On 06/24/2014 09:52 AM, Al Viro wrote:
> unsigned long __attribute__((noinline)) f(double x)
> {
> return (unsigned long)x; // SVCTQ/SVC
> }
>
> main()
> {
> unsigned long x;
> extern unsigned long __ieee_get_fp_control(void);
> printf("before:%lx\n", __ieee_get_fp_control());
> x = f(1ULL<<63);
> printf("after:%lx\n", __ieee_get_fp_control());
> printf("result:%lx\n", x);
> }
>
> On actual hardware:
> before:0
> after:20000
> result:8000000000000000
>
> On qemu:
> before:0
> after:0
> result:8000000000000000
>
>
> IOW, gen_fcvttq() is also affected, not only gen_fp_exc_raise().
Clearly a gross misunderstanding of what bits are actually computed, never mind
what gets signaled.
Thanks for the test. I've not had working hardware for a couple of years to
validate what's supposed to get set and what isn't.
> Can't we simply have separate helpers for various trap suffices, with
> all this work on getting exc, etc. taken inside them? It's not as if
> we distinguished many variants, after all... Right now we have:
> plain, /U, /V
> /S, /SU
> /SUI
> /SV
> /SVI
We used to have separate helpers... at least for the modes that had been
implemented at the time. The combinatorial explosion ugly though -- 4
different versions of add, sub, etc. I thought the partial inlining was a
decent solution, as far as maintainability, but it's not unreasonable to
disagree.
> Another thing: shouldn't arithmetics on denorms without /S raise EXC_M_INV,
> rather than EXC_M_UNF?
No idea. Should they?
r~
