Hello Curt,
first look through the code, my brain triggered on something in memory
that seemed to have survived time despite wear and beer:
lwip_standard_chksum2(void *dataptr, int len)
{
u8_t *pb = dataptr;
u16_t *ps, t = 0;
u32_t sum = 0;
int odd = ((u32_t)pb & 1);
/* Get aligned to u16_t */
if (odd && len > 0) {
((u8_t *)&t)[1] = *pb++;
This fetches a 16-bit word, possible on a non-16 bit aligned address.
Some CPU's will puke at this point and stop execution. I wish I could
name a few CPUs with 100% certainty but I cannot now, though I am sure
have come across this on some 16 and 32-bit platforms.
What should be done here is fetch a 8-bit number on a 8-bit aligned
address (which is always the case :-)).
Regards,
Leon.
_______________________________________________
lwip-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/lwip-users
