Пример #1
0
static int
ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
		   int length, struct devrequest *req)
{
	struct QH *qh;
	struct qTD *td;
	volatile struct qTD *vtd;
	unsigned long ts;
	uint32_t *tdp;
	uint32_t endpt, token, usbsts;
	uint32_t c, toggle;
	uint32_t cmd;
	int timeout;
	int ret = 0;

	debug("dev=%p, pipe=%lx, buffer=%p, length=%d, req=%p\n", dev, pipe,
	      buffer, length, req);
	if (req != NULL)
		debug("req=%u (%#x), type=%u (%#x), value=%u (%#x), index=%u\n",
		      req->request, req->request,
		      req->requesttype, req->requesttype,
		      le16_to_cpu(req->value), le16_to_cpu(req->value),
		      le16_to_cpu(req->index));

	qh = ehci_alloc(sizeof(struct QH), 32);
	if (qh == NULL) {
		debug("unable to allocate QH\n");
		return -1;
	}
	qh->qh_link = cpu_to_hc32((uint32_t)&qh_list | QH_LINK_TYPE_QH);
	c = (usb_pipespeed(pipe) != USB_SPEED_HIGH &&
	     usb_pipeendpoint(pipe) == 0) ? 1 : 0;
	endpt = (8 << 28) |
	    (c << 27) |
	    (usb_maxpacket(dev, pipe) << 16) |
	    (0 << 15) |
	    (1 << 14) |
	    (usb_pipespeed(pipe) << 12) |
	    (usb_pipeendpoint(pipe) << 8) |
	    (0 << 7) | (usb_pipedevice(pipe) << 0);
	qh->qh_endpt1 = cpu_to_hc32(endpt);
	endpt = (1 << 30) |
	    (dev->portnr << 23) |
	    (dev->parent->devnum << 16) | (0 << 8) | (0 << 0);
	qh->qh_endpt2 = cpu_to_hc32(endpt);
	qh->qh_overlay.qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);

	td = NULL;
	tdp = &qh->qh_overlay.qt_next;

	toggle =
	    usb_gettoggle(dev, usb_pipeendpoint(pipe), usb_pipeout(pipe));

	if (req != NULL) {
		td = ehci_alloc(sizeof(struct qTD), 32);
		if (td == NULL) {
			debug("unable to allocate SETUP td\n");
			goto fail;
		}
		td->qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
		td->qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
		token = (0 << 31) |
		    (sizeof(*req) << 16) |
		    (0 << 15) | (0 << 12) | (3 << 10) | (2 << 8) | (0x80 << 0);
		td->qt_token = cpu_to_hc32(token);
		if (ehci_td_buffer(td, req, sizeof(*req)) != 0) {
			debug("unable construct SETUP td\n");
			ehci_free(td, sizeof(*td));
			goto fail;
		}
		*tdp = cpu_to_hc32((uint32_t) td);
		tdp = &td->qt_next;
		toggle = 1;
	}

	if (length > 0 || req == NULL) {
		td = ehci_alloc(sizeof(struct qTD), 32);
		if (td == NULL) {
			debug("unable to allocate DATA td\n");
			goto fail;
		}
		td->qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
		td->qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
		token = (toggle << 31) |
		    (length << 16) |
		    ((req == NULL ? 1 : 0) << 15) |
		    (0 << 12) |
		    (3 << 10) |
		    ((usb_pipein(pipe) ? 1 : 0) << 8) | (0x80 << 0);
		td->qt_token = cpu_to_hc32(token);
		if (ehci_td_buffer(td, buffer, length) != 0) {
			debug("unable construct DATA td\n");
			ehci_free(td, sizeof(*td));
			goto fail;
		}
		*tdp = cpu_to_hc32((uint32_t) td);
		tdp = &td->qt_next;
	}

	if (req != NULL) {
		td = ehci_alloc(sizeof(struct qTD), 32);
		if (td == NULL) {
			debug("unable to allocate ACK td\n");
			goto fail;
		}
		td->qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
		td->qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
		token = (toggle << 31) |
		    (0 << 16) |
		    (1 << 15) |
		    (0 << 12) |
		    (3 << 10) |
		    ((usb_pipein(pipe) ? 0 : 1) << 8) | (0x80 << 0);
		td->qt_token = cpu_to_hc32(token);
		*tdp = cpu_to_hc32((uint32_t) td);
		tdp = &td->qt_next;
	}

	qh_list.qh_link = cpu_to_hc32((uint32_t) qh | QH_LINK_TYPE_QH);

	/* Flush dcache */
	ehci_flush_dcache(&qh_list);

	usbsts = ehci_readl(&hcor->or_usbsts);
	ehci_writel(&hcor->or_usbsts, (usbsts & 0x3f));

	/* Enable async. schedule. */
	cmd = ehci_readl(&hcor->or_usbcmd);
	cmd |= CMD_ASE;
	ehci_writel(&hcor->or_usbcmd, cmd);

	ret = handshake((uint32_t *)&hcor->or_usbsts, STD_ASS, STD_ASS,
			100 * 1000);
	if (ret < 0) {
		printf("EHCI fail timeout STD_ASS set\n");
		goto fail;
	}

	/* Wait for TDs to be processed. */
	ts = get_timer(0);
	vtd = td;
	timeout = USB_TIMEOUT_MS(pipe);
	do {
		/* Invalidate dcache */
		ehci_invalidate_dcache(&qh_list);
		token = hc32_to_cpu(vtd->qt_token);
		if (!(token & 0x80))
			break;
		WATCHDOG_RESET();
	} while (get_timer(ts) < timeout);

	/* Check that the TD processing happened */
	if (token & 0x80) {
		printf("EHCI timed out on TD - token=%#x\n", token);
		goto fail;
	}

	/* Disable async schedule. */
	cmd = ehci_readl(&hcor->or_usbcmd);
	cmd &= ~CMD_ASE;
	ehci_writel(&hcor->or_usbcmd, cmd);

	ret = handshake((uint32_t *)&hcor->or_usbsts, STD_ASS, 0,
			100 * 1000);
	if (ret < 0) {
		printf("EHCI fail timeout STD_ASS reset\n");
		goto fail;
	}

	qh_list.qh_link = cpu_to_hc32((uint32_t)&qh_list | QH_LINK_TYPE_QH);

	token = hc32_to_cpu(qh->qh_overlay.qt_token);
	if (!(token & 0x80)) {
		debug("TOKEN=%#x\n", token);
		switch (token & 0xfc) {
		case 0:
			toggle = token >> 31;
			usb_settoggle(dev, usb_pipeendpoint(pipe),
				       usb_pipeout(pipe), toggle);
			dev->status = 0;
			break;
		case 0x40:
			dev->status = USB_ST_STALLED;
			break;
		case 0xa0:
		case 0x20:
			dev->status = USB_ST_BUF_ERR;
			break;
		case 0x50:
		case 0x10:
			dev->status = USB_ST_BABBLE_DET;
			break;
		default:
			dev->status = USB_ST_CRC_ERR;
			if ((token & 0x40) == 0x40)
				dev->status |= USB_ST_STALLED;
			break;
		}
		dev->act_len = length - ((token >> 16) & 0x7fff);
	} else {
Пример #2
0
static int
ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
		   int length, struct devrequest *req)
{
	ALLOC_ALIGN_BUFFER(struct QH, qh, 1, USB_DMA_MINALIGN);
	struct qTD *qtd;
	int qtd_count = 0;
	int qtd_counter = 0;
	volatile struct qTD *vtd;
	unsigned long ts;
	uint32_t *tdp;
	uint32_t endpt, maxpacket, token, usbsts;
	uint32_t c, toggle;
	uint32_t cmd;
	int timeout;
	int ret = 0;
	struct ehci_ctrl *ctrl = ehci_get_ctrl(dev);

	debug("dev=%p, pipe=%lx, buffer=%p, length=%d, req=%p\n", dev, pipe,
	      buffer, length, req);
	if (req != NULL)
		debug("req=%u (%#x), type=%u (%#x), value=%u (%#x), index=%u\n",
		      req->request, req->request,
		      req->requesttype, req->requesttype,
		      le16_to_cpu(req->value), le16_to_cpu(req->value),
		      le16_to_cpu(req->index));

#define PKT_ALIGN	512
	/*
	 * The USB transfer is split into qTD transfers. Eeach qTD transfer is
	 * described by a transfer descriptor (the qTD). The qTDs form a linked
	 * list with a queue head (QH).
	 *
	 * Each qTD transfer starts with a new USB packet, i.e. a packet cannot
	 * have its beginning in a qTD transfer and its end in the following
	 * one, so the qTD transfer lengths have to be chosen accordingly.
	 *
	 * Each qTD transfer uses up to QT_BUFFER_CNT data buffers, mapped to
	 * single pages. The first data buffer can start at any offset within a
	 * page (not considering the cache-line alignment issues), while the
	 * following buffers must be page-aligned. There is no alignment
	 * constraint on the size of a qTD transfer.
	 */
	if (req != NULL)
		/* 1 qTD will be needed for SETUP, and 1 for ACK. */
		qtd_count += 1 + 1;
	if (length > 0 || req == NULL) {
		/*
		 * Determine the qTD transfer size that will be used for the
		 * data payload (not considering the first qTD transfer, which
		 * may be longer or shorter, and the final one, which may be
		 * shorter).
		 *
		 * In order to keep each packet within a qTD transfer, the qTD
		 * transfer size is aligned to PKT_ALIGN, which is a multiple of
		 * wMaxPacketSize (except in some cases for interrupt transfers,
		 * see comment in submit_int_msg()).
		 *
		 * By default, i.e. if the input buffer is aligned to PKT_ALIGN,
		 * QT_BUFFER_CNT full pages will be used.
		 */
		int xfr_sz = QT_BUFFER_CNT;
		/*
		 * However, if the input buffer is not aligned to PKT_ALIGN, the
		 * qTD transfer size will be one page shorter, and the first qTD
		 * data buffer of each transfer will be page-unaligned.
		 */
		if ((unsigned long)buffer & (PKT_ALIGN - 1))
			xfr_sz--;
		/* Convert the qTD transfer size to bytes. */
		xfr_sz *= EHCI_PAGE_SIZE;
		/*
		 * Approximate by excess the number of qTDs that will be
		 * required for the data payload. The exact formula is way more
		 * complicated and saves at most 2 qTDs, i.e. a total of 128
		 * bytes.
		 */
		qtd_count += 2 + length / xfr_sz;
	}
/*
 * Threshold value based on the worst-case total size of the allocated qTDs for
 * a mass-storage transfer of 65535 blocks of 512 bytes.
 */
#if CONFIG_SYS_MALLOC_LEN <= 64 + 128 * 1024
#warning CONFIG_SYS_MALLOC_LEN may be too small for EHCI
#endif
	qtd = memalign(USB_DMA_MINALIGN, qtd_count * sizeof(struct qTD));
	if (qtd == NULL) {
		printf("unable to allocate TDs\n");
		return -1;
	}

	memset(qh, 0, sizeof(struct QH));
	memset(qtd, 0, qtd_count * sizeof(*qtd));

	toggle = usb_gettoggle(dev, usb_pipeendpoint(pipe), usb_pipeout(pipe));

	/*
	 * Setup QH (3.6 in ehci-r10.pdf)
	 *
	 *   qh_link ................. 03-00 H
	 *   qh_endpt1 ............... 07-04 H
	 *   qh_endpt2 ............... 0B-08 H
	 * - qh_curtd
	 *   qh_overlay.qt_next ...... 13-10 H
	 * - qh_overlay.qt_altnext
	 */
	qh->qh_link = cpu_to_hc32((unsigned long)&ctrl->qh_list | QH_LINK_TYPE_QH);
	c = (dev->speed != USB_SPEED_HIGH) && !usb_pipeendpoint(pipe);
	maxpacket = usb_maxpacket(dev, pipe);
	endpt = QH_ENDPT1_RL(8) | QH_ENDPT1_C(c) |
		QH_ENDPT1_MAXPKTLEN(maxpacket) | QH_ENDPT1_H(0) |
		QH_ENDPT1_DTC(QH_ENDPT1_DTC_DT_FROM_QTD) |
		QH_ENDPT1_EPS(ehci_encode_speed(dev->speed)) |
		QH_ENDPT1_ENDPT(usb_pipeendpoint(pipe)) | QH_ENDPT1_I(0) |
		QH_ENDPT1_DEVADDR(usb_pipedevice(pipe));
	qh->qh_endpt1 = cpu_to_hc32(endpt);
	endpt = QH_ENDPT2_MULT(1) | QH_ENDPT2_UFCMASK(0) | QH_ENDPT2_UFSMASK(0);
	qh->qh_endpt2 = cpu_to_hc32(endpt);
	ehci_update_endpt2_dev_n_port(dev, qh);
	qh->qh_overlay.qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
	qh->qh_overlay.qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);

	tdp = &qh->qh_overlay.qt_next;

	if (req != NULL) {
		/*
		 * Setup request qTD (3.5 in ehci-r10.pdf)
		 *
		 *   qt_next ................ 03-00 H
		 *   qt_altnext ............. 07-04 H
		 *   qt_token ............... 0B-08 H
		 *
		 *   [ buffer, buffer_hi ] loaded with "req".
		 */
		qtd[qtd_counter].qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
		qtd[qtd_counter].qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
		token = QT_TOKEN_DT(0) | QT_TOKEN_TOTALBYTES(sizeof(*req)) |
			QT_TOKEN_IOC(0) | QT_TOKEN_CPAGE(0) | QT_TOKEN_CERR(3) |
			QT_TOKEN_PID(QT_TOKEN_PID_SETUP) |
			QT_TOKEN_STATUS(QT_TOKEN_STATUS_ACTIVE);
		qtd[qtd_counter].qt_token = cpu_to_hc32(token);
		if (ehci_td_buffer(&qtd[qtd_counter], req, sizeof(*req))) {
			printf("unable to construct SETUP TD\n");
			goto fail;
		}
		/* Update previous qTD! */
		*tdp = cpu_to_hc32((unsigned long)&qtd[qtd_counter]);
		tdp = &qtd[qtd_counter++].qt_next;
		toggle = 1;
	}

	if (length > 0 || req == NULL) {
		uint8_t *buf_ptr = buffer;
		int left_length = length;

		do {
			/*
			 * Determine the size of this qTD transfer. By default,
			 * QT_BUFFER_CNT full pages can be used.
			 */
			int xfr_bytes = QT_BUFFER_CNT * EHCI_PAGE_SIZE;
			/*
			 * However, if the input buffer is not page-aligned, the
			 * portion of the first page before the buffer start
			 * offset within that page is unusable.
			 */
			xfr_bytes -= (unsigned long)buf_ptr & (EHCI_PAGE_SIZE - 1);
			/*
			 * In order to keep each packet within a qTD transfer,
			 * align the qTD transfer size to PKT_ALIGN.
			 */
			xfr_bytes &= ~(PKT_ALIGN - 1);
			/*
			 * This transfer may be shorter than the available qTD
			 * transfer size that has just been computed.
			 */
			xfr_bytes = min(xfr_bytes, left_length);

			/*
			 * Setup request qTD (3.5 in ehci-r10.pdf)
			 *
			 *   qt_next ................ 03-00 H
			 *   qt_altnext ............. 07-04 H
			 *   qt_token ............... 0B-08 H
			 *
			 *   [ buffer, buffer_hi ] loaded with "buffer".
			 */
			qtd[qtd_counter].qt_next =
					cpu_to_hc32(QT_NEXT_TERMINATE);
			qtd[qtd_counter].qt_altnext =
					cpu_to_hc32(QT_NEXT_TERMINATE);
			token = QT_TOKEN_DT(toggle) |
				QT_TOKEN_TOTALBYTES(xfr_bytes) |
				QT_TOKEN_IOC(req == NULL) | QT_TOKEN_CPAGE(0) |
				QT_TOKEN_CERR(3) |
				QT_TOKEN_PID(usb_pipein(pipe) ?
					QT_TOKEN_PID_IN : QT_TOKEN_PID_OUT) |
				QT_TOKEN_STATUS(QT_TOKEN_STATUS_ACTIVE);
			qtd[qtd_counter].qt_token = cpu_to_hc32(token);
			if (ehci_td_buffer(&qtd[qtd_counter], buf_ptr,
						xfr_bytes)) {
				printf("unable to construct DATA TD\n");
				goto fail;
			}
			/* Update previous qTD! */
			*tdp = cpu_to_hc32((unsigned long)&qtd[qtd_counter]);
			tdp = &qtd[qtd_counter++].qt_next;
			/*
			 * Data toggle has to be adjusted since the qTD transfer
			 * size is not always an even multiple of
			 * wMaxPacketSize.
			 */
			if ((xfr_bytes / maxpacket) & 1)
				toggle ^= 1;
			buf_ptr += xfr_bytes;
			left_length -= xfr_bytes;
		} while (left_length > 0);
	}

	if (req != NULL) {
		/*
		 * Setup request qTD (3.5 in ehci-r10.pdf)
		 *
		 *   qt_next ................ 03-00 H
		 *   qt_altnext ............. 07-04 H
		 *   qt_token ............... 0B-08 H
		 */
		qtd[qtd_counter].qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
		qtd[qtd_counter].qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
		token = QT_TOKEN_DT(1) | QT_TOKEN_TOTALBYTES(0) |
			QT_TOKEN_IOC(1) | QT_TOKEN_CPAGE(0) | QT_TOKEN_CERR(3) |
			QT_TOKEN_PID(usb_pipein(pipe) ?
				QT_TOKEN_PID_OUT : QT_TOKEN_PID_IN) |
			QT_TOKEN_STATUS(QT_TOKEN_STATUS_ACTIVE);
		qtd[qtd_counter].qt_token = cpu_to_hc32(token);
		/* Update previous qTD! */
		*tdp = cpu_to_hc32((unsigned long)&qtd[qtd_counter]);
		tdp = &qtd[qtd_counter++].qt_next;
	}

	ctrl->qh_list.qh_link = cpu_to_hc32((unsigned long)qh | QH_LINK_TYPE_QH);

	/* Flush dcache */
	flush_dcache_range((unsigned long)&ctrl->qh_list,
		ALIGN_END_ADDR(struct QH, &ctrl->qh_list, 1));
	flush_dcache_range((unsigned long)qh, ALIGN_END_ADDR(struct QH, qh, 1));
	flush_dcache_range((unsigned long)qtd,
			   ALIGN_END_ADDR(struct qTD, qtd, qtd_count));

	/* Set async. queue head pointer. */
	ehci_writel(&ctrl->hcor->or_asynclistaddr, (unsigned long)&ctrl->qh_list);

	usbsts = ehci_readl(&ctrl->hcor->or_usbsts);
	ehci_writel(&ctrl->hcor->or_usbsts, (usbsts & 0x3f));

	/* Enable async. schedule. */
	cmd = ehci_readl(&ctrl->hcor->or_usbcmd);
	cmd |= CMD_ASE;
	ehci_writel(&ctrl->hcor->or_usbcmd, cmd);

	ret = handshake((uint32_t *)&ctrl->hcor->or_usbsts, STS_ASS, STS_ASS,
			100 * 1000);
	if (ret < 0) {
		printf("EHCI fail timeout STS_ASS set\n");
		goto fail;
	}

	/* Wait for TDs to be processed. */
	ts = get_timer(0);
	vtd = &qtd[qtd_counter - 1];
	timeout = USB_TIMEOUT_MS(pipe);
	do {
		/* Invalidate dcache */
		invalidate_dcache_range((unsigned long)&ctrl->qh_list,
			ALIGN_END_ADDR(struct QH, &ctrl->qh_list, 1));
		invalidate_dcache_range((unsigned long)qh,
			ALIGN_END_ADDR(struct QH, qh, 1));
		invalidate_dcache_range((unsigned long)qtd,
			ALIGN_END_ADDR(struct qTD, qtd, qtd_count));

		token = hc32_to_cpu(vtd->qt_token);
		if (!(QT_TOKEN_GET_STATUS(token) & QT_TOKEN_STATUS_ACTIVE))
			break;
		WATCHDOG_RESET();
	} while (get_timer(ts) < timeout);

	/*
	 * Invalidate the memory area occupied by buffer
	 * Don't try to fix the buffer alignment, if it isn't properly
	 * aligned it's upper layer's fault so let invalidate_dcache_range()
	 * vow about it. But we have to fix the length as it's actual
	 * transfer length and can be unaligned. This is potentially
	 * dangerous operation, it's responsibility of the calling
	 * code to make sure enough space is reserved.
	 */
	invalidate_dcache_range((unsigned long)buffer,
		ALIGN((unsigned long)buffer + length, ARCH_DMA_MINALIGN));

	/* Check that the TD processing happened */
	if (QT_TOKEN_GET_STATUS(token) & QT_TOKEN_STATUS_ACTIVE)
		printf("EHCI timed out on TD - token=%#x\n", token);

	/* Disable async schedule. */
	cmd = ehci_readl(&ctrl->hcor->or_usbcmd);
	cmd &= ~CMD_ASE;
	ehci_writel(&ctrl->hcor->or_usbcmd, cmd);

	ret = handshake((uint32_t *)&ctrl->hcor->or_usbsts, STS_ASS, 0,
			100 * 1000);
	if (ret < 0) {
		printf("EHCI fail timeout STS_ASS reset\n");
		goto fail;
	}

	token = hc32_to_cpu(qh->qh_overlay.qt_token);
	if (!(QT_TOKEN_GET_STATUS(token) & QT_TOKEN_STATUS_ACTIVE)) {
		debug("TOKEN=%#x\n", token);
		switch (QT_TOKEN_GET_STATUS(token) &
			~(QT_TOKEN_STATUS_SPLITXSTATE | QT_TOKEN_STATUS_PERR)) {
		case 0:
			toggle = QT_TOKEN_GET_DT(token);
			usb_settoggle(dev, usb_pipeendpoint(pipe),
				       usb_pipeout(pipe), toggle);
			dev->status = 0;
			break;
		case QT_TOKEN_STATUS_HALTED:
			dev->status = USB_ST_STALLED;
			break;
		case QT_TOKEN_STATUS_ACTIVE | QT_TOKEN_STATUS_DATBUFERR:
		case QT_TOKEN_STATUS_DATBUFERR:
			dev->status = USB_ST_BUF_ERR;
			break;
		case QT_TOKEN_STATUS_HALTED | QT_TOKEN_STATUS_BABBLEDET:
		case QT_TOKEN_STATUS_BABBLEDET:
			dev->status = USB_ST_BABBLE_DET;
			break;
		default:
			dev->status = USB_ST_CRC_ERR;
			if (QT_TOKEN_GET_STATUS(token) & QT_TOKEN_STATUS_HALTED)
				dev->status |= USB_ST_STALLED;
			break;
		}
		dev->act_len = length - QT_TOKEN_GET_TOTALBYTES(token);
	} else {
		dev->act_len = 0;
#ifndef CONFIG_USB_EHCI_FARADAY
		debug("dev=%u, usbsts=%#x, p[1]=%#x, p[2]=%#x\n",
		      dev->devnum, ehci_readl(&ctrl->hcor->or_usbsts),
		      ehci_readl(&ctrl->hcor->or_portsc[0]),
		      ehci_readl(&ctrl->hcor->or_portsc[1]));
#endif
	}

	free(qtd);
	return (dev->status != USB_ST_NOT_PROC) ? 0 : -1;

fail:
	free(qtd);
	return -1;
}
Пример #3
0
static int
ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
		   int length, struct devrequest *req)
{
	static struct QH qh __attribute__((aligned(32)));
	static struct qTD qtd[3] __attribute__((aligned (32)));
	int qtd_counter = 0;

	volatile struct qTD *vtd;
	unsigned long ts;
	uint32_t *tdp;
	uint32_t endpt, token, usbsts;
	uint32_t c, toggle;
	uint32_t cmd;
	int timeout;
	int ret = 0;

	debug("dev=%p, pipe=%lx, buffer=%p, length=%d, req=%p\n", dev, pipe,
	      buffer, length, req);
	if (req != NULL)
		debug("req=%u (%#x), type=%u (%#x), value=%u (%#x), index=%u\n",
		      req->request, req->request,
		      req->requesttype, req->requesttype,
		      le16_to_cpu(req->value), le16_to_cpu(req->value),
		      le16_to_cpu(req->index));

	memset(&qh, 0, sizeof(struct QH));
	memset(qtd, 0, sizeof(qtd));

	toggle = usb_gettoggle(dev, usb_pipeendpoint(pipe), usb_pipeout(pipe));

	/*
	 * Setup QH (3.6 in ehci-r10.pdf)
	 *
	 *   qh_link ................. 03-00 H
	 *   qh_endpt1 ............... 07-04 H
	 *   qh_endpt2 ............... 0B-08 H
	 * - qh_curtd
	 *   qh_overlay.qt_next ...... 13-10 H
	 * - qh_overlay.qt_altnext
	 */
	qh.qh_link = cpu_to_hc32((uint32_t)&qh_list | QH_LINK_TYPE_QH);
	c = (usb_pipespeed(pipe) != USB_SPEED_HIGH &&
	     usb_pipeendpoint(pipe) == 0) ? 1 : 0;
	endpt = (8 << 28) |
	    (c << 27) |
	    (usb_maxpacket(dev, pipe) << 16) |
	    (0 << 15) |
	    (1 << 14) |
	    (usb_pipespeed(pipe) << 12) |
	    (usb_pipeendpoint(pipe) << 8) |
	    (0 << 7) | (usb_pipedevice(pipe) << 0);
	qh.qh_endpt1 = cpu_to_hc32(endpt);
	endpt = (1 << 30) |
	    (dev->portnr << 23) |
	    (dev->parent->devnum << 16) | (0 << 8) | (0 << 0);
	qh.qh_endpt2 = cpu_to_hc32(endpt);
	qh.qh_overlay.qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);

	tdp = &qh.qh_overlay.qt_next;

	if (req != NULL) {
		/*
		 * Setup request qTD (3.5 in ehci-r10.pdf)
		 *
		 *   qt_next ................ 03-00 H
		 *   qt_altnext ............. 07-04 H
		 *   qt_token ............... 0B-08 H
		 *
		 *   [ buffer, buffer_hi ] loaded with "req".
		 */
		qtd[qtd_counter].qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
		qtd[qtd_counter].qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
		token = (0 << 31) |
		    (sizeof(*req) << 16) |
		    (0 << 15) | (0 << 12) | (3 << 10) | (2 << 8) | (0x80 << 0);
		qtd[qtd_counter].qt_token = cpu_to_hc32(token);
		if (ehci_td_buffer(&qtd[qtd_counter], req, sizeof(*req)) != 0) {
			debug("unable construct SETUP td\n");
			goto fail;
		}
		/* Update previous qTD! */
		*tdp = cpu_to_hc32((uint32_t)&qtd[qtd_counter]);
		tdp = &qtd[qtd_counter++].qt_next;
		toggle = 1;
	}

	if (length > 0 || req == NULL) {
		/*
		 * Setup request qTD (3.5 in ehci-r10.pdf)
		 *
		 *   qt_next ................ 03-00 H
		 *   qt_altnext ............. 07-04 H
		 *   qt_token ............... 0B-08 H
		 *
		 *   [ buffer, buffer_hi ] loaded with "buffer".
		 */
		qtd[qtd_counter].qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
		qtd[qtd_counter].qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
		token = (toggle << 31) |
		    (length << 16) |
		    ((req == NULL ? 1 : 0) << 15) |
		    (0 << 12) |
		    (3 << 10) |
		    ((usb_pipein(pipe) ? 1 : 0) << 8) | (0x80 << 0);
		qtd[qtd_counter].qt_token = cpu_to_hc32(token);
		if (ehci_td_buffer(&qtd[qtd_counter], buffer, length) != 0) {
			debug("unable construct DATA td\n");
			goto fail;
		}
		/* Update previous qTD! */
		*tdp = cpu_to_hc32((uint32_t)&qtd[qtd_counter]);
		tdp = &qtd[qtd_counter++].qt_next;
	}

	if (req != NULL) {
		/*
		 * Setup request qTD (3.5 in ehci-r10.pdf)
		 *
		 *   qt_next ................ 03-00 H
		 *   qt_altnext ............. 07-04 H
		 *   qt_token ............... 0B-08 H
		 */
		qtd[qtd_counter].qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
		qtd[qtd_counter].qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
		token = (toggle << 31) |
		    (0 << 16) |
		    (1 << 15) |
		    (0 << 12) |
		    (3 << 10) |
		    ((usb_pipein(pipe) ? 0 : 1) << 8) | (0x80 << 0);
		qtd[qtd_counter].qt_token = cpu_to_hc32(token);
		/* Update previous qTD! */
		*tdp = cpu_to_hc32((uint32_t)&qtd[qtd_counter]);
		tdp = &qtd[qtd_counter++].qt_next;
	}

	qh_list.qh_link = cpu_to_hc32((uint32_t)&qh | QH_LINK_TYPE_QH);

	/* Flush dcache */
	flush_dcache_range((uint32_t)&qh_list,
		(uint32_t)&qh_list + sizeof(struct QH));
	flush_dcache_range((uint32_t)&qh, (uint32_t)&qh + sizeof(struct QH));
	flush_dcache_range((uint32_t)qtd, (uint32_t)qtd + sizeof(qtd));

	usbsts = ehci_readl(&hcor->or_usbsts);
	ehci_writel(&hcor->or_usbsts, (usbsts & 0x3f));

	/* Enable async. schedule. */
	cmd = ehci_readl(&hcor->or_usbcmd);
	cmd |= CMD_ASE;
	ehci_writel(&hcor->or_usbcmd, cmd);

	ret = handshake((uint32_t *)&hcor->or_usbsts, STD_ASS, STD_ASS,
			100 * 1000);
	if (ret < 0) {
		printf("EHCI fail timeout STD_ASS set\n");
		goto fail;
	}

	/* Wait for TDs to be processed. */
	ts = get_timer(0);
	vtd = &qtd[qtd_counter - 1];
	timeout = USB_TIMEOUT_MS(pipe);
	do {
		/* Invalidate dcache */
		invalidate_dcache_range((uint32_t)&qh_list,
			(uint32_t)&qh_list + sizeof(struct QH));
		invalidate_dcache_range((uint32_t)&qh,
			(uint32_t)&qh + sizeof(struct QH));
		invalidate_dcache_range((uint32_t)qtd,
			(uint32_t)qtd + sizeof(qtd));

		token = hc32_to_cpu(vtd->qt_token);
		if (!(token & 0x80))
			break;
		WATCHDOG_RESET();
	} while (get_timer(ts) < timeout);

	/* Invalidate the memory area occupied by buffer */
	invalidate_dcache_range(((uint32_t)buffer & ~31),
		((uint32_t)buffer & ~31) + roundup(length, 32));

	/* Check that the TD processing happened */
	if (token & 0x80) {
		printf("EHCI timed out on TD - token=%#x\n", token);
	}

	/* Disable async schedule. */
	cmd = ehci_readl(&hcor->or_usbcmd);
	cmd &= ~CMD_ASE;
	ehci_writel(&hcor->or_usbcmd, cmd);

	ret = handshake((uint32_t *)&hcor->or_usbsts, STD_ASS, 0,
			100 * 1000);
	if (ret < 0) {
		printf("EHCI fail timeout STD_ASS reset\n");
		goto fail;
	}

	qh_list.qh_link = cpu_to_hc32((uint32_t)&qh_list | QH_LINK_TYPE_QH);

	token = hc32_to_cpu(qh.qh_overlay.qt_token);
	if (!(token & 0x80)) {
		debug("TOKEN=%#x\n", token);
		switch (token & 0xfc) {
		case 0:
			toggle = token >> 31;
			usb_settoggle(dev, usb_pipeendpoint(pipe),
				       usb_pipeout(pipe), toggle);
			dev->status = 0;
			break;
		case 0x40:
			dev->status = USB_ST_STALLED;
			break;
		case 0xa0:
		case 0x20:
			dev->status = USB_ST_BUF_ERR;
			break;
		case 0x50:
		case 0x10:
			dev->status = USB_ST_BABBLE_DET;
			break;
		default:
			dev->status = USB_ST_CRC_ERR;
			if ((token & 0x40) == 0x40)
				dev->status |= USB_ST_STALLED;
			break;
		}
		dev->act_len = length - ((token >> 16) & 0x7fff);
	} else {