xHCI: AMD isoc link TRB chain bit quirk

Setting the chain (CH) bit in the link TRB of isochronous transfer rings
is required by AMD 0.96 xHCI host controller to successfully transverse
multi-TRB TD that span through different memory segments.

When a Missed Service Error event occurs, if the chain bit is not set in
the link TRB and the host skips TDs which just across a link TRB, the
host may falsely recognize the link TRB as a normal TRB. You can see
this may cause big trouble - the host does not jump to the right address
which is pointed by the link TRB, but continue fetching the memory which
is after the link TRB address, which may not even belong to the host,
and the result cannot be predicted.

This causes some big problems. Without the former patch I sent: "xHCI:
prevent infinite loop when processing MSE event", the system may hang.
With that patch applied, system does not hang, but the host still access
wrong memory address and isoc transfer will fail. With this patch,
isochronous transfer works as expected.

This patch should be applied to kernels as old as 2.6.36, which was when
the first isochronous support was added for the xHCI host controller.

Signed-off-by: Andiry Xu <andiry.xu@amd.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Cc: stable@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

drivers/usb/host/xhci-mem.c

@@ -79,7 +79,7 @@ static void xhci_segment_free(struct xhci_hcd *xhci, struct xhci_segment *seg)
  * related flags, such as End TRB, Toggle Cycle, and no snoop.
  */
 static void xhci_link_segments(struct xhci_hcd *xhci, struct xhci_segment *prev,
-		struct xhci_segment *next, bool link_trbs)
+		struct xhci_segment *next, bool link_trbs, bool isoc)
 {
 	u32 val;
 
@@ -95,7 +95,9 @@ static void xhci_link_segments(struct xhci_hcd *xhci, struct xhci_segment *prev,
 		val &= ~TRB_TYPE_BITMASK;
 		val |= TRB_TYPE(TRB_LINK);
 		/* Always set the chain bit with 0.95 hardware */
-		if (xhci_link_trb_quirk(xhci))
+		/* Set chain bit for isoc rings on AMD 0.96 host */
+		if (xhci_link_trb_quirk(xhci) ||
+				(isoc && (xhci->quirks & XHCI_AMD_0x96_HOST)))
 			val |= TRB_CHAIN;
 		prev->trbs[TRBS_PER_SEGMENT-1].link.control = cpu_to_le32(val);
 	}
@@ -152,7 +154,7 @@ static void xhci_initialize_ring_info(struct xhci_ring *ring)
  * See section 4.9.1 and figures 15 and 16.
  */
 static struct xhci_ring *xhci_ring_alloc(struct xhci_hcd *xhci,
-		unsigned int num_segs, bool link_trbs, gfp_t flags)
+		unsigned int num_segs, bool link_trbs, bool isoc, gfp_t flags)
 {
 	struct xhci_ring	*ring;
 	struct xhci_segment	*prev;
@@ -178,12 +180,12 @@ static struct xhci_ring *xhci_ring_alloc(struct xhci_hcd *xhci,
 		next = xhci_segment_alloc(xhci, flags);
 		if (!next)
 			goto fail;
-		xhci_link_segments(xhci, prev, next, link_trbs);
+		xhci_link_segments(xhci, prev, next, link_trbs, isoc);
 
 		prev = next;
 		num_segs--;
 	}
-	xhci_link_segments(xhci, prev, ring->first_seg, link_trbs);
+	xhci_link_segments(xhci, prev, ring->first_seg, link_trbs, isoc);
 
 	if (link_trbs) {
 		/* See section 4.9.2.1 and 6.4.4.1 */
@@ -229,14 +231,14 @@ void xhci_free_or_cache_endpoint_ring(struct xhci_hcd *xhci,
  * pointers to the beginning of the ring.
  */
 static void xhci_reinit_cached_ring(struct xhci_hcd *xhci,
-		struct xhci_ring *ring)
+		struct xhci_ring *ring, bool isoc)
 {
 	struct xhci_segment	*seg = ring->first_seg;
 	do {
 		memset(seg->trbs, 0,
 				sizeof(union xhci_trb)*TRBS_PER_SEGMENT);
 		/* All endpoint rings have link TRBs */
-		xhci_link_segments(xhci, seg, seg->next, 1);
+		xhci_link_segments(xhci, seg, seg->next, 1, isoc);
 		seg = seg->next;
 	} while (seg != ring->first_seg);
 	xhci_initialize_ring_info(ring);
@@ -540,7 +542,7 @@ struct xhci_stream_info *xhci_alloc_stream_info(struct xhci_hcd *xhci,
 	 */
 	for (cur_stream = 1; cur_stream < num_streams; cur_stream++) {
 		stream_info->stream_rings[cur_stream] =
-			xhci_ring_alloc(xhci, 1, true, mem_flags);
+			xhci_ring_alloc(xhci, 1, true, false, mem_flags);
 		cur_ring = stream_info->stream_rings[cur_stream];
 		if (!cur_ring)
 			goto cleanup_rings;
@@ -874,7 +876,7 @@ int xhci_alloc_virt_device(struct xhci_hcd *xhci, int slot_id,
 	}
 
 	/* Allocate endpoint 0 ring */
-	dev->eps[0].ring = xhci_ring_alloc(xhci, 1, true, flags);
+	dev->eps[0].ring = xhci_ring_alloc(xhci, 1, true, false, flags);
 	if (!dev->eps[0].ring)
 		goto fail;
 
@@ -1315,10 +1317,10 @@ int xhci_endpoint_init(struct xhci_hcd *xhci,
 	 */
 	if (usb_endpoint_xfer_isoc(&ep->desc))
 		virt_dev->eps[ep_index].new_ring =
-			xhci_ring_alloc(xhci, 8, true, mem_flags);
+			xhci_ring_alloc(xhci, 8, true, true, mem_flags);
 	else
 		virt_dev->eps[ep_index].new_ring =
-			xhci_ring_alloc(xhci, 1, true, mem_flags);
+			xhci_ring_alloc(xhci, 1, true, false, mem_flags);
 	if (!virt_dev->eps[ep_index].new_ring) {
 		/* Attempt to use the ring cache */
 		if (virt_dev->num_rings_cached == 0)
@@ -1327,7 +1329,8 @@ int xhci_endpoint_init(struct xhci_hcd *xhci,
 			virt_dev->ring_cache[virt_dev->num_rings_cached];
 		virt_dev->ring_cache[virt_dev->num_rings_cached] = NULL;
 		virt_dev->num_rings_cached--;
-		xhci_reinit_cached_ring(xhci, virt_dev->eps[ep_index].new_ring);
+		xhci_reinit_cached_ring(xhci, virt_dev->eps[ep_index].new_ring,
+			usb_endpoint_xfer_isoc(&ep->desc) ? true : false);
 	}
 	virt_dev->eps[ep_index].skip = false;
 	ep_ring = virt_dev->eps[ep_index].new_ring;
@@ -2236,7 +2239,7 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
 		goto fail;
 
 	/* Set up the command ring to have one segments for now. */
-	xhci->cmd_ring = xhci_ring_alloc(xhci, 1, true, flags);
+	xhci->cmd_ring = xhci_ring_alloc(xhci, 1, true, false, flags);
 	if (!xhci->cmd_ring)
 		goto fail;
 	xhci_dbg(xhci, "Allocated command ring at %p\n", xhci->cmd_ring);
@@ -2267,7 +2270,8 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
 	 * the event ring segment table (ERST).  Section 4.9.3.
 	 */
 	xhci_dbg(xhci, "// Allocating event ring\n");
-	xhci->event_ring = xhci_ring_alloc(xhci, ERST_NUM_SEGS, false, flags);
+	xhci->event_ring = xhci_ring_alloc(xhci, ERST_NUM_SEGS, false, false,
+						flags);
 	if (!xhci->event_ring)
 		goto fail;
 	if (xhci_check_trb_in_td_math(xhci, flags) < 0)

drivers/usb/host/xhci-pci.c

@@ -128,6 +128,9 @@ static int xhci_pci_setup(struct usb_hcd *hcd)
 	if (pdev->vendor == PCI_VENDOR_ID_NEC)
 		xhci->quirks |= XHCI_NEC_HOST;
 
+	if (pdev->vendor == PCI_VENDOR_ID_AMD && xhci->hci_version == 0x96)
+		xhci->quirks |= XHCI_AMD_0x96_HOST;
+
 	/* AMD PLL quirk */
 	if (pdev->vendor == PCI_VENDOR_ID_AMD && usb_amd_find_chipset_info())
 		xhci->quirks |= XHCI_AMD_PLL_FIX;

drivers/usb/host/xhci-ring.c

@@ -185,7 +185,7 @@ static void inc_deq(struct xhci_hcd *xhci, struct xhci_ring *ring, bool consumer
  *			prepare_transfer()?
  */
 static void inc_enq(struct xhci_hcd *xhci, struct xhci_ring *ring,
-		bool consumer, bool more_trbs_coming)
+		bool consumer, bool more_trbs_coming, bool isoc)
 {
 	u32 chain;
 	union xhci_trb *next;
@@ -212,11 +212,13 @@ static void inc_enq(struct xhci_hcd *xhci, struct xhci_ring *ring,
 				if (!chain && !more_trbs_coming)
 					break;
 
-				/* If we're not dealing with 0.95 hardware,
+				/* If we're not dealing with 0.95 hardware or
+				 * isoc rings on AMD 0.96 host,
 				 * carry over the chain bit of the previous TRB
 				 * (which may mean the chain bit is cleared).
 				 */
-				if (!xhci_link_trb_quirk(xhci)) {
+				if (!(isoc && (xhci->quirks & XHCI_AMD_0x96_HOST))
+						&& !xhci_link_trb_quirk(xhci)) {
 					next->link.control &=
 						cpu_to_le32(~TRB_CHAIN);
 					next->link.control |=
@@ -2391,7 +2393,7 @@ irqreturn_t xhci_msi_irq(int irq, struct usb_hcd *hcd)
  *			prepare_transfer()?
  */
 static void queue_trb(struct xhci_hcd *xhci, struct xhci_ring *ring,
-		bool consumer, bool more_trbs_coming,
+		bool consumer, bool more_trbs_coming, bool isoc,
 		u32 field1, u32 field2, u32 field3, u32 field4)
 {
 	struct xhci_generic_trb *trb;
@@ -2401,7 +2403,7 @@ static void queue_trb(struct xhci_hcd *xhci, struct xhci_ring *ring,
 	trb->field[1] = cpu_to_le32(field2);
 	trb->field[2] = cpu_to_le32(field3);
 	trb->field[3] = cpu_to_le32(field4);
-	inc_enq(xhci, ring, consumer, more_trbs_coming);
+	inc_enq(xhci, ring, consumer, more_trbs_coming, isoc);
 }
 
 /*
@@ -2409,7 +2411,7 @@ static void queue_trb(struct xhci_hcd *xhci, struct xhci_ring *ring,
  * FIXME allocate segments if the ring is full.
  */
 static int prepare_ring(struct xhci_hcd *xhci, struct xhci_ring *ep_ring,
-		u32 ep_state, unsigned int num_trbs, gfp_t mem_flags)
+		u32 ep_state, unsigned int num_trbs, bool isoc, gfp_t mem_flags)
 {
 	/* Make sure the endpoint has been added to xHC schedule */
 	switch (ep_state) {
@@ -2451,10 +2453,11 @@ static int prepare_ring(struct xhci_hcd *xhci, struct xhci_ring *ep_ring,
 		next = ring->enqueue;
 
 		while (last_trb(xhci, ring, ring->enq_seg, next)) {
-			/* If we're not dealing with 0.95 hardware,
-			 * clear the chain bit.
+			/* If we're not dealing with 0.95 hardware or isoc rings
+			 * on AMD 0.96 host, clear the chain bit.
 			 */
-			if (!xhci_link_trb_quirk(xhci))
+			if (!xhci_link_trb_quirk(xhci) && !(isoc &&
+					(xhci->quirks & XHCI_AMD_0x96_HOST)))
 				next->link.control &= cpu_to_le32(~TRB_CHAIN);
 			else
 				next->link.control |= cpu_to_le32(TRB_CHAIN);
@@ -2487,6 +2490,7 @@ static int prepare_transfer(struct xhci_hcd *xhci,
 		unsigned int num_trbs,
 		struct urb *urb,
 		unsigned int td_index,
+		bool isoc,
 		gfp_t mem_flags)
 {
 	int ret;
@@ -2504,7 +2508,7 @@ static int prepare_transfer(struct xhci_hcd *xhci,
 
 	ret = prepare_ring(xhci, ep_ring,
 			   le32_to_cpu(ep_ctx->ep_info) & EP_STATE_MASK,
-			   num_trbs, mem_flags);
+			   num_trbs, isoc, mem_flags);
 	if (ret)
 		return ret;
 
@@ -2727,7 +2731,7 @@ static int queue_bulk_sg_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 
 	trb_buff_len = prepare_transfer(xhci, xhci->devs[slot_id],
 			ep_index, urb->stream_id,
-			num_trbs, urb, 0, mem_flags);
+			num_trbs, urb, 0, false, mem_flags);
 	if (trb_buff_len < 0)
 		return trb_buff_len;
 
@@ -2822,7 +2826,7 @@ static int queue_bulk_sg_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 			more_trbs_coming = true;
 		else
 			more_trbs_coming = false;
-		queue_trb(xhci, ep_ring, false, more_trbs_coming,
+		queue_trb(xhci, ep_ring, false, more_trbs_coming, false,
 				lower_32_bits(addr),
 				upper_32_bits(addr),
 				length_field,
@@ -2913,7 +2917,7 @@ int xhci_queue_bulk_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 
 	ret = prepare_transfer(xhci, xhci->devs[slot_id],
 			ep_index, urb->stream_id,
-			num_trbs, urb, 0, mem_flags);
+			num_trbs, urb, 0, false, mem_flags);
 	if (ret < 0)
 		return ret;
 
@@ -2985,7 +2989,7 @@ int xhci_queue_bulk_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 			more_trbs_coming = true;
 		else
 			more_trbs_coming = false;
-		queue_trb(xhci, ep_ring, false, more_trbs_coming,
+		queue_trb(xhci, ep_ring, false, more_trbs_coming, false,
 				lower_32_bits(addr),
 				upper_32_bits(addr),
 				length_field,
@@ -3045,7 +3049,7 @@ int xhci_queue_ctrl_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 		num_trbs++;
 	ret = prepare_transfer(xhci, xhci->devs[slot_id],
 			ep_index, urb->stream_id,
-			num_trbs, urb, 0, mem_flags);
+			num_trbs, urb, 0, false, mem_flags);
 	if (ret < 0)
 		return ret;
 
@@ -3078,7 +3082,7 @@ int xhci_queue_ctrl_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 		}
 	}
 
-	queue_trb(xhci, ep_ring, false, true,
+	queue_trb(xhci, ep_ring, false, true, false,
 		  setup->bRequestType | setup->bRequest << 8 | le16_to_cpu(setup->wValue) << 16,
 		  le16_to_cpu(setup->wIndex) | le16_to_cpu(setup->wLength) << 16,
 		  TRB_LEN(8) | TRB_INTR_TARGET(0),
@@ -3098,7 +3102,7 @@ int xhci_queue_ctrl_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 	if (urb->transfer_buffer_length > 0) {
 		if (setup->bRequestType & USB_DIR_IN)
 			field |= TRB_DIR_IN;
-		queue_trb(xhci, ep_ring, false, true,
+		queue_trb(xhci, ep_ring, false, true, false,
 				lower_32_bits(urb->transfer_dma),
 				upper_32_bits(urb->transfer_dma),
 				length_field,
@@ -3114,7 +3118,7 @@ int xhci_queue_ctrl_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 		field = 0;
 	else
 		field = TRB_DIR_IN;
-	queue_trb(xhci, ep_ring, false, false,
+	queue_trb(xhci, ep_ring, false, false, false,
 			0,
 			0,
 			TRB_INTR_TARGET(0),
@@ -3263,7 +3267,8 @@ static int xhci_queue_isoc_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 		trbs_per_td = count_isoc_trbs_needed(xhci, urb, i);
 
 		ret = prepare_transfer(xhci, xhci->devs[slot_id], ep_index,
-				urb->stream_id, trbs_per_td, urb, i, mem_flags);
+				urb->stream_id, trbs_per_td, urb, i, true,
+				mem_flags);
 		if (ret < 0) {
 			if (i == 0)
 				return ret;
@@ -3333,7 +3338,7 @@ static int xhci_queue_isoc_tx(struct xhci_hcd *xhci, gfp_t mem_flags,
 				remainder |
 				TRB_INTR_TARGET(0);
 
-			queue_trb(xhci, ep_ring, false, more_trbs_coming,
+			queue_trb(xhci, ep_ring, false, more_trbs_coming, true,
 				lower_32_bits(addr),
 				upper_32_bits(addr),
 				length_field,
@@ -3415,7 +3420,7 @@ int xhci_queue_isoc_tx_prepare(struct xhci_hcd *xhci, gfp_t mem_flags,
 	 * Do not insert any td of the urb to the ring if the check failed.
 	 */
 	ret = prepare_ring(xhci, ep_ring, le32_to_cpu(ep_ctx->ep_info) & EP_STATE_MASK,
-			   num_trbs, mem_flags);
+			   num_trbs, true, mem_flags);
 	if (ret)
 		return ret;
 
@@ -3474,7 +3479,7 @@ static int queue_command(struct xhci_hcd *xhci, u32 field1, u32 field2,
 		reserved_trbs++;
 
 	ret = prepare_ring(xhci, xhci->cmd_ring, EP_STATE_RUNNING,
-			reserved_trbs, GFP_ATOMIC);
+			reserved_trbs, false, GFP_ATOMIC);
 	if (ret < 0) {
 		xhci_err(xhci, "ERR: No room for command on command ring\n");
 		if (command_must_succeed)
@@ -3482,8 +3487,8 @@ static int queue_command(struct xhci_hcd *xhci, u32 field1, u32 field2,
 					"unfailable commands failed.\n");
 		return ret;
 	}
-	queue_trb(xhci, xhci->cmd_ring, false, false, field1, field2, field3,
-			field4 | xhci->cmd_ring->cycle_state);
+	queue_trb(xhci, xhci->cmd_ring, false, false, false, field1, field2,
+			field3,	field4 | xhci->cmd_ring->cycle_state);
 	return 0;
 }
 

drivers/usb/host/xhci.h

@@ -1462,6 +1462,7 @@ struct xhci_hcd {
 #define XHCI_BROKEN_MSI		(1 << 6)
 #define XHCI_RESET_ON_RESUME	(1 << 7)
 #define	XHCI_SW_BW_CHECKING	(1 << 8)
+#define XHCI_AMD_0x96_HOST	(1 << 9)
 	unsigned int		num_active_eps;
 	unsigned int		limit_active_eps;
 	/* There are two roothubs to keep track of bus suspend info for */