[NWFPE] performance improvements [Part 6]

NWFPE performance improvements, part 6.  Builds on patch 1468/1.

Break NWFPE initialization into separate function and call it from
entry.S as appropriate.  Also avoid nRc variable usage in EmulateAll().
This buys 6% speedup in NWFPE.  (Note that this is over half of the
total speedup between original RMK and the netwinder cvs version).

(note that 2.5 already has the separate initialisation, and this
cset actually makes 2.5 conform to 2.4 function naming for this
part. --rmk)

arch/arm/nwfpe/ChangeLog

@@ -3,7 +3,8 @@
 	  Used "indent -kr -i8 -ts8 -sob -l132 -ss" and a few manual fixups.
 	* Removed dead code and fixed function protypes to match definitions.
 	* Consolidated use of (opcode && MASK_ARITHMETIC_OPCODE) >> 20.
-	* Make 80-bit precision a compile-time option.
+	* Make 80-bit precision a compile-time option. (1%)
+	* Only initialize FPE state once in repeat-FP situations. (6%)
 
 2002-01-19  Russell King <rmk@arm.linux.org.uk>
 

arch/arm/nwfpe/fpa11.c

@@ -94,9 +94,12 @@ void SetRoundingPrecision(const unsigned int opcode)
 #endif
 }
 
-void nwfpe_init(union fp_state *fp)
+void nwfpe_init_fpa(union fp_state *fp)
 {
 	FPA11 *fpa11 = (FPA11 *)fp;
+#ifdef NWFPE_DEBUG
+	printk("NWFPE: setting up state.\n");
+#endif
  	memset(fpa11, 0, sizeof(FPA11));
 	resetFPA11();
 	SetRoundingMode(ROUND_TO_NEAREST);
@@ -107,8 +110,11 @@ void nwfpe_init(union fp_state *fp)
 /* Emulate the instruction in the opcode. */
 unsigned int EmulateAll(unsigned int opcode)
 {
-	unsigned int nRc = 1, code;
+	unsigned int code;
 
+#ifdef NWFPE_DEBUG
+	printk("NWFPE: emulating opcode %08x\n", opcode);
+#endif
 	code = opcode & 0x00000f00;
 	if (code == 0x00000100 || code == 0x00000200) {
 		/* For coprocessor 1 or 2 (FPA11) */
@@ -118,21 +124,19 @@ unsigned int EmulateAll(unsigned int opcode)
 				/* Emulate conversion opcodes. */
 				/* Emulate register transfer opcodes. */
 				/* Emulate comparison opcodes. */
-				nRc = EmulateCPRT(opcode);
+				return EmulateCPRT(opcode);
 			} else {
 				/* Emulate monadic arithmetic opcodes. */
 				/* Emulate dyadic arithmetic opcodes. */
-				nRc = EmulateCPDO(opcode);
+				return EmulateCPDO(opcode);
 			}
 		} else if (code == 0x0c000000) {
 			/* Emulate load/store opcodes. */
 			/* Emulate load/store multiple opcodes. */
-			nRc = EmulateCPDT(opcode);
-		} else {
-			/* Invalid instruction detected.  Return FALSE. */
-			nRc = 0;
+			return EmulateCPDT(opcode);
 		}
 	}
 
-	return (nRc);
+	/* Invalid instruction detected.  Return FALSE. */
+	return 0;
 }

arch/arm/nwfpe/fpa11.h

@@ -87,6 +87,6 @@ typedef struct tagFPA11 {
 
 extern void SetRoundingMode(const unsigned int);
 extern void SetRoundingPrecision(const unsigned int);
-extern void nwfpe_init(union fp_state *fp);
+extern void nwfpe_init_fpa(union fp_state *fp);
 
 #endif

arch/arm/nwfpe/fpmodule.c

@@ -98,7 +98,7 @@ static int __init fpe_init(void)
 	orig_fp_enter = kern_fp_enter;
 	orig_fp_init = fp_init;
 	kern_fp_enter = nwfpe_enter;
-	fp_init = nwfpe_init;
+	fp_init = nwfpe_init_fpa;
 
 	return 0;
 }