Merge branch 'master' of git://git.denx.de/u-boot-usb

* 'master' of git://git.denx.de/u-boot-usb:
  MUSB driver: Timeout is never detected as the while loop does not end
  usb: fix ulpi_set_vbus prototype
  pxa25x: Add UDC registers definitions
  USB: Fix strict aliasing in ohci-hcd
  usb: Optimize USB storage read/write
  ehci: Optimize qTD allocations
  usb_stor_BBB_transport: Do not delay when not required
  usb_storage: Remove EHCI constraints
  usb_storage: Restore non-EHCI support
  ehci-hcd: Boost transfer speed
  ehci: cosmetic: Define used constants
  ehci: Fail for multi-transaction interrupt transfers
  arm:trats: Enable g_dnl composite USB gadget with embedded DFU function on TRATS
  arm:trats: Support for USB UDC driver at TRATS board.
  dfu:cmd: Support for DFU u-boot command
  dfu: MMC specific routines for DFU operation
  dfu: DFU backend implementation
  dfu:usb: DFU USB function (f_dfu) support for g_dnl composite gadget
  dfu:usb: Support for g_dnl composite download gadget.
  ehci: cosmetic: Define the number of qt_buffers

Signed-off-by: Wolfgang Denk <wd@denx.de>

3 files changed, 301 insertions, 134 deletions

diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c
index 2a82a2912..18b4bc654 100644
--- a/drivers/usb/host/ehci-hcd.c
+++ b/drivers/usb/host/ehci-hcd.c
@@ -163,7 +163,7 @@ static int ehci_reset(void)
 
 #ifdef CONFIG_USB_EHCI_TXFIFO_THRESH
 	cmd = ehci_readl(&hcor->or_txfilltuning);
-	cmd &= ~TXFIFO_THRESH(0x3f);
+	cmd &= ~TXFIFO_THRESH_MASK;
 	cmd |= TXFIFO_THRESH(CONFIG_USB_EHCI_TXFIFO_THRESH);
 	ehci_writel(&hcor->or_txfilltuning, cmd);
 #endif
@@ -183,10 +183,10 @@ static int ehci_td_buffer(struct qTD *td, void *buf, size_t sz)
 	flush_dcache_range(addr, ALIGN(addr + sz, ARCH_DMA_MINALIGN));
 
 	idx = 0;
-	while (idx < 5) {
+	while (idx < QT_BUFFER_CNT) {
 		td->qt_buffer[idx] = cpu_to_hc32(addr);
 		td->qt_buffer_hi[idx] = 0;
-		next = (addr + 4096) & ~4095;
+		next = (addr + EHCI_PAGE_SIZE) & ~(EHCI_PAGE_SIZE - 1);
 		delta = next - addr;
 		if (delta >= sz)
 			break;
@@ -195,7 +195,7 @@ static int ehci_td_buffer(struct qTD *td, void *buf, size_t sz)
 		idx++;
 	}
 
-	if (idx == 5) {
+	if (idx == QT_BUFFER_CNT) {
 		printf("out of buffer pointers (%u bytes left)\n", sz);
 		return -1;
 	}
@@ -208,13 +208,14 @@ 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);
-	ALLOC_ALIGN_BUFFER(struct qTD, qtd, 3, 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, token, usbsts;
+	uint32_t endpt, maxpacket, token, usbsts;
 	uint32_t c, toggle;
 	uint32_t cmd;
 	int timeout;
@@ -229,8 +230,73 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 		      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 ((uint32_t)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, 3 * sizeof(*qtd));
+	memset(qtd, 0, qtd_count * sizeof(*qtd));
 
 	toggle = usb_gettoggle(dev, usb_pipeendpoint(pipe), usb_pipeout(pipe));
 
@@ -245,20 +311,18 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 	 * - 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);
+	c = usb_pipespeed(pipe) != 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(usb_pipespeed(pipe)) |
+		QH_ENDPT1_ENDPT(usb_pipeendpoint(pipe)) | QH_ENDPT1_I(0) |
+		QH_ENDPT1_DEVADDR(usb_pipedevice(pipe));
 	qh->qh_endpt1 = cpu_to_hc32(endpt);
-	endpt = (1 << 30) |
-	    (dev->portnr << 23) |
-	    (dev->parent->devnum << 16) | (0 << 8) | (0 << 0);
+	endpt = QH_ENDPT2_MULT(1) | QH_ENDPT2_PORTNUM(dev->portnr) |
+		QH_ENDPT2_HUBADDR(dev->parent->devnum) |
+		QH_ENDPT2_UFCMASK(0) | QH_ENDPT2_UFSMASK(0);
 	qh->qh_endpt2 = cpu_to_hc32(endpt);
 	qh->qh_overlay.qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
 
@@ -276,12 +340,13 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 		 */
 		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);
+		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)) != 0) {
-			printf("unable construct SETUP td\n");
+		if (ehci_td_buffer(&qtd[qtd_counter], req, sizeof(*req))) {
+			printf("unable to construct SETUP TD\n");
 			goto fail;
 		}
 		/* Update previous qTD! */
@@ -291,31 +356,71 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 	}
 
 	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) {
-			printf("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;
+		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 -= (uint32_t)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((uint32_t)&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) {
@@ -328,12 +433,11 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *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) |
-		    (0 << 16) |
-		    (1 << 15) |
-		    (0 << 12) |
-		    (3 << 10) |
-		    ((usb_pipein(pipe) ? 0 : 1) << 8) | (0x80 << 0);
+		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((uint32_t)&qtd[qtd_counter]);
@@ -346,7 +450,8 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 	flush_dcache_range((uint32_t)qh_list,
 		ALIGN_END_ADDR(struct QH, qh_list, 1));
 	flush_dcache_range((uint32_t)qh, ALIGN_END_ADDR(struct QH, qh, 1));
-	flush_dcache_range((uint32_t)qtd, ALIGN_END_ADDR(struct qTD, qtd, 3));
+	flush_dcache_range((uint32_t)qtd,
+			   ALIGN_END_ADDR(struct qTD, qtd, qtd_count));
 
 	/* Set async. queue head pointer. */
 	ehci_writel(&hcor->or_asynclistaddr, (uint32_t)qh_list);
@@ -359,10 +464,10 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 	cmd |= CMD_ASE;
 	ehci_writel(&hcor->or_usbcmd, cmd);
 
-	ret = handshake((uint32_t *)&hcor->or_usbsts, STD_ASS, STD_ASS,
+	ret = handshake((uint32_t *)&hcor->or_usbsts, STS_ASS, STS_ASS,
 			100 * 1000);
 	if (ret < 0) {
-		printf("EHCI fail timeout STD_ASS set\n");
+		printf("EHCI fail timeout STS_ASS set\n");
 		goto fail;
 	}
 
@@ -377,10 +482,10 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 		invalidate_dcache_range((uint32_t)qh,
 			ALIGN_END_ADDR(struct QH, qh, 1));
 		invalidate_dcache_range((uint32_t)qtd,
-			ALIGN_END_ADDR(struct qTD, qtd, 3));
+			ALIGN_END_ADDR(struct qTD, qtd, qtd_count));
 
 		token = hc32_to_cpu(vtd->qt_token);
-		if (!(token & 0x80))
+		if (!(QT_TOKEN_GET_STATUS(token) & QT_TOKEN_STATUS_ACTIVE))
 			break;
 		WATCHDOG_RESET();
 	} while (get_timer(ts) < timeout);
@@ -398,50 +503,50 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 		ALIGN((uint32_t)buffer + length, ARCH_DMA_MINALIGN));
 
 	/* Check that the TD processing happened */
-	if (token & 0x80) {
+	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(&hcor->or_usbcmd);
 	cmd &= ~CMD_ASE;
 	ehci_writel(&hcor->or_usbcmd, cmd);
 
-	ret = handshake((uint32_t *)&hcor->or_usbsts, STD_ASS, 0,
+	ret = handshake((uint32_t *)&hcor->or_usbsts, STS_ASS, 0,
 			100 * 1000);
 	if (ret < 0) {
-		printf("EHCI fail timeout STD_ASS reset\n");
+		printf("EHCI fail timeout STS_ASS reset\n");
 		goto fail;
 	}
 
 	token = hc32_to_cpu(qh->qh_overlay.qt_token);
-	if (!(token & 0x80)) {
+	if (!(QT_TOKEN_GET_STATUS(token) & QT_TOKEN_STATUS_ACTIVE)) {
 		debug("TOKEN=%#x\n", token);
-		switch (token & 0xfc) {
+		switch (QT_TOKEN_GET_STATUS(token) &
+			~(QT_TOKEN_STATUS_SPLITXSTATE | QT_TOKEN_STATUS_PERR)) {
 		case 0:
-			toggle = token >> 31;
+			toggle = QT_TOKEN_GET_DT(token);
 			usb_settoggle(dev, usb_pipeendpoint(pipe),
 				       usb_pipeout(pipe), toggle);
 			dev->status = 0;
 			break;
-		case 0x40:
+		case QT_TOKEN_STATUS_HALTED:
 			dev->status = USB_ST_STALLED;
 			break;
-		case 0xa0:
-		case 0x20:
+		case QT_TOKEN_STATUS_ACTIVE | QT_TOKEN_STATUS_DATBUFERR:
+		case QT_TOKEN_STATUS_DATBUFERR:
 			dev->status = USB_ST_BUF_ERR;
 			break;
-		case 0x50:
-		case 0x10:
+		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 ((token & 0x40) == 0x40)
+			if (QT_TOKEN_GET_STATUS(token) & QT_TOKEN_STATUS_HALTED)
 				dev->status |= USB_ST_STALLED;
 			break;
 		}
-		dev->act_len = length - ((token >> 16) & 0x7fff);
+		dev->act_len = length - QT_TOKEN_GET_TOTALBYTES(token);
 	} else {
 		dev->act_len = 0;
 		debug("dev=%u, usbsts=%#x, p[1]=%#x, p[2]=%#x\n",
@@ -450,9 +555,11 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer,
 		      ehci_readl(&hcor->or_portsc[1]));
 	}
 
+	free(qtd);
 	return (dev->status != USB_ST_NOT_PROC) ? 0 : -1;
 
 fail:
+	free(qtd);
 	return -1;
 }
 
@@ -499,12 +606,14 @@ ehci_submit_root(struct usb_device *dev, unsigned long pipe, void *buffer,
 		case USB_DT_DEVICE:
 			debug("USB_DT_DEVICE request\n");
 			srcptr = &descriptor.device;
-			srclen = 0x12;
+			srclen = descriptor.device.bLength;
 			break;
 		case USB_DT_CONFIG:
 			debug("USB_DT_CONFIG config\n");
 			srcptr = &descriptor.config;
-			srclen = 0x19;
+			srclen = descriptor.config.bLength +
+					descriptor.interface.bLength +
+					descriptor.endpoint.bLength;
 			break;
 		case USB_DT_STRING:
 			debug("USB_DT_STRING config\n");
@@ -539,7 +648,7 @@ ehci_submit_root(struct usb_device *dev, unsigned long pipe, void *buffer,
 		case USB_DT_HUB:
 			debug("USB_DT_HUB config\n");
 			srcptr = &descriptor.hub;
-			srclen = 0x8;
+			srclen = descriptor.hub.bLength;
 			break;
 		default:
 			debug("unknown value %x\n", le16_to_cpu(req->value));
@@ -577,13 +686,13 @@ ehci_submit_root(struct usb_device *dev, unsigned long pipe, void *buffer,
 			tmpbuf[1] |= USB_PORT_STAT_POWER >> 8;
 
 		if (ehci_is_TDI()) {
-			switch ((reg >> 26) & 3) {
-			case 0:
+			switch (PORTSC_PSPD(reg)) {
+			case PORTSC_PSPD_FS:
 				break;
-			case 1:
+			case PORTSC_PSPD_LS:
 				tmpbuf[1] |= USB_PORT_STAT_LOW_SPEED >> 8;
 				break;
-			case 2:
+			case PORTSC_PSPD_HS:
 			default:
 				tmpbuf[1] |= USB_PORT_STAT_HIGH_SPEED >> 8;
 				break;
@@ -729,26 +838,28 @@ int usb_lowlevel_init(void)
 	uint32_t reg;
 	uint32_t cmd;
 
-	if (ehci_hcd_init() != 0)
+	if (ehci_hcd_init())
 		return -1;
 
 	/* EHCI spec section 4.1 */
-	if (ehci_reset() != 0)
+	if (ehci_reset())
 		return -1;
 
 #if defined(CONFIG_EHCI_HCD_INIT_AFTER_RESET)
-	if (ehci_hcd_init() != 0)
+	if (ehci_hcd_init())
 		return -1;
 #endif
 
 	/* Set head of reclaim list */
 	memset(qh_list, 0, sizeof(*qh_list));
 	qh_list->qh_link = cpu_to_hc32((uint32_t)qh_list | QH_LINK_TYPE_QH);
-	qh_list->qh_endpt1 = cpu_to_hc32((1 << 15) | (USB_SPEED_HIGH << 12));
+	qh_list->qh_endpt1 = cpu_to_hc32(QH_ENDPT1_H(1) |
+						QH_ENDPT1_EPS(USB_SPEED_HIGH));
 	qh_list->qh_curtd = cpu_to_hc32(QT_NEXT_TERMINATE);
 	qh_list->qh_overlay.qt_next = cpu_to_hc32(QT_NEXT_TERMINATE);
 	qh_list->qh_overlay.qt_altnext = cpu_to_hc32(QT_NEXT_TERMINATE);
-	qh_list->qh_overlay.qt_token = cpu_to_hc32(0x40);
+	qh_list->qh_overlay.qt_token =
+			cpu_to_hc32(QT_TOKEN_STATUS(QT_TOKEN_STATUS_HALTED));
 
 	reg = ehci_readl(&hccr->cr_hcsparams);
 	descriptor.hub.bNbrPorts = HCS_N_PORTS(reg);
@@ -808,7 +919,7 @@ submit_control_msg(struct usb_device *dev, unsigned long pipe, void *buffer,
 	}
 
 	if (usb_pipedevice(pipe) == rootdev) {
-		if (rootdev == 0)
+		if (!rootdev)
 			dev->speed = USB_SPEED_HIGH;
 		return ehci_submit_root(dev, pipe, buffer, length, setup);
 	}
@@ -819,8 +930,24 @@ int
 submit_int_msg(struct usb_device *dev, unsigned long pipe, void *buffer,
 	       int length, int interval)
 {
-
 	debug("dev=%p, pipe=%lu, buffer=%p, length=%d, interval=%d",
 	      dev, pipe, buffer, length, interval);
+
+	/*
+	 * Interrupt transfers requiring several transactions are not supported
+	 * because bInterval is ignored.
+	 *
+	 * Also, ehci_submit_async() relies on wMaxPacketSize being a power of 2
+	 * <= PKT_ALIGN if several qTDs are required, while the USB
+	 * specification does not constrain this for interrupt transfers. That
+	 * means that ehci_submit_async() would support interrupt transfers
+	 * requiring several transactions only as long as the transfer size does
+	 * not require more than a single qTD.
+	 */
+	if (length > usb_maxpacket(dev, pipe)) {
+		printf("%s: Interrupt transfers requiring several transactions "
+			"are not supported.\n", __func__);
+		return -1;
+	}
 	return ehci_submit_async(dev, pipe, buffer, length, NULL);
 }
diff --git a/drivers/usb/host/ehci.h b/drivers/usb/host/ehci.h
index cc00ce428..39acdf965 100644
--- a/drivers/usb/host/ehci.h
+++ b/drivers/usb/host/ehci.h
@@ -68,7 +68,7 @@ struct ehci_hcor {
 #define CMD_RESET	(1 << 1)		/* reset HC not bus */
 #define CMD_RUN		(1 << 0)		/* start/stop HC */
 	uint32_t or_usbsts;
-#define	STD_ASS		(1 << 15)
+#define STS_ASS		(1 << 15)
 #define STS_HALT	(1 << 12)
 	uint32_t or_usbintr;
 #define INTR_UE         (1 << 0)                /* USB interrupt enable */
@@ -83,11 +83,16 @@ struct ehci_hcor {
 	uint32_t _reserved_0_;
 	uint32_t or_burstsize;
 	uint32_t or_txfilltuning;
+#define TXFIFO_THRESH_MASK		(0x3f << 16)
 #define TXFIFO_THRESH(p)		((p & 0x3f) << 16)
 	uint32_t _reserved_1_[6];
 	uint32_t or_configflag;
 #define FLAG_CF		(1 << 0)	/* true:  we'll support "high speed" */
 	uint32_t or_portsc[CONFIG_SYS_USB_EHCI_MAX_ROOT_PORTS];
+#define PORTSC_PSPD(x)		(((x) >> 26) & 0x3)
+#define PORTSC_PSPD_FS			0x0
+#define PORTSC_PSPD_LS			0x1
+#define PORTSC_PSPD_HS			0x2
 	uint32_t or_systune;
 } __attribute__ ((packed, aligned(4)));
 
@@ -171,16 +176,40 @@ struct usb_linux_config_descriptor {
 /* Queue Element Transfer Descriptor (qTD). */
 struct qTD {
 	/* this part defined by EHCI spec */
-	uint32_t qt_next;		/* see EHCI 3.5.1 */
+	uint32_t qt_next;			/* see EHCI 3.5.1 */
 #define	QT_NEXT_TERMINATE	1
-	uint32_t qt_altnext;		/* see EHCI 3.5.2 */
-	uint32_t qt_token;		/* see EHCI 3.5.3 */
-	uint32_t qt_buffer[5];		/* see EHCI 3.5.4 */
-	uint32_t qt_buffer_hi[5];	/* Appendix B */
+	uint32_t qt_altnext;			/* see EHCI 3.5.2 */
+	uint32_t qt_token;			/* see EHCI 3.5.3 */
+#define QT_TOKEN_DT(x)		(((x) & 0x1) << 31)	/* Data Toggle */
+#define QT_TOKEN_GET_DT(x)		(((x) >> 31) & 0x1)
+#define QT_TOKEN_TOTALBYTES(x)	(((x) & 0x7fff) << 16)	/* Total Bytes to Transfer */
+#define QT_TOKEN_GET_TOTALBYTES(x)	(((x) >> 16) & 0x7fff)
+#define QT_TOKEN_IOC(x)		(((x) & 0x1) << 15)	/* Interrupt On Complete */
+#define QT_TOKEN_CPAGE(x)	(((x) & 0x7) << 12)	/* Current Page */
+#define QT_TOKEN_CERR(x)	(((x) & 0x3) << 10)	/* Error Counter */
+#define QT_TOKEN_PID(x)		(((x) & 0x3) << 8)	/* PID Code */
+#define QT_TOKEN_PID_OUT		0x0
+#define QT_TOKEN_PID_IN			0x1
+#define QT_TOKEN_PID_SETUP		0x2
+#define QT_TOKEN_STATUS(x)	(((x) & 0xff) << 0)	/* Status */
+#define QT_TOKEN_GET_STATUS(x)		(((x) >> 0) & 0xff)
+#define QT_TOKEN_STATUS_ACTIVE		0x80
+#define QT_TOKEN_STATUS_HALTED		0x40
+#define QT_TOKEN_STATUS_DATBUFERR	0x20
+#define QT_TOKEN_STATUS_BABBLEDET	0x10
+#define QT_TOKEN_STATUS_XACTERR		0x08
+#define QT_TOKEN_STATUS_MISSEDUFRAME	0x04
+#define QT_TOKEN_STATUS_SPLITXSTATE	0x02
+#define QT_TOKEN_STATUS_PERR		0x01
+#define QT_BUFFER_CNT		5
+	uint32_t qt_buffer[QT_BUFFER_CNT];	/* see EHCI 3.5.4 */
+	uint32_t qt_buffer_hi[QT_BUFFER_CNT];	/* Appendix B */
 	/* pad struct for 32 byte alignment */
 	uint32_t unused[3];
 };
 
+#define EHCI_PAGE_SIZE		4096
+
 /* Queue Head (QH). */
 struct QH {
 	uint32_t qh_link;
@@ -190,7 +219,26 @@ struct QH {
 #define	QH_LINK_TYPE_SITD	4
 #define	QH_LINK_TYPE_FSTN	6
 	uint32_t qh_endpt1;
+#define QH_ENDPT1_RL(x)		(((x) & 0xf) << 28)	/* NAK Count Reload */
+#define QH_ENDPT1_C(x)		(((x) & 0x1) << 27)	/* Control Endpoint Flag */
+#define QH_ENDPT1_MAXPKTLEN(x)	(((x) & 0x7ff) << 16)	/* Maximum Packet Length */
+#define QH_ENDPT1_H(x)		(((x) & 0x1) << 15)	/* Head of Reclamation List Flag */
+#define QH_ENDPT1_DTC(x)	(((x) & 0x1) << 14)	/* Data Toggle Control */
+#define QH_ENDPT1_DTC_IGNORE_QTD_TD	0x0
+#define QH_ENDPT1_DTC_DT_FROM_QTD	0x1
+#define QH_ENDPT1_EPS(x)	(((x) & 0x3) << 12)	/* Endpoint Speed */
+#define QH_ENDPT1_EPS_FS		0x0
+#define QH_ENDPT1_EPS_LS		0x1
+#define QH_ENDPT1_EPS_HS		0x2
+#define QH_ENDPT1_ENDPT(x)	(((x) & 0xf) << 8)	/* Endpoint Number */
+#define QH_ENDPT1_I(x)		(((x) & 0x1) << 7)	/* Inactivate on Next Transaction */
+#define QH_ENDPT1_DEVADDR(x)	(((x) & 0x7f) << 0)	/* Device Address */
 	uint32_t qh_endpt2;
+#define QH_ENDPT2_MULT(x)	(((x) & 0x3) << 30)	/* High-Bandwidth Pipe Multiplier */
+#define QH_ENDPT2_PORTNUM(x)	(((x) & 0x7f) << 23)	/* Port Number */
+#define QH_ENDPT2_HUBADDR(x)	(((x) & 0x7f) << 16)	/* Hub Address */
+#define QH_ENDPT2_UFCMASK(x)	(((x) & 0xff) << 8)	/* Split Completion Mask */
+#define QH_ENDPT2_UFSMASK(x)	(((x) & 0xff) << 0)	/* Interrupt Schedule Mask */
 	uint32_t qh_curtd;
 	struct qTD qh_overlay;
 	/*
diff --git a/drivers/usb/host/ohci-hcd.c b/drivers/usb/host/ohci-hcd.c
index d24f2f131..9f4735167 100644
--- a/drivers/usb/host/ohci-hcd.c
+++ b/drivers/usb/host/ohci-hcd.c
@@ -1261,19 +1261,11 @@ static int ohci_submit_rh_msg(struct usb_device *dev, unsigned long pipe,
 	int leni = transfer_len;
 	int len = 0;
 	int stat = 0;
-	__u32 datab[4];
-	union {
-		void *ptr;
-		__u8 *u8;
-		__u16 *u16;
-		__u32 *u32;
-	} databuf;
 	__u16 bmRType_bReq;
 	__u16 wValue;
 	__u16 wIndex;
 	__u16 wLength;
-
-	databuf.u32 = (__u32 *)datab;
+	ALLOC_ALIGN_BUFFER(__u8, databuf, 16, sizeof(u32));
 
 #ifdef DEBUG
 pkt_print(NULL, dev, pipe, buffer, transfer_len,
@@ -1304,20 +1296,20 @@ pkt_print(NULL, dev, pipe, buffer, transfer_len,
 	*/
 
 	case RH_GET_STATUS:
-		databuf.u16[0] = cpu_to_le16(1);
+		*(u16 *)databuf = cpu_to_le16(1);
 		OK(2);
 	case RH_GET_STATUS | RH_INTERFACE:
-		databuf.u16[0] = cpu_to_le16(0);
+		*(u16 *)databuf = cpu_to_le16(0);
 		OK(2);
 	case RH_GET_STATUS | RH_ENDPOINT:
-		databuf.u16[0] = cpu_to_le16(0);
+		*(u16 *)databuf = cpu_to_le16(0);
 		OK(2);
 	case RH_GET_STATUS | RH_CLASS:
-		databuf.u32[0] = cpu_to_le32(
+		*(u32 *)databuf = cpu_to_le32(
 				RD_RH_STAT & ~(RH_HS_CRWE | RH_HS_DRWE));
 		OK(4);
 	case RH_GET_STATUS | RH_OTHER | RH_CLASS:
-		databuf.u32[0] = cpu_to_le32(RD_RH_PORTSTAT);
+		*(u32 *)databuf = cpu_to_le32(RD_RH_PORTSTAT);
 		OK(4);
 
 	case RH_CLEAR_FEATURE | RH_ENDPOINT:
@@ -1381,14 +1373,14 @@ pkt_print(NULL, dev, pipe, buffer, transfer_len,
 					min_t(unsigned int,
 					sizeof(root_hub_dev_des),
 					wLength));
-			databuf.ptr = root_hub_dev_des; OK(len);
+			databuf = root_hub_dev_des; OK(len);
 		case (0x02): /* configuration descriptor */
 			len = min_t(unsigned int,
 					leni,
 					min_t(unsigned int,
 					sizeof(root_hub_config_des),
 					wLength));
-			databuf.ptr = root_hub_config_des; OK(len);
+			databuf = root_hub_config_des; OK(len);
 		case (0x03): /* string descriptors */
 			if (wValue == 0x0300) {
 				len = min_t(unsigned int,
@@ -1396,7 +1388,7 @@ pkt_print(NULL, dev, pipe, buffer, transfer_len,
 						min_t(unsigned int,
 						sizeof(root_hub_str_index0),
 						wLength));
-				databuf.ptr = root_hub_str_index0;
+				databuf = root_hub_str_index0;
 				OK(len);
 			}
 			if (wValue == 0x0301) {
@@ -1405,7 +1397,7 @@ pkt_print(NULL, dev, pipe, buffer, transfer_len,
 						min_t(unsigned int,
 						sizeof(root_hub_str_index1),
 						wLength));
-				databuf.ptr = root_hub_str_index1;
+				databuf = root_hub_str_index1;
 				OK(len);
 		}
 		default:
@@ -1417,40 +1409,40 @@ pkt_print(NULL, dev, pipe, buffer, transfer_len,
 	{
 		__u32 temp = roothub_a(&gohci);
 
-		databuf.u8[0] = 9;		/* min length; */
-		databuf.u8[1] = 0x29;
-		databuf.u8[2] = temp & RH_A_NDP;
+		databuf[0] = 9;		/* min length; */
+		databuf[1] = 0x29;
+		databuf[2] = temp & RH_A_NDP;
 #ifdef CONFIG_AT91C_PQFP_UHPBUG
-		databuf.u8[2] = (databuf.u8[2] == 2) ? 1 : 0;
+		databuf[2] = (databuf[2] == 2) ? 1 : 0;
 #endif
-		databuf.u8[3] = 0;
+		databuf[3] = 0;
 		if (temp & RH_A_PSM)	/* per-port power switching? */
-			databuf.u8[3] |= 0x1;
+			databuf[3] |= 0x1;
 		if (temp & RH_A_NOCP)	/* no overcurrent reporting? */
-			databuf.u8[3] |= 0x10;
+			databuf[3] |= 0x10;
 		else if (temp & RH_A_OCPM)/* per-port overcurrent reporting? */
-			databuf.u8[3] |= 0x8;
+			databuf[3] |= 0x8;
 
-		/* corresponds to databuf.u8[4-7] */
-		databuf.u8[1] = 0;
-		databuf.u8[5] = (temp & RH_A_POTPGT) >> 24;
+		databuf[4] = 0;
+		databuf[5] = (temp & RH_A_POTPGT) >> 24;
+		databuf[6] = 0;
 		temp = roothub_b(&gohci);
-		databuf.u8[7] = temp & RH_B_DR;
-		if (databuf.u8[2] < 7) {
-			databuf.u8[8] = 0xff;
+		databuf[7] = temp & RH_B_DR;
+		if (databuf[2] < 7) {
+			databuf[8] = 0xff;
 		} else {
-			databuf.u8[0] += 2;
-			databuf.u8[8] = (temp & RH_B_DR) >> 8;
-			databuf.u8[10] = databuf.u8[9] = 0xff;
+			databuf[0] += 2;
+			databuf[8] = (temp & RH_B_DR) >> 8;
+			databuf[10] = databuf[9] = 0xff;
 		}
 
 		len = min_t(unsigned int, leni,
-			    min_t(unsigned int, databuf.u8[0], wLength));
+			    min_t(unsigned int, databuf[0], wLength));
 		OK(len);
 	}
 
 	case RH_GET_CONFIGURATION:
-		databuf.u8[0] = 0x01;
+		databuf[0] = 0x01;
 		OK(1);
 
 	case RH_SET_CONFIGURATION:
@@ -1469,8 +1461,8 @@ pkt_print(NULL, dev, pipe, buffer, transfer_len,
 #endif
 
 	len = min_t(int, len, leni);
-	if (data != databuf.ptr)
-		memcpy(data, databuf.ptr, len);
+	if (data != databuf)
+		memcpy(data, databuf, len);
 	dev->act_len = len;
 	dev->status = stat;