Smalltalk › Squeak › Squeak - Dev

new Cog VMs available

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

7 messages Options

Eliot Miranda-2

new Cog VMs available

at http://www.mirandabanda.org/files/Cog/VM/VM.r2697/

CogVM binaries as per VMMaker.oscog-eem.270/r2697

Fix the become issue where methods that are identical are failing

the code test because their penultimate literals are different objects.

Add a flag "cmUsesPenultimateLit" to jitted methods, stealing bits

from stackCheckOffset (which was way larger than needed).

Shrink stackCheckOffset to 12 bits (still an order of magnitude larger

than needed) and add an error check on assigning it.

Also add a check for max method size (2^16-1 bytes) and refuse to

jit a method that generates too much code.

When comparing code, use the cmUsesPenultimateLit flag to decide

if comparison includes penultimate lit or not.

This is mildly insane, but the VM really doesn't know about the penultimate

literal and it shouldn't depend on knowing it can be ignored. Note that the

CoInterpreter knows about the last literal; it uses this in supersends.

With this hack, Pharo's condenseSources works.

Fix bug in primitiveClone/cloneContext: that causes the copy to be a word short.

Use isPointerNonInt: and isContextNonInt: in a few places.

Implement unknownBytecode. Send unknownBytecode to the activeContext on

unknown bytecode if the selector is in the specialObjectsArray.

Fix page size bug in platforms/Cross/vm/sqHeapMap.c.

--
best,

Eliot

Frank Shearar-3

Re: new Cog VMs available

On 5 March 2013 22:45, Eliot Miranda <[hidden email]> wrote:

> at http://www.mirandabanda.org/files/Cog/VM/VM.r2697/
>
> CogVM binaries as per VMMaker.oscog-eem.270/r2697
>
> Fix the become issue where methods that are identical are failing
> the code test because their penultimate literals are different objects.
> Add a flag "cmUsesPenultimateLit" to jitted methods, stealing bits
> from stackCheckOffset (which was way larger than needed).
> Shrink stackCheckOffset to 12 bits (still an order of magnitude larger
> than needed) and add an error check on assigning it.
> Also add a check for max method size (2^16-1 bytes) and refuse to
> jit a method that generates too much code.
> When comparing code, use the cmUsesPenultimateLit flag to decide
> if comparison includes penultimate lit or not.
>
> This is mildly insane, but the VM really doesn't know about the penultimate
> literal and it shouldn't depend on knowing it can be ignored. Note that the
> CoInterpreter knows about the last literal; it uses this in supersends.
> With this hack, Pharo's condenseSources works.
>
> Fix bug in primitiveClone/cloneContext: that causes the copy to be a word
> short.
> Use isPointerNonInt: and isContextNonInt: in a few places.
>
> Implement unknownBytecode. Send unknownBytecode to the activeContext on
> unknown bytecode if the selector is in the specialObjectsArray.
>
> Fix page size bug in platforms/Cross/vm/sqHeapMap.c.

Build scripts duly updated.

frank

Chris Muller-3

Re: new Cog VMs available

In reply to this post by Eliot Miranda-2

Thanks Eliot. Unfortunately, for me my app still crashes instantly with:

-----------
[ ... big stack trace ... ]
stack page bytes 4096 available headroom 3300 minimum unused headroom 2596

(last object overwritten)
-----------

2640 is stable for me (as long as I don't use a write-barrier).

On Tue, Mar 5, 2013 at 4:45 PM, Eliot Miranda <[hidden email]> wrote:

timrowledge

Re: [Vm-dev] new Cog VMs available

In reply to this post by Eliot Miranda-2

But the cog is dead…
http://www.youtube.com/watch?v=tq7XG4XOJsc

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Useful random insult:- He fell out of the ugly tree and hit every branch on the way down.

Ron Teitelbaum

Re: [Vm-dev] new Cog VMs available

That's pretty funny!

Ron

> -----Original Message-----
> From: [hidden email] [mailto:squeak-dev-
> [hidden email]] On Behalf Of tim Rowledge
> Sent: Thursday, March 07, 2013 9:09 PM
> To: The general-purpose Squeak developers list
> Subject: [squeak-dev] Re: [Vm-dev] new Cog VMs available
>
>
> But the cog is dead.
> http://www.youtube.com/watch?v=tq7XG4XOJsc
>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim Useful
> random insult:- He fell out of the ugly tree and hit every branch on the

way
> down.
>
>
>

timrowledge

Re: [Vm-dev] new Cog VMs available

On 07-03-2013, at 9:32 PM, "Ron Teitelbaum" <[hidden email]> wrote:

> That's pretty funny!

They're actually a pretty good band but I prefer Abney PArk

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
On a clear disk you can seek forever...

Eliot Miranda-2

Re: new Cog VMs available

In reply to this post by Chris Muller-3

Hi Chris,

can you package up a reproducible case and point me to the image? By reproducible I mean save the image with a doit that invokes the code that creates the bug, e.g.

Smalltalk saveAs: 'crash'.

Gofer it
url: 'http://www.squeaksource.com/MetacelloRepository';
package: 'ConfigurationOfFFI';
load.
(Smalltalk at: #ConfigurationOfFFI) project stableVersion load.

On Wed, Mar 6, 2013 at 4:06 PM, Chris Muller <[hidden email]> wrote:

Thanks Eliot. Unfortunately, for me my app still crashes instantly with:

-----------
[ ... big stack trace ... ]
stack page bytes 4096 available headroom 3300 minimum unused headroom 2596

(last object overwritten)
-----------

2640 is stable for me (as long as I don't use a write-barrier).

On Tue, Mar 5, 2013 at 4:45 PM, Eliot Miranda <[hidden email]> wrote:
> at http://www.mirandabanda.org/files/Cog/VM/VM.r2697/
>
> CogVM binaries as per VMMaker.oscog-eem.270/r2697
>
> Fix the become issue where methods that are identical are failing
> the code test because their penultimate literals are different objects.
> Add a flag "cmUsesPenultimateLit" to jitted methods, stealing bits
> from stackCheckOffset (which was way larger than needed).
> Shrink stackCheckOffset to 12 bits (still an order of magnitude larger
> than needed) and add an error check on assigning it.
> Also add a check for max method size (2^16-1 bytes) and refuse to
> jit a method that generates too much code.
> When comparing code, use the cmUsesPenultimateLit flag to decide
> if comparison includes penultimate lit or not.
>
> This is mildly insane, but the VM really doesn't know about the penultimate
> literal and it shouldn't depend on knowing it can be ignored. Note that the
> CoInterpreter knows about the last literal; it uses this in supersends.
> With this hack, Pharo's condenseSources works.
>
> Fix bug in primitiveClone/cloneContext: that causes the copy to be a word
> short.
> Use isPointerNonInt: and isContextNonInt: in a few places.
>
> Implement unknownBytecode. Send unknownBytecode to the activeContext on
> unknown bytecode if the selector is in the specialObjectsArray.
>
> Fix page size bug in platforms/Cross/vm/sqHeapMap.c.
>
> --
> best,
> Eliot
>
>
>

--
best,

Eliot