Bug#822665: bumblebee: reports wrong error message from Xorg log

[ydirson at free.fr]

No I just reported it to the bts as I saw it

----- Mail original -----
> De: "Luca Boccassi" <luca.boccassi at gmail.com>
> À: ydirson at free.fr, 822665 at bugs.debian.org
> Envoyé: Mardi 26 Avril 2016 11:28:29
> Objet: Re: Bug#822665: bumblebee: reports wrong error message from Xorg log
> 
> Control: tags -1 upstream
> 
> On Tue, 2016-04-26 at 10:17 +0200, ydirson at free.fr wrote:
> > Package: bumblebee
> > Version: 3.2.1-10
> > 
> > When trying to use bumblebee those days, the reported failure is:
> > 
> > $ optirun xterm
> > [42519.843834] [ERROR]Cannot access secondary GPU - error: [XORG]
> > (EE) /dev/dri/card0: failed to set DRM interface version 1.4:
> > Permission denied
> > 
> > [42519.843864] [ERROR]Aborting because fallback start is disabled.
> > $
> > 
> > Looking in Xorg.8.log, we can see:
> > 
> > [ 42519.553] (II) xfree86: Adding drm device (/dev/dri/card1)
> > [ 42519.840] (II) xfree86: Adding drm device (/dev/dri/card0)
> > [ 42519.840] (EE) /dev/dri/card0: failed to set DRM interface
> > version 1.4: Permission denied
> > [ 42519.841] (--) PCI:*(0:1:0:0) 10de:139b:1462:1139 rev 162, Mem @
> > 0xa2000000/16777216, 0xc0000000/268435456, 0xd0000000/33554432,
> > I/O @ 0x00004000/128, BIOS @ 0x????????/524288
> > ...
> > [ 42519.843] (II) [drm] nouveau interface version: 1.3.1
> > [ 42519.843] (EE) Unknown chipset: NV117
> > [ 42519.843] (EE) No devices detected.
> > [ 42519.843] (EE)
> > Fatal server error:
> > [ 42519.843] (EE) no screens found(EE)
> > 
> > The "permission denied" is quite strange, as the device does have
> > the ACL allowing RW for the current user.
> > But anyway that card is already used by :0, could it be that the
> > error code returned by the DRI device is
> > just wrong (EPERM instead of EBUSY) ?
> > 
> > Anyway, that error does not appear to be fatal to Xorg, which does
> > list the second one as fatal,
> > and incidentally that error message looks like a much more
> > understandable reason
> 
> Hi,
> 
> Yes unfortunately very often bumblebee errors are a bit misleading or
> hard to interpret :-(
> Have you reported this upstream?
> 
> Kind regards,
> Luca Boccassi
>