usb: dwc3: gadget: fix scatter gather implementation

To work with scatter gather properly, fixes have been done in number of
functions. I will explain requirement of each fixes one by one.

start_slot: used to retrieve all request of SG during cleanup

dwc3_gadget_giveback: We need to skip link TRB if it was one of the
intermediate TRB of SG.

dwc3_prepare_one_trb: We need to track all submitted TRBs during
cleanup. Since, all TRBs would be serially allocated, so we can just
keep starting slot info and we can always find rest of them. We need to
pass sg node number, so that we cab appropriately program ISOC_FIRST/ISOC,
Chain etc.

dwc3_prepare_trbs: last_one should be set when it is last node
of SG as well as last node of request_list.

__dwc3_cleanup_done_trbs: It has been prepared after re-factorization of
dwc3_cleanup_done_reqs. It is called for each TRB of SG.

Signed-off-by: Pratyush Anand <pratyush.anand@st.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>

drivers/usb/dwc3/core.h

@@ -581,6 +581,7 @@ struct dwc3_request {
 	struct usb_request	request;
 	struct list_head	list;
 	struct dwc3_ep		*dep;
+	u32			start_slot;
 
 	u8			epnum;
 	struct dwc3_trb		*trb;

drivers/usb/dwc3/gadget.c

@@ -241,21 +241,22 @@ void dwc3_gadget_giveback(struct dwc3_ep *dep, struct dwc3_request *req,
 		int status)
 {
 	struct dwc3			*dwc = dep->dwc;
+	int				i;
 
 	if (req->queued) {
-		if (req->request.num_mapped_sgs)
-			dep->busy_slot += req->request.num_mapped_sgs;
-		else
+		i = 0;
+		do {
 			dep->busy_slot++;
-
-		/*
-		 * Skip LINK TRB. We can't use req->trb and check for
-		 * DWC3_TRBCTL_LINK_TRB because it points the TRB we just
-		 * completed (not the LINK TRB).
-		 */
-		if (((dep->busy_slot & DWC3_TRB_MASK) == DWC3_TRB_NUM - 1) &&
+			/*
+			 * Skip LINK TRB. We can't use req->trb and check for
+			 * DWC3_TRBCTL_LINK_TRB because it points the TRB we
+			 * just completed (not the LINK TRB).
+			 */
+			if (((dep->busy_slot & DWC3_TRB_MASK) ==
+				DWC3_TRB_NUM- 1) &&
 				usb_endpoint_xfer_isoc(dep->endpoint.desc))
-			dep->busy_slot++;
+				dep->busy_slot++;
+		} while(++i < req->request.num_mapped_sgs);
 	}
 	list_del(&req->list);
 	req->trb = NULL;
@@ -749,7 +750,7 @@ static void dwc3_gadget_ep_free_request(struct usb_ep *ep,
  */
 static void dwc3_prepare_one_trb(struct dwc3_ep *dep,
 		struct dwc3_request *req, dma_addr_t dma,
-		unsigned length, unsigned last, unsigned chain)
+		unsigned length, unsigned last, unsigned chain, unsigned node)
 {
 	struct dwc3		*dwc = dep->dwc;
 	struct dwc3_trb		*trb;
@@ -765,14 +766,16 @@ static void dwc3_prepare_one_trb(struct dwc3_ep *dep,
 		dep->free_slot++;
 
 	trb = &dep->trb_pool[dep->free_slot & DWC3_TRB_MASK];
-	dep->free_slot++;
 
 	if (!req->trb) {
 		dwc3_gadget_move_request_queued(req);
 		req->trb = trb;
 		req->trb_dma = dwc3_trb_dma_offset(dep, trb);
+		req->start_slot = dep->free_slot & DWC3_TRB_MASK;
 	}
 
+	dep->free_slot++;
+
 	trb->size = DWC3_TRB_SIZE_LENGTH(length);
 	trb->bpl = lower_32_bits(dma);
 	trb->bph = upper_32_bits(dma);
@@ -783,9 +786,12 @@ static void dwc3_prepare_one_trb(struct dwc3_ep *dep,
 		break;
 
 	case USB_ENDPOINT_XFER_ISOC:
-		trb->ctrl = DWC3_TRBCTL_ISOCHRONOUS_FIRST;
+		if (!node)
+			trb->ctrl = DWC3_TRBCTL_ISOCHRONOUS_FIRST;
+		else
+			trb->ctrl = DWC3_TRBCTL_ISOCHRONOUS;
 
-		if (!req->request.no_interrupt)
+		if (!req->request.no_interrupt && !chain)
 			trb->ctrl |= DWC3_TRB_CTRL_IOC;
 		break;
 
@@ -804,14 +810,13 @@ static void dwc3_prepare_one_trb(struct dwc3_ep *dep,
 	if (usb_endpoint_xfer_isoc(dep->endpoint.desc)) {
 		trb->ctrl |= DWC3_TRB_CTRL_ISP_IMI;
 		trb->ctrl |= DWC3_TRB_CTRL_CSP;
-	} else {
-		if (chain)
-			trb->ctrl |= DWC3_TRB_CTRL_CHN;
-
-		if (last)
-			trb->ctrl |= DWC3_TRB_CTRL_LST;
+	} else if (last) {
+		trb->ctrl |= DWC3_TRB_CTRL_LST;
 	}
 
+	if (chain)
+		trb->ctrl |= DWC3_TRB_CTRL_CHN;
+
 	if (usb_endpoint_xfer_bulk(dep->endpoint.desc) && dep->stream_capable)
 		trb->ctrl |= DWC3_TRB_CTRL_SID_SOFN(req->request.stream_id);
 
@@ -882,6 +887,7 @@ static void dwc3_prepare_trbs(struct dwc3_ep *dep, bool starting)
 	list_for_each_entry_safe(req, n, &dep->request_list, list) {
 		unsigned	length;
 		dma_addr_t	dma;
+		last_one = false;
 
 		if (req->request.num_mapped_sgs > 0) {
 			struct usb_request *request = &req->request;
@@ -897,7 +903,9 @@ static void dwc3_prepare_trbs(struct dwc3_ep *dep, bool starting)
 
 				if (i == (request->num_mapped_sgs - 1) ||
 						sg_is_last(s)) {
-					last_one = true;
+					if (list_is_last(&req->list,
+							&dep->request_list))
+						last_one = true;
 					chain = false;
 				}
 
@@ -909,7 +917,7 @@ static void dwc3_prepare_trbs(struct dwc3_ep *dep, bool starting)
 					chain = false;
 
 				dwc3_prepare_one_trb(dep, req, dma, length,
-						last_one, chain);
+						last_one, chain, i);
 
 				if (last_one)
 					break;
@@ -927,7 +935,7 @@ static void dwc3_prepare_trbs(struct dwc3_ep *dep, bool starting)
 				last_one = 1;
 
 			dwc3_prepare_one_trb(dep, req, dma, length,
-					last_one, false);
+					last_one, false, 0);
 
 			if (last_one)
 				break;
@@ -1642,89 +1650,115 @@ static void dwc3_gadget_release(struct device *dev)
 }
 
 /* -------------------------------------------------------------------------- */
-static int dwc3_cleanup_done_reqs(struct dwc3 *dwc, struct dwc3_ep *dep,
+static int __dwc3_cleanup_done_trbs(struct dwc3 *dwc, struct dwc3_ep *dep,
+		struct dwc3_request *req, struct dwc3_trb *trb,
 		const struct dwc3_event_depevt *event, int status)
 {
-	struct dwc3_request	*req;
-	struct dwc3_trb		*trb;
 	unsigned int		count;
 	unsigned int		s_pkt = 0;
 	unsigned int		trb_status;
 
+	if ((trb->ctrl & DWC3_TRB_CTRL_HWO) && status != -ESHUTDOWN)
+		/*
+		 * We continue despite the error. There is not much we
+		 * can do. If we don't clean it up we loop forever. If
+		 * we skip the TRB then it gets overwritten after a
+		 * while since we use them in a ring buffer. A BUG()
+		 * would help. Lets hope that if this occurs, someone
+		 * fixes the root cause instead of looking away :)
+		 */
+		dev_err(dwc->dev, "%s's TRB (%p) still owned by HW\n",
+				dep->name, trb);
+	count = trb->size & DWC3_TRB_SIZE_MASK;
+
+	if (dep->direction) {
+		if (count) {
+			trb_status = DWC3_TRB_SIZE_TRBSTS(trb->size);
+			if (trb_status == DWC3_TRBSTS_MISSED_ISOC) {
+				dev_dbg(dwc->dev, "incomplete IN transfer %s\n",
+						dep->name);
+				/*
+				 * If missed isoc occurred and there is
+				 * no request queued then issue END
+				 * TRANSFER, so that core generates
+				 * next xfernotready and we will issue
+				 * a fresh START TRANSFER.
+				 * If there are still queued request
+				 * then wait, do not issue either END
+				 * or UPDATE TRANSFER, just attach next
+				 * request in request_list during
+				 * giveback.If any future queued request
+				 * is successfully transferred then we
+				 * will issue UPDATE TRANSFER for all
+				 * request in the request_list.
+				 */
+				dep->flags |= DWC3_EP_MISSED_ISOC;
+			} else {
+				dev_err(dwc->dev, "incomplete IN transfer %s\n",
+						dep->name);
+				status = -ECONNRESET;
+			}
+		} else {
+			dep->flags &= ~DWC3_EP_MISSED_ISOC;
+		}
+	} else {
+		if (count && (event->status & DEPEVT_STATUS_SHORT))
+			s_pkt = 1;
+	}
+
+	/*
+	 * We assume here we will always receive the entire data block
+	 * which we should receive. Meaning, if we program RX to
+	 * receive 4K but we receive only 2K, we assume that's all we
+	 * should receive and we simply bounce the request back to the
+	 * gadget driver for further processing.
+	 */
+	req->request.actual += req->request.length - count;
+	if (s_pkt)
+		return 1;
+	if ((event->status & DEPEVT_STATUS_LST) &&
+			(trb->ctrl & (DWC3_TRB_CTRL_LST |
+				DWC3_TRB_CTRL_HWO)))
+		return 1;
+	if ((event->status & DEPEVT_STATUS_IOC) &&
+			(trb->ctrl & DWC3_TRB_CTRL_IOC))
+		return 1;
+	return 0;
+}
+
+static int dwc3_cleanup_done_reqs(struct dwc3 *dwc, struct dwc3_ep *dep,
+		const struct dwc3_event_depevt *event, int status)
+{
+	struct dwc3_request	*req;
+	struct dwc3_trb		*trb;
+	unsigned int		slot;
+	unsigned int		i;
+	int			ret;
+
 	do {
 		req = next_request(&dep->req_queued);
 		if (!req) {
 			WARN_ON_ONCE(1);
 			return 1;
 		}
+		i = 0;
+		do {
+			slot = req->start_slot + i;
+			if ((slot == DWC3_TRB_NUM - 1) &&
+				usb_endpoint_xfer_isoc(dep->endpoint.desc))
+				slot++;
+			slot %= DWC3_TRB_NUM;
+			trb = &dep->trb_pool[slot];
 
-		trb = req->trb;
-
-		if ((trb->ctrl & DWC3_TRB_CTRL_HWO) && status != -ESHUTDOWN)
-			/*
-			 * We continue despite the error. There is not much we
-			 * can do. If we don't clean it up we loop forever. If
-			 * we skip the TRB then it gets overwritten after a
-			 * while since we use them in a ring buffer. A BUG()
-			 * would help. Lets hope that if this occurs, someone
-			 * fixes the root cause instead of looking away :)
-			 */
-			dev_err(dwc->dev, "%s's TRB (%p) still owned by HW\n",
-					dep->name, req->trb);
-		count = trb->size & DWC3_TRB_SIZE_MASK;
-
-		if (dep->direction) {
-			if (count) {
-				trb_status = DWC3_TRB_SIZE_TRBSTS(trb->size);
-				if (trb_status == DWC3_TRBSTS_MISSED_ISOC) {
-					dev_dbg(dwc->dev, "incomplete IN transfer %s\n",
-							dep->name);
-					/*
-					 * If missed isoc occurred and there is
-					 * no request queued then issue END
-					 * TRANSFER, so that core generates
-					 * next xfernotready and we will issue
-					 * a fresh START TRANSFER.
-					 * If there are still queued request
-					 * then wait, do not issue either END
-					 * or UPDATE TRANSFER, just attach next
-					 * request in request_list during
-					 * giveback.If any future queued request
-					 * is successfully transferred then we
-					 * will issue UPDATE TRANSFER for all
-					 * request in the request_list.
-					 */
-					dep->flags |= DWC3_EP_MISSED_ISOC;
-				} else {
-					dev_err(dwc->dev, "incomplete IN transfer %s\n",
-							dep->name);
-					status = -ECONNRESET;
-				}
-			} else {
-				dep->flags &= ~DWC3_EP_MISSED_ISOC;
-			}
-		} else {
-			if (count && (event->status & DEPEVT_STATUS_SHORT))
-				s_pkt = 1;
-		}
+			ret = __dwc3_cleanup_done_trbs(dwc, dep, req, trb,
+					event, status);
+			if (ret)
+				break;
+		}while (++i < req->request.num_mapped_sgs);
 
-		/*
-		 * We assume here we will always receive the entire data block
-		 * which we should receive. Meaning, if we program RX to
-		 * receive 4K but we receive only 2K, we assume that's all we
-		 * should receive and we simply bounce the request back to the
-		 * gadget driver for further processing.
-		 */
-		req->request.actual += req->request.length - count;
 		dwc3_gadget_giveback(dep, req, status);
-		if (s_pkt)
-			break;
-		if ((event->status & DEPEVT_STATUS_LST) &&
-				(trb->ctrl & (DWC3_TRB_CTRL_LST |
-						DWC3_TRB_CTRL_HWO)))
-			break;
-		if ((event->status & DEPEVT_STATUS_IOC) &&
-				(trb->ctrl & DWC3_TRB_CTRL_IOC))
+
+		if (ret)
 			break;
 	} while (1);