[Vm-dev] VM Maker: VMMaker.oscog-eem.1426.mcz

Eliot Miranda eliot.miranda at gmail.com
Sun Jul 19 01:09:22 UTC 2015


Hi David,

On Sat, Jul 18, 2015 at 3:49 PM, David T. Lewis <lewis at mail.msen.com> wrote:

>
> On Sat, Jul 18, 2015 at 08:55:27PM +0000, commits at source.squeak.org wrote:
> >
> > Eliot Miranda uploaded a new version of VMMaker to project VM Maker:
> > http://source.squeak.org/VMMaker/VMMaker.oscog-eem.1426.mcz
> >
> > ==================== Summary ====================
> >
> > Name: VMMaker.oscog-eem.1426
> > Author: eem
> > Time: 18 July 2015, 1:54:29.051 pm
> > UUID: 94ab92ba-c5c4-4953-8566-a4cd9c38df1f
> > Ancestors: VMMaker.oscog-eem.1425
> >
> > Fix at least one set of 64-bit issues caused by Slang.  In particular
> the integerObjectOf: code resulted in (objectMemory integerObjectOf:
> MillisecondClockMask) evaluating to the -1 object, instead of the
> 16r1FFFFFFF object, which was the cause of the initially nresponsive 64-bit
> image on the real VM (no problem in the simulator).
> >
>
> I can't test now to verify, but I wonder if this change is fixing the
> right problem.
>

I'm pretty sure it is.  It only bites in a 64-bit implementation with >
31-bit integers.  In the "standard" 64-bit image SmallIntegers are still
only 31-bits so the issue never occurs.  The issue is that the default type
of an integer constant in C is int.  So if one has to shift any constant
such that a non-zero bit will occupy bit 31 (0 relative), it must be cast
to a long type to avoid sign extension.

Now of course I could generate all Integer constants with the L or UL
suffix, e.g.

#define MillisecondClockMask 0x1FFFFFFFL

instead of

#define MillisecondClockMask 0x1FFFFFFF

but that's a much more pervasive change than only generating the cast in
integerObjectOf when on 64-bits.  So I'm happy with the change that I've
made.  Experience can of course prove me wrong...



> The CCodeGenerator>>generateIntegerObjectOf:on:indent: in VMM trunk has
> been in use long enough that it has no author initials. I have found it
> to work correctly on all combinations of 32/64 bit image and VM. If it
> does not work correctly, I would be inclined to suspect type declaration
> issues elsewhere.
>
> Original implementation:
>
> CCodeGenerator>>generateIntegerObjectOf: msgNode on: aStream indent: level
>         "Generate the C code for this message onto the given stream."
>
>         aStream nextPutAll: '(('.
>         self emitCExpression: msgNode args first on: aStream.
>         aStream nextPutAll: ' << 1) | 1)'.
>
> Dave
>
>
> > =============== Diff against VMMaker.oscog-eem.1425 ===============
> >
> > Item was changed:
> >   ----- Method: CCodeGenerator>>generateIntegerObjectOf:on:indent: (in
> category 'C translation') -----
> >   generateIntegerObjectOf: msgNode on: aStream indent: level
> >       "Generate the C code for this message onto the given stream."
> > +     | expr castToSqint |
> > +     expr := msgNode args first.
> > +     aStream nextPutAll: '(('.
> > +     "Note that the default type of an integer constant in C is int.
> Hence we /must/
> > +      cast constants to long if in the 64-bit world, since e.g. in
> 64-bits
> > +             (int)(16r1FFFFF << 3) = (int)16rFFFFFFF8 = -8
> > +      whereas
> > +             (long)(16r1FFFFF << 3) = (long) 16rFFFFFFF8 = 4294967288."
> > +     castToSqint := expr isConstant and: [vmClass isNil or: [vmClass
> objectMemoryClass wordSize = 8]].
> > +     castToSqint ifTrue:
> > +             [aStream nextPutAll: '(sqInt)'].
> > +     self emitCExpression: expr on: aStream.
> > -
> >       aStream
> > -             nextPutAll: '(('.
> > -     self emitCExpression: msgNode args first on: aStream.
> > -     aStream
> >               nextPutAll: ' << ';
> >               print: vmClass objectMemoryClass numSmallIntegerTagBits;
> >               nextPutAll: ') | 1)'!
>



-- 
_,,,^..^,,,_
best, Eliot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20150718/24d22bc7/attachment.htm


More information about the Vm-dev mailing list