USB: EHCI: AMD periodic frame list table quirk

On AMD SB700/SB800/Hudson-2/3 platforms, USB EHCI controller may read/write
to memory space not allocated to USB controller if there is longer than
normal latency on DMA read encountered. In this condition the exposure will
be encountered only if the driver has following format of Periodic Frame
List link pointer structure:

For any idle periodic schedule, the Frame List link pointers that have the
T-bit set to 1 intending to terminate the use of frame list link pointer
as a physical memory pointer.

Idle periodic schedule Frame List Link pointer shoule be in the following
format to avoid the issue:

Frame list link pointer should be always contains a valid pointer to a
inactive QHead with T-bit set to 0.

Signed-off-by: Andiry Xu <andiry.xu@amd.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

1 files changed, 13 insertions, 0 deletions

diff --git a/drivers/usb/host/ehci-pci.c b/drivers/usb/host/ehci-pci.c
index a1e8d27..01bb72b 100644
--- a/drivers/usb/host/ehci-pci.c
+++ b/drivers/usb/host/ehci-pci.c
@@ -103,6 +103,19 @@ static int ehci_pci_setup(struct usb_hcd *hcd)
 	if (retval)
 		return retval;
 
+	if ((pdev->vendor == PCI_VENDOR_ID_AMD && pdev->device == 0x7808) ||
+	    (pdev->vendor == PCI_VENDOR_ID_ATI && pdev->device == 0x4396)) {
+		/* EHCI controller on AMD SB700/SB800/Hudson-2/3 platforms may
+		 * read/write memory space which does not belong to it when
+		 * there is NULL pointer with T-bit set to 1 in the frame list
+		 * table. To avoid the issue, the frame list link pointer
+		 * should always contain a valid pointer to a inactive qh.
+		 */
+		ehci->use_dummy_qh = 1;
+		ehci_info(ehci, "applying AMD SB700/SB800/Hudson-2/3 EHCI "
+				"dummy qh workaround\n");
+	}
+
 	/* data structure init */
 	retval = ehci_init(hcd);
 	if (retval)