Re: Bug#65491: lpd generates unaligned traps
Peter Samuelson <email@example.com> writes:
> If I'm not mistaken, "unaligned trap" means someone is doing something
> void foo(char *x)
> int *c = (int *)x;
> and the char* passed in is not necessarily 32-bit-aligned. Some CPUs
> (like the Alpha) care deeply about this sort of thing, others (like the
> Intel) don't really, except for some loss of efficiency.
> Is that correct? Can someone definitively support or refute my
> understanding here?
Well, sometimes it is more subtle than that (in particular see the bug
report I submitted against ghostcript), but essentially, yes.
Linuxcare. Support for the revolution.