A real 256x256 CRTC mode, derived not recalled; palette ceiling was 2 dB low

Session 3 left the harness on the IPL's 768x512 text timing because no CRTC
values had been derived and guessing them was the failure mode to avoid. This
derives them from MAME 0.277's divisor ladder instead, and the derivation is
self-checking: the 256-wide mode runs at div 6 against the 768 mode's div 2, so
htotal is exactly 1104/3 = 368 dots and every horizontal register divides by
three with no remainder. Only the blanking split rounds. Verified by snapshot:
native 256x512, active area pixel-exact, x=512 wrap gone.

Two things fell out that change numbers elsewhere:

- The palette's shared LSB I must be chosen per entry, not hardcoded to 1.
  Doing so lifts the display ceiling from 38.85 to 40.81 dB and is the only way
  to reach true black at all, since pal6bit(1) = 4. 102 of 256 entries want
  I = 0, so this is not a corner case. Supersedes FINDINGS 22.4; scsi has ~2 dB
  more headroom than that section claimed. The encoder does not do this yet.

- Letterboxing costs a palette entry: GVRAM cleared to zero shows entry 0, and
  a free mediancut palette puts a real image colour there. 255 colours plus a
  reserved black, via prep_frame.py --reserve-black.

MAME's graphics double-scan is phase-shifted one raster line (it halves the
absolute scanline and vbegin is odd), which produced a false failure before it
was understood; the regression test now asserts the shifted pairing explicitly.

Still Lua-side. No 68000 instruction has drawn a pixel; the 38% blit estimate
remains unvalidated. What this buys is a defined geometry for the decoder to
write into: 256 words per row, 1024-byte stride, rows 32..223.

Claude-Session: https://claude.ai/code/session_01194oWYW8DQXK1SZ2DnChW6

tools/bench/crtc_mode.lua

@@ -0,0 +1,60 @@
+-- CRTC mode table: 256x256, 256 colours, 31.5kHz, graphics double-scanned.
+--
+-- DERIVED from MAME 0.277 src/mame/sharp/x68k_crtc.cpp, not recalled. The
+-- derivation is self-checking, which is why it is trustworthy:
+--
+--   refresh_mode() picks the dot clock as (reg20 bit4 ? 69.55199MHz : 38.86363MHz)/div
+--   with div from the (reg20 & 0x1f) ladder:  0x16 -> 2   (IPL's 768x512)
+--                                             0x11 -> 3   (512 wide)
+--                                             0x10 -> 6   (256 wide, double-scan)
+--
+--   IPL 768 mode: div 2 -> 34.776 MHz, m_htotal = (137+1)*8 = 1104 dots
+--                 34.776e6 / 1104 = 31500.0 Hz  exactly.
+--   256 mode:     div 6 -> 11.592 MHz.  Same 31.5kHz line rate requires
+--                 11.592e6 / 31500 = 368 dots = 46 chars -> R00 = 45.
+--
+--   368 = 1104/3 exactly, so EVERY horizontal register is the 768-mode value
+--   divided by 3 -- no rounding for the active window:
+--       visible chars   (124-28) / 3 = 32  ->  32*8 = 256 dots  exact
+--   Only the blanking split needs rounding. 768 mode is sync/back/front =
+--   14/14/14 chars; /3 = 4.67 each; the closest integer triple summing to
+--   46-32 = 14 is 5/5/4.  ->  R01=5, R02=10, R03=42.
+--
+--   Total blanking time is identical to the 768 mode (112 dots @ 11.592MHz =
+--   336 dots @ 34.776MHz = 9.66us), which is what a real monitor needs.
+--
+-- VERTICAL registers are NOT halved. The CRTC still generates a 568-line
+-- 31.5kHz raster (31500/568 = 55.46 Hz); "256 lines" is a graphics-layer
+-- double-scan (draw_gfx() halves gfxrect, x68k_v.cpp:401). Halving them would
+-- ask the monitor for 110 Hz. So R04-R07 keep the 31kHz text-mode values.
+-- MAME logerrors "visarea larger then reg[20]" for this; it is cosmetic.
+
+local M = {}
+
+M.regs = {
+  [0] = 45,   -- H total      (46 chars = 368 dots @ 11.592MHz = 31500.0 Hz)
+  [1] = 5,    -- H sync end   (5 chars = 3.45us)
+  [2] = 10,   -- H disp begin (hbegin = 10*8+1 = 81)
+  [3] = 42,   -- H disp end   (hend = 336; inclusive width = 336-81+1 = 256)
+  [4] = 567,  -- V total      (568 scanlines -> 55.46 Hz)
+  [5] = 5,    -- V sync end
+  [6] = 40,   -- V disp begin (vbegin = 41)
+  [7] = 552,  -- V disp end   (512 scanlines -> 256 gfx rows, double-scanned)
+  [8] = 27,   -- H sync adjust (MAME stores but does not use it; IPL value)
+}
+
+-- R20: bit11=0 display (not buffer), bits9-8=01 256-colour,
+--      bit4=1 31.5kHz, bits3-2=00 256 lines, bits1-0=00 256 dots
+M.r20 = 0x0110
+
+M.width, M.height = 256, 256
+
+function M.apply(SP)
+  for r, v in pairs(M.regs) do SP:write_u16(0xE80000 + r*2, v) end
+  SP:write_u16(0xE80000 + 20*2, M.r20)
+  SP:write_u16(0xE82400, 0x0001)   -- video ctrl reg 0: 256 colours
+  SP:write_u16(0xE82600, 0x001F)   -- reg 2: graphics on, all 4 pages, text/PCG off
+  SP:write_u8 (0xE8E001, 15)       -- monitor contrast (IPL leaves 14 = 7% dark)
+end
+
+return M