forked from OSchip/llvm-project
[APFloat] Improve asserts in isSignificandAllOnes and isSignificandAllZeros so they protect shift operations from undefined behavior.
For example, the assert in isSignificandAllZeros allowed NumHighBits to be integerPartWidth. But since it is used directly as a shift amount it must be less than integerPartWidth.
This commit is contained in:
parent
d4a1db4f3f
commit
12bdd427b3
|
@ -850,8 +850,8 @@ bool IEEEFloat::isSignificandAllOnes() const {
|
|||
// Set the unused high bits to all ones when we compare.
|
||||
const unsigned NumHighBits =
|
||||
PartCount*integerPartWidth - semantics->precision + 1;
|
||||
assert(NumHighBits <= integerPartWidth && "Can not have more high bits to "
|
||||
"fill than integerPartWidth");
|
||||
assert(NumHighBits <= integerPartWidth && NumHighBits > 0 &&
|
||||
"Can not have more high bits to fill than integerPartWidth");
|
||||
const integerPart HighBitFill =
|
||||
~integerPart(0) << (integerPartWidth - NumHighBits);
|
||||
if (~(Parts[PartCount - 1] | HighBitFill))
|
||||
|
@ -870,9 +870,10 @@ bool IEEEFloat::isSignificandAllZeros() const {
|
|||
if (Parts[i])
|
||||
return false;
|
||||
|
||||
// Compute how many bits are used in the final word.
|
||||
const unsigned NumHighBits =
|
||||
PartCount*integerPartWidth - semantics->precision + 1;
|
||||
assert(NumHighBits <= integerPartWidth && "Can not have more high bits to "
|
||||
assert(NumHighBits < integerPartWidth && "Can not have more high bits to "
|
||||
"clear than integerPartWidth");
|
||||
const integerPart HighBitMask = ~integerPart(0) >> NumHighBits;
|
||||
|
||||
|
|
Loading…
Reference in New Issue