lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20110928220127.971419646@clark.kroah.org>
Date:	Wed, 28 Sep 2011 14:59:50 -0700
From:	Greg KH <gregkh@...e.de>
To:	linux-kernel@...r.kernel.org, stable@...nel.org
Cc:	stable-review@...nel.org, torvalds@...ux-foundation.org,
	akpm@...ux-foundation.org, alan@...rguk.ukuu.org.uk,
	Sarah Sharp <sarah.a.sharp@...ux.intel.com>,
	Andiry Xu <andiry.xu@....com>
Subject: [026/244] xhci: Remove TDs from TD lists when URBs are canceled.

3.0-stable review patch.  If anyone has any objections, please let us know.

------------------

From: Sarah Sharp <sarah.a.sharp@...ux.intel.com>

commit 585df1d90cb07a02ca6c7a7d339e56e46d50dafb upstream.

When a driver tries to cancel an URB, and the host controller is dying,
xhci_urb_dequeue will giveback the URB without removing the xhci_tds
that comprise that URB from the td_list or the cancelled_td_list.  This
can cause a race condition between the driver calling URB dequeue and
the stop endpoint command watchdog timer.

If the timer fires on a dying host, and a driver attempts to resubmit
while the watchdog timer has dropped the xhci->lock to giveback a
cancelled URB, URBs may be given back by the xhci_urb_dequeue() function.
At that point, the URB's priv pointer will be freed and set to NULL, but
the TDs will remain on the td_list.  This will cause an oops in
xhci_giveback_urb_in_irq() when the watchdog timer attempts to loop
through the endpoints' td_lists, giving back killed URBs.

Make sure that xhci_urb_dequeue() removes TDs from the TD lists and
canceled TD lists before it gives back the URB.

This patch should be backported to kernels as old as 2.6.36.

Signed-off-by: Sarah Sharp <sarah.a.sharp@...ux.intel.com>
Cc: Andiry Xu <andiry.xu@....com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...e.de>

---
 drivers/usb/host/xhci-ring.c |   16 ++++++++--------
 drivers/usb/host/xhci.c      |    7 +++++++
 2 files changed, 15 insertions(+), 8 deletions(-)

--- a/drivers/usb/host/xhci-ring.c
+++ b/drivers/usb/host/xhci-ring.c
@@ -744,7 +744,7 @@ remove_finished_td:
 		 * so remove it from the endpoint ring's TD list.  Keep it in
 		 * the cancelled TD list for URB completion later.
 		 */
-		list_del(&cur_td->td_list);
+		list_del_init(&cur_td->td_list);
 	}
 	last_unlinked_td = cur_td;
 	xhci_stop_watchdog_timer_in_irq(xhci, ep);
@@ -772,7 +772,7 @@ remove_finished_td:
 	do {
 		cur_td = list_entry(ep->cancelled_td_list.next,
 				struct xhci_td, cancelled_td_list);
-		list_del(&cur_td->cancelled_td_list);
+		list_del_init(&cur_td->cancelled_td_list);
 
 		/* Clean up the cancelled URB */
 		/* Doesn't matter what we pass for status, since the core will
@@ -880,9 +880,9 @@ void xhci_stop_endpoint_command_watchdog
 				cur_td = list_first_entry(&ring->td_list,
 						struct xhci_td,
 						td_list);
-				list_del(&cur_td->td_list);
+				list_del_init(&cur_td->td_list);
 				if (!list_empty(&cur_td->cancelled_td_list))
-					list_del(&cur_td->cancelled_td_list);
+					list_del_init(&cur_td->cancelled_td_list);
 				xhci_giveback_urb_in_irq(xhci, cur_td,
 						-ESHUTDOWN, "killed");
 			}
@@ -891,7 +891,7 @@ void xhci_stop_endpoint_command_watchdog
 						&temp_ep->cancelled_td_list,
 						struct xhci_td,
 						cancelled_td_list);
-				list_del(&cur_td->cancelled_td_list);
+				list_del_init(&cur_td->cancelled_td_list);
 				xhci_giveback_urb_in_irq(xhci, cur_td,
 						-ESHUTDOWN, "killed");
 			}
@@ -1582,10 +1582,10 @@ td_cleanup:
 			else
 				*status = 0;
 		}
-		list_del(&td->td_list);
+		list_del_init(&td->td_list);
 		/* Was this TD slated to be cancelled but completed anyway? */
 		if (!list_empty(&td->cancelled_td_list))
-			list_del(&td->cancelled_td_list);
+			list_del_init(&td->cancelled_td_list);
 
 		urb_priv->td_cnt++;
 		/* Giveback the urb when all the tds are completed */
@@ -3370,7 +3370,7 @@ cleanup:
 	/* Clean up a partially enqueued isoc transfer. */
 
 	for (i--; i >= 0; i--)
-		list_del(&urb_priv->td[i]->td_list);
+		list_del_init(&urb_priv->td[i]->td_list);
 
 	/* Use the first TD as a temporary variable to turn the TDs we've queued
 	 * into No-ops with a software-owned cycle bit. That way the hardware
--- a/drivers/usb/host/xhci.c
+++ b/drivers/usb/host/xhci.c
@@ -1252,6 +1252,13 @@ int xhci_urb_dequeue(struct usb_hcd *hcd
 	if (temp == 0xffffffff || (xhci->xhc_state & XHCI_STATE_HALTED)) {
 		xhci_dbg(xhci, "HW died, freeing TD.\n");
 		urb_priv = urb->hcpriv;
+		for (i = urb_priv->td_cnt; i < urb_priv->length; i++) {
+			td = urb_priv->td[i];
+			if (!list_empty(&td->td_list))
+				list_del_init(&td->td_list);
+			if (!list_empty(&td->cancelled_td_list))
+				list_del_init(&td->cancelled_td_list);
+		}
 
 		usb_hcd_unlink_urb_from_ep(hcd, urb);
 		spin_unlock_irqrestore(&xhci->lock, flags);


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ