Quoting Lucas De Marchi (2020-11-24 19:13:14) > If drm_dev_register() fails there is no reason to continue registering > the driver and initializing. > > Signed-off-by: Lucas De Marchi <[email protected]> > --- > drivers/gpu/drm/i915/i915_drv.c | 20 +++++++++++--------- > 1 file changed, 11 insertions(+), 9 deletions(-) > > diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c > index 320856b665a1..c3fad01ce26f 100644 > --- a/drivers/gpu/drm/i915/i915_drv.c > +++ b/drivers/gpu/drm/i915/i915_drv.c > @@ -669,17 +669,19 @@ static void i915_driver_register(struct > drm_i915_private *dev_priv) > intel_vgpu_register(dev_priv); > > /* Reveal our presence to userspace */ > - if (drm_dev_register(dev, 0) == 0) { > - i915_debugfs_register(dev_priv); > - if (HAS_DISPLAY(dev_priv)) > - intel_display_debugfs_register(dev_priv); > - i915_setup_sysfs(dev_priv); > - > - /* Depends on sysfs having been initialized */ > - i915_perf_register(dev_priv); > - } else > + if (drm_dev_register(dev, 0) != 0) { > drm_err(&dev_priv->drm, > "Failed to register driver for userspace access!\n"); > + return; > + }
s/!= 0// Ok, we keep the driver loaded to do powersaving if not registered, so there is some advantage in the face of failure. And there is little point in trying to register some auxiliary interfaces if the primary inodes are not exposed. Reviewed-by: Chris Wilson <[email protected]> -Chris _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
