On Fri, Nov 23 2018, Segher Boessenkool wrote:
> On Fri, Nov 23, 2018 at 09:01:56PM +0100, Helmut Eller wrote:
>> 2.) %rsp is adjusted before calling __addvdi3. Why is that needed?
>
> To keep the stack aligned (to 16 bytes).
Hmm... I guess %rsp is initially 16 byte aligned and we need to add 8
because the call instruction also pushes the return address.
>> 3.) Obviously __addvdi3 is not implemented as sibling-call even though
>> -O2 should enable that.
>
> It calls via the PLT, do sibling calls via the PLT work in your ABI?
I think so. At least
extern long f (long x, long y);
long g (long x, long y) { return f(x, y); }
is compiled to:
g:
jmp f@PLT
>> Where should I start, if I wanted to teach GCC how to produce the same
>> code for foo as for bar? Would it be enough to add a pattern to
>> i386.md? There is already a pattern for "addv<mode>4", but apparently
>> it's not used in this case.
>
> As Marc says, -ftrapv is probably not the way to go.
>
> Adding an addv<mode>3 to the i386 backend might help.
Maybe I'll try it; out of interest.
> You do *not* want exactly the same code, btw; addv3 calls abort on
> overflow, that's not the same as executing an ud2 instruction.
Hmm, yes. I would prefer ud2, as that is more compact.
Helmut
