Re: [gentoo-user] Network failed and weird error message

[Alan McKinnon <alan.mckinnon@gmail.com>]

On 06/10/2013 21:24, Dale wrote:
> Alan McKinnon wrote:
>> On 06/10/2013 20:36, Dale wrote:
>>> Alan McKinnon wrote:
>>>> These days all you need is ehci for usb2 and xhci for usb3 (unless you
>>>> are using ancient hardware with physical usb1 ports) 
>>> Well, I rebuilt the kernel and removed the OHCI and UHCI.  When I
>>> rebooted, it couldn't see my UPS and nut couldn't start its services. 
>>> So, it appears that mine must be "ancient" hardware.  My messages file
>>> is still full of the same error after this change.  That would be adding
>>> back the OHCI part.
>> lsusb, lshw, dmideciode and friends will tell you what hardware you
>> really have
>>
> 
> Yep, they say it needs OHCI.  I also checked here and it says the same
> thing.
> 
> http://kmuto.jp/debian/hcl/Giga-byte/GA-770T-USB3 
> 
> I guess my hardware is just a little out of date.  ;-) 

I thought your hardware was new enough to have dropped USB1 ports. Oh well.


> 
> 
> 
>>> BTW, I didn't have XHCI enabled so maybe now some things will be faster
>>> when using USB ports.  ;-) 
>> Nope. The hardware only runs at whatever speed it runs at.
>>
>> A USB2 device plugged into a USB3 port runs at USB2 speeds.
>> A USB1 and a USB2 device plugged into the same USB port makes both runs
>> at USB1 speeds
>>
>> There's no magic software to change that.
>>
>> But if you plug a USB3 drive into a USB3 port controlled by an OHCI
>> driver, it will run at USB2 speeds. Switching to XHCI is the only thing
>> you could do to improve speeds
> 
> That's what I meant tho.  I have USB3 ports but it seems they have been
> running at USB2 speeds since I never enabled USB3 drivers.  I sort of
> missed that.  No clue if the stuff I am plugging in supports USB3 or not
> tho.  Maybe my USB sticks do tho. 

Your USB sticks are not USB3. I have yet to see one anywhere that is. I
don;t thing they are even remotely fast enough to warrant it

If you have USB3 drives, you already know all about it. It would have
had USB3 logos emblazened all over the box, it would have cost more than
a comparable USB2 drive of the same capacity, I will be newish (last 2
years?) and the connectors are different:

The full-size classic USB plug has a blue insulator and if you look
inside it has extra pins at the rear. The end that plugs into the drive
is usually micro-USB3 and it clearly consists of 2 sections - a regular
micro-usb set of pins (that does accept old micro-usb cables) and a
second set that is slightly shorter.

You can't get this wrong, the cables are very different and yet still
backwards compatible. If your drives don't have such unusual sockets,
they are not USB3



> 
>>> Is it safe to disable this and will this kill the messages:  USB verbose
>>> debug messages
>> Well I have no idea. We haven't established yet what we are dealing with
>>
> 
> Those pesky errors filling up my message file right now.  No clue on the
> network part since it hasn't happened again. 
> 
>>> This is a grep of USB stuff. 
>>>
>>> root@fireball / # zcat /proc/config.gz | grep -i hci
>>> CONFIG_SATA_AHCI=y
>>> # CONFIG_SATA_AHCI_PLATFORM is not set
>>> # CONFIG_SATA_ACARD_AHCI is not set
>>> # CONFIG_FIREWIRE_OHCI is not set
>>> CONFIG_USB_ARCH_HAS_OHCI=y
>>> CONFIG_USB_ARCH_HAS_EHCI=y
>>> CONFIG_USB_ARCH_HAS_XHCI=y
>>> CONFIG_USB_XHCI_HCD=y
>>> # CONFIG_USB_XHCI_HCD_DEBUGGING is not set
>>> CONFIG_USB_EHCI_HCD=y
>>> # CONFIG_USB_EHCI_ROOT_HUB_TT is not set
>>> # CONFIG_USB_EHCI_TT_NEWSCHED is not set
>>> CONFIG_USB_EHCI_PCI=y
>>> CONFIG_USB_OHCI_HCD=y
>>> # CONFIG_USB_OHCI_HCD_PLATFORM is not set
>>> # CONFIG_USB_EHCI_HCD_PLATFORM is not set
>>> # CONFIG_USB_OHCI_BIG_ENDIAN_DESC is not set
>>> # CONFIG_USB_OHCI_BIG_ENDIAN_MMIO is not set
>>> CONFIG_USB_OHCI_LITTLE_ENDIAN=y
>>> # CONFIG_USB_UHCI_HCD is not set
>>> CONFIG_PROVIDE_OHCI1394_DMA_INIT=y
>>> root@fireball / #
>>>
>>> So, now what?  Can I tell syslog to ignore that error or do I need to
>>> beat something into the kernel?  
>> First find out what those errors mean. Then and only then can you decide
>> if they are ignorable or not
>>
>>
> 
> I have googled the error and there is very little info about it.  Most
> of the hits now is my posts here about the error.  I found something on
> the kernel list but it didn't appear to be the same error but somewhat
> close enough for google to grasp at straws. 
> 
> Basically, everything works that I can tell.  If I can't change
> something to fix the error, I'd just rather get rid of the error. 
> 
> I'm going to boot a new kernel and see if that helps.  I been sticking
> with this one because nvidia works well with this version but has some
> hiccups with other versions. 
> 
> Dale
> 
> :-)  :-)
> 
> P. S.  I may be slow to reply at times.  I may be out of town a lot this
> next week depending on what is up with my brother.  I know I will be out
> of town tomorrow tho. 
> 


-- 
Alan McKinnon
alan.mckinnon@gmail.com