M-Labs
/
compiler-builtins-zynq
8
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Auto merge of #132 - parched:master, r=japaric 

mul.rs: use wrapping_mul to avoid infinite recursion

rustc in debug mode with a plain multiplication will call @llvm.umul.with.overflow.* which may call the builtin resulting in infinite recursion.

Fixes #127.

Cc @nezza this should be the fix you can try. Note I made this patch in github's web Ui so it's completely untested.
master
bors 2017-01-02 16:41:18 +00:00
parent 0fd5e7549f ee0aacf4be
commit 764557f0b6
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
src/int/mul.rs
View File

@ -9,18 +9,18 @@ macro_rules! mul {
pub extern "C" fn $intrinsic(a: $ty, b: $ty) -> $ty {
let half_bits = <$ty>::bits() / 4;
let lower_mask = !0 >> half_bits;
let mut low = (a.low() & lower_mask) * (b.low() & lower_mask);
let mut low = (a.low() & lower_mask).wrapping_mul(b.low() & lower_mask);
let mut t = low >> half_bits;
low &= lower_mask;
t += (a.low() >> half_bits) * (b.low() & lower_mask);
t += (a.low() >> half_bits).wrapping_mul(b.low() & lower_mask);
low += (t & lower_mask) << half_bits;
let mut high = t >> half_bits;
t = low >> half_bits;
low &= lower_mask;
t += (b.low() >> half_bits) * (a.low() & lower_mask);
t += (b.low() >> half_bits).wrapping_mul(a.low() & lower_mask);
low += (t & lower_mask) << half_bits;
high += t >> half_bits;
high += (a.low() >> half_bits) * (b.low() >> half_bits);
high += (a.low() >> half_bits).wrapping_mul(b.low() >> half_bits);
high = high.wrapping_add(a.high().wrapping_mul(b.low()).wrapping_add(a.low().wrapping_mul(b.high())));
<$ty>::from_parts(low, high)
}