linux-kernel - [068/156] USB: EHCI: fix ITD list order

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: <20100330224804.789202835@linux.site>
Date:	Tue, 30 Mar 2010 15:41:42 -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,
	Clemens Ladisch <clemens@...isch.de>,
	Greg Kroah-Hartman <gregkh@...e.de>
Subject: [068/156] USB: EHCI: fix ITD list order

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

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

From: Clemens Ladisch <clemens@...isch.de>

commit 92bc3648e6027384479852b770a542722fadee7c upstream.

When isochronous URBs are shorter than one frame and when more than one
ITD in a frame has been completed before the interrupt can be handled,
scan_periodic() completes the URBs in the order in which they are found
in the descriptor list.  Therefore, the descriptor list must contain the
ITDs in the correct order, i.e., a new ITD must be linked in after any
previous ITDs of the same endpoint.

This should fix garbled capture data in the USB audio drivers.

Signed-off-by: Clemens Ladisch <clemens@...isch.de>
Reported-by: Colin Fletcher <colin.m.fletcher@...glemail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...e.de>

---
 drivers/usb/host/ehci-sched.c |   24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)

--- a/drivers/usb/host/ehci-sched.c
+++ b/drivers/usb/host/ehci-sched.c
@@ -1563,13 +1563,27 @@ itd_patch(
 static inline void
 itd_link (struct ehci_hcd *ehci, unsigned frame, struct ehci_itd *itd)
 {
-	/* always prepend ITD/SITD ... only QH tree is order-sensitive */
-	itd->itd_next = ehci->pshadow [frame];
-	itd->hw_next = ehci->periodic [frame];
-	ehci->pshadow [frame].itd = itd;
+	union ehci_shadow	*prev = &ehci->pshadow[frame];
+	__hc32			*hw_p = &ehci->periodic[frame];
+	union ehci_shadow	here = *prev;
+	__hc32			type = 0;
+
+	/* skip any iso nodes which might belong to previous microframes */
+	while (here.ptr) {
+		type = Q_NEXT_TYPE(ehci, *hw_p);
+		if (type == cpu_to_hc32(ehci, Q_TYPE_QH))
+			break;
+		prev = periodic_next_shadow(ehci, prev, type);
+		hw_p = shadow_next_periodic(ehci, &here, type);
+		here = *prev;
+	}
+
+	itd->itd_next = here;
+	itd->hw_next = *hw_p;
+	prev->itd = itd;
 	itd->frame = frame;
 	wmb ();
-	ehci->periodic[frame] = cpu_to_hc32(ehci, itd->itd_dma | Q_TYPE_ITD);
+	*hw_p = cpu_to_hc32(ehci, itd->itd_dma | Q_TYPE_ITD);
 }
 
 /* fit urb's itds into the selected schedule slot; activate as needed */


--
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/