[BUG] FL1009: xHCI host not responding to stop endpoint command.

[Arnaud Ebalard]

Hi Sarah,

Sarah Sharp <sarah.a.sharp at linux.intel.com> writes:

> On Sat, Jan 18, 2014 at 10:49:17PM +0100, Arnaud Ebalard wrote:
>> Hi,
>> 
>> I have added Thomas in the recipients, because I guess he may be of some
>> help debugging the issue further. Thomas, the beginning of the thread is
>> here: http://thread.gmane.org/gmane.linux.usb.general/101531
>
> ...
>
>> I started suspecting the introduction of MSI support in Marvell PCIe
>> host controller driver (FL1009 is on the PCIe bus) and compiled a
>> a 3.13.0-rc8 w/ CONFIG_PCI_MSI disabled (it was enabled in all my
>> previous tests): I did not manage to reproduce the issue with this
>> kernel. As a side note, commits 5b4deb6526bd, 31f614edb726 and
>> 627dfcc249e2 are
>> 
>> ATM, I do not know if the problem is related to a bug in introduced MSI
>> support or some weird incompatibility of that functionality with the
>> FL1009 which would require some quirk in XHCI stack.
>
> We've actually had issues in the past with Fresco Logic hosts not
> supporting MSI properly, even though the PCI devices claim to have MSI
> support.  So turning off CONFIG_PCI_MSI may actually mean the Fresco
> Logic host is to blame, rather than the Marvell patches.  I assume MSI
> wouldn't have been turned on for the Fresco Logic host unless the parent
> PCI host controller supported it.
>
> Let's see if the Fresco Logic host is really the root cause.  Please
> apply the this patch to 3.13.0-rc8 and recompile with CONFIG_PCI_MSI
> enabled:
>
> diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c
> index 6c03584ac15f..74748444c040 100644
> --- a/drivers/usb/host/xhci-pci.c
> +++ b/drivers/usb/host/xhci-pci.c
> @@ -30,6 +30,7 @@
>  /* Device for a quirk */
>  #define PCI_VENDOR_ID_FRESCO_LOGIC	0x1b73
>  #define PCI_DEVICE_ID_FRESCO_LOGIC_PDK	0x1000
> +#define PCI_DEVICE_ID_FRESCO_LOGIC_FL1009	0x1009
>  #define PCI_DEVICE_ID_FRESCO_LOGIC_FL1400	0x1400
>  
>  #define PCI_VENDOR_ID_ETRON		0x1b6f
> @@ -63,6 +64,9 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci)
>  
>  	/* Look for vendor-specific quirks */
>  	if (pdev->vendor == PCI_VENDOR_ID_FRESCO_LOGIC &&
> +			pdev->device == PCI_DEVICE_ID_FRESCO_LOGIC_FL1009)
> +		xhci->quirks |= XHCI_BROKEN_MSI;
> +	if (pdev->vendor == PCI_VENDOR_ID_FRESCO_LOGIC &&
>  			(pdev->device == PCI_DEVICE_ID_FRESCO_LOGIC_PDK ||
>  			 pdev->device == PCI_DEVICE_ID_FRESCO_LOGIC_FL1400)) {
>  		if (pdev->device == PCI_DEVICE_ID_FRESCO_LOGIC_PDK &&

With the patch applied on top of 3.13.0 kernel recompiled w/
CONFIG_PCI_MSI enabled, I cannot reproduce the bug. I guess
you can add my:

 Reported-and-tested-By: Arnaud Ebalard <arno at natisbad.org>

Since you'll have to push the patch to -stable team at least for 3.13,
I wonder if it would not make sense to extend that at least to 3.12.
and possibly 3.10 (3.2 is still widely used but I wonder if it makes
sense to go that far).

Cheers,

a+