CELL: use variant-length fragment ops programs

This is a set of changes that optimizes the memory use of fragment
operation programs (by using and transmitting only as much memory as is
needed for the fragment ops programs, instead of maximal sizes), as well
as eliminate the dependency on hard-coded maximal program sizes.  State
that is not dependent on fragment facing (i.e. that isn't using
two-sided stenciling) will only save and transmit a single
fragment operation program, instead of two identical programs.

- Added the ability to emit a LNOP (No Operation (Load)) instruction.
  This is used to pad the generated fragment operations programs to
  a multiple of 8 bytes, which is necessary for proper operation of
  the dual instruction pipeline, and also required for proper SPU-side
  decoding.

- Added the ability to allocate and manage a variant-length
  struct cell_command_fragment_ops.  This structure now puts the
  generated function field at the end, where it can be as large
  as necessary.

- On the PPU side, we now combine the generated front-facing and
  back-facing code into a single variant-length buffer (and only use one
  if the two sets of code are identical) for transmission to the SPU.

- On the SPU side, we pull the correct sizes out of the buffer,
  allocate a new code buffer if the one we have isn't large enough,
  and save the code to that buffer.  The buffer is deallocated when
  the SPU exits.

- Commented out the emit_fetch() static function, which was not being used.

3 files changed, 16 insertions, 4 deletions

diff --git a/src/gallium/auxiliary/rtasm/rtasm_ppc_spe.c b/src/gallium/auxiliary/rtasm/rtasm_ppc_spe.c
index 1bd9f1c8dd0..b9a75ae5591 100644
--- a/src/gallium/auxiliary/rtasm/rtasm_ppc_spe.c
+++ b/src/gallium/auxiliary/rtasm/rtasm_ppc_spe.c
@@ -341,7 +341,11 @@ static void emit_RI18(struct spe_function *p, unsigned op, unsigned rT,
 }
 
 
-
+#define EMIT(_name, _op) \
+void _name (struct spe_function *p) \
+{ \
+   emit_RR(p, _op, 0, 0, 0, __FUNCTION__); \
+}
 
 #define EMIT_(_name, _op) \
 void _name (struct spe_function *p, unsigned rT) \
@@ -713,7 +717,6 @@ hbrr;
 #if 0
 stop;
 EMIT_RR  (spe_stopd, 0x140);
-EMIT_    (spe_lnop,  0x001);
 EMIT_    (spe_nop,   0x201);
 sync;
 EMIT_    (spe_dsync, 0x003);
diff --git a/src/gallium/auxiliary/rtasm/rtasm_ppc_spe.h b/src/gallium/auxiliary/rtasm/rtasm_ppc_spe.h
index 7c211ffc51b..f9ad2acacdd 100644
--- a/src/gallium/auxiliary/rtasm/rtasm_ppc_spe.h
+++ b/src/gallium/auxiliary/rtasm/rtasm_ppc_spe.h
@@ -99,7 +99,9 @@ extern void spe_comment(struct spe_function *p, int rel_indent, const char *s);
 
 #endif /* RTASM_PPC_SPE_H */
 
-#ifndef EMIT_
+#ifndef EMIT
+#define EMIT(_name, _op) \
+    extern void _name (struct spe_function *p);
 #define EMIT_(_name, _op) \
     extern void _name (struct spe_function *p, unsigned rT);
 #define EMIT_R(_name, _op) \
@@ -129,7 +131,7 @@ extern void spe_comment(struct spe_function *p, int rel_indent, const char *s);
 #define EMIT_I16(_name, _op) \
     extern void _name (struct spe_function *p, int imm);
 #define UNDEF_EMIT_MACROS
-#endif /* EMIT_ */
+#endif /* EMIT */
 
 
 /* Memory load / store instructions
@@ -294,6 +296,10 @@ EMIT_RI16(spe_brz,       0x040)
 EMIT_RI16(spe_brhnz,     0x046)
 EMIT_RI16(spe_brhz,      0x044)
 
+/* Control instructions
+ */
+EMIT     (spe_lnop,      0x001)
+
 extern void
 spe_lqd(struct spe_function *p, unsigned rT, unsigned rA, int offset);
 
@@ -418,6 +424,7 @@ EMIT_R   (spe_wrch,       0x10d)
 
 
 #ifdef UNDEF_EMIT_MACROS
+#undef EMIT
 #undef EMIT_
 #undef EMIT_R
 #undef EMIT_RR
diff --git a/src/gallium/auxiliary/util/u_memory.h b/src/gallium/auxiliary/util/u_memory.h
index 857102719dc..1a6b596421f 100644
--- a/src/gallium/auxiliary/util/u_memory.h
+++ b/src/gallium/auxiliary/util/u_memory.h
@@ -151,6 +151,8 @@ REALLOC( void *old_ptr, unsigned old_size, unsigned new_size )
 
 #define CALLOC_STRUCT(T)   (struct T *) CALLOC(1, sizeof(struct T))
 
+#define CALLOC_VARIANT_LENGTH_STRUCT(T,more_size)   ((struct T *) CALLOC(1, sizeof(struct T) + more_size))
+
 
 /**
  * Return memory on given byte alignment