Core Panic
MagnusThome opened this issue · 6 comments
Hi!
Using an MCP2515 with interrupt connected to the ESP32 and also writing to EEPROM in my code I get this crash below. It depends a bit on timing but if an interrupt arrives around the time I'm writing to EEPROM it crashes. Everything else works beautifully 👍
21:21:55.295 -> Guru Meditation Error: Core 0 panic'ed (Cache disabled but cached memory region accessed)
21:21:55.330 -> Core 0 register dump:
21:21:55.330 -> PC : 0x400d15b0 PS : 0x00060034 A0 : 0x40084764 A1 : 0x3ffbe1b0
21:21:55.330 -> A2 : 0x0000001a A3 : 0x3ffc06c4 A4 : 0x00000000 A5 : 0x04000000
21:21:55.330 -> A6 : 0x00000000 A7 : 0x13001248 A8 : 0x80081138 A9 : 0x00000001
21:21:55.330 -> A10 : 0x00000000 A11 : 0x00000000 A12 : 0x3ffc08e0 A13 : 0x00002004
21:21:55.364 -> A14 : 0x3ffc08e0 A15 : 0xffffdffb SAR : 0x0000001f EXCCAUSE: 0x00000007
21:21:55.364 -> EXCVADDR: 0x00000000 LBEG : 0x00000000 LEND : 0x00000000 LCOUNT : 0x00000000
21:21:55.364 -> Core 0 was running in ISR context:
21:21:55.364 -> EPC1 : 0x400867f7 EPC2 : 0x00000000 EPC3 : 0x00000000 EPC4 : 0x400d15b0
21:21:55.364 ->
21:21:55.364 -> Backtrace: 0x400d15b0:0x3ffbe1b0 0x40084761:0x3ffbe1d0 0x400867f4:0x3ffb9d80 0x40082b5f:0x3ffb9da0 0x40088379:0x3ffb9dc0
21:21:55.398 ->
21:21:55.398 -> Rebooting...
21:21:55.398 -> ets Jun 8 2016 00:22:57
21:21:55.398 ->
21:21:55.398 -> rst:0x3 (SW_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
21:21:55.398 -> configsip: 0, SPIWP:0xee
21:21:55.398 -> clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
21:21:55.398 -> mode:DIO, clock div:1
21:21:55.398 -> load:0x3fff0018,len:4
21:21:55.398 -> load:0x3fff001c,len:1044
21:21:55.398 -> load:0x40078000,len:8896
21:21:55.398 -> load:0x40080400,len:5816
21:21:55.431 -> entry 0x400806ac
Are there any possible workarounds? Doing CAN1.disable and enable before and after the EEPROM write doesn't help.
I run the ESP32 as a bridge with two CAN interfaces. The ONLY changes I've done to the library source is this, works great!
////////////////////////////////////////////////////////////
Edit in library:
library\esp32_can\src\esp32_can.cpp
ESP32CAN CAN0(GPIO_NUM_16, GPIO_NUM_17);
//MCP2517FD CAN1(5, 27);
MCP2515 CAN1(5, 26);
library\esp32_can\src\esp32_can.h
//extern MCP2517FD CAN1;
extern MCP2515 CAN1;
////////////////////////////////////////////////////////////
Connect:
ESP32 MCP2515
----- -------
GPIO5 CS
GPIO23 SI
GPIO19 SO
GPIO18 SCK
GPIO26 INT
5V VCC
GND GND
ESP32 SN65HVD230 (or similar transceiver)
----- ----------
GPIO16 CRX
GPIO17 CTX
3V3 3V3
GND GND
////////////////////////////////////////////////////////////
This test code below crashes every time when I have an OBD2 display transmitting traffic to the ESP32 but of course its a bit time sensitive. Seems to only crash when there is an interrupt around the time of the EEPROM write. With my OBD2 display the timing hits the mark and I get a crash 100% of the time.
#include "esp32_can.h"
#include "EEPROM.h"
CAN_FRAME frame;
void setup() {
EEPROM.begin(401);
CAN1.begin(500000);
CAN1.watchFor();
}
void loop() {
if (CAN1.read(frame)) {
frame.id = 0x7E8;
frame.data.uint8[0] = 0x03;
frame.data.uint8[1] = 0x41;
frame.data.uint8[3] = 0xFF;
CAN1.sendFrame(frame);
EEPROM.write(0, 1); // beware of wear of the eeprom...
EEPROM.commit();
}
}
Can you try power mcp2515 with 3.3v? Esp32 gpio`s only support 3.3v. If you power mcp2515 with 5v, it should give you problems in theory. And transceiver can be powered by 3.3v or 5v, depends on model/datasheet specs. If that doesnt help, try use mcp2517 or use SPIFFS instead of eeprom
Thanks for the input! Using a level shifter doesn't help so I doubt running on 3.3V will help but SPIFFS is a good idea!!! I'll look into that :-)
Just a wild guess from my side but I was wondering if it rather has to do with some irq calls not being in IRAM?
If you tested this code without eeprom and you are 100% sure that its eeprom library, problem not mcp2515 wiring, it could be that writing eeprom(flash) at very high speeds causes somehow core panic. Your code above, constantly writes eeprom at high speeds. Eeprom examples has delay(20); after every eeprom write.I dont think that irq calls are your issue here
Again, thanks for your input!
-
If I add a lot of delay between writes it still crashes like this.
-
It only crashes when the irq pin has an input signal triggering it. No trigger on the irq pin = no crashes at all.
This might be relevant. TL;DR: There are two limitations on the usable SPI clock speed. One is a yet unknown issue with the mcp25xxfd: it doesn't properly work with SYSCLOCK / 2 speed. Based on trial and error, the SPI clock speed was reduced to 92.5% of SYSCLOCK / 2 for the Raspberry Pi 4. This might be different for the ESP32 though.
Hi !!!
I got the same Problem. And use Exception Decoder ! It shows :
Guru Meditation Error: Core 0 panic'ed (Cache disabled but cached memory region accessed)
Core 0 register dump:
PC : 0x400d7540 PS : 0x00060034 A0 : 0x40084888 A1 : 0x3ffbe1b0
A2 : 0x0000001b A3 : 0x3ffc0980 A4 : 0x00000000 A5 : 0x08000000
A6 : 0x00000000 A7 : 0x13001248 A8 : 0x80081148 A9 : 0x00000001
A10 : 0x00000000 A11 : 0x00000000 A12 : 0x800885ce A13 : 0x3ffbe180
A14 : 0x00000000 A15 : 0x00000000 SAR : 0x0000001a EXCCAUSE: 0x00000007
EXCVADDR: 0x00000000 LBEG : 0x00000000 LEND : 0x00000000 LCOUNT : 0x00000000
Core 0 was running in ISR context:
EPC1 : 0x40086923 EPC2 : 0x00000000 EPC3 : 0x00000000 EPC4 : 0x400d7540
Backtrace: 0x400d7540:0x3ffbe1b0 0x40084885:0x3ffbe1d0 0x40086920:0x3ffb9d80 0x40082c83:0x3ffb9da0 0x400884dd:0x3ffb9dc0
Rebooting...
ets Jun 8 2016 00:22:57
rst:0x3 (SW_RESET),boot:0x17 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:1
load:0x3fff0018,len:4
load:0x3fff001c,len:1216
ho 0 tail 12 room 4
load:0x40078000,len:9720
ho 0 tail 12 room 4
load:0x40080400,len:6352
entry 0x400806b8
PC: 0x400d7540: MCP_INTHandler() at C:\Users\telem\OneDrive\Documents\Arduino\libraries\esp32_can-master\src\mcp2515.cpp line 43
EXCVADDR: 0x00000000
Decoding stack results
0x400d7540: MCP_INTHandler() at C:\Users\telem\OneDrive\Documents\Arduino\libraries\esp32_can-master\src\mcp2515.cpp line 43
0x40086920: spi_flash_op_block_func at /home/runner/work/esp32-arduino-lib-builder/esp32-arduino-lib-builder/esp-idf/components/spi_flash/cache_utils.c line 82
0x40082c83: ipc_task at /home/runner/work/esp32-arduino-lib-builder/esp32-arduino-lib-builder/esp-idf/components/esp32/ipc.c line 62
0x400884dd: vPortTaskWrapper at /home/runner/work/esp32-arduino-lib-builder/esp32-arduino-lib-builder/esp-idf/components/freertos/port.c line 143
And i did try to use CAN0 instead of CAN1 .There's no problem. What's wrong with MCP_INTHandler() at C:\Users\telem\OneDrive\Documents\Arduino\libraries\esp32_can-master\src\mcp2515.cpp line 43 ????