[PATCH] Possible bug fix in the ehci-hcd

[PATCH] Possible bug fix in the ehci-hcd

[Andy Shevchenko]

Hi!

The Sequans company who produce WiMAX chips provides the USB driver and kthp
addon driver. However the USB device doesn't work correctly.

Here is more true hack for Sequans SQN11x0 based modem.

P.S. For more details, please, ask Sequans staff (see Signed-off-by lines).

Signed-off-by: Andy Shevchenko <andy@smile.org.ua>
Signed-off-by: Gosne Stephane <sgosne@sequans.com>

--- a/drivers/usb/host/ehci-q.c.orig	2008-02-07 16:53:18.000000000 +0200
+++ b/drivers/usb/host/ehci-q.c	2008-02-07 16:53:18.000000000 +0200
@@ -750,8 +750,14 @@
 			info1 |= 1 << 14;	/* toggle from qtd */
 			info2 |= (EHCI_TUNE_MULT_HS << 30);
 		} else if (type == PIPE_BULK) {
+			struct usb_device_descriptor *desc = &urb->dev->descriptor;
 			info1 |= (EHCI_TUNE_RL_HS << 28);
-			info1 |= 512 << 16;	/* usb2 fixed maxpacket */
+			/* Special hook for SQN11x0 */
+			if (desc->idVendor == 0x148e && desc->idProduct == 0x0900) {
+				info1 |= max_packet (maxp) << 16;
+			} else {
+				info1 |= 512 << 16;	/* usb2 fixed maxpacket */
+			}
 			info2 |= (EHCI_TUNE_MULT_HS << 30);
 		} else {		/* PIPE_INTERRUPT */
 			info1 |= max_packet (maxp) << 16;

-- 
With best regards,
Andy Shevchenko.      mailto: andy@smile.org.ua

Re: [PATCH] Possible bug fix in the ehci-hcd

[Greg KH]

On Thu, Feb 07, 2008 at 06:48:32PM +0200, Andy Shevchenko wrote:
> Hi!
> 
> The Sequans company who produce WiMAX chips provides the USB driver and kthp
> addon driver. However the USB device doesn't work correctly.
> 
> Here is more true hack for Sequans SQN11x0 based modem.
> 
> P.S. For more details, please, ask Sequans staff (see Signed-off-by lines).
> 
> Signed-off-by: Andy Shevchenko <andy@smile.org.ua>
> Signed-off-by: Gosne Stephane <sgosne@sequans.com>
> 
> --- a/drivers/usb/host/ehci-q.c.orig	2008-02-07 16:53:18.000000000 +0200
> +++ b/drivers/usb/host/ehci-q.c	2008-02-07 16:53:18.000000000 +0200
> @@ -750,8 +750,14 @@
>  			info1 |= 1 << 14;	/* toggle from qtd */
>  			info2 |= (EHCI_TUNE_MULT_HS << 30);
>  		} else if (type == PIPE_BULK) {
> +			struct usb_device_descriptor *desc = &urb->dev->descriptor;
>  			info1 |= (EHCI_TUNE_RL_HS << 28);
> -			info1 |= 512 << 16;	/* usb2 fixed maxpacket */
> +			/* Special hook for SQN11x0 */
> +			if (desc->idVendor == 0x148e && desc->idProduct == 0x0900) {
> +				info1 |= max_packet (maxp) << 16;
> +			} else {
> +				info1 |= 512 << 16;	/* usb2 fixed maxpacket */
> +			}

We should not be adding hacks for a specific USB device to the host
driver, that's just wrong.

The driver for the device itself should be fixed.

Do you, or anyone else have a pointer to the driver that is causing this
problem?

Andrew, please drop this from your queue.

thanks,

greg k-h