CoreS3-SE FT6336U identity 64/03/20 triggers misleading version-read warning

[marosell]

On a physical M5Stack CoreS3-SE, M5GFX 0.2.27 logs:

CoreS3 touch version read failed (CIPHER:0x64 / FIRMID:0x03 / VENDID:0x20), panel:ILI9342C

Environment:

• M5GFX 0.2.27
• M5Unified 0.2.20
• ESP-IDF 5.5.4
• board autodetected as board_M5StackCoreS3SE

The reads themselves succeed and touch remains fully functional. The warning occurs because Panel_M5StackCoreS3::initPanelByTouchVersion accepts only VENDID 0x11 with FIRMID 0x10 or 0x12. It then takes the ILI9342C fallback.

M5Stack documentation identifies the CoreS3-SE display as ILI9342C and touch controller as FT6336U, so treating identity 0x64/0x03/0x20 as the ILI9342C variant appears appropriate:
https://docs.m5stack.com/en/core/M5CoreS3%20SE

Could M5Stack confirm that this tuple is an expected CoreS3-SE hardware revision? If so, I can submit a focused PR that recognizes it as ILI9342C while retaining the warning for genuinely unknown identities.

[ainyan03]

Thanks for the report and the measurements. The touch-firmware tuple turned out to be independent of the panel (three different tuples were seen on ILI9342C units), so instead of extending the whitelist, #296 (now in develop) asks the LCD controller itself which command set it has and creates the matching panel class from the start. The touch-based check, its retries and the "version read failed" warning are gone, so a CoreS3-SE boots without the warning and without the 280 ms delay. It will be in the next release.

[ainyan03]

M5GFX 0.2.30 is out with the panel-based detection from #296 (and a follow-up, #300, that reads the panel before its reset on the Core2 / Tough). Please update and, if the warning and the start-up delay are gone on your CoreS3-SE, close this issue.

What changed: the ILI9342C / ILI9342E variant is now identified from the panel's own registers at detection time, so the touch-firmware tuple is no longer consulted, the "version read failed" warning and its retries are gone, and a CoreS3-SE with an ILI9342C takes the C path directly.

[marosell]

Confirmed fixed with M5GFX 0.2.30 on the same physical CoreS3-SE originally reported.
Environment:

• M5GFX 0.2.30
• M5Unified 0.2.20
• ESP-IDF 5.5.4
• CoreS3-SE with ILI9342C and FT6336U
  Serial startup now reports:
  M5GFX: [Autodetect] board_M5StackCoreS3SE
  M5GFX: [Autodetect] ILI9342 read-back DDh:00 CBh:14 ID4:1342 -> ILI9342C
  The previous CoreS3 touch version read failed warning and retry delay are gone. The display and touchscreen are operating correctly.
  Thank you for implementing the panel-based detection. This issue can be closed.