RFR: 8267332: xor value should handle bounded values [v3]

[Claes Redestad]

On Thu, 20 May 2021 21:50:46 GMT, Nils Eliasson <neliasso at openjdk.org> wrote:

>> src/hotspot/share/opto/addnode.cpp line 929:
>> 
>>> 927:   if ((t1i->_lo >= 0)        &&
>>> 928:       (t1i->_hi > 0)         &&
>>> 929:       (t1i->_hi <= max_power_of_2<jint>()) &&
>> 
>> I think you could use `<= std::numeric_limits<T>::max()` as the upper bound condition, and use `(jint)(next_power_of_2((uint)t1i->_hi) - 1)` to produce the mask in an overflow-conscious way.
>
> max_power_of_2<jint>() is less than std::numeric_limits<T>::max(). 
> 
> But I noticed I got the bounds wrong - is should be strictly less:
> t1i->_hi < max_power_of_2<jint>()

Right, that's why I widened to an `uint` (since `next_power_of_2((uint)jint::max())` is well-defined) then cast back to a `jint` after subtracting 1.

-------------

PR: https://git.openjdk.java.net/jdk/pull/4136