Hi all,
First things first, I'm using:
- CY8C5468AXI-LP106
- PSoC Creator 4.3
- ARM-GCC 5.4-2016-q2-update
I've got a weird problem with strtok() that has suddenly arisen. I have a Tokenize() function that is written as follows:
/** Tokenize
* - Splits incoming string into individual substrings
* - Splits on comma, space, or underscore characters
* - This should normally break out upon finding strtok() finding the end of the input string
* For some reason, this function can continue running even when pch==NULL.
* - Added a break if count >= MAX_NUM_CMDS to protect from this overrun.
*/
static void Tokenize(char* src, char dest[MAX_NUM_CMDS][CMD_MAX_CHAR]){
uint8_t count = 0;
char * pch;
pch = strtok(src, ", _");
while (pch != NULL){
memcpy(dest[count++], pch, CMD_MAX_CHAR);
pch = strtok(NULL, ", _");
if (count >= MAX_NUM_CMDS){
break;
}
}
}
The purpose is to take a string (input over serial) and split it into arguments of <5 characters that I can then parse and act on. The problem I am having is that when I send a command, strtok() is getting stuck and never returning.
The string I am sending to reproduce this is simple, "00\r\n". The first pch=strtok(src, ", _"); line performs as expected, I drop into the while() loop, and the 2nd call pch=strtok(NULL, ", _") is what gets stuck. I can confirm the entire string "00\r\n" is getting copied to dest[0], so the call should just give back null (the rest of src is just null characters, 0x00), and I should exit the while loop. Instead I am getting stuck inside strtok() on this call.
It is getting stuck in looping endlessly inside strtok_r when I step through in the disassembly:
0x000077AC ldrb.w r6, [r7], #1
0x000077B0 cmp r5, r6
0x000077B2 beq.n 77c4 <__strtok_r+0x48>
0x000077B4 cmp r6, #0
0x000077B6 bne.n 77ac <__strtok_r+0x30>
After banging my head against it for a while, I tried disabling newlib-nano in the compiler settings, and suddenly strtok() works great and I have no problems. I guess I can just disable newlib-nano as a workaround, but I really should not have to.
Has anyone else had an issue like this? I've used this code many times before with newlib and have not had this problem until now. Even this identical firmware that this is taken from worked absolutely fine until today I got a new revision of the PCB and now this is suddenly an issue. One thing that did change was the exact PSoC being used. Due to shortages all we have available is the CY8C5468AXI-LP106. Originally this system/firmware used a CY8C5667AXI-LP040.
Any thoughts/ideas? it's a bit of a head scratcher.
Thanks!
