/*************************************************************************** * __________ __ ___. * Open \______ \ ____ ____ | | _\_ |__ _______ ___ * Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ / * Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < < * Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \ * \/ \/ \/ \/ \/ * $Id$ * * Copyright (C) 2007 by Andree Buschmann * * All files in this archive are subject to the GNU General Public License. * See the file COPYING in the source tree root for full license agreement. * * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY * KIND, either express or implied. * ****************************************************************************/ /**************************************************************************** * void lcd_write_yuv_420_lines(unsigned char const * const src[3], * int width, * int stride); * * |R| |1.000000 -0.000001 1.402000| |Y'| * |G| = |1.000000 -0.334136 -0.714136| |Pb| * |B| |1.000000 1.772000 0.000000| |Pr| * Scaled, normalized, rounded and tweaked to yield RGB 565: * |R| |74 0 101| |Y' - 16| >> 9 * |G| = |74 -24 -51| |Cb - 128| >> 8 * |B| |74 128 0| |Cr - 128| >> 9 */ .section .icode, "ax", %progbits .align 2 .global lcd_write_yuv420_lines .type lcd_write_yuv420_lines, %function lcd_write_yuv420_lines: /* r0 = yuv_src */ /* r1 = width */ /* r2 = stride */ stmfd sp!, { r4-r12 } /* save non-scratch */ ldmia r0, { r10, r11, r12 } /* r10 = yuv_src[0] = Y'_p */ /* r11 = yuv_src[1] = Cb_p */ /* r12 = yuv_src[2] = Cr_p */ add r2, r10, r2 /* r2 = &ysrc[stride] */ add r3, r1, r1, asl #1 /* number of bytes for chroma buffer */ add r3, r3, #15 /* plus room for 3 additional words, */ bic r3, r3, #3 /* rounded up to multiples of 4 byte */ sub sp, sp, r3 /* and allocate on stack */ stmia sp, {r1, r2, r3} /* width, &ysrc[stride], stack_alloc */ mov r7, r1 /* r7 = loop count */ add r8, sp, #12 /* chroma buffer */ mov r9, #0x30000000 /* LCD data port */ /* 1st loop start */ 10: /* loop start */ ldrb r0, [r11], #1 /* r0 = *usrc++ = *Cb_p++ */ ldrb r1, [r12], #1 /* r1 = *vsrc++ = *Cr_p++ */ sub r0, r0, #128 /* r0 = Cb-128 */ sub r1, r1, #128 /* r1 = Cr-128 */ add r3, r1, r1, asl #1 /* r3 = Cr*51 + Cb*24 */ add r3, r3, r3, asl #4 add r3, r3, r0, asl #3 add r3, r3, r0, asl #4 add r4, r1, r1, asl #2 /* r1 = Cr*101 */ add r4, r4, r1, asl #5 add r1, r4, r1, asl #6 add r1, r1, #256 /* r1 = rv = (r1 + 256) >> 9 */ mov r1, r1, asr #9 strb r1, [r8], #1 /* store r1 to chroma_buf */ rsb r3, r3, #128 /* r3 = guv = (-r3 + 128) >> 8 */ mov r3, r3, asr #8 strb r3, [r8], #1 /* store r3 to chroma_buf */ add r0, r0, #2 /* r0 = bu = (Cb*128 + 256) >> 9 */ mov r0, r0, asr #2 strb r0, [r8], #1 /* store r0 to chroma_buf */ /* 1st loop, first pixel */ ldrb r5, [r10], #1 /* r5 = *ysrc++ = *Y'_p++ */ sub r5, r5, #16 /* r5 = (Y'-16) * 74 */ add r2, r5, r5, asl #2 add r5, r2, r5, asl #5 add r6, r1, r5, asr #8 /* r6 = r = (Y >> 9) + rv */ add r2, r3, r5, asr #7 /* r2 = g = (Y >> 8) + guv */ add r4, r0, r5, asr #8 /* r4 = b = (Y >> 9) + bu */ orr r5, r6, r4 /* check if clamping is needed... */ orr r5, r5, r2, asr #1 /* ...at all */ cmp r5, #31 bls 15f /* -> no clamp */ cmp r6, #31 /* clamp r */ mvnhi r6, r6, asr #31 andhi r6, r6, #31 cmp r2, #63 /* clamp g */ mvnhi r2, r2, asr #31 andhi r2, r2, #63 cmp r4, #31 /* clamp b */ mvnhi r4, r4, asr #31 andhi r4, r4, #31 15: /* no clamp */ orr r4, r4, r2, lsl #5 /* pixel = r<<11 | g<<5 | b */ orr r4, r4, r6, lsl #11 strh r4, [r9] /* write pixel */ /* 1st loop, second pixel */ ldrb r5, [r10], #1 /* r5 = *ysrc++ = *Y'_p++ */ sub r5, r5, #16 /* r5 = (Y'-16) * 74 */ add r2, r5, r5, asl #2 add r5, r2, r5, asl #5 add r6, r1, r5, asr #8 /* r6 = r = (Y >> 9) + rv */ add r2, r3, r5, asr #7 /* r2 = g = (Y >> 8) + guv */ add r4, r0, r5, asr #8 /* r4 = b = (Y >> 9) + bu */ orr r5, r6, r4 /* check if clamping is needed... */ orr r5, r5, r2, asr #1 /* ...at all */ cmp r5, #31 bls 15f /* -> no clamp */ cmp r6, #31 /* clamp r */ mvnhi r6, r6, asr #31 andhi r6, r6, #31 cmp r2, #63 /* clamp g */ mvnhi r2, r2, asr #31 andhi r2, r2, #63 cmp r4, #31 /* clamp b */ mvnhi r4, r4, asr #31 andhi r4, r4, #31 15: /* no clamp */ orr r4, r4, r2, lsl #5 /* pixel = r<<11 | g<<5 | b */ orr r4, r4, r6, lsl #11 strh r4, [r9] /* write pixel */ subs r7, r7, #2 /* check for loop end */ bgt 10b /* back to beginning */ /* 1st loop end */ add r8, sp, #12 /* chroma buffer */ ldmia sp, { r7, r10 } /* r7 = loop count */ /* r10 = &ysrc[stride] */ /* 2nd loop start */ 20: /* loop start */ /* restore r1, r3 and r0 from chroma buffer */ ldrsb r1, [r8], #1 ldrsb r3, [r8], #1 ldrsb r0, [r8], #1 /* 2nd loop, first pixel */ ldrb r5, [r10], #1 /* r5 = *ysrc++ = *Y'_p++ */ sub r5, r5, #16 /* r5 = (Y'-16) * 74 */ add r2, r5, r5, asl #2 add r5, r2, r5, asl #5 add r6, r1, r5, asr #8 /* r6 = r = (Y >> 9) + rv */ add r2, r3, r5, asr #7 /* r2 = g = (Y >> 8) + guv */ add r4, r0, r5, asr #8 /* r4 = b = (Y >> 9) + bu */ orr r5, r6, r4 /* check if clamping is needed... */ orr r5, r5, r2, asr #1 /* ...at all */ cmp r5, #31 bls 15f /* -> no clamp */ cmp r6, #31 /* clamp r */ mvnhi r6, r6, asr #31 andhi r6, r6, #31 cmp r2, #63 /* clamp g */ mvnhi r2, r2, asr #31 andhi r2, r2, #63 cmp r4, #31 /* clamp b */ mvnhi r4, r4, asr #31 andhi r4, r4, #31 15: /* no clamp */ orr r4, r4, r2, lsl #5 /* pixel = r<<11 | g<<5 | b */ orr r4, r4, r6, lsl #11 strh r4, [r9] /* write pixel */ /* 2nd loop, second pixel */ ldrb r5, [r10], #1 /* r5 = *ysrc++ = *Y'_p++ */ sub r5, r5, #16 /* r5 = (Y'-16) * 74 */ add r2, r5, r5, asl #2 add r5, r2, r5, asl #5 add r6, r1, r5, asr #8 /* r6 = r = (Y >> 9) + rv */ add r2, r3, r5, asr #7 /* r2 = g = (Y >> 8) + guv */ add r4, r0, r5, asr #8 /* r4 = b = (Y >> 9) + bu */ orr r5, r6, r4 /* check if clamping is needed... */ orr r5, r5, r2, asr #1 /* ...at all */ cmp r5, #31 bls 15f /* -> no clamp */ cmp r6, #31 /* clamp r */ mvnhi r6, r6, asr #31 andhi r6, r6, #31 cmp r2, #63 /* clamp g */ mvnhi r2, r2, asr #31 andhi r2, r2, #63 cmp r4, #31 /* clamp b */ mvnhi r4, r4, asr #31 andhi r4, r4, #31 15: /* no clamp */ orr r4, r4, r2, lsl #5 /* pixel = r<<11 | g<<5 | b */ orr r4, r4, r6, lsl #11 strh r4, [r9] /* write pixel */ subs r7, r7, #2 /* check for loop end */ bgt 20b /* back to beginning */ /* 2nd loop end */ ldr r3, [sp, #8] add sp, sp, r3 /* deallocate buffer */ ldmfd sp!, { r4-r12 } /* restore registers */ bx lr .ltorg .size lcd_write_yuv420_lines, .-lcd_write_yuv420_lines