| Message ID | 000101d7a29e$8de86870$a9b93950$@nextmovesoftware.com |
|---|---|
| State | New |
| Headers | show |
| Series | Correct implementation of wi::clz | expand |
On 9/5/2021 3:39 PM, Roger Sayle wrote: > As diagnosed with Jakub and Richard in the analysis of PR 102134, the > current implementation of wi::clz has incorrect/inconsistent behaviour. > As mentioned by Richard in comment #7, clz should (always) return zero > for negative values, but the current implementation can only return 0 > when precision is a multiple of HOST_BITS_PER_WIDE_INT. The fix is > simply to reorder/shuffle the existing tests. > > This patch has been tested on x86_64-pc-linux-gnu with "make bootstrap" > and "make -k check" with no new failures. > > Ok for mainline? > > 2021-09-05 Roger Sayle <roger@nextmovesoftware.com> > > gcc/ChangeLog > * gcc/wide-int.cc (wi::clz): Reorder tests to ensure the result > is zero for all negative values. OK jeff
diff --git a/gcc/wide-int.cc b/gcc/wide-int.cc index 906f4ea..a142151 100644 --- a/gcc/wide-int.cc +++ b/gcc/wide-int.cc @@ -2050,6 +2050,10 @@ wi::arshift_large (HOST_WIDE_INT *val, const HOST_WIDE_INT *xval, int wi::clz (const wide_int_ref &x) { + if (x.sign_mask () < 0) + /* The upper bit is set, so there are no leading zeros. */ + return 0; + /* Calculate how many bits there above the highest represented block. */ int count = x.precision - x.len * HOST_BITS_PER_WIDE_INT; @@ -2058,9 +2062,6 @@ wi::clz (const wide_int_ref &x) /* The upper -COUNT bits of HIGH are not part of the value. Clear them out. */ high = (high << -count) >> -count; - else if (x.sign_mask () < 0) - /* The upper bit is set, so there are no leading zeros. */ - return 0; /* We don't need to look below HIGH. Either HIGH is nonzero, or the top bit of the block below is nonzero; clz_hwi is