[lug] RHEL6 USB and network issue

Maxwell Spangler maxlists at maxwellspangler.com
Mon Nov 11 20:12:53 MST 2013


I'm not sure why the IOMMU would have been disabled other than some
early AMD cpus had IOMMU bugs in them, but that was a while ago.

If you've got an IOMMU option, enable it.  For Virtualization, they
really help performance of mapping memory among other things.

Cheers,

On Mon, 2013-11-11 at 19:34 -0700, Steven Hart wrote:

> Thanks for the info Maxwell!  It's definitely pointed me in the right
> direction.  I could find no option to turn off the 3x power for the
> usb in the bios and upon further reading, for the 3x to work, you are
> supposed to load up a specific driver. I'm going to assume that
> feature is turned off for now.  But I did find this link:
> 
> http://ubuntuforums.org/archive/index.php/t-2111223.html
> 
> And it seems that if you set  IOMMU to ENABLED in the bios the
> problems disappear using Ubuntu.  They even mention the board I have
> so I'm hopeful but when working with RHEL I've found that a little
> skepticism is warranted.  
> 
> I'm trying this now but need to head out soon so I'll let the install
> run and test it tomorrow morning.  A 12 hour work day is long enough
> for me.
> 
> Thanks for the help and I'll be sure to post my progress in case
> anyone else runs into this similar issue.
> 
> cheers
> 
> Steve
> 
> 
> On 11/11/2013 06:27 PM, Maxwell Spangler wrote:
> 
> > 
> > Here are two suggestions, both related to BIOS/UEFI configuration:
> > 
> > First, is exactly your issue on exactly your motherboard:
> > 
> > http://unix.stackexchange.com/questions/72625/why-is-usb-not-working-in-linux-when-it-works-in-uefi-bios
> > 
> > Second, my guess related to USB power:
> > 
> > This motherboard says it has proprietary features that let it
> > provide 3x USB power (presumably 1500ma instead of 500ma).  
> > 
> > So maybe this guys onto something?  The USB system might think it's
> > got an issue with the power use.
> > 
> > http://paulphilippov.com/articles/how-to-fix-device-not-accepting-address-error
> > 
> > I would try turning off any 3x USB Power features in the BIOS,
> > rebooting, and see what happens as well.
> > 
> > Keep us posted.
> > 
> > On Mon, 2013-11-11 at 17:42 -0700, Steven Hart wrote:
> > 
> > > Tried it out and got no change in behavior.  Exact same errors and
> > > still no working keyboard and mouse.
> > > 
> > > Steve
> > > 
> > > On 11/11/2013 05:32 PM, Vishal Verma wrote:
> > > 
> > > 
> > > > I'd try disabling the autosuspend at this point: 
> > > > 
> > > > 
> > > > # echo -1 > /sys/module/usbcore/parameters/autosuspend 
> > > > 
> > > > 
> > > > And reboot. 
> > > > 
> > > > 
> > > > -Vishal 
> > > > 
> > > > 
> > > > On Mon, Nov 11, 2013 at 5:23 PM, Steven Hart
> > > > <steven.hart at colorado.edu> wrote: 
> > > > 
> > > >         # while :; do lsusb -v > /dev/null; done
> > > >         
> > > >         
> > > >         results in repeated messages like before:  "device
> > > >         descriptor read/64, error -32"
> > > >         
> > > >         the keyboard and mouse still do not work.
> > > >         
> > > >         # cat /sys/module/usbcore/parameters/autosuspend
> > > >         
> > > >         returns:  "2"
> > > >         
> > > >         Thanks for the help!  This one is driving me nuts.
> > > >         
> > > >         cheers
> > > >         
> > > >         Steve 
> > > >         
> > > >         
> > > >         On 11/11/2013 05:16 PM, Maxwell Spangler wrote:
> > > >         
> > > >         
> > > >         > On this motherboard, the AMD SB950 chip provides USB
> > > >         > 2.0/1.1 ports and two separate Etron EJ168 chips
> > > >         > provide USB3.0/2.0 ports.
> > > >         > 
> > > >         > Your problem could be the RHEL 6 kernel's drivers
> > > >         > putting the AMD USB2 ports into suspend mode.  It's
> > > >         > probably not doing so on the Etron ports which makes
> > > >         > the Etron ports work while the AMD ports don't.  If
> > > >         > you boot into BIOS and the USB2 ports that further
> > > >         > suggests the hardware works but that it's the kernel
> > > >         > driver that is the source of your problems.
> > > >         > 
> > > >         > If you execute this
> > > >         > 
> > > >         > # while :; do lsusb -v > /dev/null; done
> > > >         > 
> > > >         > Do the USB2 keyboard and mouse peripherals work
> > > >         > properly?
> > > >         > 
> > > >         > What does
> > > >         > 
> > > >         > # cat /sys/module/usbcore/parameters/autosuspend
> > > >         > 
> > > >         > return?
> > > >         > 
> > > >         > If the loop above resolves your problem, boot with a
> > > >         > kernel or module option to disable the usbcore
> > > >         > autosupend function and things should work better. 
> > > >         > 
> > > >         > If not, tell us more.
> > > >         > 
> > > >         > Cheers,
> > > >         > 
> > > >         > On Mon, 2013-11-11 at 16:37 -0700, Steven Hart wrote:
> > > >         > 
> > > >         > > Another oddity is that when the keyboard and mouse
> > > >         > > are plugged into USB2 ports, they work fine in the
> > > >         > > system BIOS and on the grub screen. Once booted into
> > > >         > > the OS (even single user) I get the above message
> > > >         > > and they stop working.
> > > >         > > 
> > > >         > > Cheers
> > > >         > > 
> > > >         > > Steve
> > > >         > > 
> > > >         > > On 11/11/2013 03:55 PM, Steven Hart wrote:
> > > >         > > 
> > > >         > > 
> > > >         > > > Motherboard = Gigabyte GA-990FXA-UD3
> > > >         > > > OS = RHEL 6 workstation 64 bit
> > > >         > > > 
> > > >         > > > PROBLEM 1:  When I have the USB keyboard and mouse
> > > >         > > > plugged into any USB2 port, they do not work and
> > > >         > > > RHEL6 says:
> > > >         > > > 
> > > >         > > > device descriptor read/64, error -32
> > > >         > > > 
> > > >         > > > When I plug the keyboard into the USB3 slot, it
> > > >         > > > works normally.  I need the USB2 ports to work.  I
> > > >         > > > get the same thing when I try a USB flash drive as
> > > >         > > > well.
> > > >         > > > 
> > > >         > > > PROBLEM 2: I can't get the on-board NIC to work.
> > > >         > > > I have the network settings set manually and
> > > >         > > > correctly as they work on my other 55 RHEL6
> > > >         > > > clients for my subnet.  ifconfig shows the correct
> > > >         > > > IP but no packets visible.  ethtool sees a link
> > > >         > > > connected, that it's gig speed, and that it's full
> > > >         > > > duplex.  That being said, you can't ping the
> > > >         > > > system and the system can't ping out.
> > > >         > > > 
> > > >         > > > I usually have very good luck with Gigabyte
> > > >         > > > motherboards right out of the box with RHEL6 but
> > > >         > > > this one is pissing me off.  I have a feeling it's
> > > >         > > > a setting in the BIOS but I've tried the optimized
> > > >         > > > defaults as well as just using the legacy settings
> > > >         > > > instead of the UEFI settings.
> > > >         > > > 
> > > >         > > > Any ideas on this would be greatly appreciated.
> > > >         > > > 
> > > >         > > > Cheers
> > > >         > > > 
> > > >         > > > Steve 
> > > >         > > > -- 
> > > >         > > > Steve Hart
> > > >         > > > Systems Administrator
> > > >         > > > Colorado Center for Astrodynamics Research
> > > >         > > > University of Colorado Boulder
> > > >         > > > Steven.Hart at colorado.edu
> > > >         > > > (303)492-8109 
> > > >         > > > 
> > > >         > > > 
> > > >         > > > 
> > > >         > > > _______________________________________________
> > > >         > > > Web Page:  http://lug.boulder.co.us
> > > >         > > > Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> > > >         > > > Join us on IRC: irc.hackingsociety.org port=6667 channel=#hackingsociety
> > > >         > > 
> > > >         > > 
> > > >         > > -- 
> > > >         > > Steve Hart
> > > >         > > Systems Administrator
> > > >         > > Colorado Center for Astrodynamics Research
> > > >         > > University of Colorado Boulder
> > > >         > > Steven.Hart at colorado.edu
> > > >         > > (303)492-8109 
> > > >         > > 
> > > >         > > _______________________________________________
> > > >         > > Web Page:  http://lug.boulder.co.us
> > > >         > > Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> > > >         > > Join us on IRC: irc.hackingsociety.org port=6667 channel=#hackingsociety
> > > >         > 
> > > >         > 
> > > >         > -- 
> > > >         > Maxwell Spangler
> > > >         > ========================================================================
> > > >         > Linux System Administration / Virtualization /
> > > >         > Development / Computing Services
> > > >         > Photography / Graphics Design / Writing
> > > >         > Fort Collins, Colorado
> > > >         > http://www.maxwellspangler.com 
> > > >         > 
> > > >         > 
> > > >         > 
> > > >         > _______________________________________________
> > > >         > Web Page:  http://lug.boulder.co.us
> > > >         > Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> > > >         > Join us on IRC: irc.hackingsociety.org port=6667 channel=#hackingsociety
> > > >         
> > > >         
> > > >         -- 
> > > >         Steve Hart
> > > >         Systems Administrator
> > > >         Colorado Center for Astrodynamics Research
> > > >         University of Colorado Boulder
> > > >         Steven.Hart at colorado.edu
> > > >         (303)492-8109 
> > > >         
> > > >         _______________________________________________
> > > >         Web Page:  http://lug.boulder.co.us
> > > >         Mailing List:
> > > >         http://lists.lug.boulder.co.us/mailman/listinfo/lug
> > > >         Join us on IRC: irc.hackingsociety.org port=6667
> > > >         channel=#hackingsociety 
> > > > 
> > > > 
> > > > 
> > > > 
> > > > 
> > > > 
> > > > _______________________________________________
> > > > Web Page:  http://lug.boulder.co.us
> > > > Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> > > > Join us on IRC: irc.hackingsociety.org port=6667 channel=#hackingsociety
> > > 
> > > 
> > > -- 
> > > Steve Hart
> > > Systems Administrator
> > > Colorado Center for Astrodynamics Research
> > > University of Colorado Boulder
> > > Steven.Hart at colorado.edu
> > > (303)492-8109 
> > > 
> > > _______________________________________________
> > > Web Page:  http://lug.boulder.co.us
> > > Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> > > Join us on IRC: irc.hackingsociety.org port=6667 channel=#hackingsociety
> > 
> > 
> > -- 
> > Maxwell Spangler
> > ========================================================================
> > Linux System Administration / Virtualization / Development /
> > Computing Services
> > Photography / Graphics Design / Writing
> > Fort Collins, Colorado
> > http://www.maxwellspangler.com 
> 
> 
> 
> -- 
> Steve Hart
> Systems Administrator
> Colorado Center for Astrodynamics Research
> University of Colorado Boulder
> Steven.Hart at colorado.edu
> (303)492-8109
> 
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug
> Join us on IRC: irc.hackingsociety.org port=6667 channel=#hackingsociety


-- 
Maxwell Spangler
========================================================================
Linux System Administration / Virtualization / Development / Computing
Services
Photography / Graphics Design / Writing
Fort Collins, Colorado
http://www.maxwellspangler.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lug.boulder.co.us/pipermail/lug/attachments/20131111/40eca40a/attachment.html>


More information about the LUG mailing list