[Intel-gfx] [PATCH igt 1/2] lib: Attempt to load the module for a missing device

[Chris Wilson]

Quoting Petri Latvala (2017-11-14 14:10:27)
> On Tue, Nov 14, 2017 at 01:56:22PM +0000, Chris Wilson wrote:
> > Quoting Chris Wilson (2017-11-14 13:18:44)
> > > If we asked to open a particular chipset and we find no matching device,
> > > try again after attempting to load its module. Previously we only did
> > > this for vgem, which is not automatically probed during boot, but if we
> > > want to leave the module unloaded we have to try harder when we need the
> > > device.
> > > 
> > > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > > ---
> > >  lib/drmtest.c | 68 +++++++++++++++++++++++++++++++++++++++++++----------------
> > >  1 file changed, 50 insertions(+), 18 deletions(-)
> > > 
> > > diff --git a/lib/drmtest.c b/lib/drmtest.c
> > > index e6bdbc35..3a2a6343 100644
> > > --- a/lib/drmtest.c
> > > +++ b/lib/drmtest.c
> > > @@ -235,20 +235,14 @@ static int modprobe(const char *driver)
> > >         return igt_kmod_load(driver, "");
> > >  }
> > >  
> > > -/**
> > > - * __drm_open_driver:
> > > - * @chipset: OR'd flags for each chipset to search, eg. #DRIVER_INTEL
> > > - *
> > > - * Open the first DRM device we can find, searching up to 16 device nodes
> > > - *
> > > - * Returns:
> > > - * An open DRM fd or -1 on error
> > > - */
> > > -int __drm_open_driver(int chipset)
> > > +static void modprobe_i915(const char *name)
> > >  {
> > > -       if (chipset & DRIVER_VGEM)
> > > -               modprobe("vgem");
> > > +       /* When loading i915, we also want to load snd-hda et al */
> > > +       igt_i915_driver_load(NULL);
> > > +}
> > >  
> > > +static int __open_device(unsigned int chipset)
> > > +{
> > >         for (int i = 0; i < 16; i++) {
> > >                 char name[80];
> > >                 int fd;
> > > @@ -262,16 +256,13 @@ int __drm_open_driver(int chipset)
> > >                     has_known_intel_chipset(fd))
> > >                         return fd;
> > >  
> > > -               if (chipset & DRIVER_VC4 &&
> > > -                   is_vc4_device(fd))
> > > +               if (chipset & DRIVER_VC4 && is_vc4_device(fd))
> > >                         return fd;
> > >  
> > > -               if (chipset & DRIVER_VGEM &&
> > > -                   is_vgem_device(fd))
> > > +               if (chipset & DRIVER_VGEM && is_vgem_device(fd))
> > >                         return fd;
> > >  
> > > -               if (chipset & DRIVER_VIRTIO &&
> > > -                   is_virtio_device(fd))
> > > +               if (chipset & DRIVER_VIRTIO && is_virtio_device(fd))
> > >                         return fd;
> > >  
> > >                 if (chipset & DRIVER_AMDGPU && is_amd_device(fd))
> > > @@ -287,6 +278,47 @@ int __drm_open_driver(int chipset)
> > >         return -1;
> > >  }
> > >  
> > > +/**
> > > + * __drm_open_driver:
> > > + * @chipset: OR'd flags for each chipset to search, eg. #DRIVER_INTEL
> > > + *
> > > + * Open the first DRM device we can find, searching up to 16 device nodes
> > > + *
> > > + * Returns:
> > > + * An open DRM fd or -1 on error
> > > + */
> > > +int __drm_open_driver(int chipset)
> > > +{
> > > +       static const struct {
> > > +               unsigned int bit;
> > > +               const char *module;
> > > +               void (*modprobe)(const char *name);
> > > +       } modules[] = {
> > > +               { DRIVER_AMDGPU, "amdgpu" },
> > > +               { DRIVER_INTEL, "i915", modprobe_i915 },
> > > +               { DRIVER_VC4, "vc4" },
> > > +               { DRIVER_VGEM, "vgem" },
> > > +               { DRIVER_VIRTIO, "virtio-gpu" },
> > > +               {}
> > > +       }, *m;
> > > +       int fd;
> > > +
> > > +       fd = __open_device(chipset);
> > > +       if (fd != -1)
> > > +               return fd;
> > > +
> > > +       for (m = modules; m->module; m++) {
> > > +               if (chipset & (1ul << m->bit)) {
> > 
> > Eek, m->bit is already a bit and not a shift.
> > 
> > Still puzzled where the device went between tests though, no sign of it
> > being unloaded.
> 
> 
> I don't see any signs of it being loaded in the first
> place. Assuming I'm reading right and this is the correct line to look
> for (from the base run):
> 
> 
> <6>[  265.406984] [drm] Initialized vgem 1.0.0 20120112 for virtual device on minor 1

Ah, right; looking for the wrong thing and didn't click that it was
erroring out on vgem. In which case, fixing m->bit should explain it.
-chris