Hi all:
At present,I used am1808 usb-host-controler to link usb-to-serial device(using to chips all 8 serial console.)
But I only used three console, the fourth console blocked.
I reviewed the code and found the reason which urb bloked.
I readed the datasheet which supporting NAKTIMEOUT for bulk transport.
But the function : musb_urb_enqueue()
>>case USB_ENDPOINT_XFER_ISOC:
/* ISO always uses logarithmic encoding */
interval = min_t(u8, epd->bInterval, 16);
break;
default:
/* REVISIT we actually want to use NAK limits, hinting to the
* transfer scheduling logic to try some other qh, e.g. try
* for 2 msec first:
*
* interval = (USB_SPEED_HIGH == urb->dev->speed) ? 16 : 2;
*
* The downside of disabling this is that transfer scheduling
* gets VERY unfair for nonperiodic transfers; a misbehaving
* peripheral could make that hurt. That's perfectly normal
* for reads from network or serial adapters ... so we have
* partial NAKlimit support for bulk RX.
*
* The upside of disabling it is simpler transfer scheduling.
*/
>> interval = 0;
>>}
>>qh->intv_reg = interval;
For bulk tran,set interval is 0,so no NAKtimeout interrupt can hanppen? Is it?
But I readed the NAKLIMINT handle: musb_h_ep0_irq() :
>>} else if (csr & MUSB_CSR0_H_NAKTIMEOUT) {
dev_dbg(musb->controller, "control NAK timeout\n");
/* NOTE: this code path would be a good place to PAUSE a
* control transfer, if another one is queued, so that
* ep0 is more likely to stay busy. That's already done
* for bulk RX transfers.
*
* if (qh->ring.next != &musb->control), then
* we have a candidate... NAKing is *NOT* an error
*/
>> musb_writew(epio, MUSB_CSR0, 0);
>> retval = IRQ_HANDLED;
}
But did not do anything.
Is it the code error or I?
Thanks for !
--------------
majianpeng
2012-05-18
--
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
