Vm-dev May 2006

vm-dev@lists.squeakfoundation.org

17 participants
25 discussions

Re: method dictionary lookups and walking up the inheritance tree
by piumarta＠speakeasy.net 16 May '06

16 May '06

> -----Original Message----- > From: Andreas Raab [mailto:andreas.raab@gmx.de] > Subject: Re: method dictionary lookups and walking up the inheritance tree > > But isn't the global method cache polymorphic by definition? It maps > class+selector -> method (actual class of object; not class where the > method is defined in) and unless I got the meaning of the word wrong > this is polymorphic, is it not? ;-) You got the meaning wrong. A send site with a monomorphic inline cache looks like this: if (receiverClass == prevClass) call prevMethod else call relink(&prevMethod, prevClass := receiverClass, mySelector) where relink() looks up mySelector in receiverClass and overwrites *prevMethod with the target method, then jumps into the method. (You can also put this test+relink in the method prologue if it's cheap enough to reconstruct &prevMethod and &prevClass from the call return address.) prevClass and prevMethod can be inline constants overwritten by relink() (provided your architecture makes it easy to poke constants into code and it is really cheap to flush the affected i-cache lines) otherwise they can be a private data structure owned exclusively by the send site whose address is supplied to relink() when necessary. A send site with a polymorphic inline cache looks like this: if (receiverClass == prevClass[0[) call prevMethod[0] elif (receiverClass == prevClass[1]) call prevMethod[1] ... elif receiverClass == prevClass[N]) call prevMethod[N] else call relink(&prevMethod[firstFreeOrLeastRecentlyUsed, prevClass[firsFreeOrLeastRecentlyused] := receiverClass, mySelector) but otherwise works just like the monomorphic case. Obviously relink() can use a global method cache to refill the inline caches. The clever way to organise the code is like this: - When a monomorphic inline cache starts thrashing, promote it into a polyorhic inline cache. - When a polymorphic inline cache starts thrashing, either turn it off (and go straight to the global method cache) or copy mdict entries down into subclasses until the thrashing stops. Cheers, Ian

1 0

class formats
by Michael Haupt 16 May '06

16 May '06

Hi, apologies for the cross-post to two lists, but I wasn't too sure where to ask this. I've just been browsing the VM sources for details on the encoding of class formats used in object instantiation. Apparently, it seems as if the class format word that can be retrieved by sending #format to a class has the same layout as the standard header word of all objects. The only difference seems to be that the format word is shifted left by 1. Can anyone confirm this? Or am I wrong? Thanks, Michael

2 3

3.9-4 VM on cygwin: check that /usr/local/lib/squeak/3.9-4/vm-display-X11.so exists
by Klaus D. Witzel 16 May '06

16 May '06

List, we've compiled the 3.9-4 VM on the cygwin platform, only minor "*nix'en are platform[tm] independent" problems to solve: FIONBIO is in <sys/socket.h> # define O_ASYNC O_NONBLOCK timezone is _timezone (define is not part of cygwin) These three issues where solved using www.google.com :-) other people have already had this cygwin experience. #include "debug.h" (moved further down, dprintf must be redefined) # define dprintf(...) __sq_dprintf(__VA_ARGS__) There is a nice overview, on what gcc supports when it comes to variadic macros, at - http://gcc.gnu.org/onlinedocs/gcc/Variadic-Macros.html But the main problem is, compared to the subject line, ls -l /usr/local/lib/squeak/3.9-4/ -rw-r--r-- 1 kWitzel None 310484 May 15 11:56 B3DAcceleratorPlugin.a -rw-r--r-- 1 kWitzel None 57442 May 15 11:56 PseudoTTYPlugin.a -rw-r--r-- 1 kWitzel None 97012 May 15 11:56 UnixOSProcessPlugin.a -rw-r--r-- 1 kWitzel None 50022 May 15 11:56 XDisplayControlPlugin.a -rwxr-xr-x 1 kWitzel None 2386463 May 15 12:33 squeak.exe -rw-r--r-- 1 kWitzel None 326214 May 15 11:56 vm-display-X11.a -rw-r--r-- 1 kWitzel None 98856 May 15 11:56 vm-display-null.a -rw-r--r-- 1 kWitzel None 77634 May 15 11:56 vm-sound-OSS.a -rw-r--r-- 1 kWitzel None 26792 May 15 11:56 vm-sound-null.a A similiar problem is mentioned at - http://macos.tuwien.ac.at:9009/519430413.asHtml and because the scripts balks when linking npsqueak, we run --without-npsqueak - http://bugs.impara.de/view.php?id=559 Question: how can vm-display-X11.so be made, what's the trick? Suggestions, pointers, URLs, all warmly appreciated. /Klaus P.S. captured stdout and stderr from make and from ./config.status, if that is needed.

3 7

Re: 3.9-4 VM on cygwin: check that /usr/local/lib/squeak/3.9-4/vm-display-X11.so exists
by Bert Freudenberg 16 May '06

16 May '06

Am 15.05.2006 um 22:09 schrieb Klaus D. Witzel: > Hi Bert, > > on Mon, 15 May 2006 21:18:33 +0200, you <bert(a)impara.de> wrote: > >> Am 15.05.2006 um 17:59 schrieb Klaus D. Witzel: >> >>> At the moment we'd be happy if it would read script files. >> >> Which "it" doesn't? What are you trying to do? > > As I wrote earlier, the main problem is, compared to the subject > line, ls -l /usr/local/lib/squeak/3.9-4/ > -rw-r--r-- 1 kWitzel None 310484 May 15 11:56 B3DAcceleratorPlugin.a > -rw-r--r-- 1 kWitzel None 57442 May 15 11:56 PseudoTTYPlugin.a > -rw-r--r-- 1 kWitzel None 97012 May 15 11:56 UnixOSProcessPlugin.a > -rw-r--r-- 1 kWitzel None 50022 May 15 11:56 XDisplayControlPlugin.a > -rwxr-xr-x 1 kWitzel None 2386463 May 15 12:33 squeak.exe > -rw-r--r-- 1 kWitzel None 326214 May 15 11:56 vm-display-X11.a > -rw-r--r-- 1 kWitzel None 98856 May 15 11:56 vm-display-null.a > -rw-r--r-- 1 kWitzel None 77634 May 15 11:56 vm-sound-OSS.a > -rw-r--r-- 1 kWitzel None 26792 May 15 11:56 vm-sound-null.a > > That is, neither the X11 nor the null display *.so drivers are > found (and the message in the subject line, the reason for my post, > is printed). > > Ironically, > ./squeak -headless -nodisplay Squeak3.9b-7032.image > and only -headless and only -nodisplay, attempts to use the X11 > display driver... "-headless" is an option of the X11 driver, so that's not that surprising. As for loading the display modules, they have to be shared libraries suitable to be used by dlopen(). You only have static libraries, at least that's what I guess from the ".a" extension. I guess on Windows that would have to e a DLL. And even if you compile the modules as Windows DLL you might have to adjust the module loader to look for a ".dll" extension. - Bert -

2 1

Re: 3.9-4 VM on cygwin: check that /usr/local/lib/squeak/3.9-4/vm-display-X11.so exists
by Wolfgang Helbig 15 May '06

15 May '06

Hi /Klaus, You asked a question, >Question: how can vm-display-X11.so be made, what's the trick? And here is an answer: Make sure, your gcc knows about the X11 includes and your library path includes the X11 shared libs before you run configure. Check the makefile generated by config (in directory bld) for the target: vm-display-X11${la} : @$(SHELL) -ec 'cd vm-display-X11; $(MAKE) vm-display-X11${la}' And then check the file bld/libtool which is a shell script generated by config which is supposed to generate the shared objects. Hope, that helps. Wolfgang -- Weniger, aber besser.

2 1

searchable archive of mailing list?
by Michael Kohout 10 May '06

10 May '06

Is there a searchable archive of this mailing list somewhere on the web? I have some questions on how to build a plug in( like what library files I need to include, and from where ) that I'm sure have been answered before and I don't want to bog down the list with them unless it's necessary.... thanks Mike Kohout

2 1

building a unix vm with src32
by Martin Kuball 07 May '06

07 May '06

Hi! I'm trying to build a unix vm on 64 bit AMD (system is debian sid) for use with 32 bit images. I'm using the current svn trunk and build with the src32 directory. Things go well until the link step: gcc -g -O2 -DLSB_FIRST=1 -o squeak disabledPlugins.o version.o -Wl,--export-dynamic vm/vm.a B2DPlugin/B2DPlugin.a BitBltPlugin/BitBltPlugin.a FilePlugin/FilePlugin.a SocketPlugin/SocketPlugin.a -lutil -ldl -lm -lnsl vm/vm.a -Wl,--rpath -Wl,/usr/lib vm/vm.a(sqVirtualMachine.o): In function `sqGetInterpreterProxy':/home/kuball/SMT/vm-svn/trunk/platforms/Cross/vm/sqVirtualMachine.c:202: undefined reference to `obsoleteDontUseThisFetchWordofObject' :/home/kuball/SMT/vm-svn/trunk/platforms/Cross/vm/sqVirtualMachine.c:333: undefined reference to `fetchLong32ofObject' :/home/kuball/SMT/vm-svn/trunk/platforms/Cross/vm/sqVirtualMachine.c:334: undefined reference to `getThisSessionID' :/home/kuball/SMT/vm-svn/trunk/platforms/Cross/vm/sqVirtualMachine.c:335: undefined reference to `ioFilenamefromStringofLengthresolveAliases' :/home/kuball/SMT/vm-svn/trunk/platforms/Cross/vm/sqVirtualMachine.c:336: undefined reference to `vmEndianness' As you can see some methods are missing from interp.c. Can anybody help me fixing this problem? Martin

5 8

test post
by tim Rowledge 06 May '06

06 May '06

apparently my mail server is bouncing some messages.. tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Dreams are free, but you get soaked on the connect time.

1 0

VMM patch to help with cross-platform plugin generation
by David T. Lewis 06 May '06

06 May '06

Hi Tim, VMM supports generating sources for target platforms other than the one it's running on, but the plugins have no way to tell VMM if they should be translated in that case. Trivial enhancement attached. Dave

1 1

[jim.p.davis@gmail.com: [Newbies] A squeak machine?]
by David T. Lewis 05 May '06

05 May '06

This question was posted on the "beginners" list, but I suspect that the inhabitants of vm-dev might be better able to provide an answer. Dave ----- Forwarded message from Jim Davis <jim.p.davis(a)gmail.com> ----- X-Sent-To: lewis(a)conch-old.msen.com Delivered-To: list-beginners(a)lists.squeakfoundation.org DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:mime-version:content-type; b=ky1cBxXe+0sK+4mJpjvcg+c4v7NWhUcLnoh7b32YUq6e4PtA+in7DLTw9/y7vICqogn1KG3mhiM+0EURfB8fVbminWAglAA9pEf8n2XrwXXxl95xPz1wW+cyrm18zsyrsAyHolQyQdUvxUw6H65aXHGKAKlgA6nUzPQAVHVF+ks= Date: Fri, 5 May 2006 01:20:23 -0700 From: "Jim Davis" <jim.p.davis(a)gmail.com> To: beginners(a)lists.squeakfoundation.org Subject: [Newbies] A squeak machine? X-BeenThere: beginners(a)lists.squeakfoundation.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: "A friendly place to get answers to even the most basic questions about Squeak." <beginners(a)lists.squeakfoundation.org> List-Id: "A friendly place to get answers to even the most basic questions about Squeak." <beginners.lists.squeakfoundation.org> List-Unsubscribe: <http://lists.squeakfoundation.org/mailman/listinfo/beginners>, <mailto:beginners-request@lists.squeakfoundation.org?subject=unsubscribe> List-Archive: <http://lists.squeakfoundation.org/pipermail/beginners> List-Post: <mailto:beginners@lists.squeakfoundation.org> List-Help: <mailto:beginners-request@lists.squeakfoundation.org?subject=help> List-Subscribe: <http://lists.squeakfoundation.org/mailman/listinfo/beginners>, <mailto:beginners-request@lists.squeakfoundation.org?subject=subscribe> Errors-To: beginners-bounces(a)lists.squeakfoundation.org X-Scanned-By: milter-7bit/0.11.137 (shell.msen.com [148.59.86.2]); Fri, 05 May 2006 04:20:30 -0400 X-milter-7bit-Pass: YES X-Spamilter-SPF: pass (policy result: [pass] from rule [a/24]) reciever=shell.msen.com; client-ip=85.10.195.197; envelope-from=<beginners-bounces(a)lists.squeakfoundation.org>; helo=box2.squeakfoundation.org;) X-Milter: Spamilter (Reciever: shell.msen.com; Sender-ip: 85.10.195.197; Sender-helo: box2.squeakfoundation.org;) Hi, I'm just a post silicon test engineer for a company that is developing a object oriented medium grain parallel processor on a chip. It's quite a bit like a FPGA, but the chip implements ALU state machines and table lookup logic. It has multiply / accumulate (dot product) units, datapath muxes and lots of memory. The kicker is it's a tagged architecture, with 5 bits riding shotgun on the 16 bits of data. That's 8 bits of tag and 2 ready bits for a real-world 32 bit object. I've heard talk in the office that a harvard/princeton machine demo hack implementation would be nice, but that seems really 1950's .My first thought was to build a FORTH machine, or a bunch of forth machines, But Squeak seems really attractive and it goes well with the object oriented mindset of the company and the chip. So? do you think I should burn a couple hundred hours of my own time trying to cook this up? BTW: this thing runs at 1 GHz and has 256 alu's, 64 macs, tons of registers and scratch and 2X266 Mhz 40 bit ddr2 dram interfaces with 256 Mwords of store.. AKA: Mr Jones. Just how do I implement a parallel or massivly pipeline(d) squeak machine? _______________________________________________ Beginners mailing list Beginners(a)lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners ----- End forwarded message -----

3 2

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev May 2006