USBPD Analyzer often lags with I2C timeout

[tom-van]

Two FreeRTOS threads call at least BSP_PWRMON_GetVoltage() without locking I2C by a mutex. When another I2C operation is re-entered while the first one is not finished, I2C fails and waits till 3 sec timeout. Mainly happens at https://github.com/STMicroelectronics/STM32CubeG0/blob/fcc63f394b69d25d005c431be60960d2c22acab8/Projects/STM32G071B-DISCO/Demonstrations/USBPD_Analyzer/Src/demo_disco.c#L1347

Mutex guarded version BSP_USBPD_PWR_VBUSGetVoltage() should be used instead of BSP_PWRMON_GetVoltage(ALERT_VBUS, ...). Similarly BSP_USBPD_PWR_VBUSGetCurrent() should replace BSP_PWRMON_GetCurrent(ALERT_VBUS, ...). Also other I2C should be guarded similarly, e.g. BSP_PWRMON_GetVoltage(ALERT_CC1 / ALERT_CC2 ...) but there is no direct replacement in BSP_USBPD_PWR_ module.

[ALABSTM]

Hi @tom-van,

Your point looks relevant. Let's ask the development teams and see their feedback.

With regards,

[ALABSTM]

ST Internal Reference: 201898

[tom-van]

After writing this issue I found that in spy mode the mutex is probably not initialized and therefore BSP_USBPD_PWR_VBUSGetVoltage() doesn't work. I fixed the problem using condition similar to https://github.com/STMicroelectronics/STM32CubeG0/blob/fcc63f394b69d25d005c431be60960d2c22acab8/Projects/STM32G071B-DISCO/Demonstrations/USBPD_Analyzer/Src/demo_disco.c#L1241-L1244 however I didn't go to the depth code to find if mutex locking is or isn't necessary in spy mode. Perhaps collisions are not so frequent in spy mode. I would recommend use a mutex always.