Nice Day! All the days of New Year''s Festivities were spent on updating linux dom0 Happy new year I have: Ubuntu 10.10 Server Kernel Jeremy 2.6.32.48 (with pvusb patch & usbip modules) Xen 4.1.2 (4.2 untable freeze vm) Windows7 (32 & 64 in testsigning mode) Pv driver 0.11.0.238 (308, 251) After xm usb-hc-create v1 2 8 in win exists Xen PVUSB device driver xm usb-list-assignable-devices 1-4.1 : ID 046d:c05b Logitech USB Optical Mouse 1-4.4 : ID 04f3:0a01 Elan USB Phone 3-3 : ID 04f2:0200 Chicony USB Wireless HID Receiver 5-3 : ID 07d1:f101 Broadcom Corp D-Link DBT-122 xm usb-attach v3 0 1 1-4.1 xm usb-list v3 Idx BE state usb-ver BE-path 0 0 4 USB2.0 /local/domain/0/backend/vusb/5/0 port 1: 1-4.1 [ID 046d:c05b Logitech USB Optical Mouse] port 2: port 3: gplpv_Vista2008x64_0.11.0.351.msi But in HVM Win7x64 no action gplpv_Vista2008x32_0.11.0.351.msi In HVM win7x32 - unknown devices In gplpv 308 - driver not installed How to make that work properly? -- Likarpenkov ALex
I visited this maybe a year ago and discovered that it was only supported under Linux. I have copied James on this email to confirm, but afaik the GPLPV drivers do not have support for PVUSB On Fri, Jan 6, 2012 at 9:57 AM, Likarpenkov Alexander <al@ohosting.org.ua>wrote:> Nice Day! > > All the days of New Year''s Festivities were spent on updating linux dom0 > > Happy new year > > I have: > Ubuntu 10.10 Server > Kernel Jeremy 2.6.32.48 (with pvusb patch & usbip modules) > Xen 4.1.2 (4.2 untable freeze vm) > > Windows7 (32 & 64 in testsigning mode) > Pv driver 0.11.0.238 (308, 251) > > After xm usb-hc-create v1 2 8 > in win exists > Xen PVUSB device driver > > xm usb-list-assignable-devices > 1-4.1 : ID 046d:c05b Logitech USB Optical Mouse > 1-4.4 : ID 04f3:0a01 Elan USB Phone > 3-3 : ID 04f2:0200 Chicony USB Wireless HID Receiver > 5-3 : ID 07d1:f101 Broadcom Corp D-Link DBT-122 > > xm usb-attach v3 0 1 1-4.1 > > xm usb-list v3 > Idx BE state usb-ver BE-path > 0 0 4 USB2.0 /local/domain/0/backend/vusb/**5/0 > port 1: 1-4.1 [ID 046d:c05b Logitech USB Optical Mouse] > port 2: > port 3: > > gplpv_Vista2008x64_0.11.0.351.**msi > But in HVM Win7x64 no action > gplpv_Vista2008x32_0.11.0.351.**msi > In HVM win7x32 - unknown devices > > In gplpv 308 - driver not installed > > How to make that work properly? > > > > -- > Likarpenkov ALex > > ______________________________**_________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/**xen-users<http://lists.xensource.com/xen-users> >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> I have: > Ubuntu 10.10 Server > Kernel Jeremy 2.6.32.48 (with pvusb patch & usbip modules) Xen 4.1.2 (4.2 > untable freeze vm) > > Windows7 (32 & 64 in testsigning mode) > Pv driver 0.11.0.238 (308, 251) > > After xm usb-hc-create v1 2 8 > in win exists > Xen PVUSB device driver > > xm usb-list-assignable-devices > 1-4.1 : ID 046d:c05b Logitech USB Optical Mouse > 1-4.4 : ID 04f3:0a01 Elan USB Phone > 3-3 : ID 04f2:0200 Chicony USB Wireless HID Receiver > 5-3 : ID 07d1:f101 Broadcom Corp D-Link DBT-122 > > xm usb-attach v3 0 1 1-4.1 > > xm usb-list v3 > Idx BE state usb-ver BE-path > 0 0 4 USB2.0 /local/domain/0/backend/vusb/5/0 > port 1: 1-4.1 [ID 046d:c05b Logitech USB Optical Mouse] port 2: > port 3: > > gplpv_Vista2008x64_0.11.0.351.msi > But in HVM Win7x64 no action > gplpv_Vista2008x32_0.11.0.351.msi > In HVM win7x32 - unknown devices > > In gplpv 308 - driver not installed > > How to make that work properly? >pvusb under GPLPV is only partially implemented. It seems to work for some mass-storage devices, and at one time it was working for a keyboard too but I think that broke again somewhere along the way. I don''t have any servers at the moment that support the usb backend driver (using Linux 3.1.x kernel from Debain), so I can''t even confirm that for you at the moment. I do plan to get back to it one day but just don''t have time at the moment. James
How can I help. It is very necessary? What debug can help? Usb 1.1 is already drawn in HVM windows. USB mass storage by the way also does not work I''m on the same platform two cards and two independent OS windows and each is allocated by the system controller with usb and Linux is not possible to leave them ? ???????, ???????? ?????? 2012 ????, ? 02:52:44 ?? ??????: ??>> I have: ??>> Ubuntu 10.10 Server ??>> Kernel Jeremy 2.6.32.48 (with pvusb patch & usbip modules) Xen 4.1.2 ??>> (4.2 untable freeze vm) ??>> ??>> Windows7 (32 & 64 in testsigning mode) ??>> Pv driver 0.11.0.238 (308, 251) ??>> ??>> After xm usb-hc-create v1 2 8 ??>> in win exists ??>> Xen PVUSB device driver ??>> ??>> xm usb-list-assignable-devices ??>> 1-4.1 : ID 046d:c05b Logitech USB Optical Mouse ??>> 1-4.4 : ID 04f3:0a01 Elan USB Phone ??>> 3-3 : ID 04f2:0200 Chicony USB Wireless HID Receiver ??>> 5-3 : ID 07d1:f101 Broadcom Corp D-Link DBT-122 ??>> ??>> xm usb-attach v3 0 1 1-4.1 ??>> ??>> xm usb-list v3 ??>> Idx BE state usb-ver BE-path ??>> 0 0 4 USB2.0 /local/domain/0/backend/vusb/5/0 ??>> port 1: 1-4.1 [ID 046d:c05b Logitech USB Optical Mouse] port 2: ??>> port 3: ??>> ??>> gplpv_Vista2008x64_0.11.0.351.msi ??>> But in HVM Win7x64 no action ??>> gplpv_Vista2008x32_0.11.0.351.msi ??>> In HVM win7x32 - unknown devices ??>> ??>> In gplpv 308 - driver not installed ??>> ??>> How to make that work properly? ??>> JH> pvusb under GPLPV is only partially implemented. It seems to work for JH> some mass-storage devices, and at one time it was working for a JH> keyboard too but I think that broke again somewhere along the way. JH> I don''t have any servers at the moment that support the usb backend JH> driver (using Linux 3.1.x kernel from Debain), so I can''t even confirm JH> that for you at the moment. JH> I do plan to get back to it one day but just don''t have time at the JH> moment. JH> James
> How can I help. It is very necessary? > What debug can help? > > Usb 1.1 is already drawn in HVM windows. USB mass storage by the way also > does not work I''m on the same platform two cards and two independent OS > windows and each is allocated by the system controller with usb and Linux is > not possible to leave them >I have just pushed some updates that get a USB mouse and keyboard working via pvusb, and possibly other devices. Only INTR and BULK USB transfers are supported, and with the debug builds of GPLPV it is _very_ noisy (lots of output to /var/log/xen/qemu-dm-<domu name>.log If you can build GPLPV yourself you can try it now, otherwise I''ll upload some builds in the next day or two, but it''s only tested under 2008R2 at the moment so there may be some additional bugs to find. James
Of course, I will build GPLPV if you give instructions how to do it right. ??>> How can I help. It is very necessary? ??>> What debug can help? ??>> ??>> Usb 1.1 is already drawn in HVM windows. USB mass storage by the way ??>> also does not work I''m on the same platform two cards and two ??>> independent OS windows and each is allocated by the system controller ??>> with usb and Linux is not possible to leave them ??>> JH> I have just pushed some updates that get a USB mouse and keyboard JH> working via pvusb, and possibly other devices. Only INTR and BULK USB JH> transfers are supported, and with the debug builds of GPLPV it is JH> _very_ noisy (lots of output to /var/log/xen/qemu-dm-<domu name>.log JH> If you can build GPLPV yourself you can try it now, otherwise I''ll JH> upload some builds in the next day or two, but it''s only tested under JH> 2008R2 at the moment so there may be some additional bugs to find. JH> James
[This email is either empty or too large to be displayed at this time]
It looks like you are testing without the debug build of the drivers. With log throttling turned off you should be able to run the debug build without too much slowing down. If you can do that and just try adding the first usb device (the wireless HID should be fine) then shut down and send me the logfile I''ll have a look at it. Can you then start up again and add the mouse? I have definitely had the mouse working properly before but maybe the wireless HID device crashed it preventing adding the second device from working. I''ve basically only implemented the USB functions I''ve been able to test - there are lots more functions probably missing but they should be easy enough to add. James> -----Original Message----- > From: Likarpenkov Alexander [mailto:al@ohosting.org.ua] > Sent: Monday, 16 January 2012 11:54 PM > To: James Harper; xen-users@lists.xensource.com > Subject: Re: [Xen-users] PVUSB - how to? > > RESULTS: NOT WORKED Properly 0.11.0.357 > > About process: > I''m changed startup config from: > device_model = ''/usr/lib/xen/bin/qemu-dm'' > to > device_model = ''/usr/lib64/xen/bin/qemu-dm.nothrottle'' > See in attach qemu-dm.nothrottle > ANd log: > qemu: ignoring not-understood drive `/local/domain/10/log-throttling'' > medium change watch on `/local/domain/10/log-throttling'' - unknown > device, ignored > > > after loading domU: > See in attach 01.jpg > # xm usb-hc-create v3 2 8 > See in attach 02.jpg > > # xm usb-list-assignable-devices > 1-4.3 : ID 04f3:0a01 Elan USB Phone > 1-4.4 : ID 046d:c05b Logitech USB Optical Mouse > 3-3 : ID 04f2:0200 Chicony USB Wireless HID Receiver > 5-3 : ID 07d1:f101 Broadcom Corp D-Link DBT-122 > > # xm usb-attach v3 0 1 3-3 > See in attach 03.jpg > > # xm usb-attach v3 0 2 1-4.3 > NO ACTION, but: > # xm usb-list v3 > Idx BE state usb-ver BE-path > 0 0 4 USB2.0 /local/domain/0/backend/vusb/10/0 > port 1: 3-3 [ID 04f2:0200 Chicony USB Wireless HID Receiver] > port 2: 1-4.3 [ID 04f3:0a01 Elan USB Phone] > port 3: > port 4: > port 5: > > # xm usb-attach v3 0 3 1-4.4 > See in attach 04.jpg > > > # xm usb-attach v3 0 4 5-3 > BlueTooth added properly > See in attach 05.jpg > And really search devices > See in attach 06.jpg > But after adding BSOD: > See in attach 07.jpg
[This email is either empty or too large to be displayed at this time]
> > ????????????! > ? ???????????, ????????????? ?????? 2012 ????, ? 15:05:25 ?? ??????: > > JH> It looks like you are testing without the debug build of the drivers. > JH> With log throttling turned off you should be able to run the debug JH> > build without too much slowing down. > > JH> If you can do that and just try adding the first usb device (the JH> > wireless HID should be fine) then shut down and send me the logfile JH> I''ll > have a look at it. > > In an attachment, the log > Interesting fact: in debug mode, HID working properly, but very slowlyHmmm... logging is very expensive in terms of CPU power (basically a VMEXIT + work by qemu for every character output to the logs.) This would work to slow down the processing in the usb driver immensely so maybe I have a race or something in the code. How many vcpu''s do you have in your config? If more than 1, please change it to 1 and try again (without debug) and see what happens.> > JH> Can you then start up again and add the mouse? I have definitely had > JH> the mouse working properly before but maybe the wireless HID device > JH> crashed it preventing adding the second device from working. > > JH> I''ve basically only implemented the USB functions I''ve been able to JH> > test - there are lots more functions probably missing but they should JH> be > easy enough to add. > > How about usb flash drive and usb sound (as Elan USB Phone)?I''m not implementing ISOC functions. So any audio or video device probably won''t work. I need to do a bit of reading about those though. A question about your logs - the first one shows a crash at the end... was this after shutdown? I think I might have a bug in the cancellation code... James
JH> Hmmm... logging is very expensive in terms of CPU power (basically a JH> VMEXIT + work by qemu for every character output to the logs.) This JH> would work to slow down the processing in the usb driver immensely so JH> maybe I have a race or something in the code. JH> How many vcpu''s do you have in your config? If more than 1, please JH> change it to 1 and try again (without debug) and see what happens. During the test used 3 processors. There is also a system that waits for a successful outcome, with 6 processors, which need to passing usb device. Result from a one processor will give later. Uncomfortable every time to shuting down 15 DomU systems to boot the system without usb devices forwarding in the virtual system. You accidentally do not know how to make pciback unhide dinamycally? JH>>> Can you then start up again and add the mouse? I have definitely had JH>>> the mouse working properly before but maybe the wireless HID device JH>>> crashed it preventing adding the second device from working. ??>> JH>>> I''ve basically only implemented the USB functions I''ve been able to JH>>> JH> ??>> test - there are lots more functions probably missing but they should ??>> JH> be easy enough to add. ??>> ??>> How about usb flash drive and usb sound (as Elan USB Phone)? JH> I''m not implementing ISOC functions. So any audio or video device JH> probably won''t work. I need to do a bit of reading about those though. Maybe you can do by analogy with usbip, when traffic is encapsulated in all usb commands, and do not need to translate and transmit directly? JH> A question about your logs - the first one shows a crash at the end... JH> was this after shutdown? I think I might have a bug in the cancellation JH> code... Via VNC was seen BSOD and reboot after dumping system. In each of the options, when it was connected HID Chicony. Why forwarding pci device (usb controller) via pciback in a virtual environment - only works in 1.1 instead of 2.0 or 3.0
> > JH> Hmmm... logging is very expensive in terms of CPU power (basically a > JH> VMEXIT + work by qemu for every character output to the logs.) This > JH> would work to slow down the processing in the usb driver immensely so > JH> maybe I have a race or something in the code. > > JH> How many vcpu''s do you have in your config? If more than 1, please > JH> change it to 1 and try again (without debug) and see what happens. > > During the test used 3 processors. There is also a system that waits for a > successful outcome, with 6 processors, which need to passing usb device. > Result from a one processor will give later. Uncomfortable every time to > shuting down 15 DomU systems to boot the system without usb devices > forwarding in the virtual system.So you don''t have a test system you can use with just a couple of usb devices?> You accidentally do not know how to make pciback unhide dinamycally?I''ve never used pci passthrough since xen 2.x (or was it 1.x?)> > JH> I''m not implementing ISOC functions. So any audio or video device > JH> probably won''t work. I need to do a bit of reading about those though. > > Maybe you can do by analogy with usbip, when traffic is encapsulated in all > usb commands, and do not need to translate and transmit directly?Usbip does things a bit differently but I have had a look at that code before.> > JH> A question about your logs - the first one shows a crash at the end... > JH> was this after shutdown? I think I might have a bug in the cancellation > JH> code... > > > Via VNC was seen BSOD and reboot after dumping system. > In each of the options, when it was connected HID Chicony. > > Why forwarding pci device (usb controller) via pciback in a virtual > environment - only works in 1.1 instead of 2.0 or 3.0Not sure. It should work if you are passing through the whole host controller. James
JH>>> Hmmm... logging is very expensive in terms of CPU power (basically a JH>>> VMEXIT + work by qemu for every character output to the logs.) This JH>>> would work to slow down the processing in the usb driver immensely so JH>>> maybe I have a race or something in the code. ??>> JH>>> How many vcpu''s do you have in your config? If more than 1, please JH>>> change it to 1 and try again (without debug) and see what happens. ??>> ??>> During the test used 3 processors. There is also a system that waits ??>> for a successful outcome, with 6 processors, which need to passing usb ??>> device. Result from a one processor will give later. Uncomfortable ??>> every time to shuting down 15 DomU systems to boot the system without ??>> usb devices forwarding in the virtual system. JH> So you don''t have a test system you can use with just a couple of usb JH> devices? I test it on those machines that can move a mouse or see the reaction of the usb device. And also do not want to do it remotely ??>> You accidentally do not know how to make pciback unhide dinamycally? JH> I''ve never used pci passthrough since xen 2.x (or was it 1.x?) Xen 4.0.2 rc3 and 4.1.2. JH>>> I''m not implementing ISOC functions. So any audio or video device JH>>> probably won''t work. I need to do a bit of reading about those JH>>> though. ??>> ??>> Maybe you can do by analogy with usbip, when traffic is encapsulated ??>> in all usb commands, and do not need to translate and transmit ??>> directly? JH> Usbip does things a bit differently but I have had a look at that code JH> before. Maybe even take a look? It works for audio devices. In this application, the problem with hanging up in an emergency disconnection or disabling parent device. usbip.exe - impossible close handle of application JH>>> A question about your logs - the first one shows a crash at the JH>>> end... was this after shutdown? I think I might have a bug in the JH>>> cancellation code... ??>> ??>> Via VNC was seen BSOD and reboot after dumping system. ??>> In each of the options, when it was connected HID Chicony. ??>> ??>> Why forwarding pci device (usb controller) via pciback in a virtual ??>> environment - only works in 1.1 instead of 2.0 or 3.0 JH> Not sure. It should work if you are passing through the whole host JH> controller. Forwarding usb not working correctly. In the system a lot of usb controllers 00:12.0 USB Controller: ATI Technologies Inc SB700/SB800 USB OHCI0 Controller 00:12.2 USB Controller: ATI Technologies Inc SB700/SB800 USB EHCI Controller 00:13.0 USB Controller: ATI Technologies Inc SB700/SB800 USB OHCI0 Controller 00:13.2 USB Controller: ATI Technologies Inc SB700/SB800 USB EHCI Controller 00:14.5 USB Controller: ATI Technologies Inc SB700/SB800 USB OHCI2 Controller 00:16.0 USB Controller: ATI Technologies Inc SB700/SB800 USB OHCI0 Controller 00:16.2 USB Controller: ATI Technologies Inc SB700/SB800 USB EHCI Controller 03:00.0 USB Controller: NEC Corporation uPD720200 USB 3.0 Host Controller (rev 03) In one of the VMs connected device 00:13.0. This system has control of usb ports usb 3.0(withost passing) with speed usb 1.1. The motherboard M4A89TD PRO USB3 all your usb controller can be divided only in two different hub for pci passing
??>> You accidentally do not know how to make pciback unhide dinamycally? Maybe someone knows? JH> I''ve never used pci passthrough since xen 2.x (or was it 1.x?) Worth a try. Now my usb hub and windows does not know anything about the existence of Linux. The connected USB flash drive as soon as determined by the system, as well as printer, sound card, an FM tuner, hdd video camera, a foto camera. Only works in this mode, all usb 1.1, which is very inconvenient JH>>> I''m not implementing ISOC functions. So any audio or video device JH>>> probably won''t work. I need to do a bit of reading about those JH>>> though.
I hope to have an update for you to test late tonight (AEDT). There does seem to be some races in some allocation routines and also in the cancellation routines. James
When will test the new version? JH> I hope to have an update for you to test late tonight (AEDT). There JH> does seem to be some races in some allocation routines and also in the JH> cancellation routines.