Three fixes in chip.py:

1. Add early device-tree check for Broadcom BCM chips. Previously,
   BCM detection relied solely on /proc/cpuinfo Hardware field, which
   is absent on aarch64 Alpine Linux (and potentially other distros).
   The new check matches 'brcm,bcm2' in /proc/device-tree/compatible,
   which is present on all Raspberry Pi boards regardless of distro.

2. Guard against NoneType crash when compatible is None. If both
   /proc/cpuinfo Hardware and /proc/device-tree/compatible are
   unavailable, compatible.split() would raise AttributeError,
   which gets silently swallowed by __getattr__ and surfaces as
   a confusing bare AttributeError.

3. Include attribute name in __getattr__ AttributeError for better
   debugging.

Fixes #342

Add _orange_pi_id() to H616 and H6 chip detection paths so that
Orange Pi boards are detected via device-tree model strings on
non-Armbian OS distributions (e.g. Debian).

Expand _orange_pi_id() to handle Orange Pi Zero 2, Zero 3, Zero 2W
(with space variant), and 3 LTS model strings from /proc/device-tree/model.

Fixes #285

- Fix boards.py: Remove stray blank line inside BEAGLEBONE tuple that
  caused black formatting CI failure.
- Clean up board.py comment: The special-case EEPROM check is the
  correct approach for new-format Beagle boards (like BeaglePlay),
  so replaced the uncertain comment with an explanatory one.

Signed-off-by: David ML Brown Jr <dmlb2000@gmail.com>

Signed-off-by: David ML Brown Jr <dmlb2000@gmail.com>

Signed-off-by: David ML Brown Jr <dmlb2000@gmail.com>

This adds support for BananaPi P2 Pro, at least how I need it to work...

Signed-off-by: David ML Brown Jr <dmlb2000@gmail.com>

Those boards identify with detector.chip.ZYNQMP == True.

Adafruit Platform Detect is missing Beaglebone Black rev D boards id

Add Board name to list

Vicharak has introduced Vaaman and Axon board which uses
rk3399 and rk3588 respectively.

Signed-off-by: djkabutar <d.kabutarwala@yahoo.com>

JetPack 6.2 Jetson Nano Super compatibility

- Use glob builtin to find device on i2c address 0x50 matching `0-0050*`
  instead of checking paths one at a time