title: I2C; fix issues when using the slave mode
team: IO
key: cd177762-bc48-11ea-af02-1fe19e7a99a4
assignee: Wolfram
status: Done

upstream:
  - torvalds: 314139f9f0abdba61ed9a8463bbcb0bf900ac5a2 # i2c: rcar: slave: only send STOP event when we have been addressed
  - torvalds: 914a7b3563b8fb92f976619bbd0fa3a4a708baae # i2c: rcar: in slave mode, clear NACK earlier
  - torvalds: eb01597158ffb1853a7a7fc2c57d4c844640f75e # i2c: rcar: always clear ICSAR to avoid side effects
  - torvalds: c7c9e914f9a0478fba4dc6f227cfd69cf84a4063 # i2c: rcar: avoid race when unregistering slave

comments:
  - "Issue 1: we should not monitor every STOP but only when we have been addressed"
  - "v1 https://patchwork.kernel.org/patch/11632397/"
  - "v2 https://patchwork.kernel.org/patch/11632107/"
  - "v1 merged, v2 was not really better"

  - "Issue 2: how to properly handle the FNA bit"
  - "working prototype exists, waiting for confirmation from the HW team"
  - "HW team acknowledged the prototype"
  - "v1 https://patchwork.kernel.org/patch/11718127/"
  - "v1 merged"

  - "Issue 3: do we need to clear SAR when disabling the slave interface?"
  - "working prototype exists, needs confirmation from the HW team"
  - "only applies to Gen2, Gen3 is fine"
  - "v1: https://patchwork.kernel.org/patch/11643625/"
  - "v1 merged"

  - "Issue 4: race when unregistering slave"
  - "v1: https://patchwork.kernel.org/patch/11685989/"
  - "v1 merged"