On Sun, Nov 13, 2011 at 10:12 PM, Daniel van Vugt <email address hidden> wrote:
> And it's still happening in kernel 3.1. Logs attached.
Looks like ehci controller has DMA-read a invalid zero address
from the below lines:
[94596.962587] DRHD: handling fault status reg 2
[94596.962592] DMAR:[DMA Read] Request device [00:1d.0] fault addr 0
[94596.962593] DMAR:[fault reason 06] PTE Read access is not set
then which caused ehci host controller to stop working.
Generally speaking, DMA address of transfer buffer is from usb interface
driver if flag URB_NO_TRANSFER_DMA_MAP is set. Considered that you
said it may be triggered by your PlayTV usb device, please apply the
attchment debug patch to see if something insane can be dumped out.
If you can't build kernel, please let me know and I will upload a
built kernel image for you.
Hi Daniel,
On Sun, Nov 13, 2011 at 10:12 PM, Daniel van Vugt <email address hidden> wrote:
> And it's still happening in kernel 3.1. Logs attached.
Looks like ehci controller has DMA-read a invalid zero address
from the below lines:
[94596.962587] DRHD: handling fault status reg 2
[94596.962592] DMAR:[DMA Read] Request device [00:1d.0] fault addr 0
[94596.962593] DMAR:[fault reason 06] PTE Read access is not set
then which caused ehci host controller to stop working.
Generally speaking, DMA address of transfer buffer is from usb interface TRANSFER_ DMA_MAP is set. Considered that you
driver if flag URB_NO_
said it may be triggered by your PlayTV usb device, please apply the
attchment debug patch to see if something insane can be dumped out.
If you can't build kernel, please let me know and I will upload a
built kernel image for you.
thanks,
--
Ming Lei