[Intel-gfx] [PATCH v11 12/19] drm_print: wrap drm_dbg in _dynamic_func_call_no_desc

[Jim Cromie]

Change __drm_dbg() macro to be CONFIG_DRM_USE_DYNAMIC_DEBUG dependent:
  N- keep straight mapping to exported ___drm_dbg()
  Y- wrap ___drm_dbg() inside _dynamic_func_call_no_desc()

This patch places ~1/2 of drm.debug API messages behind dyndbg's
JUMP_LABEL/NOOP optimization.

The CONFIG_DRM_USE_DYNAMIC_DEBUG dependence is due to the .data
footprint cost of per-callsite control; 56 bytes/site * ~2k,3k
callsites (for i915, amdgpu), which is significant enough to make
optional.

bash-5.1# drms_load
[    7.489844] dyndbg: 239 debug prints in module drm
[    7.494010] ACPI: bus type drm_connector registered
[    7.546076] dyndbg:  81 debug prints in module drm_kms_helper
[    7.555723] dyndbg:   2 debug prints in module ttm
[    7.558920] dyndbg:   8 debug prints in module video
[    8.074699] dyndbg: 431 debug prints in module i915
[    8.158682] AMD-Vi: AMD IOMMUv2 functionality not available on this system - This is not a bug.
[    8.574456] dyndbg: 3817 debug prints in module amdgpu
[    8.589962] [drm] amdgpu kernel modesetting enabled.
[    8.590548] amdgpu: CRAT table not found
[    8.590998] amdgpu: Virtual CRAT table created for CPU
[    8.591634] amdgpu: Topology: Add CPU node
[    8.636446] dyndbg:   3 debug prints in module wmi
[    8.768667] dyndbg:  24 debug prints in module nouveau

Signed-off-by: Jim Cromie <jim.cromie at gmail.com>
---
 include/drm/drm_print.h | 34 +++++++++++++++++++++++++++++++++-
 1 file changed, 33 insertions(+), 1 deletion(-)

diff --git a/include/drm/drm_print.h b/include/drm/drm_print.h
index 1eef315a0a65..8d6b74270c50 100644
--- a/include/drm/drm_print.h
+++ b/include/drm/drm_print.h
@@ -319,10 +319,36 @@ enum drm_debug_category {
 	DRM_UT_DRMRES
 };
 
+/*
+ * 3 name flavors of drm_debug_enabled:
+ *   drm_debug_enabled - public/legacy, always checks bits
+ *  _drm_debug_enabled - instrumented to observe call-rates, est overheads.
+ * __drm_debug_enabled - privileged - knows jump-label state, can short-circuit
+ */
 static inline bool drm_debug_enabled(enum drm_debug_category category)
 {
 	return unlikely(__drm_debug & BIT(category));
 }
+/*
+ * Wrap fn in macro, so that the pr_debug sees the actual caller, not
+ * the inline fn.  Using this name creates a callsite entry / control
+ * point in /proc/dynamic_debug/control.
+ */
+#define _drm_debug_enabled(category)				\
+	({							\
+		pr_debug("todo: maybe avoid via dyndbg\n");     \
+		drm_debug_enabled(category);			\
+	})
+#if defined(CONFIG_DRM_USE_DYNAMIC_DEBUG)
+/*
+ * dyndbg is wrapping the drm.debug API, so as to avoid the runtime
+ * bit-test overheads of drm_debug_enabled() in those api calls.
+ * In this case, executed callsites are known enabled, so true.
+ */
+#define __drm_debug_enabled(category)	true
+#else
+#define __drm_debug_enabled(category)	drm_debug_enabled(category)
+#endif
 
 /*
  * struct device based logging
@@ -497,7 +523,13 @@ void ___drm_dbg(enum drm_debug_category category, const char *format, ...);
 __printf(1, 2)
 void __drm_err(const char *format, ...);
 
+#if !defined(CONFIG_DRM_USE_DYNAMIC_DEBUG)
 #define __drm_dbg(fmt, ...)		___drm_dbg(fmt, ##__VA_ARGS__)
+#else
+#define __drm_dbg(eCat, fmt, ...)					\
+	_dynamic_func_call_no_desc(fmt, ___drm_dbg,			\
+				   eCat, fmt, ##__VA_ARGS__)
+#endif
 
 /* Macros to make printk easier */
 
@@ -569,7 +601,7 @@ void __drm_err(const char *format, ...);
 	static DEFINE_RATELIMIT_STATE(rs_, DEFAULT_RATELIMIT_INTERVAL, DEFAULT_RATELIMIT_BURST);\
 	const struct drm_device *drm_ = (drm);							\
 												\
-	if (drm_debug_enabled(DRM_UT_ ## category) && __ratelimit(&rs_))			\
+	if (__drm_debug_enabled(DRM_UT_ ## category) && __ratelimit(&rs_))			\
 		drm_dev_printk(drm_ ? drm_->dev : NULL, KERN_DEBUG, fmt, ## __VA_ARGS__);	\
 })
 
-- 
2.33.1