render error detected, EIR: 0x00000010

Jaswinder Singh Rajput jaswinderlinux at gmail.com
Fri May 14 00:35:00 PDT 2010


Hello Chris,

On Fri, May 14, 2010 at 4:16 AM, Chris Wilson <chris at chris-wilson.co.uk> wrote:
> On Fri, 14 May 2010 03:59:08 +0530, Jaswinder Singh Rajput <jaswinderlinux at gmail.com> wrote:
>> Hello Chris,
>>
>> I am also getting display error on another PC :
>>
>> [   34.198957] render error detected, EIR: 0x00000010
>> [   34.198972] [drm:i915_handle_error] *ERROR* EIR stuck: 0x00000010, masking
>> [   34.198998] render error detected, EIR: 0x00000010
>
> Wow. Things are really amiss there. MI_SET_CONTEXT (the first instruction
> in the batch) was removed in
>
>  commit 45f45c73469f1bd46a1b6fb206f2e9e5e4fd66b3
>  Author: Eric Anholt <eric at anholt.net>
>  Date:   Mon Dec 29 11:26:11 2008 -0800
>
>      Remove logical context setup.
>
> which explains why the destination bo is not on the active list, you
> have a pre-GEM UMS xf86-video-intel (so the batchbuffer is using absolute
> addressing rather than relocs and the kernel memory manager). The hardware
> is reporting a PTE error (in the overlay plane!) and the single active
> command is to clear the absolute buffer. As it is not being tracked by the
> kernel, I can't tell you what's wrong with the buffer, but most probably
> either the size or its stride is illegal in the context of that command.
> Or maybe even the GTT hasn't been fully initialised for the buffer, it is
> impossible to tell.
>
> In short, you can trade this bug against the risk of GTT incoherency that
> is endemic on i8xx by updating your driver.
>

I am getting XWindows on this old machine so this is not so serious
issue, I can ignore the error message ;-)

I am more worried about Acer machine where I am not getting XWindows.

Thanks,
--
Jaswinder Singh.


More information about the dri-devel mailing list