From 4f5a78a8b9e9cade9fa17afdc6157fa29fbe688a Mon Sep 17 00:00:00 2001 From: TC1995 Date: Mon, 6 Sep 2021 13:03:50 +0200 Subject: [PATCH] Added a force byte mode variable to prevent dword mode from operating due to glitches in the S3 acceleration using said mode. --- src/include/86box/vid_svga.h | 1 + src/include/86box/vid_svga_render_remap.h | 2 +- src/video/vid_ht216.c | 4 +- src/video/vid_s3.c | 375 +++++++++++++--------- 4 files changed, 225 insertions(+), 157 deletions(-) diff --git a/src/include/86box/vid_svga.h b/src/include/86box/vid_svga.h index eed2d2ad2..93d4aeffc 100644 --- a/src/include/86box/vid_svga.h +++ b/src/include/86box/vid_svga.h @@ -153,6 +153,7 @@ typedef struct svga_t /*Force CRTC to dword mode, regardless of CR14/CR17. Required for S3 enhanced mode*/ int force_dword_mode; + int force_byte_mode; int remap_required; uint32_t (*remap_func)(struct svga_t *svga, uint32_t in_addr); diff --git a/src/include/86box/vid_svga_render_remap.h b/src/include/86box/vid_svga_render_remap.h index c2e48f1b2..5b7f0fe18 100644 --- a/src/include/86box/vid_svga_render_remap.h +++ b/src/include/86box/vid_svga_render_remap.h @@ -101,7 +101,7 @@ void svga_recalc_remap_func(svga_t *svga) { int func_nr; - if (svga->fb_only) + if (svga->fb_only || svga->force_byte_mode) func_nr = 0; else { if (svga->force_dword_mode) diff --git a/src/video/vid_ht216.c b/src/video/vid_ht216.c index eafc7400c..3f12bdfdc 100644 --- a/src/video/vid_ht216.c +++ b/src/video/vid_ht216.c @@ -649,7 +649,9 @@ ht216_recalctimings(svga_t *svga) svga->hdisp *= (svga->seqregs[1] & 8) ? 16 : 8; svga->rowoffset <<= 1; if ((svga->crtc[0x17] & 0x60) == 0x20) /*Would result in a garbled screen with trailing cursor glitches*/ - svga->crtc[0x17] |= 0x40; + svga->force_byte_mode = 1; + else + svga->force_byte_mode = 0; } svga->render = svga_render_8bpp_highres; } diff --git a/src/video/vid_s3.c b/src/video/vid_s3.c index bb68540d0..0e251ba74 100644 --- a/src/video/vid_s3.c +++ b/src/video/vid_s3.c @@ -490,26 +490,45 @@ s3_accel_out_pixtrans_w(s3_t *s3, uint16_t val) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if ((s3->accel.cmd & 0x1000) && (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40))) - val = (val >> 8) | (val << 8); - s3_accel_start(8, 1, val | (val << 16), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = (val >> 8) | (val << 8); + s3_accel_start(8, 1, val | (val << 16), 0, s3); + } else + s3_accel_start(1, 1, 0xffffffff, val | (val << 16), s3); } else s3_accel_start(1, 1, 0xffffffff, val | (val << 16), s3); break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if ((s3->accel.cmd & 0x1000) && (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40))) - val = (val >> 8) | (val << 8); - s3_accel_start(16, 1, val | (val << 16), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = (val >> 8) | (val << 8); + s3_accel_start(16, 1, val | (val << 16), 0, s3); + } else + s3_accel_start(2, 1, 0xffffffff, val | (val << 16), s3); } else s3_accel_start(2, 1, 0xffffffff, val | (val << 16), s3); break; case 0x400: if (svga->crtc[0x53] & 0x08) { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if ((s3->accel.cmd & 0x1000) && (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40))) - val = (val >> 8) | (val << 8); - s3_accel_start(32, 1, val | (val << 16), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = (val >> 8) | (val << 8); + s3_accel_start(32, 1, val | (val << 16), 0, s3); + } else + s3_accel_start(4, 1, 0xffffffff, val | (val << 16), s3); + } else + s3_accel_start(4, 1, 0xffffffff, val | (val << 16), s3); + } else { + if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = (val >> 8) | (val << 8); + s3_accel_start(16, 1, val | (val << 16), 0, s3); + } else + s3_accel_start(4, 1, 0xffffffff, val | (val << 16), s3); } else s3_accel_start(4, 1, 0xffffffff, val | (val << 16), s3); } @@ -517,10 +536,12 @@ s3_accel_out_pixtrans_w(s3_t *s3, uint16_t val) case 0x600: if (s3->chip == S3_TRIO32 || s3->chip == S3_VISION968 || s3->chip == S3_VISION868 || s3->chip >= S3_TRIO64V) { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if ((s3->accel.cmd & 0x1000) && (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40))) - val = (val >> 8) | (val << 8); - s3_accel_start(8, 1, (val >> 8) & 0xff, 0, s3); - s3_accel_start(8, 1, val & 0xff, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = (val >> 8) | (val << 8); + s3_accel_start(8, 1, (val >> 8) & 0xff, 0, s3); + s3_accel_start(8, 1, val & 0xff, 0, s3); + } } } break; @@ -535,21 +556,31 @@ s3_accel_out_pixtrans_l(s3_t *s3, uint32_t val) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if ((s3->accel.cmd & 0x1000) && (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40))) - val = ((val & 0xff00ff00) >> 8) | ((val & 0x00ff00ff) << 8); - s3_accel_start(8, 1, val, 0, s3); - s3_accel_start(8, 1, val >> 16, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = ((val & 0xff00ff00) >> 8) | ((val & 0x00ff00ff) << 8); + s3_accel_start(8, 1, val, 0, s3); + s3_accel_start(8, 1, val >> 16, 0, s3); + } else { + s3_accel_start(1, 1, 0xffffffff, val, s3); + s3_accel_start(1, 1, 0xffffffff, val >> 16, s3); + } } else { - s3_accel_start(2, 1, 0xffffffff, val, s3); - s3_accel_start(2, 1, 0xffffffff, val >> 16, s3); + s3_accel_start(1, 1, 0xffffffff, val, s3); + s3_accel_start(1, 1, 0xffffffff, val >> 16, s3); } break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if ((s3->accel.cmd & 0x1000) && (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40))) - val = ((val & 0xff00ff00) >> 8) | ((val & 0x00ff00ff) << 8); - s3_accel_start(16, 1, val, 0, s3); - s3_accel_start(16, 1, val >> 16, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = ((val & 0xff00ff00) >> 8) | ((val & 0x00ff00ff) << 8); + s3_accel_start(16, 1, val, 0, s3); + s3_accel_start(16, 1, val >> 16, 0, s3); + } else { + s3_accel_start(2, 1, 0xffffffff, val, s3); + s3_accel_start(2, 1, 0xffffffff, val >> 16, s3); + } } else { s3_accel_start(2, 1, 0xffffffff, val, s3); s3_accel_start(2, 1, 0xffffffff, val >> 16, s3); @@ -557,21 +588,26 @@ s3_accel_out_pixtrans_l(s3_t *s3, uint32_t val) break; case 0x400: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if ((s3->accel.cmd & 0x1000) && (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40))) - val = ((val & 0xff000000) >> 24) | ((val & 0x00ff0000) >> 8) | ((val & 0x0000ff00) << 8) | ((val & 0x000000ff) << 24); - s3_accel_start(32, 1, val, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = ((val & 0xff000000) >> 24) | ((val & 0x00ff0000) >> 8) | ((val & 0x0000ff00) << 8) | ((val & 0x000000ff) << 24); + s3_accel_start(32, 1, val, 0, s3); + } else + s3_accel_start(4, 1, 0xffffffff, val, s3); } else s3_accel_start(4, 1, 0xffffffff, val, s3); break; case 0x600: if (s3->chip == S3_TRIO32 || s3->chip == S3_VISION968 || s3->chip == S3_VISION868 || s3->chip >= S3_TRIO64V) { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if ((s3->accel.cmd & 0x1000) && (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40))) - val = ((val & 0xff000000) >> 24) | ((val & 0x00ff0000) >> 8) | ((val & 0x0000ff00) << 8) | ((val & 0x000000ff) << 24); - s3_accel_start(8, 1, (val >> 24) & 0xff, 0, s3); - s3_accel_start(8, 1, (val >> 16) & 0xff, 0, s3); - s3_accel_start(8, 1, (val >> 8) & 0xff, 0, s3); - s3_accel_start(8, 1, val & 0xff, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) + val = ((val & 0xff000000) >> 24) | ((val & 0x00ff0000) >> 8) | ((val & 0x0000ff00) << 8) | ((val & 0x000000ff) << 24); + s3_accel_start(8, 1, (val >> 24) & 0xff, 0, s3); + s3_accel_start(8, 1, (val >> 16) & 0xff, 0, s3); + s3_accel_start(8, 1, (val >> 8) & 0xff, 0, s3); + s3_accel_start(8, 1, val & 0xff, 0, s3); + } } } break; @@ -1023,7 +1059,10 @@ s3_accel_out_fifo(s3_t *s3, uint16_t port, uint8_t val) if (s3->accel.cmd & 0x100) { if (!(s3->accel.cmd & 0x600)) { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, s3->accel.pix_trans[0], 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, s3->accel.pix_trans[0], 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0], s3); } else s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0], s3); } @@ -1038,59 +1077,51 @@ s3_accel_out_fifo(s3_t *s3, uint16_t port, uint8_t val) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); } else s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); break; case 0x200: /*Windows 95's built-in driver expects this to be loaded regardless of the byte swap bit (0xE2E9) in the 86c928*/ - if (s3->accel.cmd & 0x1000) { - if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) s3_accel_start(16, 1, s3->accel.pix_trans[1] | (s3->accel.pix_trans[0] << 8), 0, s3); - else { - if (s3->chip == S3_86C928) - s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8)); - else - s3_accel_start(16, 1, s3->accel.pix_trans[1] | (s3->accel.pix_trans[0] << 8), 0, s3); - } - } else { - if (s3->chip == S3_86C928) - s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8)); else - s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[1] | (s3->accel.pix_trans[0] << 8), s3); - } - } else { - if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) s3_accel_start(16, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); - else { - if (s3->chip == S3_86C928) - s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8)); - else - s3_accel_start(16, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); - } } else { if (s3->chip == S3_86C928) s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8)); - else - s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); - } + else { + if (s3->accel.cmd & 0x1000) + s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[1] | (s3->accel.pix_trans[0] << 8), s3); + else + s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); + } + } } break; case 0x400: if (svga->crtc[0x53] & 0x08) { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(32, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(32, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); + else + s3_accel_start(4, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); } else - s3_accel_start(4, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); + s3_accel_start(4, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); } break; case 0x600: if (s3->chip == S3_TRIO32 || s3->chip == S3_VISION968 || s3->chip >= S3_TRIO64V) { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, s3->accel.pix_trans[1], 0, s3); - s3_accel_start(8, 1, s3->accel.pix_trans[0], 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + s3_accel_start(8, 1, s3->accel.pix_trans[1], 0, s3); + s3_accel_start(8, 1, s3->accel.pix_trans[0], 0, s3); + } } } break; @@ -1110,59 +1141,51 @@ s3_accel_out_fifo(s3_t *s3, uint16_t port, uint8_t val) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); } else s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); break; case 0x200: /*Windows 95's built-in driver expects the upper 16 bits to be loaded instead of the whole 32-bit one, regardless of the byte swap bit (0xE2EB) in the 86c928*/ - if (s3->accel.cmd & 0x1000) { - if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + if (s3->accel.cmd & 0x1000) s3_accel_start(16, 1, s3->accel.pix_trans[3] | (s3->accel.pix_trans[2] << 8) | (s3->accel.pix_trans[1] << 16) | (s3->accel.pix_trans[0] << 24), 0, s3); - else { - if (s3->chip == S3_86C928) - s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[2] | (s3->accel.pix_trans[3] << 8)); - else - s3_accel_start(16, 1, s3->accel.pix_trans[3] | (s3->accel.pix_trans[2] << 8) | (s3->accel.pix_trans[1] << 16) | (s3->accel.pix_trans[0] << 24), 0, s3); - } - } else { - if (s3->chip == S3_86C928) - s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8)); else - s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[3] | (s3->accel.pix_trans[2] << 8) | (s3->accel.pix_trans[1] << 16) | (s3->accel.pix_trans[0] << 24), s3); - } - } else { - if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) s3_accel_start(16, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); - else { - if (s3->chip == S3_86C928) - s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[2] | (s3->accel.pix_trans[3] << 8)); - else - s3_accel_start(16, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); - } } else { if (s3->chip == S3_86C928) - s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8)); - else - s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); - } + s3_accel_out_pixtrans_w(s3, s3->accel.pix_trans[2] | (s3->accel.pix_trans[3] << 8)); + else { + if (s3->accel.cmd & 0x1000) + s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[3] | (s3->accel.pix_trans[2] << 8) | (s3->accel.pix_trans[1] << 16) | (s3->accel.pix_trans[0] << 24), s3); + else + s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); + } + } } break; case 0x400: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(32, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(32, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); + else + s3_accel_start(4, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); } else s3_accel_start(4, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); break; case 0x600: if (s3->chip == S3_TRIO32 || s3->chip == S3_VISION968 || s3->chip >= S3_TRIO64V) { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, s3->accel.pix_trans[3], 0, s3); - s3_accel_start(8, 1, s3->accel.pix_trans[2], 0, s3); - s3_accel_start(8, 1, s3->accel.pix_trans[1], 0, s3); - s3_accel_start(8, 1, s3->accel.pix_trans[0], 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + s3_accel_start(8, 1, s3->accel.pix_trans[3], 0, s3); + s3_accel_start(8, 1, s3->accel.pix_trans[2], 0, s3); + s3_accel_start(8, 1, s3->accel.pix_trans[1], 0, s3); + s3_accel_start(8, 1, s3->accel.pix_trans[0], 0, s3); + } } } break; @@ -1352,14 +1375,20 @@ s3_accel_write_fifo(s3_t *s3, uint32_t addr, uint8_t val) if (s3->accel.cmd & 0x100) { if ((s3->accel.cmd & 0x600) == 0x200) { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(16, 1, val, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(16, 1, val | (val << 8) | (val << 16) | (val << 24), 0, s3); + else + s3_accel_start(2, 1, 0xffffffff, val | (val << 8) | (val << 16) | (val << 24), s3); } else - s3_accel_start(2, 1, 0xffffffff, val, s3); + s3_accel_start(2, 1, 0xffffffff, val | (val << 8) | (val << 16) | (val << 24), s3); } else { if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, val, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, val | (val << 8) | (val << 16) | (val << 24), 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, val | (val << 8) | (val << 16) | (val << 24), s3); } else - s3_accel_start(1, 1, 0xffffffff, val, s3); + s3_accel_start(1, 1, 0xffffffff, val | (val << 8) | (val << 16) | (val << 24), s3); } } } @@ -2713,8 +2742,14 @@ static void s3_recalctimings(svga_t *svga) svga->lowres = !((svga->gdcreg[5] & 0x40) && (svga->crtc[0x3a] & 0x10)); if (((svga->gdcreg[5] & 0x40) && (svga->crtc[0x3a] & 0x10)) || (svga->crtc[0x3a] & 0x10)) { - if (((svga->crtc[0x17] & 0x60) == 0x20) && (svga->crtc[0x31] & 0x08)) - svga->crtc[0x17] |= 0x40; + pclog("CRTC14 bit 6 byte mode = %02x, CRTC17 bit 6 byte mode = %02x, CRTC17 bit 5 word mode ma15 = %02x, CRTC31 bit 3 dword mode = %02x\n", + svga->crtc[0x14] & 0x40, svga->crtc[0x17] & 0x40, svga->crtc[0x17] & 0x20, svga->crtc[0x31] & 0x08); + + if (svga->crtc[0x31] & 0x08) /*This would typically force dword mode, but we are encountering accel bugs with it, so force byte mode instead*/ + svga->force_byte_mode = 1; + else + svga->force_byte_mode = 0; + switch (svga->bpp) { case 8: svga->render = svga_render_8bpp_highres; @@ -2851,8 +2886,11 @@ static void s3_trio64v_recalctimings(svga_t *svga) svga->lowres = !((svga->gdcreg[5] & 0x40) && (svga->crtc[0x3a] & 0x10)); if ((svga->gdcreg[5] & 0x40) && (svga->crtc[0x3a] & 0x10)) { - if (((svga->crtc[0x17] & 0x60) == 0x20) && (svga->crtc[0x31] & 0x08)) - svga->crtc[0x17] |= 0x40; + if (svga->crtc[0x31] & 0x08) /*This would typically force dword mode, but we are encountering accel bugs with it, so force byte mode instead*/ + svga->force_byte_mode = 1; + else + svga->force_byte_mode = 0; + switch (svga->bpp) { case 8: svga->render = svga_render_8bpp_highres; @@ -3478,7 +3516,10 @@ s3_accel_in(uint16_t port, void *p) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, s3->accel.pix_trans[0], 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, s3->accel.pix_trans[0], 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); } else s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); break; @@ -3503,13 +3544,19 @@ s3_accel_in(uint16_t port, void *p) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); } else s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(16, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(16, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), 0, s3); + else + s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); } else s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8), s3); break; @@ -3531,13 +3578,19 @@ s3_accel_in(uint16_t port, void *p) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); } else s3_accel_start(1, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(16, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(16, 1, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), 0, s3); + else + s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); } else s3_accel_start(2, 1, 0xffffffff, s3->accel.pix_trans[0] | (s3->accel.pix_trans[1] << 8) | (s3->accel.pix_trans[2] << 16) | (s3->accel.pix_trans[3] << 24), s3); break; @@ -3574,13 +3627,19 @@ s3_accel_in_w(uint16_t port, void *p) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, temp | (temp << 16), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, temp | (temp << 16), 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, temp | (temp << 16), s3); } else s3_accel_start(1, 1, 0xffffffff, temp | (temp << 16), s3); break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(16, 1, temp | (temp << 16), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(16, 1, temp | (temp << 16), 0, s3); + else + s3_accel_start(2, 1, 0xffffffff, temp | (temp << 16), s3); } else s3_accel_start(2, 1, 0xffffffff, temp | (temp << 16), s3); break; @@ -3606,8 +3665,13 @@ s3_accel_in_l(uint16_t port, void *p) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, temp, 0, s3); - s3_accel_start(8, 1, temp >> 16, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + s3_accel_start(8, 1, temp, 0, s3); + s3_accel_start(8, 1, temp >> 16, 0, s3); + } else { + s3_accel_start(1, 1, 0xffffffff, temp, s3); + s3_accel_start(1, 1, 0xffffffff, temp >> 16, s3); + } } else { s3_accel_start(1, 1, 0xffffffff, temp, s3); s3_accel_start(1, 1, 0xffffffff, temp >> 16, s3); @@ -3615,8 +3679,13 @@ s3_accel_in_l(uint16_t port, void *p) break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(16, 1, temp, 0, s3); - s3_accel_start(16, 1, temp >> 16, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + s3_accel_start(16, 1, temp, 0, s3); + s3_accel_start(16, 1, temp >> 16, 0, s3); + } else { + s3_accel_start(2, 1, 0xffffffff, temp, s3); + s3_accel_start(2, 1, 0xffffffff, temp >> 16, s3); + } } else { s3_accel_start(2, 1, 0xffffffff, temp, s3); s3_accel_start(2, 1, 0xffffffff, temp >> 16, s3); @@ -3728,15 +3797,21 @@ s3_accel_read(uint32_t addr, void *p) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, temp, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, temp | (temp << 8) | (temp << 16) | (temp << 24), 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, temp | (temp << 8) | (temp << 16) | (temp << 24), s3); } else - s3_accel_start(1, 1, 0xffffffff, temp, s3); + s3_accel_start(1, 1, 0xffffffff, temp | (temp << 8) | (temp << 16) | (temp << 24), s3); break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(16, 1, temp, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(16, 1, temp | (temp << 8) | (temp << 16) | (temp << 24), 0, s3); + else + s3_accel_start(2, 1, 0xffffffff, temp | (temp << 8) | (temp << 16) | (temp << 24), s3); } else - s3_accel_start(2, 1, 0xffffffff, temp, s3); + s3_accel_start(2, 1, 0xffffffff, temp | (temp << 8) | (temp << 16) | (temp << 24), s3); break; } } @@ -3773,13 +3848,19 @@ s3_accel_read_w(uint32_t addr, void *p) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, temp | (temp << 16), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(8, 1, temp | (temp << 16), 0, s3); + else + s3_accel_start(1, 1, 0xffffffff, temp | (temp << 16), s3); } else s3_accel_start(1, 1, 0xffffffff, temp | (temp << 16), s3); break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(16, 1, temp | (temp << 16), 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) + s3_accel_start(16, 1, temp | (temp << 16), 0, s3); + else + s3_accel_start(2, 1, 0xffffffff, temp | (temp << 16), s3); } else s3_accel_start(2, 1, 0xffffffff, temp | (temp << 16), s3); break; @@ -3925,8 +4006,13 @@ s3_accel_read_l(uint32_t addr, void *p) switch (s3->accel.cmd & 0x600) { case 0x000: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(8, 1, temp, 0, s3); - s3_accel_start(8, 1, temp >> 16, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + s3_accel_start(8, 1, temp, 0, s3); + s3_accel_start(8, 1, temp >> 16, 0, s3); + } else { + s3_accel_start(1, 1, 0xffffffff, temp, s3); + s3_accel_start(1, 1, 0xffffffff, temp >> 16, s3); + } } else { s3_accel_start(1, 1, 0xffffffff, temp, s3); s3_accel_start(1, 1, 0xffffffff, temp >> 16, s3); @@ -3934,8 +4020,13 @@ s3_accel_read_l(uint32_t addr, void *p) break; case 0x200: if (((s3->accel.multifunc[0xa] & 0xc0) == 0x80) || (s3->accel.cmd & 2)) { - s3_accel_start(16, 1, temp, 0, s3); - s3_accel_start(16, 1, temp >> 16, 0, s3); + if (((s3->accel.frgd_mix & 0x60) != 0x40) || ((s3->accel.bkgd_mix & 0x60) != 0x40)) { + s3_accel_start(16, 1, temp, 0, s3); + s3_accel_start(16, 1, temp >> 16, 0, s3); + } else { + s3_accel_start(2, 1, 0xffffffff, temp, s3); + s3_accel_start(2, 1, 0xffffffff, temp >> 16, s3); + } } else { s3_accel_start(2, 1, 0xffffffff, temp, s3); s3_accel_start(2, 1, 0xffffffff, temp >> 16, s3); @@ -4728,32 +4819,6 @@ s3_accel_start(int count, int cpu_input, uint32_t mix_dat, uint32_t cpu_dat, s3_ if (!cpu_input) s3->accel.dat_count = 0; - - /*If the mask source is the CPU, the color source can't be the same as the mask, or vice versa*/ - if (cpu_input) { - if (mix_dat & mix_mask) { - if ((s3->accel.cmd & 0x600) != 0x600) { - if (((s3->accel.frgd_mix & 0x60) == 0x40) && ((s3->accel.multifunc[0xa] & 0xc0) == 0x80)) { - cpu_dat = mix_dat; - count >>= 3; - } - } - } else { - if (cmd == 3) { - if (((s3->accel.frgd_mix & 0x60) == 0x40) && ((s3->accel.multifunc[0xa] & 0xc0) == 0x80)) { - cpu_dat = mix_dat; - count >>= 3; - } - } else { - if ((s3->accel.cmd & 0x600) != 0x600) { - if (((s3->accel.bkgd_mix & 0x60) == 0x40) && ((s3->accel.multifunc[0xa] & 0xc0) == 0x80)) { - cpu_dat = mix_dat; - count >>= 3; - } - } - } - } - } if (cpu_input && (((s3->accel.multifunc[0xa] & 0xc0) != 0x80) || (!(s3->accel.cmd & 2)))) { if ((s3->bpp == 3) && count == 2) { @@ -4777,6 +4842,9 @@ s3_accel_start(int count, int cpu_input, uint32_t mix_dat, uint32_t cpu_dat, s3_ else if (s3->bpp == 1) rd_mask &= 0xffff; + if (s3->bpp == 0) compare &= 0xff; + if (s3->bpp == 1) compare &= 0xffff; + switch (s3->accel.cmd & 0x600) { case 0x000: mix_mask = 0x80; break; @@ -4785,9 +4853,6 @@ s3_accel_start(int count, int cpu_input, uint32_t mix_dat, uint32_t cpu_dat, s3_ case 0x600: mix_mask = (s3->chip == S3_TRIO32 || s3->chip >= S3_TRIO64V || s3->chip == S3_VISION968 || s3->chip == S3_VISION868) ? 0x80 : 0x80000000; break; } - if (s3->bpp == 0) compare &= 0xff; - if (s3->bpp == 1) compare &= 0xffff; - /*Bit 4 of the Command register is the draw yes bit, which enables writing to memory/reading from memory when enabled. When this bit is disabled, no writing to memory/reading from memory is allowed. (This bit is almost meaningless on the NOP command)*/