drm/gr/nvc0-: merge nvc0/nve0 ucode, and use cpp instead of m4

No code changes, proven by envyas producing identical binaries.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>

drivers/gpu/drm/nouveau/core/engine/graph/fuc/nvc0.fuc → drivers/gpu/drm/nouveau/core/engine/graph/fuc/com.fuc

@@ -23,42 +23,7 @@
  * Authors: Ben Skeggs
  */
 
-define(`mmctx_data', `.b32 eval((($2 - 1) << 26) | $1)')
-define(`queue_init', `.skip eval((2 * 4) + ((8 * 4) * 2))')
-
-ifdef(`include_code', `
-// Error codes
-define(`E_BAD_COMMAND', 0x01)
-define(`E_CMD_OVERFLOW', 0x02)
-
-// Util macros to help with debugging ucode hangs etc
-define(`T_WAIT', 0)
-define(`T_MMCTX', 1)
-define(`T_STRWAIT', 2)
-define(`T_STRINIT', 3)
-define(`T_AUTO', 4)
-define(`T_CHAN', 5)
-define(`T_LOAD', 6)
-define(`T_SAVE', 7)
-define(`T_LCHAN', 8)
-define(`T_LCTXH', 9)
-
-define(`trace_set', `
-	mov $r8 0x83c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-define(`trace_clr', `
-	mov $r8 0x85c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
+#ifdef INCLUDE_CODE
 // queue_put - add request to queue
 //
 // In : $r13 queue pointer
@@ -178,27 +143,41 @@ watchdog_clear:
 	iowr I[$r8 + 0x000] $r0
 	ret
 
-// wait_done{z,o} - wait on FUC_DONE bit to become clear/set
+// wait_donez - wait on FUC_DONE bit to become clear
+//
+// In : $r10 bit to wait on
+//
+wait_donez:
+	trace_set(T_WAIT);
+	mov $r8 0x818
+	shl b32 $r8 6
+	iowr I[$r8 + 0x000] $r10
+	wait_donez_ne:
+		mov $r8 0x400
+		shl b32 $r8 6
+		iord $r8 I[$r8 + 0x000]
+		xbit $r8 $r8 $r10
+		bra ne #wait_donez_ne
+	trace_clr(T_WAIT)
+	ret
+
+// wait_doneo - wait on FUC_DONE bit to become set
 //
 // In : $r10 bit to wait on
 //
-define(`wait_done', `
-$1:
+wait_doneo:
 	trace_set(T_WAIT);
 	mov $r8 0x818
 	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r10	// CC_SCRATCH[6] = wait bit
-	wait_done_$1:
+	iowr I[$r8 + 0x000] $r10
+	wait_doneo_e:
 		mov $r8 0x400
 		shl b32 $r8 6
-		iord $r8 I[$r8 + 0x000]	// DONE
+		iord $r8 I[$r8 + 0x000]
 		xbit $r8 $r8 $r10
-		bra $2 #wait_done_$1
+		bra e #wait_doneo_e
 	trace_clr(T_WAIT)
 	ret
-')
-wait_done(wait_donez, ne)
-wait_done(wait_doneo, e)
 
 // mmctx_size - determine size of a mmio list transfer
 //
@@ -397,4 +376,4 @@ strand_ctx_init:
 	sub b32 $r15 $r14 $r15
 	trace_clr(T_STRINIT)
 	ret
-')
+#endif