opencode.energia.arti.ee/vt1200
1
0 Fork 0
forked from arti/vt1200
Code Pull Requests Activity

Commit Graph

  • a5769d3d2e dwt: GUI/X architecture note; Xlib callable but XOpenDisplay hangs SDD 2026-10-18 14:08:31 +03:00
  • c0f142300b dwt/screenshot.py: network screenshot over rdebug peek (VRAM -> PNG), verified against the live OS SDD 2026-10-18 13:49:19 +03:00
  • 2cc3ab1c79 dwt/exploit.py: default to the computed dwt_prf answer -- echo is the known-failing mode and made a bare --peek look like a closed socket SDD 2026-10-18 11:28:00 +03:00
  • 0e4075958b dwt: rdebug opcode 8 puts attacker text in the TM Messages box SDD 2026-10-18 11:03:01 +03:00
  • 619aff0a93 dwt: visual injection -- confirm the persistent box on the live OS, and record that the blink is not established (toggle period is ~1536 frames, not 30) SDD 2026-10-18 03:24:38 +03:00
  • 05f1895bf5 dwt: visual injection experiments via the returning rdebug jump SDD 2026-10-18 03:22:59 +03:00
  • 1e5c095b72 dwt: reverse the dwt_prf challenge; remote poke/jump now works SDD 2026-10-18 02:36:23 +03:00
  • 041c82cb73 dwt: runtime oracle and ground-truth vectors for the dwt_prf challenge transform SDD 2026-10-18 02:30:49 +03:00
  • 5645bb00b3 dwt: exploit client and gate analysis; remote execution blocked at the challenge SDD 2026-10-18 02:20:16 +03:00
  • 2c7b2e8c2e dwt: gate B is real; challenge args must follow the answer SDD 2026-10-18 02:19:54 +03:00
  • bd58af3569 dwt: gate A is a 2-bit field and is open; correct the static reading SDD 2026-10-18 02:07:18 +03:00
  • a18881f45d dwt: MOP console carrier client, protocol notes, and emulator evidence SDD 2026-10-18 02:02:17 +03:00
  • 0a1042e81f dwt: correct the Done\n reply length (payload includes its NUL, len field = payload+1) SDD 2026-10-18 01:55:56 +03:00
  • 68a1aeabdc dwt: static analysis of rdebug / DWT console -- poke+jump exist, gated by a config word and a 16-byte handshake SDD 2026-10-18 01:55:41 +03:00
  • be32e8a138 harness: regenerate duart2_default_tm_frame3000.png -- the cited evidence image predated the LANCE and contradicted the README (no MAC on screen) SDD 2026-10-18 01:21:41 +03:00
  • 1608a68cdb hello-world: the TMS34010 SDK assembles code that runs on the emulated VT1200 SDD 2026-10-18 01:20:38 +03:00
  • 0c3d24b93b tools/tms34010-sdk: the two dis34 bugs are fixed (af7d3565a5) -- record it, with the 2623/6235-line scale of the correction SDD 2026-10-18 00:55:12 +03:00
  • af7d3565a5 dis34: print long relative JR as jr and surface B-file SP as b15 SDD 2026-10-18 00:54:34 +03:00
  • 8bc8d90ea6 tools/tms34010-sdk: record the results -- GSPA (not as34), byte-exact round trip at bit 0x3F208000, 1990 tools are OS/2, dosemu 5x faster, plus the traps SDD 2026-10-18 00:40:25 +03:00
  • 7d299451da reference: TMS34010 SDK investigation; prove GSPA round trip SDD 2026-10-18 00:39:32 +03:00
  • 960b256b00 tools/tms34010-sdk: add DOSBOX.md -- how to run the SDK's DOS tools headless and at speed, with measurements SDD 2026-10-18 00:12:18 +03:00
  • 1ec84abf03 tools: TMS34010 SDK (curated) -- the 1987 ASM/LNK package and the 1990 Code Generation Tools, so the project has an assembler/linker/compiler it was missing; scanned docs excluded SDD 2026-10-17 23:59:24 +03:00
  • c21cab05f8 harness: net-setup.sh -- recreate the tap+bridge the harness needs (they do not survive a reboot) SDD 2026-10-17 14:15:17 +03:00
  • 307df67f9d harness/mame: the TFTP PCF loader skips BITMAPS, so the n*4 overflow is not on that path arti 2026-10-17 14:14:07 +03:00
  • 3a51b95317 harness/fontwork: reproduction scripts for the TFTP Disk-full investigation arti 2026-10-16 22:03:49 +03:00
  • 7efb4880f0 harness: the TFTP font 'Disk full' is the terminal's own tftp_close, not a storage or map bug arti 2026-10-16 22:03:16 +03:00
  • 1051bf2940 patches: MAME is now pushed as a full repo (branch vt1200-local) -- correct the README, which said it was not, and say why the patch is still kept SDD 2026-10-16 17:50:58 +03:00
  • 1561e34967 patches: capture the MAME local edits as a 50-line patch (the MAME remote rejects shallow updates, so a mirror would cost GBs for a 50-line diff) SDD 2026-10-16 17:42:13 +03:00
  • 264d5330ad harness/serve.py: optional BOOTP responder (--bootp) so the VT1200 gets an IP at boot SDD 2026-10-16 17:27:00 +03:00
  • d4a3ccc42c harness: model the battery-backed NVR at bit 0x6000000 arti 2026-10-16 17:25:04 +03:00
  • b6da987aae harness: dump the LANCE DMA; the C - LANCE TEST's failing frame carries no FCS by design arti 2026-10-16 17:12:02 +03:00
  • 5ed4048720 gitignore: ignore root RAM dumps (ROM stays tracked) and the Ghidra transient lock; untrack the lock file SDD 2026-10-16 17:11:29 +03:00
  • f05200e44e gitignore: cover test frame dumps written into the project root, not just harness/mame/ SDD 2026-10-16 17:11:01 +03:00
  • 82ab65bc82 run.sh: document ScrollLock (UI_TOGGLE_UI) as the fix when MAME's menu stops responding, and that closing the window is a clean exit SDD 2026-10-16 17:01:04 +03:00
  • 14f78f3cbe plan: the network provider must be selected per device (interface=-1 is none); index 0 = taptun on this build; NVRAM must not be isolated SDD 2026-10-16 16:56:24 +03:00
  • b65d9ef670 run.sh net: the per-device interface selection must PERSIST (an empty cfg dir produces no tap); -networkprovider alone does not choose one SDD 2026-10-16 16:54:08 +03:00
  • b356553825 run.sh net: use a private cfg directory -- a saved per-device interface=-1 silently overrides -networkprovider and no tap is ever opened SDD 2026-10-16 16:53:14 +03:00
  • a0d752062f run.sh: add a net mode (skip + tablet + networkprovider taptun) for the crash-work configuration SDD 2026-10-16 16:48:40 +03:00
  • 7612fcbda9 reference: LANCE loopback-FCS datasheet evidence (MAME condition is right) SDD 2026-10-16 16:46:47 +03:00
  • 60b0b02e3c run.sh+README: fix the stale tablet comment; amend rule 2 with the orchestrator's own collision (uncommitted edits get swept too) SDD 2026-10-16 16:41:16 +03:00
  • d1be7da1f1 run.sh: replace the now-redundant tablet mode with a faithful mouse mode (tablet is the default) SDD 2026-10-16 16:40:52 +03:00
  • 634246ecb1 harness: LANCE Ethernet -- register file, 8-byte address ROM, INT1 and byte<->bit DMA SDD 2026-10-16 16:39:21 +03:00
  • 7bf195e615 reference: commit skip_frame3000.png (the X11 weave, cited as evidence in the README); ignore worker scratch SDD 2026-10-16 16:23:15 +03:00
  • 97dbea7fed reference: commit the real-boot frame evidence (33MB) with a NOTES.md; gitignore the 201MB source recording SDD 2026-10-16 16:22:59 +03:00
  • f46363b473 AGENTS.md: symlink to harness/mame/README.md so agent harnesses auto-read the working conventions (single source, no drift) SDD 2026-10-16 16:20:37 +03:00
  • 6d8097360c harness: the VSXXX-AB tablet is the default pointing device; VT1200_MOUSE=1 forces the faithful mouse back SDD 2026-10-16 16:19:46 +03:00
  • 7f0b2ff27c plan: record host-side tap/bridge plumbing for the LANCE network milestone (MAME's tap is tap-mess-<uid>-0, not tap0) SDD 2026-10-16 16:17:43 +03:00
  • 0e2d7d9d0d docs: PROTOCOL.md tablet coordinate range (0..2199) and the device-code discriminator, confirmed against the running OS SDD 2026-10-16 16:10:11 +03:00
  • 942a45800b run.sh: tablet mode (skip + VSXXX-AB absolute pointing); lightgun options in COMMON, verified inert by frame-md5 A/B SDD 2026-10-16 16:09:51 +03:00
  • 065f946e2f harness: VSXXX-AB tablet on DUART2, driven by the absolute host pointer arti 2026-10-16 16:07:36 +03:00
  • 02acabb7ef harness: real LK401 keyboard + VSXXX-AA mouse devices on DUART2 arti 2026-10-16 15:33:30 +03:00
  • 7e65692b54 reference: DUART counter/timer period when CTUR/CTLR unprogrammed -- datasheet does not settle it SDD 2026-10-16 15:00:20 +03:00
  • a6b6ae82ef docs: correct Q4 of the level-1 bypass analysis -- a level-1 failure is NOT fatal (superseded answer kept as record) SDD 2026-10-16 14:57:54 +03:00
  • 41454d5d4c harness: the stub level-1 run continues past both DUART failures; correct the "fatal/LED loop" claim arti 2026-10-16 14:56:47 +03:00
  • 7296a86034 run.sh: add the stub mode and document VT1200_IO_PATH_STUB as a non-faithful opt-in SDD 2026-10-16 14:32:07 +03:00
  • f7caf9aed8 harness: opt-in VT1200_IO_PATH_STUB stand-in for the level-0 keyboard-port test fixture SDD 2026-10-16 14:31:20 +03:00
  • 215fd573db docs: level-0 I/O test is not honestly satisfiable by any documented device; record why and the only honest route to close it SDD 2026-10-16 14:17:10 +03:00
  • a1e3ebfaff reference: the level-0 10-byte puzzle -- count-only, no loopback, no 10-byte device message SDD 2026-10-16 14:16:37 +03:00
  • d017953d52 docs: plan the LANCE Ethernet model + network font delivery SDD 2026-10-16 14:16:05 +03:00
  • 47dac6b283 docs: DUART2 modelled; Terminal Manager reached; the bare run's old level-1 pass was a latch artefact (honest failure now) SDD 2026-10-16 14:00:23 +03:00
  • 1ca4817745 harness: model DUART2 by default (drop the VT1200_DUART2 gate) SDD 2026-10-16 13:59:09 +03:00
  • 6ce84b86f2 run.sh: skip mode also models DUART2 -- required for the OS (phantom ISR read at 0x06200005); reaches the Terminal Manager SDD 2026-10-16 13:40:21 +03:00
  • c06cf7096b harness: skip run is STUCK at the X11 weave -- wall is unmodelled DUART2 (phantom ISR read at 0x06200005); add multi-timestamp dumps + console string resolver SDD 2026-10-16 13:39:37 +03:00
  • 855f3e85e9 docs: the OS DOES paint -- skip run boots to a running X11 server (root weave confirmed in a frame dump); correct the claim I wrote SDD 2026-10-16 13:26:27 +03:00
  • b99b7de62e harness: add run.sh launcher (faithful/skip/pie/headless modes, verified end to end) SDD 2026-10-16 13:21:51 +03:00
  • c3faa7729d docs: correct stale README - DUARTs/VRAM are modelled, fix memory_access template args, add gate list and new baseline SDD 2026-10-16 13:19:25 +03:00
  • 5abf4f33be harness: promote the working VT1200 config to the default; retire settled env gates SDD 2026-10-16 13:18:14 +03:00
  • 9d912ae084 harness: l1patch.lua -- runtime level-1 host-port patch + test-map probe SDD 2026-10-16 13:09:45 +03:00
  • 55cf122edd docs: correct the stale 256 KiB RAM claim in BRINGUP_LOG SDD 2026-10-16 13:09:37 +03:00
  • db535c2adc harness: explicit monochrome palette (pen 0 black, pen 1 white) SDD 2026-10-16 13:02:50 +03:00
  • ac60ad2f89 docs: plan to retire the settled VT1200_* env gates (promote working config to default, delete finished experiments) SDD 2026-10-16 13:02:20 +03:00
  • 6cbb362bb9 harness: VT1200_DUART1_INT2 -- route DUART1's IRQ to INT2 (A/B knob) SDD 2026-10-16 13:01:40 +03:00
  • c19c603d77 harness: VT1200_DUART_IP3 -- drive DUART1 IP3 high to take the OS cold-start gate SDD 2026-10-16 12:55:36 +03:00
  • c69d2c295f reference: level-1 host-port test — can it be skipped? (analysis, no skip) SDD 2026-10-16 12:42:35 +03:00
  • 571850e564 harness: render the TMS34010 active scanlines from video RAM SDD 2026-10-16 12:42:33 +03:00
  • bcbd7924af docs: last authorised timing change tested (idle-path transmit start) -- real deviation, not the cause; reverted SDD 2026-10-16 12:26:01 +03:00
  • 88d006730a docs: identify the writers of +0x3c0/+0x3d0 (the DUART ISR) and what L3FA71BC0 checks SDD 2026-10-16 12:23:14 +03:00
  • d7725faa1e docs: enumerate the three remaining failing checks; record the transmit-start citation SDD 2026-10-16 12:21:16 +03:00
  • 03e6a88b22 docs: establish causally what raises HOST PORT - 15; resolve the sub-test polarity; rule out the loopback-clock deviation as the cause SDD 2026-10-16 12:18:30 +03:00
  • 70e2486d57 mc68681: revert the TxRDY edit (REFUTED by the fitted part's datasheet); keep the overrun edit SDD 2026-10-16 12:15:47 +03:00
  • 72a96862c7 docs: causality audit -- the host-port failure is pre-existing, not caused by the two MAME edits SDD 2026-10-16 12:13:39 +03:00
  • 031d30da67 reference: DUART datasheet verification of the two mc68681 edits (claim 1 refuted) SDD 2026-10-16 12:13:07 +03:00
  • 9902722165 harness: wire DUART IRQ to the TMS34010 (host-port test transmit); fix two DUART timing deviations SDD 2026-10-16 12:07:27 +03:00
  • 3fd1e13b2e docs: keyboard/mouse protocol spec — LK401 (LK201 protocol) + VSXXX-AA mouse, sources, MAME lk201 device SDD 2026-10-16 11:38:52 +03:00
  • 23e508f075 docs: level-1 test sequence, port tests vs device tests (keyboard/mouse device models needed for tests 26/27) SDD 2026-10-16 11:31:56 +03:00
  • b38da304c3 docs: correct DUART report - host port test is local loopback (service guide), wiring verified sound, missing THRA write is the real bug SDD 2026-10-16 11:27:11 +03:00
  • 2fd9286c1a harness: wire MAME SCN2681 at the DUART windows (host-port investigation) SDD 2026-10-16 11:22:56 +03:00
  • 1c0a9b2b64 docs: Ghidra decompiler unreliable for this ROM (~85% of sampled functions flagged); SLEIGH instruction coverage incomplete SDD 2026-10-16 11:08:46 +03:00
  • 77c918f15c docs: next level-1 gate is HOST PORT (window 0x6400000); evidence and approach recorded SDD 2026-10-16 11:01:39 +03:00
  • 49378827e8 fix(tms34010): drop addr32 &0xFFFFF, fix ADDI imm; rebuild .sla, repair program, validate 31691 sites SDD 2026-10-16 10:59:40 +03:00
  • 5000b8bc07 harness: level-1 VIDEO MEMORY test passes (video RAM mapped + REFCNT advances); next gate is HOST PORT SDD 2026-10-16 10:57:06 +03:00
  • 5ef8712960 harness: identify the level-1 video-memory test that raises code 0x100 (VIDEO MEMORY-22) SDD 2026-10-16 10:55:44 +03:00
  • f46b4a44a4 docs: root-cause Ghidra TMS34010 CALLA bug to addr32 & 0xFFFFF (empty call graph) SDD 2026-10-16 10:48:02 +03:00
  • 3a1d10e9bb docs: consolidate level-1 investigation state; stack walk reaches the framework not the test SDD 2026-10-16 10:40:24 +03:00
  • 8a00cef556 docs: Ghidra CALLA targets do not resolve (empty call graph); smart34010 is authoritative for call targets SDD 2026-10-16 10:38:16 +03:00
  • 9a2363a68f docs: level-1 errors are status-code dispatched; withdraw the unsound failing-address claim SDD 2026-10-16 10:36:27 +03:00
  • cca49dc27d harness: VRAM transparent logging proxy; video RAM test is a walking-bit test and still fails SDD 2026-10-16 10:33:00 +03:00
  • 07997022ad docs: the video RAM test draws characters then verifies; mapping alone cannot satisfy it SDD 2026-10-16 10:30:52 +03:00
  • 52744cbd03 harness: correct video memory to bit 0x04000000; log the failing address; level 1 still fails SDD 2026-10-16 10:27:32 +03:00
  • f49f9984bb harness: close the instrument disagreement -- catch-all trustworthy, off = bit>>4 SDD 2026-10-16 10:25:17 +03:00