Auto-vectorizer and (mis-)alignment support assumptions

[Eric Botcazou]

> may_trap_or_fault_p() does the right thing by taking the misalignement
> of stack accesses into account on STRICT_ALIGN targets. Would it be a
> solution to call that instead may_trap_p() from
> haifa-sched.c:may_trap_exp() ?

That wouldn't be the first time that this replacement is done.

> I'm not clear why may_trap_p() and may_trap_or_fault_p() are different
> functions.

And yet there is a lenghty explanation. :-)

> When would we want to disallow a trap, but allow a misalignment fault?

You use may_trap_p when you're querying the semantical properties of an 
instruction in a valid program: if it returns true, then the instruction may 
generate a trap in a valid program.

You use may_trap_or_fault_p when you're querying the actual properties of an 
instruction whatever the context: if it returns true, then the instruction may 
generate a trap in a valid program or may generate a fault if it appears in a 
context that isn't appropriate.

-- 
Eric Botcazou