[Vm-dev] VM Maker: VMMaker.oscog-nice.1843.mcz

[Esteban Lorenzano]

and I confirm it works :)

Esteban

> On 25 Apr 2016, at 21:41, Nicolas Cellier <nicolas.cellier.aka.nice at gmail.com> wrote:
> 
> And thanks to Esteban for pointing me to the bug.
> 
> 
> 2016-04-25 21:22 GMT+02:00 <commits at source.squeak.org <mailto:commits at source.squeak.org>>:
> 
> Nicolas Cellier uploaded a new version of VMMaker to project VM Maker:
> http://source.squeak.org/VMMaker/VMMaker.oscog-nice.1843.mcz <http://source.squeak.org/VMMaker/VMMaker.oscog-nice.1843.mcz>
> 
> ==================== Summary ====================
> 
> Name: VMMaker.oscog-nice.1843
> Author: nice
> Time: 25 April 2016, 8:24:03.731 pm
> UUID: 9ef5d374-ae5e-437a-9116-cde9d708c0d7
> Ancestors: VMMaker.oscog-cb.1842
> 
> Fix my recent bug for signed 32bit long access (introduced beginning of April).
> 
> One symptom is:
>   (Alien newGC: 4)
>         signedLongAt: 1 put: -16r7287E552;
>         signedLongAt: 1
> 
> I incorrecly used the signed value (integerValue) in one branch, and the magnitude (value) in the other branch...
> This is an incorrect copy/paste from maybeInlinePositive32BitIntegerFor: in my ultimate #if SPURVM refactoring.
> 
> While at it, avoid the confusion by using a better variable name (magnitude), and eliminate UB related to computing opposite of INT_MIN by explicitely using unsigned type.
> 
> =============== Diff against VMMaker.oscog-cb.1842 ===============
> 
> Item was changed:
>   ----- Method: StackInterpreter>>noInlineSigned32BitIntegerFor: (in category 'primitive support') -----
>   noInlineSigned32BitIntegerFor: integerValue
>         "Answer a full 32 bit integer object for the given integer value."
>         <notOption: #Spur64BitMemoryManager>
> +       | newLargeInteger magnitude largeClass |
> -       | newLargeInteger value largeClass |
>         <inline: false>
> +       <var: 'magnitude' type: 'unsigned int'>
>         (objectMemory isIntegerValue: integerValue) ifTrue:
>                 [^objectMemory integerObjectOf: integerValue].
>         self deny: objectMemory hasSixtyFourBitImmediates.
>          integerValue < 0
>                 ifTrue: [largeClass := ClassLargeNegativeIntegerCompactIndex.
> +                               magnitude := 0 asUnsignedInteger - integerValue]
> -                               value := 0 - integerValue]
>                 ifFalse: [largeClass := ClassLargePositiveIntegerCompactIndex.
> +                               magnitude := integerValue].
> -                               value := integerValue].
>         newLargeInteger := objectMemory
>                                                         eeInstantiateSmallClassIndex: largeClass
>                                                         format: (objectMemory byteFormatForNumBytes: 4)
>                                                         numSlots: 1.
>         self cppIf: SPURVM
>                 ifTrue:
>                         ["Memory is 8 byte aligned in Spur, make sure that oversized bytes are set to zero"
> +                       objectMemory storeLong32: 0 ofObject: newLargeInteger withValue: (objectMemory byteSwapped32IfBigEndian: magnitude).
> -                       objectMemory storeLong32: 0 ofObject: newLargeInteger withValue: (objectMemory byteSwapped32IfBigEndian: integerValue).
>                         objectMemory storeLong32: 1 ofObject: newLargeInteger withValue: 0]
>                 ifFalse:
> +                       [objectMemory storeLong32: 0 ofObject: newLargeInteger withValue: (objectMemory byteSwapped32IfBigEndian: magnitude)].
> -                       [objectMemory storeLong32: 0 ofObject: newLargeInteger withValue: (objectMemory byteSwapped32IfBigEndian: value)].
>         ^newLargeInteger!
> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20160425/c34e7d0f/attachment-0001.htm