From 9bb92eba9e1b3298d77aecaf7919c57b18255790 Mon Sep 17 00:00:00 2001
From: Robin Watts
Date: Sat, 17 Apr 2010 15:45:06 +0000
Subject: Fix for ARM Normal2xAspect scaler that was causing crashes when the
 screen 'shakes' (such as when the head/bomb goes of at the start of
 sam'n'max). Was overwriting the end of the screen when asked to plot a
 non-multiple of 5 lines.

svn-id: r48690
---
 graphics/scaler/Normal2xARM.s | 138 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 138 insertions(+)

diff --git a/graphics/scaler/Normal2xARM.s b/graphics/scaler/Normal2xARM.s
index 2dfb9e703d..e75d917567 100644
--- a/graphics/scaler/Normal2xARM.s
+++ b/graphics/scaler/Normal2xARM.s
@@ -112,6 +112,8 @@ Normal2xAspectMask:
 	SUB	r14,r14,r3, LSL #3
 	SUB	r14,r14,r3, LSL #1
 	SUB	r14,r14,r3		@ r14 = 4-dstPitch*11
+	SUBS	r5, r5, #4
+	BLE	last_few_lines
 yloop_aspect:
 xloop_aspect:
 	LDRH	r6, [r0], r1		@ r6 = A
@@ -170,5 +172,141 @@ xloop_aspect:
 	SUB	r2, r2, r4, LSL #2
 	SUBS	r5, r5, #5
 	BGT	yloop_aspect
+last_few_lines:
+	@ r5 = -4 => done, -3 => 1 left, -2 = 2 left, -1 = 3 left, 0 = 4 left
+	ADDS	r5, r5, #3
+	@ r5 = -1 => done, 0 => 1 left, 1 = 2 left, 2 = 3 left, 3 = 4 left
+	LDMLTFD	r13!,{r4-r11,PC}
+	BEQ	last1lines
+	CMP	r5, #2
+	BEQ	last3lines
+	BGT	last4lines
+last2lines:
+	ADD	r11,r11,r1
+	ADD	r14,r14,r3
+	ADD	r11,r11,r1, LSL #1
+	ADD	r14,r14,r3, LSL #3
+xloop_aspect2:
+	LDRH	r6, [r0], r1		@ r6 = A
+	LDRH	r7, [r0], r11		@ r7 = B
+	SUBS	r4, r4, #1
+	ORR	r6, r6, r6, LSL #16	@ r6 = output 0, 1
+	ORR	r7, r7, r7, LSL #16	@ r7 = output 3
+	STR	r6, [r2], r3		@ output 0 (A)
+	STR	r6, [r2], r3		@ output 1 (A)
+	AND	r6, r6, r12		@ r6 = A split
+	ADD	r6, r6, r6, LSL #1	@ r6 = A*3
+	STR	r7, [r2, r3]		@ output 3 (B)
+	AND	r7, r7, r12		@ r7 = B split
+	ADD	r6, r6, r7		@ r6 = A*3 + B
+	ADD	r6, r6, r7, LSL #2	@ r6 = A*3 + B*5
+	AND	r6, r12,r6, LSR #3	@ r6 = (A*3 + B*5)>>3
+	ORR	r6, r6, r6, ROR #16	@ r6 = output 2
+	STR	r6, [r2], r14		@ output 2 (A*3+B*5)>>3
+	BGT	xloop_aspect2
+
+	LDMFD	r13!,{r4-r11,PC}
+
+last1lines:
+xloop_aspect1:
+	LDRH	r6, [r0], #2		@ r6 = A
+	ORR	r6, r6, r6, LSL #16	@ r6 = output 0, 1
+	STR	r6, [r2, r3]		@ output 0 (A)
+	STR	r6, [r2], #4		@ output 1 (A)
+	SUBS	r4, r4, #1
+	BGT	xloop_aspect1
+
+	LDMFD	r13!,{r4-r11,PC}
+
+last3lines:
+	ADD	r14,r14,r3, LSL #2
+	ADD	r11,r11,r1
+xloop_aspect3:
+	LDRH	r6, [r0], r1		@ r6 = A
+	LDRH	r7, [r0], r1		@ r7 = B
+	LDRH	r8, [r0], r1		@ r8 = C
+	LDRH	r9, [r0], r11		@ r9 = D
+	ORR	r6, r6, r6, LSL #16	@ r6 = output 0, 1
+	ORR	r7, r7, r7, LSL #16	@ r7 = output 3
+	ORR	r8, r8, r8, LSL #16	@ r8 = output 5,6
+	ORR	r9, r9, r9, LSL #16	@ r9 = output 8
+	STR	r6, [r2], r3		@ output 0 (A)
+	STR	r6, [r2], r3		@ output 1 (A)
+	AND	r6, r6, r12		@ r6 = A split
+	ADD	r6, r6, r6, LSL #1	@ r6 = A*3
+	STR	r7, [r2, r3]		@ output 3 (B)
+	AND	r7, r7, r12		@ r7 = B split
+	ADD	r6, r6, r7		@ r6 = A*3 + B
+	ADD	r6, r6, r7, LSL #2	@ r6 = A*3 + B*5
+	AND	r6, r12,r6, LSR #3	@ r6 = (A*3 + B*5)>>3
+	ORR	r6, r6, r6, ROR #16	@ r6 = output 2
+	STR	r6, [r2], r3, LSL #1	@ output 2 (A*3+B*5)>>3
+	RSB	r7, r7, r7, LSL #3	@ r7 = B*7
+	AND	r6, r8, r12		@ r6 = C split
+	ADD	r7, r7, r6		@ r7 = B*7+C
+	AND	r7, r12,r7, LSR #3	@ r7 = (B*7 + C)>>3
+	ORR	r7, r7, r7, ROR #16	@ r7 = output 4
+	STR	r7, [r2], r3		@ output 4 (B*7+C)>>3
+	STR	r8, [r2], r3		@ output 5 (C)
+	STR	r8, [r2], r3		@ output 6 (C)
+	AND	r9, r9, r12		@ r9 = D split
+	RSB	r7, r9, r9, LSL #3	@ r7 = D*7
+	ADD	r6, r6, r7		@ r6 = C+D*7
+	AND	r6, r12,r6, LSR #3	@ r6 = (C + D*7)>>3
+	ORR	r6, r6, r6, ROR #16	@ r6 = output 7
+	STR	r6, [r2], r14		@ output 7 (C+D*7)>>3
+	SUBS	r4, r4, #1
+	BGT	xloop_aspect3
+
+	LDMFD	r13!,{r4-r11,PC}
+
+last4lines:
+	ADD	r14,r14,r3, LSL #1
+xloop_aspect4:
+	LDRH	r6, [r0], r1		@ r6 = A
+	LDRH	r7, [r0], r1		@ r7 = B
+	LDRH	r8, [r0], r1		@ r8 = C
+	LDRH	r9, [r0], r1		@ r9 = D
+	LDRH	r10,[r0], r11		@ r10= E
+	ORR	r6, r6, r6, LSL #16	@ r6 = output 0, 1
+	ORR	r7, r7, r7, LSL #16	@ r7 = output 3
+	ORR	r8, r8, r8, LSL #16	@ r8 = output 5,6
+	ORR	r9, r9, r9, LSL #16	@ r9 = output 8
+	ORR	r10,r10,r10,LSL #16	@ r10= output 10, 11
+	STR	r6, [r2], r3		@ output 0 (A)
+	STR	r6, [r2], r3		@ output 1 (A)
+	AND	r6, r6, r12		@ r6 = A split
+	ADD	r6, r6, r6, LSL #1	@ r6 = A*3
+	STR	r7, [r2, r3]		@ output 3 (B)
+	AND	r7, r7, r12		@ r7 = B split
+	ADD	r6, r6, r7		@ r6 = A*3 + B
+	ADD	r6, r6, r7, LSL #2	@ r6 = A*3 + B*5
+	AND	r6, r12,r6, LSR #3	@ r6 = (A*3 + B*5)>>3
+	ORR	r6, r6, r6, ROR #16	@ r6 = output 2
+	STR	r6, [r2], r3, LSL #1	@ output 2 (A*3+B*5)>>3
+	RSB	r7, r7, r7, LSL #3	@ r7 = B*7
+	AND	r6, r8, r12		@ r6 = C split
+	ADD	r7, r7, r6		@ r7 = B*7+C
+	AND	r7, r12,r7, LSR #3	@ r7 = (B*7 + C)>>3
+	ORR	r7, r7, r7, ROR #16	@ r7 = output 4
+	STR	r7, [r2], r3		@ output 4 (B*7+C)>>3
+	STR	r8, [r2], r3		@ output 5 (C)
+	STR	r8, [r2], r3		@ output 6 (C)
+	STR	r9, [r2, r3]		@ output 8 (D)
+	AND	r9, r9, r12		@ r9 = D split
+	RSB	r7, r9, r9, LSL #3	@ r7 = D*7
+	ADD	r6, r6, r7		@ r6 = C+D*7
+	AND	r6, r12,r6, LSR #3	@ r6 = (C + D*7)>>3
+	ORR	r6, r6, r6, ROR #16	@ r6 = output 7
+	STR	r6, [r2], r3, LSL #1	@ output 7 (C+D*7)>>3
+	ADD	r9, r9, r9, LSL #2	@ r9 = D*5
+	AND	r6, r10,r12		@ r6 = E split
+	ADD	r9, r9, r6		@ r9 = D*5+E
+	ADD	r9, r9, r6, LSL #1	@ r9 = D*5+E*3
+	AND	r9, r12,r9, LSR #3	@ r9 = (D*5 + E*3)>>3
+	ORR	r9, r9, r9, ROR #16	@ r9 = output 9
+	STR	r9, [r2], r14		@ output 9 (D*5+E*3)>>3
+	SUBS	r4, r4, #1
+	BGT	xloop_aspect4
 
 	LDMFD	r13!,{r4-r11,PC}
-- 
cgit v1.2.3