Slow performance for right half of split board if using by itself and I2C

[nooges]

Describe the Bug

If using a right half of a split keyboard that does not have I2C pullup resistors on that half and doing #define USE_I2C, the performance on that half is slow. This is because the TIMEOUT value of 100ms gets hit during transport_master() in split_common/transport.c. Adding external pullups makes performance normal again.

Also, adding the following to keyboard_master_setup() in split_util.c to set the internal pullups seems to work as well:

setPinInputHigh(D0);
setPinInputHigh(D1);

Initial proposed solution, to prevent any impact on I2C rise time issues:

• Add flag indicating if right half has no external I2C pullups (NO_I2C_PULLUPS_RIGHT_HALF)
• Update keyboard_master_setup() to set internal pullups on I2C pins if right half & NO_I2C_PULLUPS_RIGHT_HALF is defined

System Information

• Keyboard:
  • Iris
  • Levinson
  • Others (w/no I2C pullups on right half)

Additional Context

TODO:

• Test with ARM split to verify if I2C pullups should be set for that

[omscarr]

Any update on this issue?
I forked this project and pulled @nooges change from PR #7532 and that fixed my issue on my Iris rev 2.5.

I'd like to move to a build from this repo instead of my fork at some point

[spoter368]

Experienced this bug today, it would be nice if there was a way for a the right half to detect it was on its own and enable the internal pullups, but if it detects a left half, do not enable the internal pullups.