[USRP-users] B210 with VMware

Ralph A. Schmid, dk5ras ralph at schmid.xxx
Fri Dec 5 02:22:36 EST 2014


Hi Neel, 

 

So far I did not undertake deeper research, but at least VMware says on the website that they reworked the xHCI USB3 stuff, so this should have to do with it. When I find more time I will make a bare metal test again, to see if It is gone there, too. This would prove that all your work on the uhd stuff was also useful to improve things. The versions I used always are the latest master, built from source. There is one machine at the office with still VMware 10 on it, so at least I should be able to collect these outputs later this day.

 

Ralph.

 

 

 

 

Hello Ralph:

Thanks for sharing this. I'm glad to hear that you were able to get everything working reliably. Do you have an idea why the upgrade to VMware 11 helped so much? What changed in VMware 11? There must be a difference in the way the USB controllers appear/behave between VMware 10 and 11. What does "lspci -v" return on your Kubuntu system when run under VMware 10 and 11? Which version of UHD have you been using?

--Neel

 

On Wed, Dec 3, 2014 at 11:44 PM, Ralph A. Schmid, dk5ras via USRP-users <usrp-users at lists.ettus.com <mailto:usrp-users at lists.ettus.com> > wrote:

Hi,

sorry for posting this to three lists, but it affects them all. I kindly ask
you being wise when sending a reply, if it really should go to all lists,
too :)

With my relative new USRP B210 I was not very lucky by means of stability,
especially at higher sampling rates USB3.0 totally crashed, pulling the
whole virtual USB subsystem in some non-responsive black hole, when using
Kubuntu 14.04 with VMware 10 on a Windows 8.1 machine. Similar crashes
occurred when using Kubuntu 14.04 and Debian 7 on the bare metal (same
hardware, directly booted from µSD card). I tested this many weeks ago, so I
don't know if this would be better now with all the changes made in uhd in
the meantime. In the VM it also got a bit better over the time, at least at
lower sampling rates, but it did not completely vanish.

The real breakthrough came yesterday with an upgrade from VMware 10 to 11!
For the very first time OpenBTS ran through the whole night using the B210,
keeping a phone call alive between to test phones all the time. Gqrx ran in
the old config only for minutes, yesterday after this change it ran through
the whole afternoon at full 32MHz bandwidth. This morning I was able to
start OpenLTE with a 1.4MHz wide LTE carrier, and it ran for 15 minutes,
crashing only when I opened Outlook in the host system, probably stealing
too much cycles from the guest system. In the previous setup it only ran for
seconds up to a minute.

So although many factors seemed to influence the issue (see my various
frustrating bare metal experiences), one main factor was VMware what got
much, much better now. I must check the release notes if they mention the
overworked USB subsystem. Still some sour taste remains as the very similar
BladeRF with its identical FX3 USB controller never ever showed such
annoying behavior :)

Ralph.



_______________________________________________
USRP-users mailing list
USRP-users at lists.ettus.com <mailto:USRP-users at lists.ettus.com> 
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20141205/401cd330/attachment-0002.html>


More information about the USRP-users mailing list