Re: xhci and other woes

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Ok it appears I was wrong about the NEC quirk.


Here is what i'm seeing at the moment:

The ax88179 is only working for me on the Renesas uPD720201 via a USB 3.0 HUB. 
However for some reason once I use the hub I get :

xhci_hcd 0000:02:00.0: WARN Successful completion on short TX: needs 
XHCI_TRUST_TX_LENGTH quirk?

With the power port v4 patch set for some reason it seemed like all that was 
needed was the NEC quirk to get them to stop.

I don't get the TRUST warning when using the Vantec Nexstar Dual Bay disk dock 
connected directly, it works fine no warnings, very fast transfers.

So ...

ax88179 via hub = works but warning
ax88179 direct connect = can bring interface up but can only xfer a very small 
amount of data before the nic stops responding



When using the hub the performance of the ax88179 is great, it's completely 
reliable and achieves 117.8MB/sec both directions.


It's looking like whatever is stopping the ax88179 from working with the 
Asmedia 1042 may also be stopping it working on the uPD720201  BUT the 
uPD720201 has a workaround where a HUB with it works with the TRUST warning.


The kernel i'm currently running matches your for-linus branch except I set 
the sg size to ~0u


Regards,

Will Trives




On Thursday 06 February 2014 08:53:52 Sarah Sharp wrote:
> On Thu, Feb 06, 2014 at 07:42:33PM +1100, renevant@xxxxxxxxxxxxxxxx wrote:
> > Hello guys,
> > 
> > Currently i've done this in xhci-pci.c
> > 
> > #define PCI_VENDOR_ID_RENESAS   0x1912
> > #define PCI_DEVICE_ID_RENESAS_UPD720201      0x0014
> > 
> >       if (pdev->vendor == PCI_VENDOR_ID_RENESAS &&
> >       
> >                        pdev->device == PCI_DEVICE_ID_RENESAS_UPD720201 &&
> >                        
> >                                pdev->revision == 0x3)
> >                
> >                xhci->quirks |= XHCI_NEC_HOST;
> > 
> > The warnings are gone.
> 
> Which warnings are gone?
> 
> > Perhaps this quirk needs to apply to all Renesas controllers ? This
> > addition i've made obviously specifically matches the card I've got but
> > the NEC match in xhci-pci.c obviously isn't matching the Renesas vendor
> > ID only NEC.> 
> >        if (pdev->vendor == PCI_VENDOR_ID_NEC)
> >        
> >                 xhci->quirks |= XHCI_NEC_HOST;
> > 
> > Maybe this needs an && PCI_VENDOR_ID_RENESAS
> 
> The only thing this quirk does is submit a command to get the NEC
> firmware version, and then print it out.  It should have no impact
> on host controller behavior.
> 
> So, what warnings disappeared when you turned this quirk on?
> 
> Sarah Sharp

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux Media]     [Linux Input]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Old Linux USB Devel Archive]

  Powered by Linux