[Intel-gfx] [PATCH] drm/i915: Reword warning for missing cases

[Lucas De Marchi]

On Tue, Mar 13, 2018 at 03:23:54PM +0200, Ville Syrjälä wrote:
> On Tue, Mar 13, 2018 at 01:06:49PM +0000, Chris Wilson wrote:
> > Quoting Ville Syrjälä (2018-03-13 13:01:42)
> > > On Tue, Mar 13, 2018 at 12:17:13AM +0000, Chris Wilson wrote:
> > > > Quoting Lucas De Marchi (2018-03-13 00:03:12)
> > > > > In some places we end up converting switch statements to a series of
> > > > > if/else, particularly when introducing helper functions to handle a
> > > > > group of cases. It's tempting to either leave a wrong warning (since now
> > > > > we don't have a switch case anymore) or to convert to WARN(1, ...),
> > > > > losing what MISSING_CASE() provides: source location and id number.
> > > > > 
> > > > > Fix the message to allow reusing MISSING_CASE() - it may not always be
> > > > > correct (e.g. if you are not checking an id anymore), but it avoids
> > > > > useless conversions. A quick grep reveals at least a few users in
> > > > > drivers/gpu/drm/i915/intel_csr.c and drivers/gpu/drm/i915/intel_ddi.c.
> > > > > 
> > > > > Signed-off-by: Lucas De Marchi <lucas.demarchi at intel.com>
> > > > > ---
> > > > >  drivers/gpu/drm/i915/i915_utils.h | 2 +-
> > > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > > 
> > > > > diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h
> > > > > index 51dbfe5bb418..8cdc21b92f5f 100644
> > > > > --- a/drivers/gpu/drm/i915/i915_utils.h
> > > > > +++ b/drivers/gpu/drm/i915/i915_utils.h
> > > > > @@ -40,7 +40,7 @@
> > > > >  #undef WARN_ON_ONCE
> > > > >  #define WARN_ON_ONCE(x) WARN_ONCE((x), "%s", "WARN_ON_ONCE(" __stringify(x) ")")
> > > > >  
> > > > > -#define MISSING_CASE(x) WARN(1, "Missing switch case (%lu) in %s\n", \
> > > > > +#define MISSING_CASE(x) WARN(1, "Missing case (%lu) in %s\n", \
> > > > >                              (long)(x), __func__)
> > > > 
> > > > Whilst here you could make this more informative by:
> > > > "Missing case (%s = %lu) in %s\n", __stringify(x), (long)(x), __func__
> > > 
> > > The backtrace isn't enough?
> > 
> > Not if we have more than one in a function.
> 
> I was just commenting on the __func__ part actually. That seems pretty
> much redundant to me. The stringify part does seem like a decent idea,
> and matches our WARN() trickery pretty well.

Humn... indeed. On a quick look I thought it was there so we wouldn't need a addr2line.
But we are not even printing the line number, just the function name.
The warning is from the kernel already does that and more.

What about:

#define MISSING_CASE(x) WARN(1, "Missing case (%s == %lu)\n", \
			     __stringify(x), (unsigned long)(x))

I added a simple hack on probe and I get this:

[ 4535.233717] Missing case (ret == 0)
[ 4535.233868] WARNING: CPU: 1 PID: 795 at drivers/gpu/drm/i915/i915_drv.c:1341 i915_driver_load+0x42/0x10e0 [i915]


Lucas De Marchi

> 
> > (Why would we do that, you
> > might ask, and I'd answer if the point is to make this more generic and
> > versatile, then do so. We already have the value, why not then explain
> > what that value is.) And give me a single sentence identifying the missing
> > case makes it much more pleasant.
> > -Chris
> 
> -- 
> Ville Syrjälä
> Intel OTC