diff options
author | Kito Cheng <kito.cheng@sifive.com> | 2021-07-19 11:50:22 +0300 |
---|---|---|
committer | Corinna Vinschen <corinna@vinschen.de> | 2021-07-21 10:56:04 +0300 |
commit | ca7b4bd23676a9cb6d82974c2ed452e08abb7ecf (patch) | |
tree | 6ede2d9e4b248c9e5796e2af172239d0f3d97706 /newlib | |
parent | ff9c2b3ab996b359645954f22647eb1c60c9cefa (diff) |
libm: Fixing overflow handling issue for scalbnf and scalbn
cc Aldy Hernandez <aldyh@redhat.com> and Andrew MacLeod <amacleod@redhat.com>,
they are author of new VRP analysis for GCC, just to make sure I didn't
mis-understanding or mis-interpreting anything on GCC site.
GCC 11 have better value range analysis, that give GCC more confidence
to perform more aggressive optimization, but it cause scalbn/scalbnf get
wrong result.
Using scalbn to demostrate what happened on GCC 11, see comments with VRP
prefix:
```c
double scalbn (double x, int n)
{
/* VRP RESULT: n = [-INF, +INF] */
__int32_t k,hx,lx;
...
k = (hx&0x7ff00000)>>20;
/* VRP RESULT: k = [0, 2047] */
if (k==0) {
/* VRP RESULT: k = 0 */
...
k = ((hx&0x7ff00000)>>20) - 54;
if (n< -50000) return tiny*x; /*underflow*/
/* VRP RESULT: k = -54 */
}
/* VRP RESULT: k = [-54, 2047] */
if (k==0x7ff) return x+x; /* NaN or Inf */
/* VRP RESULT: k = [-54, 2046] */
k = k+n;
if (k > 0x7fe) return huge*copysign(huge,x); /* overflow */
/* VRP RESULT: k = [-INF, 2046] */
/* VRP RESULT: n = [-INF, 2100],
because k + n <= 0x7fe is false, so:
1. -INF < [-54, 2046] + n <= 0x7fe(2046) < INF
2. -INF < [-54, 2046] + n <= 2046 < INF
3. -INF < n <= 2046 - [-54, 2046] < INF
4. -INF < n <= [0, 2100] < INF
5. n = [-INF, 2100] */
if (k > 0) /* normal result */
{SET_HIGH_WORD(x,(hx&0x800fffff)|(k<<20)); return x;}
if (k <= -54) {
/* VRP OPT: Evaluate n > 50000 as true...*/
if (n > 50000) /* in case integer overflow in n+k */
return huge*copysign(huge,x); /*overflow*/
else return tiny*copysign(tiny,x); /*underflow*/
}
k += 54; /* subnormal result */
SET_HIGH_WORD(x,(hx&0x800fffff)|(k<<20));
return x*twom54;
}
```
However give the input n = INT32_MAX, k = k+n will overflow, and then we
expect got `huge*copysign(huge,x)`, but new VRP optimization think
`n > 50000` is never be true, so optimize that into `tiny*copysign(tiny,x)`.
so the solution here is to moving the overflow handle logic before `k = k + n`.
Diffstat (limited to 'newlib')
-rw-r--r-- | newlib/libm/common/s_scalbn.c | 9 | ||||
-rw-r--r-- | newlib/libm/common/sf_scalbn.c | 9 |
2 files changed, 8 insertions, 10 deletions
diff --git a/newlib/libm/common/s_scalbn.c b/newlib/libm/common/s_scalbn.c index e9dbb16ac..72da0e4a1 100644 --- a/newlib/libm/common/s_scalbn.c +++ b/newlib/libm/common/s_scalbn.c @@ -93,15 +93,14 @@ tiny = 1.0e-300; if (n< -50000) return tiny*x; /*underflow*/ } if (k==0x7ff) return x+x; /* NaN or Inf */ + if (n > 50000) /* in case integer overflow in n+k */ + return huge*copysign(huge,x); /*overflow*/ k = k+n; if (k > 0x7fe) return huge*copysign(huge,x); /* overflow */ if (k > 0) /* normal result */ {SET_HIGH_WORD(x,(hx&0x800fffff)|(k<<20)); return x;} - if (k <= -54) { - if (n > 50000) /* in case integer overflow in n+k */ - return huge*copysign(huge,x); /*overflow*/ - else return tiny*copysign(tiny,x); /*underflow*/ - } + if (k <= -54) + return tiny*copysign(tiny,x); /*underflow*/ k += 54; /* subnormal result */ SET_HIGH_WORD(x,(hx&0x800fffff)|(k<<20)); return x*twom54; diff --git a/newlib/libm/common/sf_scalbn.c b/newlib/libm/common/sf_scalbn.c index 700060010..747a421ef 100644 --- a/newlib/libm/common/sf_scalbn.c +++ b/newlib/libm/common/sf_scalbn.c @@ -56,15 +56,14 @@ tiny = 1.0e-30; k = ((ix&0x7f800000)>>23) - 25; if (n< -50000) return tiny*x; /*underflow*/ } + if (n > OVERFLOW_INT) /* in case integer overflow in n+k */ + return huge*copysignf(huge,x); /*overflow*/ k = k+n; if (k > FLT_LARGEST_EXP) return huge*copysignf(huge,x); /* overflow */ if (k > 0) /* normal result */ {SET_FLOAT_WORD(x,(ix&0x807fffff)|(k<<23)); return x;} - if (k < FLT_SMALLEST_EXP) { - if (n > OVERFLOW_INT) /* in case integer overflow in n+k */ - return huge*copysignf(huge,x); /*overflow*/ - else return tiny*copysignf(tiny,x); /*underflow*/ - } + if (k < FLT_SMALLEST_EXP) + return tiny*copysignf(tiny,x); /*underflow*/ k += 25; /* subnormal result */ SET_FLOAT_WORD(x,(ix&0x807fffff)|(k<<23)); return x*twom25; |