Skip to contentSkip to conversation

xhci: No XHCI_TRUST_TX_LENGTH check in the absence of matching TD

PATCH In review [email protected] 2 participants 3 messages Nov 13, 2019 – Nov 19, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 xhci: No XHCI_TRUST_TX_LENGTH check in the absence of matching TD

From: Eli Billauer <[email protected]>

When an IN transfer ends with a short packet, the xHCI controller is required to submit an event TRB with Completion Code COMP_SHORT_PACKET against the data TRB that was in effect when the short packet arrived, as well as any event TRBs it submits on behalf of this transfer.

Alas, some controllers (e.g. Renesas) mark the subsequent events TRBs (if any) with COMP_SUCCESS. As these subsequent event TRBs are useless, they are ignored on the basis that they have no matching TD queued (it was dequeued in response to the first COMP_SHORT_PACKET event TRB).

Accordingly, the quirk handling and kernel log warning is moved to after the TD match check, in particular in order to avoid unnecessary warnings messages.

---
 drivers/usb/host/xhci-ring.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)
drivers/usb/host/xhci-ring.c +11 -8
index 9741cdeea9d7..96680eb71a45 100644
@@ -2376,14 +2376,6 @@ static int handle_tx_event(struct xhci_hcd *xhci,
 	 * transfer type
 	 */
 	case COMP_SUCCESS:
-		if (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) == 0)
-			break;
-		if (xhci->quirks & XHCI_TRUST_TX_LENGTH)
-			trb_comp_code = COMP_SHORT_PACKET;
-		else
-			xhci_warn_ratelimited(xhci,
-					      "WARN Successful completion on short TX for slot %u ep %u: needs XHCI_TRUST_TX_LENGTH quirk?\n",
-					      slot_id, ep_index);
 	case COMP_SHORT_PACKET:
 		break;
 	/* Completion codes for endpoint stopped state */
@@ -2586,6 +2578,17 @@ static int handle_tx_event(struct xhci_hcd *xhci,
 			skip_isoc_td(xhci, td, event, ep, &status);
 			goto cleanup;
 		}
+
+		if ((trb_comp_code == COMP_SUCCESS) &&
+		    (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) != 0)) {
+			if (xhci->quirks & XHCI_TRUST_TX_LENGTH)
+				trb_comp_code = COMP_SHORT_PACKET;
+			else
+				xhci_warn_ratelimited(xhci,
+						      "WARN Successful completion on short TX for slot %u ep %u: needs XHCI_TRUST_TX_LENGTH quirk?\n",
+						      slot_id, ep_index);
+		}
+
 		if (trb_comp_code == COMP_SHORT_PACKET)
 			ep_ring->last_td_was_short = true;
 		else
Signature · 1 line
2.17.1
#1 lore raw
Quoted 49 lines · On 13.11.2019 15.06, [email protected] wrote:

From: Eli Billauer <[email protected]>

When an IN transfer ends with a short packet, the xHCI controller is required to submit an event TRB with Completion Code COMP_SHORT_PACKET against the data TRB that was in effect when the short packet arrived, as well as any event TRBs it submits on behalf of this transfer.

Alas, some controllers (e.g. Renesas) mark the subsequent events TRBs (if any) with COMP_SUCCESS. As these subsequent event TRBs are useless, they are ignored on the basis that they have no matching TD queued (it was dequeued in response to the first COMP_SHORT_PACKET event TRB).

Accordingly, the quirk handling and kernel log warning is moved to after the TD match check, in particular in order to avoid unnecessary warnings messages.

---
  drivers/usb/host/xhci-ring.c | 19 +++++++++++--------
  1 file changed, 11 insertions(+), 8 deletions(-)
drivers/usb/host/xhci-ring.c +4 -8
index 9741cdeea9d7..96680eb71a45 100644
@@ -2376,14 +2376,6 @@ static int handle_tx_event(struct xhci_hcd *xhci,
  	 * transfer type
  	 */
  	case COMP_SUCCESS:
-		if (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) == 0)
-			break;
-		if (xhci->quirks & XHCI_TRUST_TX_LENGTH)
-			trb_comp_code = COMP_SHORT_PACKET;
-		else
-			xhci_warn_ratelimited(xhci,
-					      "WARN Successful completion on short TX for slot %u ep %u: needs XHCI_TRUST_TX_LENGTH quirk?\n",
-					      slot_id, ep_index);
  	case COMP_SHORT_PACKET:
  		break;
  	/* Completion codes for endpoint stopped state */
@@ -2586,6 +2578,17 @@ static int handle_tx_event(struct xhci_hcd *xhci,
  			skip_isoc_td(xhci, td, event, ep, &status);
  			goto cleanup;
  		}
+
+		if ((trb_comp_code == COMP_SUCCESS) &&
+		    (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) != 0)) {
+			if (xhci->quirks & XHCI_TRUST_TX_LENGTH)

drivers/usb/host/xhci-ring.c static int handle_tx_event(struct xhci_hcd *xhci, · line 2594

Quoted 10 lines
drivers/usb/host/xhci-ring.c +7 -0
+				trb_comp_code = COMP_SHORT_PACKET;
+			else
+				xhci_warn_ratelimited(xhci,
+						      "WARN Successful completion on short TX for slot %u ep %u: needs XHCI_TRUST_TX_LENGTH quirk?\n",
+						      slot_id, ep_index);
+		}
+
  		if (trb_comp_code == COMP_SHORT_PACKET)
  			ep_ring->last_td_was_short = true;
  		else

I'd hate to rip out the success case from the switch statement where the other completion codes are handled. We're still only making a choice about a warning message

How about handling all COMP_SUCCESS cases with remaining data after a short transfer as COMP_SHORT_PACKET by default?

The code below won't behave exactly the same as in your patch, but should do the trick in your Renesas case as well. Can you try it out?

drivers/usb/host/xhci-ring.c +2 -1
index 9ebaa8e132a9..d23f7408c81f 100644
@@ -2381,7 +2381,8 @@ static int handle_tx_event(struct xhci_hcd *xhci,
        case COMP_SUCCESS:
                if (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) == 0)
                        break;
-               if (xhci->quirks & XHCI_TRUST_TX_LENGTH)
+               if (xhci->quirks & XHCI_TRUST_TX_LENGTH ||
+                   ep_ring->last_td_was_short)
                        trb_comp_code = COMP_SHORT_PACKET;
                else
                        xhci_warn_ratelimited(xhci,
#2 lore raw

Hello,

I've taken the liberty to add parentheses for sake of clarity, so I tested the following patch (against kernel 5.3.0) with my Resesas USB controller. And as expected, it does the job: The warnings are not printed.

drivers/usb/host/xhci-ring.c +2 -1
index 9741cdeea9d7..b062e3a19e95 100644
@@ -2378,7 +2378,8 @@ static int handle_tx_event(struct xhci_hcd *xhci,
     case COMP_SUCCESS:
         if (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) == 0)
             break;
-        if (xhci->quirks & XHCI_TRUST_TX_LENGTH)
+        if ((xhci->quirks & XHCI_TRUST_TX_LENGTH) ||
+            ep_ring->last_td_was_short)
             trb_comp_code = COMP_SHORT_PACKET;
         else
             xhci_warn_ratelimited(xhci,

So it solves the problem for me, and hopefully for future Renesas controllers with the same issue.

Regards,
Eli

Quoted 89 lines · Mathias Nyman
Quoted 59 lines · On 13.11.2019 15.06, [email protected] wrote:

From: Eli Billauer <[email protected]>

When an IN transfer ends with a short packet, the xHCI controller is required to submit an event TRB with Completion Code COMP_SHORT_PACKET against the data TRB that was in effect when the short packet arrived, as well as any event TRBs it submits on behalf of this transfer.

Alas, some controllers (e.g. Renesas) mark the subsequent events TRBs (if any) with COMP_SUCCESS. As these subsequent event TRBs are useless, they are ignored on the basis that they have no matching TD queued (it was dequeued in response to the first COMP_SHORT_PACKET event TRB).

Accordingly, the quirk handling and kernel log warning is moved to after the TD match check, in particular in order to avoid unnecessary warnings messages.

---
  drivers/usb/host/xhci-ring.c | 19 +++++++++++--------
  1 file changed, 11 insertions(+), 8 deletions(-)
drivers/usb/host/xhci-ring.c +0 -8
index 9741cdeea9d7..96680eb71a45 100644
@@ -2376,14 +2376,6 @@ static int handle_tx_event(struct xhci_hcd *xhci,
       * transfer type
       */
      case COMP_SUCCESS:
-        if (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) == 0)
-            break;
-        if (xhci->quirks & XHCI_TRUST_TX_LENGTH)
-            trb_comp_code = COMP_SHORT_PACKET;
-        else
-            xhci_warn_ratelimited(xhci,
-                          "WARN Successful completion on short TX 
 for slot %u ep %u: needs XHCI_TRUST_TX_LENGTH quirk?\n",
-                          slot_id, ep_index);
      case COMP_SHORT_PACKET:
          break;
      /* Completion codes for endpoint stopped state */
drivers/usb/host/xhci-ring.c +11 -0
@@ -2586,6 +2578,17 @@ static int handle_tx_event(struct xhci_hcd *xhci,
              skip_isoc_td(xhci, td, event, ep, &status);
              goto cleanup;
          }
+
+        if ((trb_comp_code == COMP_SUCCESS) &&
+            (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) != 0)) {
+            if (xhci->quirks & XHCI_TRUST_TX_LENGTH)
+                trb_comp_code = COMP_SHORT_PACKET;
+            else
+                xhci_warn_ratelimited(xhci,
+                              "WARN Successful completion on short 
 TX for slot %u ep %u: needs XHCI_TRUST_TX_LENGTH quirk?\n",
+                              slot_id, ep_index);
+        }
+
          if (trb_comp_code == COMP_SHORT_PACKET)
              ep_ring->last_td_was_short = true;
Quoted 2 lines

else

I'd hate to rip out the success case from the switch statement where the other completion codes are handled. We're still only making a choice about a warning message

How about handling all COMP_SUCCESS cases with remaining data after a short transfer as COMP_SHORT_PACKET by default?

The code below won't behave exactly the same as in your patch, but should do the trick in your Renesas case as well. Can you try it out?

drivers/usb/host/xhci-ring.c +2 -1
index 9ebaa8e132a9..d23f7408c81f 100644
@@ -2381,7 +2381,8 @@ static int handle_tx_event(struct xhci_hcd *xhci,
        case COMP_SUCCESS:
                if (EVENT_TRB_LEN(le32_to_cpu(event->transfer_len)) == 0)
                        break;
-               if (xhci->quirks & XHCI_TRUST_TX_LENGTH)
+               if (xhci->quirks & XHCI_TRUST_TX_LENGTH ||
+                   ep_ring->last_td_was_short)
                        trb_comp_code = COMP_SHORT_PACKET;
                else
                        xhci_warn_ratelimited(xhci,