Re: [PATCH 08/11] drm: Pull the test on drm_debug in the logging macros

[Daniel Vetter <daniel@ffwll.ch>]

On Tue, Mar 25, 2014 at 11:56:24AM +0200, Jani Nikula wrote:
> On Mon, 24 Mar 2014, Damien Lespiau <damien.lespiau@intel.com> wrote:
> > In the logging code, we are currently checking is we need to output in
> 
> s/is/if/
> 
> > drm_ut_debug_printk(). This is too late. The problem is that when we write
> > something like:
> >
> >     DRM_DEBUG_DRIVER("ELD on [CONNECTOR:%d:%s], [ENCODER:%d:%s]\n",
> >                      connector->base.id,
> >                      drm_get_connector_name(connector),
> >                      connector->encoder->base.id,
> >                      drm_get_encoder_name(connector->encoder));
> >
> > We start by evaluating the arguments (so call drm_get_connector_name() and
> > drm_get_connector_name()) before ending up in drm_ut_debug_printk() which will
> > then does nothing.
> 
> s/does/do/
> 
> > This means we execute a lot of instructions (drm_get_connector_name(), in turn,
> > calls snprintf() for example) to happily discard them in the normal case,
> > drm.debug=0.
> >
> > So, let's put the test on drm_debug earlier, in the macros themselves.
> > Sprinkle an unlikely() as well for good measure.
> 
> Bikeshed, is it likely that the unlikely matters all that much?
> https://lwn.net/Articles/70473/
> 
> I won't insist on removing them, it's more the casual nature of the
> "sprinkling unlikely for good measure" that bugs me.

I've considered this and since most users don't set debug every we should
be easily able to hit the 1:1M ratio seemingly required to make debug
work.

And once you enable debugging the actual printk overhead will wash out
anything you pay in branch mispredicts anyway.
-Daniel

> 
> 
> BR,
> Jani.
> 
> 
> >
> > Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
> > ---
> >  drivers/gpu/drm/drm_stub.c | 26 ++++++++++++--------------
> >  include/drm/drmP.h         | 27 +++++++++++++++------------
> >  2 files changed, 27 insertions(+), 26 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c
> > index dc2c609..81ed832 100644
> > --- a/drivers/gpu/drm/drm_stub.c
> > +++ b/drivers/gpu/drm/drm_stub.c
> > @@ -97,26 +97,24 @@ int drm_err(const char *func, const char *format, ...)
> >  }
> >  EXPORT_SYMBOL(drm_err);
> >  
> > -void drm_ut_debug_printk(unsigned int request_level,
> > -			 const char *prefix,
> > +void drm_ut_debug_printk(const char *prefix,
> >  			 const char *function_name,
> >  			 const char *format, ...)
> >  {
> >  	struct va_format vaf;
> >  	va_list args;
> >  
> > -	if (drm_debug & request_level) {
> > -		va_start(args, format);
> > -		vaf.fmt = format;
> > -		vaf.va = &args;
> > -
> > -		if (function_name)
> > -			printk(KERN_DEBUG "[%s:%s], %pV", prefix,
> > -			       function_name, &vaf);
> > -		else
> > -			printk(KERN_DEBUG "%pV", &vaf);
> > -		va_end(args);
> > -	}
> > +	va_start(args, format);
> > +	vaf.fmt = format;
> > +	vaf.va = &args;
> > +
> > +	if (function_name)
> > +		printk(KERN_DEBUG "[%s:%s], %pV", prefix,
> > +		       function_name, &vaf);
> > +	else
> > +		printk(KERN_DEBUG "%pV", &vaf);
> > +
> > +	va_end(args);
> >  }
> >  EXPORT_SYMBOL(drm_ut_debug_printk);
> >  
> > diff --git a/include/drm/drmP.h b/include/drm/drmP.h
> > index 3055b36..87b558a 100644
> > --- a/include/drm/drmP.h
> > +++ b/include/drm/drmP.h
> > @@ -121,9 +121,8 @@ struct videomode;
> >  #define DRM_UT_KMS		0x04
> >  #define DRM_UT_PRIME		0x08
> >  
> > -extern __printf(4, 5)
> > -void drm_ut_debug_printk(unsigned int request_level,
> > -			 const char *prefix,
> > +extern __printf(3, 4)
> > +void drm_ut_debug_printk(const char *prefix,
> >  			 const char *function_name,
> >  			 const char *format, ...);
> >  extern __printf(2, 3)
> > @@ -209,24 +208,28 @@ int drm_err(const char *func, const char *format, ...);
> >  #if DRM_DEBUG_CODE
> >  #define DRM_DEBUG(fmt, args...)						\
> >  	do {								\
> > -		drm_ut_debug_printk(DRM_UT_CORE, DRM_NAME, 		\
> > -					__func__, fmt, ##args);		\
> > +		if (unlikely(drm_debug & DRM_UT_CORE))			\
> > +			drm_ut_debug_printk(DRM_NAME, __func__,		\
> > +					    fmt, ##args);		\
> >  	} while (0)
> >  
> >  #define DRM_DEBUG_DRIVER(fmt, args...)					\
> >  	do {								\
> > -		drm_ut_debug_printk(DRM_UT_DRIVER, DRM_NAME,		\
> > -					__func__, fmt, ##args);		\
> > +		if (unlikely(drm_debug & DRM_UT_DRIVER))		\
> > +			drm_ut_debug_printk(DRM_NAME, __func__,		\
> > +					    fmt, ##args);		\
> >  	} while (0)
> > -#define DRM_DEBUG_KMS(fmt, args...)				\
> > +#define DRM_DEBUG_KMS(fmt, args...)					\
> >  	do {								\
> > -		drm_ut_debug_printk(DRM_UT_KMS, DRM_NAME, 		\
> > -					 __func__, fmt, ##args);	\
> > +		if (unlikely(drm_debug & DRM_UT_KMS))			\
> > +			drm_ut_debug_printk(DRM_NAME, __func__,		\
> > +					    fmt, ##args);		\
> >  	} while (0)
> >  #define DRM_DEBUG_PRIME(fmt, args...)					\
> >  	do {								\
> > -		drm_ut_debug_printk(DRM_UT_PRIME, DRM_NAME,		\
> > -					__func__, fmt, ##args);		\
> > +		if (unlikely(drm_debug & DRM_UT_PRIME))			\
> > +			drm_ut_debug_printk(DRM_NAME, __func__,		\
> > +					    fmt, ##args);		\
> >  	} while (0)
> >  #else
> >  #define DRM_DEBUG_DRIVER(fmt, args...) do { } while (0)
> > -- 
> > 1.8.3.1
> >
> > _______________________________________________
> > dri-devel mailing list
> > dri-devel@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/dri-devel
> 
> -- 
> Jani Nikula, Intel Open Source Technology Center
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch