[PATCH 0/2] xhci fixes for usb-linus

[PATCH 0/2] xhci fixes for usb-linus

[Mathias Nyman]

Hi Greg

two patches for usb-linus and 5.12 stable
xhci changes in 5.12 caused a regression in stall handling.
Due to this some usb card readers failed to work with 5.12

These two patches fix that regression.

Thanks
-Mathias

Mathias Nyman (2):
  xhci: fix giving back URB with incorrect status regression in 5.12
  xhci: Fix 5.12 regression of missing xHC cache clearing command after
    a Stall

 drivers/usb/host/xhci-ring.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

-- 
2.25.1

[PATCH 1/2] xhci: fix giving back URB with incorrect status regression in 5.12

[Mathias Nyman]

5.12 kernel changes how xhci handles cancelled URBs and halted
endpoints. Among these changes cancelled and stalled URBs are no longer
given back before they are cleared from xHC hardware cache.

These changes unfortunately cleared the -EPIPE status of a stalled
transfer in one case before giving bak the URB, causing a USB card reader
to fail from working.

Reported-by: Peter Ganzhorn <peter.ganzhorn@googlemail.com>
Tested-by: Peter Ganzhorn <peter.ganzhorn@googlemail.com>
Fixes: 674f8438c121 ("xhci: split handling halted endpoints into two steps")
Cc: <stable@vger.kernel.org> # 5.12
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
---
 drivers/usb/host/xhci-ring.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/drivers/usb/host/xhci-ring.c b/drivers/usb/host/xhci-ring.c
index a8e4189277da..256d336354a0 100644
--- a/drivers/usb/host/xhci-ring.c
+++ b/drivers/usb/host/xhci-ring.c
@@ -828,14 +828,10 @@ static void xhci_giveback_invalidated_tds(struct xhci_virt_ep *ep)
 	list_for_each_entry_safe(td, tmp_td, &ep->cancelled_td_list,
 				 cancelled_td_list) {
 
-		/*
-		 * Doesn't matter what we pass for status, since the core will
-		 * just overwrite it (because the URB has been unlinked).
-		 */
 		ring = xhci_urb_to_transfer_ring(ep->xhci, td->urb);
 
 		if (td->cancel_status == TD_CLEARED)
-			xhci_td_cleanup(ep->xhci, td, ring, 0);
+			xhci_td_cleanup(ep->xhci, td, ring, td->status);
 
 		if (ep->xhci->xhc_state & XHCI_STATE_DYING)
 			return;
-- 
2.25.1

[PATCH 2/2] xhci: Fix 5.12 regression of missing xHC cache clearing command after a Stall

[Mathias Nyman]

If endpoints halts due to a stall then the dequeue pointer read from
hardware may already be set ahead of the stalled TRB.
After commit 674f8438c121 ("xhci: split handling halted endpoints into two
steps") in 5.12 xhci driver won't issue a Set TR Dequeue if hardware
dequeue pointer is already in the right place.

Turns out the "Set TR Dequeue pointer" command is anyway needed as it in
addition to moving the dequeue pointer also clears endpoint state and
cache.

Reported-by: Peter Ganzhorn <peter.ganzhorn@googlemail.com>
Tested-by: Peter Ganzhorn <peter.ganzhorn@googlemail.com>
Fixes: 674f8438c121 ("xhci: split handling halted endpoints into two steps")
Cc: <stable@vger.kernel.org> # 5.12
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
---
 drivers/usb/host/xhci-ring.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/usb/host/xhci-ring.c b/drivers/usb/host/xhci-ring.c
index 256d336354a0..6acd2329e08d 100644
--- a/drivers/usb/host/xhci-ring.c
+++ b/drivers/usb/host/xhci-ring.c
@@ -933,14 +933,18 @@ static int xhci_invalidate_cancelled_tds(struct xhci_virt_ep *ep)
 			continue;
 		}
 		/*
-		 * If ring stopped on the TD we need to cancel, then we have to
+		 * If a ring stopped on the TD we need to cancel then we have to
 		 * move the xHC endpoint ring dequeue pointer past this TD.
+		 * Rings halted due to STALL may show hw_deq is past the stalled
+		 * TD, but still require a set TR Deq command to flush xHC cache.
 		 */
 		hw_deq = xhci_get_hw_deq(xhci, ep->vdev, ep->ep_index,
 					 td->urb->stream_id);
 		hw_deq &= ~0xf;
 
-		if (trb_in_td(xhci, td->start_seg, td->first_trb,
+		if (td->cancel_status == TD_HALTED) {
+			cached_td = td;
+		} else if (trb_in_td(xhci, td->start_seg, td->first_trb,
 			      td->last_trb, hw_deq, false)) {
 			switch (td->cancel_status) {
 			case TD_CLEARED: /* TD is already no-op */
-- 
2.25.1