 GB USB thumb drive is recognized when the write-protect switch is in either the locked or unlocked position. Previously, the drive would only be recognized when write-protected (locked).

comment:70 Changed 10 months ago by blueytoo

Have another problem with USB2 support on Mac OSX host. See ticket https://www.virtualbox.org/ticket/11907

comment:71 Changed 10 months ago by erny1803

I'm really surprised that I also encountered this bug and that it seems very recently fixed:

  • Host: Mac OS X 10.8.4 on Mid 2012 Mac (Core i7)
  • Guest: Ubuntu 12.04/64, 3 VCPU
  • USB Device: SmartCard Reader (Gemplus GemPC Twin SmartCard Reader)

Now I can see devices with lsusb. Thanx.

Note: See TracTickets for help on using tickets.

www.oracle.com
ContactPrivacy policyTerms of Use