Clemens Ladisch wrote:
[PATCH] USB: EHCI: adjust ehci_iso_stream for changes in ehci_qh The EHCI driver stores in usb_host_endpoint.hcpriv a pointer to either an ehci_qh or an ehci_iso_stream structure, and uses the contents of the hw_info1 field to distinguish the two cases. After ehci_qh was split into hw and sw parts, ehci_iso_stream must also be adjusted so that it again looks like an ehci_qh structure. This fixes a NULL pointer access in ehci_endpoint_disable() when it tries to access qh->hw->hw_info1. Signed-off-by: Clemens Ladisch <clemens@xxxxxxxxxx> Reported-by: Colin Fletcher <colin.m.fletcher@xxxxxxxxxxxxxx> Cc: <stable@xxxxxxxxxx> --- drivers/usb/host/ehci-hcd.c | 2 +- drivers/usb/host/ehci-sched.c | 4 ++-- drivers/usb/host/ehci.h | 5 ++--- 3 files changed, 5 insertions(+), 6 deletions(-) --- a/drivers/usb/host/ehci.h +++ b/drivers/usb/host/ehci.h @@ -394,9 +394,8 @@ struct ehci_iso_sched { * acts like a qh would, if EHCI had them for ISO. */ struct ehci_iso_stream { - /* first two fields match QH, but info1 == 0 */ - __hc32 hw_next; - __hc32 hw_info1; + /* first field matches ehci_hq, but is NULL */ + struct ehci_qh_hw *hw;u32 refcount;u8 bEndpointAddress; --- a/drivers/usb/host/ehci-hcd.c +++ b/drivers/usb/host/ehci-hcd.c @@ -995,7 +995,7 @@ rescan: /* endpoints can be iso streams. for now, we don't * accelerate iso completions ... so spin a while. */ - if (qh->hw->hw_info1 == 0) { + if (qh->hw == NULL) { ehci_vdbg (ehci, "iso delay\n"); goto idle_timeout; } --- a/drivers/usb/host/ehci-sched.c +++ b/drivers/usb/host/ehci-sched.c @@ -1121,8 +1121,8 @@ iso_stream_find (struct ehci_hcd *ehci, struct urb *urb) urb->interval); }- /* if dev->ep [epnum] is a QH, info1.maxpacket is nonzero */- } else if (unlikely (stream->hw_info1 != 0)) { + /* if dev->ep [epnum] is a QH, hw is set */ + } else if (unlikely (stream->hw != NULL)) { ehci_dbg (ehci, "dev %s ep%d%s, not iso??\n", urb->dev->devpath, epnum, usb_pipein(urb->pipe) ? "in" : "out");
Just to confirm that this patch applied to kernel 2.6.32.9 cures the NULL pointer dereference that I previously saw.
Cheers, Colin. -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
