[Libva] vainfo segfault on GM965

[Nicholas Andrade]

Thanks, I didn't realize until doing some more research today that my GM965 is actually part of the X3100 chipset (I had assumed it was using the X4500HD like my other C2D systems).  Anyways, the failure does make sense; I appreciate the support.

Thanks,Nick
 

     On Wednesday, October 29, 2014 6:28 PM, "Zhao, Yakui" <yakui.zhao at intel.com> wrote:
   

 On Wed, 2014-10-29 at 19:09 -0600, Nicholas Andrade wrote:
> Hi Yakui,
> 
> 
> Thanks for the info & the patch.  The patch worked in that it got rid
> of the segfault and returned the previous error:
> <<<<<
> libva info: VA-API version 0.36.0
> libva info: va_getDriverName() returns 0
> libva info: Trying to
> open /home/neandrad/libva-intel-driver-1.4.1/src/.libs/i965_drv_video.so
> libva info: Found init function __vaDriverInit_0_36
> libva
> error: /home/neandrad/libva-intel-driver-1.4.1/src/.libs/i965_drv_video.so init failed
> libva info: va_openDriver() returns -1
> vaInitialize failed with error code -1 (unknown libva error),exit
> 

Thanks for the testing.

As intel-vaapi doesn't support GM965, IMO it makes sense that the above
error is returned on GM965/GL960.

> >>>>>
> I did experience bug 79478 when complining, however I applied the
> workaround in common 6 to get past it.
> 
> 
> 
> Just so I'm clear, does the GM965 have any va-api/libva support?  If
> so, which driver supports the GM965?  Did any previous version of
> libva support it?  I don't mind using an unstable or deprecated branch
> for my use case.

AFAIK, GM965 has no support of libva/intel-vaapi. (The previous version
doesn't support it).

If the HW accelerated decoding/encoding is expected, you can use the
system on the Sandybridge or later. (The Ivy bridge is OK).

Thanks.
    Yakui
> 
> 
> Also, on a side note, I have an Ivy Bridge system where libva works
> very well.
> 
> 
> 
> Thanks,
> Nick
> 
> 
> 
> 
> 
> On Wednesday, October 29, 2014 5:35 PM, "Zhao, Yakui"
> <yakui.zhao at intel.com> wrote:
> 
> 
> 
> On Wed, 2014-10-29 at 02:47 -0600, Nicholas Andrade wrote:
> > Hi all,
> > 
> > 
> > I am running into segmentation fault with vainfo from libva-1.4.1 on
> > an Intel GM965/GL960 integrated graphics card using the
> > libva-intel-driver-1.4.1    .  
> > 
> > 
> Hi, 
> 
>  Will you please try the patch in 
> http://lists.freedesktop.org/archives/libva/2014-October/002801.html
> and see whether it is helpful on your platform?
> 
> Thanks.
>    Yakui 
> 
> 
> > 
> > Running vainfo, I see the following:
> > <<<<<
> > libva info: VA-API version 0.36.0
> > libva info: va_getDriverName() returns 0
> > libva info: Trying to open /usr/lib64/va/drivers/i965_drv_video.so
> > libva info: Found init function __vaDriverInit_0_36
> > Segmentation fault
> > 
> > >>>>>
> > With the following in my syslog:
> > <<<<<
> > vainfo[30098]: segfault at 28 ip 00007f1cca7b2b57 sp
> 00007fff8a64f750
> > error 4 in i965_drv_video.so[7f1cca771000+16c000]
> > 
> > >>>>>
> > 
> > 
> > Downgrading both libva & libva-intel-driver to 1.4.0, I no longer
> see
> > the segfault however vainfo still fails:
> > <<<<<
> > libva info: VA-API version 0.36.0
> > libva info: va_getDriverName() returns 0
> > libva info: Trying to open /usr/lib64/va/drivers/i965_drv_video.so
> > libva info: Found init function __vaDriverInit_0_36
> > libva error: /usr/lib64/va/drivers/i965_drv_video.so init failed
> > libva info: va_openDriver() returns -1
> > vaInitialize failed with error code -1 (unknown libva error),exit
> > 
> > >>>>>
> > 
> > 
> > Using libva-1.3.1 & libva-intel-driver-1.3.2 had the same result as
> > 1.4.0 (i.e. no segfault, same error).
> > 
> > 
> > 
> > If it helps, here's some additional details about the hardware:
> > <<<<<
> > $ /usr/sbin/lspci -xvvv -s 00:02.0
> > 00:02.0 VGA compatible controller: Intel Corporation Mobile
> > GM965/GL960 Integrated Graphics Controller (primary) (rev 03)
> (prog-if
> > 00 [VGA controller])
> >    Subsystem: Dell Device 0279
> >    Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
> > ParErr- Stepping- SERR- FastB2B- DisINTx+
> >    Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort-
> > <TAbort- <MAbort- >SERR- <PERR- INTx-
> >    Latency: 0
> >    Interrupt: pin A routed to IRQ 47
> >    Region 0: Memory at fe600000 (64-bit, non-prefetchable) [size=1M]
> >    Region 2: Memory at d0000000 (64-bit, prefetchable) [size=256M]
> >    Region 4: I/O ports at c400 [size=8]
> >    Expansion ROM at <unassigned> [disabled]
> >    Capabilities: <access denied>
> >    Kernel driver in use: i915
> >    Kernel modules: i915
> > 00: 86 80 02 2a 07 04 90 00 03 00 00 03 00 00 80 00
> > 10: 04 00 60 fe 00 00 00 00 0c 00 00 d0 00 00 00 00
> > 20: 01 c4 00 00 00 00 00 00 00 00 00 00 28 10 79 02
> > 30: 00 00 00 00 90 00 00 00 00 00 00 00 0f 01 00 00
> > 
> > >>>>>
> > 
> > 
> > 
> > I'm not sure which other package versions are relevant; but in case
> it
> > matters, this is on a Gentoo Linux (~amd64) system using:
> > kernel 3.17.1
> > gcc 4.7.3
> > glibc 2.19
> > 
> > libdrm 2.4.58 
> > xorg-server 1.16.1
> > xf86-video-intel 2.99.916
> > 
> > mesa 10.3.1
> > 
> > 
> > 
> > Any help/suggestions would be appreciated on this.
> > 
> > 
> > Thanks,
> > Nick
> > 
> 
> 
> 
> 
> 
> 




   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/libva/attachments/20141030/cc3a525a/attachment.html>