Issues with I2C deadlocks

[addshore]

Case 1 & 2: I2C device has been communicated with before, but stops part way through operations for a period of time

Potential dead-lock detected:
  Process: 3
  Program: 0f3f261e-541b-c138-fef7-199b44a7a1dc
  BCI: 0x3db
  Primitive: 6:2
fatal: Potential dead-lock

abort() was called at PC 0x4202e367 on core 0
Core  0 register dump:
MEPC    : 0x40806520  RA      : 0x408142ae  SP      : 0x4082fd90  GP      : 0x40822e84
TP      : 0x4082ffe0  T0      : 0x37363534  T1      : 0x7271706f  T2      : 0x33323130
S0/FP   : 0x4082fdcc  S1      : 0x4082fdcc  A0      : 0x4082fdcc  A1      : 0x4082fdae
A2      : 0x00000000  A3      : 0x4082fdf9  A4      : 0x00000001  A5      : 0x4082b000
A6      : 0x00000000  A7      : 0x76757473  S2      : 0x4082fdb0  S3      : 0x00000001
S4      : 0x027e1356  S5      : 0x00000005  S6      : 0x40836140  S7      : 0x004c4b40
S8      : 0x00000000  S9      : 0x00000000  S10     : 0x00000000  S11     : 0x00000000
T3      : 0x6e6d6c6b  T4      : 0x6a696867  T5      : 0x66656463  T6      : 0x62613938
MSTATUS : 0x00001881  MTVEC   : 0x40800001  MCAUSE  : 0x00000007  MTVAL   : 0x00000000
MHARTID : 0x00000000

Potential dead-lock detected:
  Process: 3
  Program: 8ba5f2ca-8927-4edc-5056-5ba029f2eb90
  BCI: 0x3e5
  Primitive: 6:7
fatal: Potential dead-lock

abort() was called at PC 0x4202e367 on core 0
Core  0 register dump:
MEPC    : 0x40806520  RA      : 0x408142ae  SP      : 0x4082fd90  GP      : 0x40822e84
TP      : 0x4082ffe0  T0      : 0x37363534  T1      : 0x7271706f  T2      : 0x33323130
S0/FP   : 0x4082fdcc  S1      : 0x4082fdcc  A0      : 0x4082fdcc  A1      : 0x4082fdae
A2      : 0x00000000  A3      : 0x4082fdf9  A4      : 0x00000001  A5      : 0x4082b000
A6      : 0x00000000  A7      : 0x76757473  S2      : 0x4082fdb0  S3      : 0x00000002
S4      : 0x02c49e66  S5      : 0x00000005  S6      : 0x40836208  S7      : 0x004c4b40
S8      : 0x00000000  S9      : 0x00000000  S10     : 0x00000000  S11     : 0x00000000
T3      : 0x6e6d6c6b  T4      : 0x6a696867  T5      : 0x66656463  T6      : 0x62613938
MSTATUS : 0x00001881  MTVEC   : 0x40800001  MCAUSE  : 0x00000007  MTVAL   : 0x00000000
MHARTID : 0x00000000

Case 3: I2C device I am talking to will not respond to anything.

[jaguar] INFO: program 9d8d7d0a-1109-640c-b897-53a2b2f25735 starPotential dead-lock detected:
  Process: 3
  Program: 9d8d7d0a-1109-640c-b897-53a2b2f25735
  BCI: 0x377a
  Primitive: 6:4
fatal: Potential dead-lock

abort() was called at PC 0x4202e367 on core 0
Core  0 register dump:
MEPC    : 0x40806520  RA      : 0x408142ae  SP      : 0x4082fd90  GP      : 0x40822e84
TP      : 0x4082ffe0  T0      : 0x37363534  T1      : 0x7271706f  T2      : 0x33323130
S0/FP   : 0x4082fdcc  S1      : 0x4082fdcc  A0      : 0x4082fdcc  A1      : 0x4082fdae
A2      : 0x00000000  A3      : 0x4082fdf9  A4      : 0x00000001  A5      : 0x4082b000
A6      : 0x00000000  A7      : 0x76757473  S2      : 0x4082fdb0  S3      : 0x00000001
S4      : 0x021b3f67  S5      : 0x00000005  S6      : 0x40836140  S7      : 0x004c4b40
S8      : 0x00000000  S9      : 0x00000000  S10     : 0x00000000  S11     : 0x00000000
T3      : 0x6e6d6c6b  T4      : 0x6a696867  T5      : 0x66656463  T6      : 0x62613938
MSTATUS : 0x00001881  MTVEC   : 0x40800001  MCAUSE  : 0x00000007  MTVAL   : 0x00000000
MHARTID : 0x00000000

[floitsch]

Normally, primitive operations are not allowed to block for longer periods of time.
The i2c operation is currently blocking, which usually isn't a problem, since the operation shouldn't take long. However, if the transaction takes too long (potentially because of a bad device, clock stretching, ...) a watchdog triggers and reboots the device.

We need to make the i2c operation asynchronous to fix this. I already need to update the driver to use the newer ESP-IDF APIs. When doing so, I will try to change it so it isn't blocking anymore.

[addshore]

Another one today, and I noticed its Primitive: 6:5 instead, so figured dumping it here might be useful?

Potential dead-lock detected:
  Process: 8
  Program: bae9bfbc-56ff-a982-a636-e0d62c461578
  BCI: 0x3c8d
  Primitive: 6:5
fatal: Potential dead-lock

abort() was called at PC 0x4202e001 on core 0
Core  0 register dump:
MEPC    : 0x408065de  RA      : 0x408145e8  SP      : 0x4082fd90  GP      : 0x40823204
TP      : 0x4082ffe0  T0      : 0x37363534  T1      : 0x7271706f  T2      : 0x33323130
S0/FP   : 0x4082fdcc  S1      : 0x4082fdcc  A0      : 0x4082fdcc  A1      : 0x4082fdae
A2      : 0x00000000  A3      : 0x4082fdf9  A4      : 0x00000001  A5      : 0x4082c000
A6      : 0x00000000  A7      : 0x76757473  S2      : 0x4082fdb0  S3      : 0x00000002
S4      : 0x5ce7e162  S5      : 0x00000005  S6      : 0x408316f0  S7      : 0x004c4b40
S8      : 0x00000000  S9      : 0x00000000  S10     : 0x00000000  S11     : 0x00000000
T3      : 0x6e6d6c6b  T4      : 0x6a696867  T5      : 0x66656463  T6      : 0x62613938
MSTATUS : 0x00001881  MTVEC   : 0x40800001  MCAUSE  : 0x00000007  MTVAL   : 0x00000000
MHARTID : 0x00000000

[floitsch]

6:5 is I2C-read.
Almost certainly the same reason: the device pulls the clock low and delays the I2C transaction. -> we run into a timeout.