til: introduce til_frame_plan_t and .cpu_affinity

modules/checkers w/fill_module=$module requires a consistent
mapping of cpu to fragnum since it creates a per-cpu
til_module_context_t for the fill_module.

The existing implementation for threaded rendering maximizes
performance by letting *any* scheduled to run thread advance
fragnum atomically and render the acquired fragnum
indiscriminately.  A side effect of this is any given frame, even
rendered by the same module, will have a random mapping of
cpus/threads to fragnums.

With this change, the simple til_module_t.prepare_frame() API of
returning a bare fragmenter function is changed to instead return
a "frame plan" in til_frame_plan_t.  Right now til_frame_plan_t
just contains the same fragmenter as before, but also has a
.cpu_affinity member for setting if the frame requires a stable
relationship of cpu/thread to fragnum.

Setting .cpu_affinity should be avoided if unnecessary, and that
is the default if you don't mention .cpu_affinity at all when
initializing the plan in the ergonomic manner w/designated
initializers.  This is because the way .cpu_affinity is
implemented will leave threads spinning while they poll for
*their* next fragnum using atomic intrinsics.  There's probably
some room for improvement here, but this is good enough for now
to get things working and correct.

1 files changed, 8 insertions, 2 deletions

diff --git a/src/til.h b/src/til.h
index ab8f6b8..e38ad9c 100644
--- a/src/til.h
+++ b/src/til.h
@@ -5,10 +5,16 @@
 #include "til_module_context.h"
 #include "til_setup.h"
 
-/* til_fragmenter produces fragments from an input fragment, num being the desired fragment for the current call.
+/* til_fragmenter_t produces fragments from an input fragment, num being the desired fragment for the current call.
  * return value of 1 means a fragment has been produced, 0 means num is beyond the end of fragments. */
 typedef int (*til_fragmenter_t)(til_module_context_t *context, const til_fb_fragment_t *fragment, unsigned number, til_fb_fragment_t *res_fragment);
 
+/* til_frame_plan_t is what til_module_t.prepare_frame() populates to return a fragmenter and any flags/rules */
+typedef struct til_frame_plan_t {
+	unsigned		cpu_affinity:1;	/* maintain a stable fragnum:cpu/thread mapping? (slower) */
+	til_fragmenter_t	fragmenter;	/* fragmenter to use in rendering the frame */
+} til_frame_plan_t;
+
 typedef struct til_settings_t settings;
 typedef struct til_setting_desc_t til_setting_desc_t;
 typedef struct til_knob_t til_knob_t;
@@ -18,7 +24,7 @@ typedef struct til_knob_t til_knob_t;
 typedef struct til_module_t {
 	til_module_context_t *	(*create_context)(unsigned seed, unsigned ticks, unsigned n_cpus, til_setup_t *setup);
 	void			(*destroy_context)(til_module_context_t *context);
-	void			(*prepare_frame)(til_module_context_t *context, unsigned ticks, til_fb_fragment_t *fragment, til_fragmenter_t *res_fragmenter);
+	void			(*prepare_frame)(til_module_context_t *context, unsigned ticks, til_fb_fragment_t *fragment, til_frame_plan_t *res_frame_plan);
 	void			(*render_fragment)(til_module_context_t *context, unsigned ticks, unsigned cpu, til_fb_fragment_t *fragment);
 	void			(*finish_frame)(til_module_context_t *context, unsigned ticks, til_fb_fragment_t *fragment);
 	int			(*setup)(const til_settings_t *settings, til_setting_t **res_setting, const til_setting_desc_t **res_desc, til_setup_t **res_setup);