From caae53a25a15983e98936f9b6f182b49d3062ccf Mon Sep 17 00:00:00 2001
From: Robin Watts
Date: Wed, 15 Jul 2009 20:51:31 +0000
Subject: New ARM code blending Smartphone landscale scaler for WinCE port.
 Seems to work in my mockup - I don't have a smartphone resolution device to
 test it.

svn-id: r42516
---
 backends/platform/wince/CEScaler.cpp     |  25 ++++-
 backends/platform/wince/Makefile         |   1 +
 backends/platform/wince/smartLandScale.s | 187 +++++++++++++++++++++++++++++++
 backends/platform/wince/wince-sdl.cpp    |   5 +
 4 files changed, 216 insertions(+), 2 deletions(-)
 create mode 100755 backends/platform/wince/smartLandScale.s

(limited to 'backends')

diff --git a/backends/platform/wince/CEScaler.cpp b/backends/platform/wince/CEScaler.cpp
index 4cc675e006..bfdb74319d 100644
--- a/backends/platform/wince/CEScaler.cpp
+++ b/backends/platform/wince/CEScaler.cpp
@@ -53,6 +53,15 @@ void PocketPCPortraitTemplate(const uint8 *srcPtr, uint32 srcPitch, uint8 *dstPt
 }
 MAKE_WRAPPER(PocketPCPortrait)
 
+void PocketPCRawPortrait(const uint8 *srcPtr, uint32 srcPitch, uint8 *dstPtr, uint32 dstPitch, int width, int height) {
+
+	while (height--) {
+		memcpy(dstPtr, srcPtr, width*sizeof(uint16_t));
+		srcPtr += srcPitch;
+		dstPtr += dstPitch;
+	}
+}
+
 // Our version of an aspect scaler. Main difference is the out-of-place
 // operation, omitting a straight blit step the sdl backend does. Also,
 // tests show unaligned access errors with the stock aspect scaler.
@@ -117,6 +126,7 @@ void PocketPCLandscapeAspect(const uint8 *srcPtr, uint32 srcPitch, uint8 *dstPtr
 #ifdef ARM
 extern "C" {
 	void PocketPCHalfARM(const uint8 *srcPtr, uint32 srcPitch, uint8 *dstPtr, uint32 dstPitch, int width, int height, int mask, int round);
+	void SmartphoneLandscapeARM(const uint8 *srcPtr, uint32 srcPitch, uint8 *dstPtr, uint32 dstPitch, int width, int height, int mask);
 	// Rounding constants and masks used for different pixel formats
 	int roundingconstants[] = { 0x00200802, 0x00201002 };
 	int redbluegreenMasks[] = { 0x03E07C1F, 0x07E0F81F };
@@ -156,7 +166,7 @@ void PocketPCHalf(const uint8 *srcPtr, uint32 srcPitch, uint8 *dstPtr, uint32 ds
 	if (gBitFormat == 565)
 		PocketPCHalfTemplate<565>(srcPtr, srcPitch, dstPtr, dstPitch, width, height);
 	else
-		PocketPCHalfTemplate<565>(srcPtr, srcPitch, dstPtr, dstPitch, width, height);
+		PocketPCHalfTemplate<555>(srcPtr, srcPitch, dstPtr, dstPitch, width, height);
 #endif
 }
 
@@ -217,4 +227,15 @@ void SmartphoneLandscapeTemplate(const uint8 *srcPtr, uint32 srcPitch, uint8 *ds
 		}
 	}
 }
-MAKE_WRAPPER(SmartphoneLandscape)
+
+void SmartphoneLandscape(const uint8 *srcPtr, uint32 srcPitch, uint8 *dstPtr, uint32 dstPitch, int width, int height) {
+#ifdef ARM
+	int maskUsed = (gBitFormat == 565);
+	SmartphoneLandscapeARM(srcPtr, srcPitch, dstPtr, dstPitch, width, height, redbluegreenMasks[maskUsed]);
+#else
+	if (gBitFormat == 565)
+		SmartphoneLandscape<565>(srcPtr, srcPitch, dstPtr, dstPitch, width, height);
+	else
+		SmartphoneLandscape<555>(srcPtr, srcPitch, dstPtr, dstPitch, width, height);
+#endif
+}
diff --git a/backends/platform/wince/Makefile b/backends/platform/wince/Makefile
index 532bc69ef4..1846029469 100644
--- a/backends/platform/wince/Makefile
+++ b/backends/platform/wince/Makefile
@@ -186,6 +186,7 @@ OBJS += $(srcdir)/gui/Actions.o $(srcdir)/gui/Key.o $(srcdir)/gui/KeysDialog.o
 OBJS += ../sdl/sdl.o ../sdl/graphics.o ../sdl/events.o ../sdl/hardwarekeys.o
 OBJS += missing/missing.o
 OBJS += ARMscaler.o
+OBJS += smartLandScale.o
 ifndef DYNAMIC_MODULES
 OBJS += PocketSCUMM.o
 endif
diff --git a/backends/platform/wince/smartLandScale.s b/backends/platform/wince/smartLandScale.s
new file mode 100755
index 0000000000..deeb093329
--- /dev/null
+++ b/backends/platform/wince/smartLandScale.s
@@ -0,0 +1,187 @@
+@ ScummVM Scumm Interpreter
+@ Copyright (C) 2009 The ScummVM project
+@
+@ This program is free software; you can redistribute it and/or
+@ modify it under the terms of the GNU General Public License
+@ as published by the Free Software Foundation; either version 2
+@ of the License, or (at your option) any later version.
+@
+@ This program is distributed in the hope that it will be useful,
+@ but WITHOUT ANY WARRANTY; without even the implied warranty of
+@ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+@ GNU General Public License for more details.
+@
+@ You should have received a copy of the GNU General Public License
+@ along with this program; if not, write to the Free Software
+@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+@
+@ $URL$
+@ $Id$
+@
+@ @author Robin Watts (robin@wss.co.uk)
+
+	@ For 16 source pixels 0123456789ABCDEF, we want to produce 11 output
+	@ pixels.
+
+	@0000000000011111111111222222222223333333333344444444444555555555
+	@<------||------><------||------><------||------><------||------>
+
+	@5566666666666777777777778888888888899999999999AAAAAAAAAAABBBBBBB
+	@<------||------><------||------><------||------><------||------>
+
+	@BBBBCCCCCCCCCCCDDDDDDDDDDDEEEEEEEEEEEFFFFFFFFFFF
+	@<------||------><------||------><------||------>
+
+	@ So, use the following weights (approximately right)
+
+	@ d0 = (3*s0 + 1*s1)>>2                Every source pixel constitutes
+	@ d1 = (2*s1 + 2*s2)>>2                3/4 of a destination pixel,
+	@ d2 = (1*s2 + 3*s3)>>2                except for s4,s5,sA and sB which
+	@ d3 = (2*s4 + 2*s5)>>2                constitute 1/2 each.
+	@ d4 = (3*s6 + 1*s7)>>2
+	@ d5 = (2*s7 + 2*s8)>>2
+	@ d6 = (1*s8 + 3*s9)>>2
+	@ d7 = (2*sA + 2*sB)>>2
+	@ d8 = (3*sC + 1*sD)>>2
+	@ d9 = (2*sD + 2*sE)>>2
+	@ dA = (1*sE + 3*sF)>>2
+
+	.text
+
+	.global	SmartphoneLandscapeARM
+
+	@ scales a width x height block of 16bpp pixels from srcPtr to dstPtr,
+	@ scaling each scanline down by 11/16ths. Every 8th scanline is dropped
+	@ srcPitch and dstPitch identify how to reach subsequent lines.
+	@ mask allows for one routine to do both 565 and 565 formats.
+
+SmartphoneLandscapeARM:
+	@ r0 = srcPtr
+	@ r1 = srcSpan
+	@ r2 = dstPtr
+	@ r3 = dstSpan
+	@ <> = width
+	@ <> = height
+	@ <> = mask
+	MOV	r12,r13
+	STMFD	r13!,{r4-r11,r14}
+	LDMFD	r12,{r4,r5,r11}		@ r4 = width
+					@ r5 = height
+					@ r11= mask
+	MOV	r7, #7			@ r7 = line
+	SUB	r8, r1, r4, LSL #1	@ r8 = srcSpan - width*2
+y_loop:
+	MOV	r6, r4			@ r6 = i
+	MOV	r9, r2			@ r9 = dstPtr
+x_loop:
+	LDRH	r14,[r0],#2		@ r14 = s0
+	LDRH	r12,[r0],#2		@ r12 = s1
+	LDRH	r10,[r0],#2		@ r10 = s2
+	ORR	r14,r14,r14,LSL #16	@ r14 = s0s0
+	ORR	r12,r12,r12,LSL #16	@ r12 = s1s1
+	AND	r14,r14,r11		@ r14 = s0 as g_b_r
+	AND	r12,r12,r11		@ r12 = s1 as g_b_r
+	ADD	r14,r14,r14,LSL #1	@ r14 = s0*3 as g_b_r
+	ORR	r10,r10,r10,LSL #16	@ r10 = s2s2
+	ADD	r14,r14,r12		@ r14 = (s0*3 + s1) as g_b_r
+	AND	r10,r10,r11		@ r10 = s2 as g_b_r
+	AND	r14,r11,r14,LSR #2	@ r14 = d0 as g_b_r
+	ORR	r14,r14,r14,LSR #16	@ r14 = d0
+	STRH	r14,[r9],#2		@ store d0
+	ADD	r12,r12,r10		@ r12 = (s1 + s2) as g_b_r
+	LDRH	r14,[r0],#2		@ r14 = s3
+	AND	r12,r11,r12,LSR #1	@ r12 = d1 as g_b_r
+	ORR	r12,r12,r12,LSR #16	@ r12 = d1
+	STRH	r12,[r9],#2		@ store d1
+	ORR	r14,r14,r14,LSL #16	@ r14 = s3s3
+	AND	r14,r14,r11		@ r14 = s3 as g_b_r
+	ADD	r10,r10,r14		@ r10 = (s2 + s3) as g_b_r
+	ADD	r10,r10,r14,LSL #1	@ r10 = (s2 + s3*3) as g_b_r
+	LDRH	r14,[r0],#2		@ r14 = s4
+	LDRH	r12,[r0],#2		@ r12 = s5
+	AND	r10,r11,r10,LSR #2	@ r10 = d2 as g_b_r
+	ORR	r10,r10,r10,LSR #16	@ r10 = d2
+	STRH	r10,[r9],#2		@ store d2
+	ORR	r14,r14,r14,LSL #16	@ r14 = s4s4
+	ORR	r12,r12,r12,LSL #16	@ r12 = s5s5
+	AND	r14,r14,r11		@ r14 = s4 as g_b_r
+	AND	r12,r12,r11		@ r12 = s5 as g_b_r
+	ADD	r14,r14,r12		@ r14 = (s4 + s5) as g_b_r
+	LDRH	r12,[r0],#2		@ r12 = s6
+	LDRH	r10,[r0],#2		@ r10 = s7
+	AND	r14,r11,r14,LSR #1	@ r14 = d3 as g_b_r
+	ORR	r14,r14,r14,LSR #16	@ r14 = d3
+	STRH	r14,[r9],#2		@ store d3
+	ORR	r12,r12,r12,LSL #16	@ r12 = s6s6
+	ORR	r10,r10,r10,LSL #16	@ r10 = s7s7
+	LDRH	r14,[r0],#2		@ r14 = s8
+	AND	r12,r12,r11		@ r12 = s6 as g_b_r
+	AND	r10,r10,r11		@ r10 = s7 as g_b_r
+	ORR	r14,r14,r14,LSL #16	@ r14 = s8s8
+	ADD	r12,r12,r12,LSL #1	@ r12 = 3*s6 as g_b_r
+	AND	r14,r14,r11		@ r14 = s8 as g_b_r
+	ADD	r12,r12,r10		@ r12 = (3*s6+s7) as g_b_r
+	AND	r12,r11,r12,LSR #2	@ r12 = d4 as g_b_r
+	ORR	r12,r12,r12,LSR #16	@ r12 = d4
+	STRH	r12,[r9],#2		@ store d4
+	ADD	r10,r10,r14		@ r10 = (s7+s8) as g_b_r
+	AND	r10,r11,r10,LSR #1	@ r10 = d5 as g_b_r
+	LDRH	r12,[r0],#2		@ r12 = s9
+	ORR	r10,r10,r10,LSR #16	@ r10 = d5
+	STRH	r10,[r9],#2		@ store d5
+	ORR	r12,r12,r12,LSL #16	@ r12 = s9s9
+	AND	r12,r12,r11		@ r12 = s9 as g_b_r
+	ADD	r12,r12,r12,LSL #1	@ r12 = s9*3 as g_b_r
+	ADD	r12,r12,r14		@ r12 = (s8+s9*3) as g_b_r
+	AND	r12,r11,r12,LSR #2	@ r12 = d6 as g_b_r
+	LDRH	r14,[r0],#2		@ r14 = sA
+	LDRH	r10,[r0],#2		@ r10 = sB
+	ORR	r12,r12,r12,LSR #16	@ r12 = d6
+	STRH	r12,[r9],#2		@ store d6
+	ORR	r14,r14,r14,LSL #16	@ r14 = sAsA
+	ORR	r10,r10,r10,LSL #16	@ r10 = sBsB
+	LDRH	r12,[r0],#2		@ r12 = sC
+	AND	r14,r14,r11		@ r14 = sA as g_b_r
+	AND	r10,r10,r11		@ r10 = sB as g_b_r
+	ORR	r12,r12,r12,LSL #16	@ r12 = sCsC
+	ADD	r14,r14,r10		@ r14 = (sA + sB) as g_b_r
+	LDRH	r10,[r0],#2		@ r10 = sD
+	AND	r14,r11,r14,LSR #1	@ r14 = d7 as g_b_r
+	AND	r12,r12,r11		@ r12 = sC as g_b_r
+	ORR	r14,r14,r14,LSR #16	@ r14 = d7
+	ORR	r10,r10,r10,LSL #16	@ r10 = sDsD
+	STRH	r14,[r9],#2		@ store d7
+	AND	r10,r10,r11		@ r10 = sD as g_b_r
+	ADD	r12,r12,r12,LSL #1	@ r12 = 3*sC as g_b_r
+	LDRH	r14,[r0],#2		@ r14 = sE
+	ADD	r12,r12,r10		@ r12 = (3*sC+sD) as g_b_r
+	AND	r12,r11,r12,LSR #2	@ r12 = d8 as g_b_r
+	ORR	r14,r14,r14,LSL #16	@ r14 = sEsE
+	ORR	r12,r12,r12,LSR #16	@ r12 = d8
+	AND	r14,r14,r11		@ r14 = sE as g_b_r
+	STRH	r12,[r9],#2		@ store d8
+	ADD	r10,r10,r14		@ r10 = (sD+sE) as g_b_r
+	LDRH	r12,[r0],#2		@ r12 = sF
+	AND	r10,r11,r10,LSR #1	@ r10 = d9 as g_b_r
+	ORR	r10,r10,r10,LSR #16	@ r10 = d9
+	STRH	r10,[r9],#2		@ store d9
+	ORR	r12,r12,r12,LSL #16	@ r12 = sFsF
+	AND	r12,r12,r11		@ r12 = sF as g_b_r
+	ADD	r12,r12,r12,LSL #1	@ r12 = 3*sF as g_b_r
+	ADD	r12,r12,r14		@ r12 = (sE+3*sF) as g_b_r
+	AND	r12,r11,r12,LSR #2	@ r12 = dA as g_b_r
+	ORR	r12,r12,r12,LSR #16	@ r12 = dA
+	SUBS	r6,r6,#16		@ width -= 16
+	STRH	r12,[r9],#2		@ store dA
+	BGT	x_loop
+	
+	ADD	r0, r0, r8		@ srcPtr += srcSpan
+	ADD	r2, r2, r3		@ dstPtr += dstSpan
+	SUBS	r7, r7, #1
+	ADDEQ	r0, r0, r1
+	MOVEQ	r7, #7
+	SUBEQ	r5, r5, #1
+	SUBS	r5, r5, #1
+	BGT	y_loop
+
+	LDMFD	r13!,{r4-r11,PC}
diff --git a/backends/platform/wince/wince-sdl.cpp b/backends/platform/wince/wince-sdl.cpp
index 9a4b312631..8b2e0848f8 100644
--- a/backends/platform/wince/wince-sdl.cpp
+++ b/backends/platform/wince/wince-sdl.cpp
@@ -1209,8 +1209,13 @@ bool OSystem_WINCE3::update_scalers() {
 	if (CEDevice::hasSmartphoneResolution()) {
 		if (_videoMode.screenWidth > 320)
 			error("Game resolution not supported on Smartphone");
+#ifdef ARM
+		_scaleFactorXm = 11;
+		_scaleFactorXd = 16;
+#else
 		_scaleFactorXm = 2;
 		_scaleFactorXd = 3;
+#endif
 		_scaleFactorYm = 7;
 		_scaleFactorYd = 8;
 		_scalerProc = SmartphoneLandscape;
-- 
cgit v1.2.3