CB3.0 F100 - I2C issues

[pito]

Code: Select all

Thread [1] (Suspended: Signal 'SIGINT' received. Description: Interrupt.)   
   5 chSysHalt() core_cmFunc.h:342 0x08001018   
   4 i2c_lld_set_clock() i2c_lld.c:188 0x08001018   
   3 i2c_lld_start() i2c_lld.c:629 0x08001018   
   2 i2cStart() i2c.c:105 0x08001018   
   1 main() main.c:253 0x08001018   

at

Code: Select all

      osalDbgAssert((STM32_PCLK1 % (clock_speed * 25)) == 0,
                    "PCLK1 must be divided without remainder");

with

Code: Select all

static const I2CConfig i2cfg1 = { OPMODE_I2C, 400000,  FAST_DUTY_CYCLE_16_9 }

@24MHz clock

With:

Code: Select all

static const I2CConfig i2cfg1 = { OPMODE_I2C, 400000,  FAST_DUTY_CYCLE_2 }

I get always:

Code: Select all

Thread [1] (Suspended: Signal 'SIGINT' received. Description: Interrupt.)   
   7 chSysHalt() core_cmFunc.h:342 0x08002dfc   
   6 _dbg_check_lock() chdebug.c:119 0x08002dfc   
   5 chSysLock() chsys.h:290 0x0800322c   
   4 i2c_lld_master_transmit_timeout() osal.h:377 0x0800322c   
   3 i2cMasterTransmitTimeout() i2c.c:189 0x080009ca   
   2 RTC_PCF8563_isrunning() chTinyRTC8563.c:165 0x080009ca   
   1 main() main.c:260 0x080009ca   

[Giovanni]

I don't see errors, you cannot select that specific combination of parameters, you have to respect constraints when specifying the working parameters.

See the documentation of the I2C CCR register.

Giovanni

[pito]

I run with

Code: Select all

static const I2CConfig i2cfg1 = { OPMODE_I2C, 400000,  FAST_DUTY_CYCLE_16_9 }

and 40MHz clock my code fine under trunk.
The same with 3.0 (and w/ the changes we discussed earlier) stops. So there must be something different, I do not see on the third glance
I see a new parameter

Code: Select all

#define STM32_I2C_BUSY_TIMEOUT              50

there. Maybe something related to?

[Giovanni]

Nope, that is unrelated.

Probably you just run without assertions enabled on trunk code, that part of the code didn't change at all.

Giovanni

[pito]

It hangs on

Code: Select all

RTC_PCF8563_status = i2cMasterTransmitTimeout(&I2CD1, PCF8563_ADDRESS, txbuf, 1, rxbuf, 9, TMO);

[Giovanni]

Hangs, how? halt, exception, assertion, simply no return? more info would help, I have no HW to test this driver right now.

Giovanni

[pito]

As indicated above:

Code: Select all

Thread [1] (Suspended: Signal 'SIGINT' received. Description: Interrupt.)   
   7 chSysHalt() core_cmFunc.h:342 0x08002e1c   
   6 _dbg_check_lock() chdebug.c:119 0x08002e1c   
   5 chSysLock() chsys.h:290 0x0800324c   
   4 i2c_lld_master_transmit_timeout() osal.h:377 0x0800324c   
   3 i2cMasterTransmitTimeout() i2c.c:189 0x080009dc   
   2 RTC_PCF8563_isrunning() chTinyRTC8563.c:165 0x080009dc   
   1 main() main.c:263 0x080009dc   

Code: Select all

uint8_t RTC_PCF8563_isrunning(void) {

   uint8_t rxbuf[RTC_PCF8563_RX_DEPTH];
   uint8_t txbuf[1];
   txbuf[0]= 0x00;
   i2cAcquireBus(&I2CD1);
>>>   RTC_PCF8563_status = i2cMasterTransmitTimeout(&I2CD1, PCF8563_ADDRESS, txbuf, 1, rxbuf, 9, TMO);
   i2cReleaseBus(&I2CD1);

   if (RTC_PCF8563_status==RDY_OK) return TRUE;
   else return FALSE;
}

TMO is: #define TMO MS2ST(20)

[pito]

In RT-STMF100-DISCOVERY there is probably wrong def

Code: Select all

MCU  = cortex-m4

What I do not understand yet is my compiled code under 3.0 is 5kB smaller than with trunk..

Code: Select all

 
3.0:
  text      data       bss       dec       hex   filename
  14384       428      4528     19340      4b8c   build/ch.elf
trunk:
   text      data       bss       dec       hex   filename
  19492       356      4416     24264      5ec8   build/ch.elf

The mcuconf, cbconf, halconf, Makefile almost identical (plus minus changes with 3.0).

[Giovanni]

Probably because LTO is now enabled by default. I will fix the makefiles but it should not affect anything.

Giovanni

[pito]

LTO disabled:

Code: Select all

   text      data       bss       dec       hex   filename
  17364       428      4528     22320      5730   build/ch.elf

2kB less, do I miss something there?
Still hangs on the I2C..
I removed the I2C stuff, printing some adc values and systime, ( with spi, adc, and pwm chunks from demos) in a loop. It does few loops, prints, then stops at 80msec with unhandled exception.. Probably there is more changes as I am aware..