[Vm-dev] about the usage of bitShift: for directed shifts in VMMaker

[Nicolas Cellier]

Hi again,

I see this horrible generated code:

    /* Spur64BitMemoryManager>>#integerValueOf: */
sqInt
integerValueOf(sqInt oop)
{
    return ((((usqInt) oop >> 63)) == 1
        ? ((((((-(numTagBits())) < 0) ? ((usqInt) oop >>
-(-(numTagBits()))) : ((usqInt) oop << (-(numTagBits()))))) &
0x1FFFFFFFFFFFFFFFLL) - 0x1FFFFFFFFFFFFFFFLL) - 1
        : (((-(numTagBits())) < 0) ? ((usqInt) oop >> -(-(numTagBits()))) :
((usqInt) oop << (-(numTagBits())))));
}

which of course generates a (false positive) warning:

Avertissement    C4293    '<<' : compteur de décalage négatif ou trop
important, comportement non défini    SqueakCogSpur
X:\Smalltalk\opensmalltalk-vm\spur64src\vm\cointerp.c    41564

Slang is:

integerValueOf: oop
    "Translator produces 'oop >> 3'"
    ^(oop bitShift: -63) = 1 "tests top bit"
        ifTrue: "negative"
            [((oop bitShift: self numTagBits negated) bitAnd:
16r1FFFFFFFFFFFFFFF) - 16r1FFFFFFFFFFFFFFF - 1  "Faster than
-16r4000000000000000 (a LgInt)"]
        ifFalse: "positive"
            [oop bitShift: self numTagBits negated]

We see that the constant -63 has been translated into a directed (right)
shift, but the other constant (numTagBits) not so, either because constant
substitution is not aggressive enough, or because negated made an
expression of the constant...

However, the intention is ALWAYS to perform a directed shift (to the right)
so the code could be expressed to make both intention AND generated code
clear (and warning-free):

integerValueOf: oop
    "Translator produces 'oop >> 3'"
    ^(oop >> 63) = 1 "tests top bit"
        ifTrue: "negative"
            [((oop >> self numTagBits) bitAnd: 16r1FFFFFFFFFFFFFFF) -
16r1FFFFFFFFFFFFFFF - 1  "Faster than -16r4000000000000000 (a LgInt)"]
        ifFalse: "positive"
            [oop >> self numTagBits]

Why use bitShift: then? Is it a problem of speed for the VM simulator?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20181225/ba874573/attachment.html>