Re: [PATCH v2] drm/drm_vblank: Change EINVAL by the correct errno

From: Ville Syrjälä
Date: Tue Oct 16 2018 - 13:28:13 EST


On Tue, Oct 16, 2018 at 06:38:31PM +0200, Daniel Vetter wrote:
> On Tue, Oct 16, 2018 at 03:36:20PM +0200, Maarten Lankhorst wrote:
> > Op 15-10-18 om 19:05 schreef Rodrigo Siqueira:
> > > For historical reason, the function drm_wait_vblank_ioctl always return
> > > -EINVAL if something gets wrong. This scenario limits the flexibility
> > > for the userspace make detailed verification of the problem and take
> > > some action. In particular, the validation of âif (!dev->irq_enabled)â
> > > in the drm_wait_vblank_ioctl is responsible for checking if the driver
> > > support vblank or not. If the driver does not support VBlank, the
> > > function drm_wait_vblank_ioctl returns EINVAL which does not represent
> > > the real issue; this patch changes this behavior by return EOPNOTSUPP.
> > > Additionally, some operations are unsupported by this function, and
> > > returns EINVAL; this patch also changes the return value to EOPNOTSUPP
> > > in this case. Lastly, the function drm_wait_vblank_ioctl is invoked by
> > > libdrm, which is used by many compositors; because of this, it is
> > > important to check if this change breaks any compositor. In this sense,
> > > the following projects were examined:
> > >
> > > * Drm-hwcomposer
> > > * Kwin
> > > * Sway
> > > * Wlroots
> > > * Wayland-core
> > > * Weston
> > > * Xorg (67 different drivers)
> > >
> > > For each repository the verification happened in three steps:
> > >
> > > * Update the main branch
> > > * Look for any occurrence "drmWaitVBlank" with the command:
> > > git grep -n "drmWaitVBlank"
> > > * Look in the git history of the project with the command:
> > > git log -SdrmWaitVBlank
> > >
> > > Finally, none of the above projects validate the use of EINVAL which
> > > make safe, at least for these projects, to change the return values.
> > >
> > > Change since V1:
> > > Daniel Vetter and Chris Wilson
> > > - Replace ENOTTY by EOPNOTSUPP
> > > - Return EINVAL if the parameters are wrong
> > >
> > > Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@xxxxxxxxx>
> > > ---
> > > drivers/gpu/drm/drm_vblank.c | 4 ++--
> > > 1 file changed, 2 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c
> > > index 98e091175921..80f5a3bb427e 100644
> > > --- a/drivers/gpu/drm/drm_vblank.c
> > > +++ b/drivers/gpu/drm/drm_vblank.c
> > > @@ -1533,10 +1533,10 @@ int drm_wait_vblank_ioctl(struct drm_device *dev, void *data,
> > > unsigned int flags, pipe, high_pipe;
> > >
> > > if (!dev->irq_enabled)
> > > - return -EINVAL;
> > > + return -EOPNOTSUPP;
> > Change to -EIO?
> >
> > If userspace would ever print this out, it would print the following
> > confusing message to userspace:
> > "Operation not supported on transport endpoint"
>
> You're a bit late, EOPNOTSUPP is not established already in upstream for
> this. And -EIO is taken already for "the gpu is dead".
>
> > >
> > > if (vblwait->request.type & _DRM_VBLANK_SIGNAL)
> > > - return -EINVAL;
> > > + return -EOPNOTSUPP;
> > I would keep this -EINVAL, tbh and making it part of the below if statement..
>
> We discussed this, it's different: This here is an ioctl flag that's no
> longer supported, the below is just an invalid request. Hence different
> errno.
>
> I think you missed a bit with your bikeshed :-)

I think I too agree with the -EINVAL here as this flag is never
supported, whereas -EOPNOTSUPP seems to mean "this flag is still
valid, but not supported by your current hardware/driver
configuration".

Also drm_invalid_op() uses -EINVAL for deprecated features as well.

--
Ville SyrjÃlÃ
Intel