- squirrelfish-dev - lists.macosforge.org

http://trac.webkit.org/changeset/45039
by Geoffrey Garen 24 Jun '09

24 Jun '09

Traditionally, if we find a dictionary in the prototype chain when we're trying to cache, we do a fromDictionary transition, so the cache can succeed. For example, in tryCacheGetById: // Since we're accessing a prototype in a loop, it's a good bet that it // should not be treated as a dictionary. if (slotBaseObject->structure()->isDictionary()) slotBaseObject- >setStructure(Structure::fromDictionaryTransition(slotBaseObject- >structure())); Is there a good reason not to do that in this case? Geoff

3 2

http://trac.webkit.org/changeset/44797
by Geoffrey Garen 18 Jun '09

18 Jun '09

> Roll out r43849 as it appears that we cannot rely on the address of > an objects property storage being constant even if the structure is > unchanged. Did you figure out why we can't rely on the address of an object's property storage being constant even if its structure is unchanged? It's fine to roll out the change to fix the crash for now, but I'd rather not lose this speedup. Do we need a bug for figuring out what's going on here and rolling back in the optimization? Geoff

3 2

X86 instructions in JITArithmetic?
by Gabor Loki 04 Jun '09

04 Jun '09

Hi, I was wondering why JITArithmetic.cpp uses so many functions from X86Assembler. This should not be happened. MacroAssembler was introduced to be the 'common interface' for every back ends (not only for x86). Am I right? For example: In 'emit_op_rshift' function http://trac.webkit.org/browser/trunk/JavaScriptCore/jit/JITArithmetic.cpp#L… 121#if USE(ALTERNATE_JSIMMEDIATE) 122 addSlowCase(emitJumpIfNotImmediateNumber(regT0)); 123 __ movq_rr(regT0, X86::xmm0); 124#else 125 emitJumpSlowCaseIfNotJSCell(regT0, op1); 126 addSlowCase(checkStructure(regT0, m_globalData->numberStructure.get())); ******************* 127 __ movsd_mr(FIELD_OFFSET(JSNumberCell, m_value), regT0, X86::xmm0); ******************* 128#endif IMHO, if a platform specific solution have to be used, creates a separate file which will contain those solutions/optimizations. What do you think about this issue? --Gabor

4 7

Is it possible to compile from javascript to native code without bytecode?
by Gyuyoung Kim 13 May '09

13 May '09

Hello, According to my knowledge, SFX makes bytecode using javascript, then the bytecode is translated to native code via JITc. I was told that V8 doesn't make bytecode. It makes native code from javascript directly. If SFX also makes native code from javascript directly, in my opinion, the performace of SFX can be better than now. Are there any movements about my question? If I have misunderstandings, please let me know. Thanks, Gyuyoung Kim.

2 1

Announce the ARM port of JIT
by Gabor Loki 01 Apr '09

01 Apr '09

Hi guys, we are pleased to announce that the ARM port of JIT is finally released. The source can be checked out from 'loki/arm-port' branch of git://code.staikos.net/webkit : - git clone git://code.staikos.net/webkit && git checkout loki/arm-port - or see at http://code.staikos.net/cgi-bin/gitweb.cgi?p=webkit;a=shortlog;h=loki/arm-p… The aim of this development is to add JIT support for ARM architectures. The port itself is developed from scratch, but we reused the ideas of x86 JIT. So we implemented property caches, stub functions, etc. in a similar way, but the code is optimized for ARM architectures. We used Qt4 build environment for the development, but we feel that the other build platforms can be easily extended with this ARM port. The current version is stable. It passes the regression tests. We did measurements on a Nokia N810, and we gained 13% performance improvement. The memory consumption is increased by 3.3% which is a small trade-off for this performance improvement. Our work has not done yet. We plan to add several other features in the future. Please, feel free to ask any questions about the ARM JIT here at the mail list or at the following Bugzilla entry: https://bugs.webkit.org/show_bug.cgi?id=24986 --Gabor

1 0

How to build squirrelfish shell ?
by Gyuyoung Kim 01 Apr '09

01 Apr '09

Hello, I am trying to build squirrelfish on my linux box. However, I could get the squirrelfish shell(jsc) together with other executable files(e.g. GtkLauncher, UnitTests, minidom) when I built webkitGTK on linux box. I don't know how to build only squirrelfish shell yet. Does anyone know how to build squirrelfish shell ? Thanks, Gyuyoung Kim

1 0

Can Squirrelfish Extream supports ARM?
by Gyuyoung Kim 11 Feb '09

11 Feb '09

Hello I wonder if SFX's JITc can generate native codes for ARM architecture, or not. According to my googling, it seems to me that SFX doesn't support ARM architecture yet. If it is true, are there any plans or on-going works for ARM? Thanks, Gyuyoung

3 3

Info on Squirrelfish Garbage Collector
by k.kuchenbecker＠gmx.net 29 Jan '09

29 Jan '09

Hi, I want to use Squirrelfish in an application. But I'm lacking information on the Garbage Collector to decide whether I can use Squirrelfish or not. What I found out so far: The GC is not (yet) generational, it is partly conservative and uses mark-sweep. What I'm not sure about is whether it is concurrent, incremental or "stop-the-world". I'm sorry if this turns out to be a stupid question or was already answered somewhere. Your help is greatly appreciated! Thanks, Kai -- NUR NOCH BIS 31.01.! GMX FreeDSL - Telefonanschluss + DSL für nur 16,37 EURO/mtl.!* http://dsl.gmx.de/?ac=OM.AD.PD003K11308T4569a

2 1

CallFrame::thisValue(), CallData::call() for host functions
by Kent Hansen 07 Jan '09

07 Jan '09

Hi, First a bit of context: I'm working on using JSC as the backend for Qt Script, the embedded script engine that's part of the Qt application development framework. This means wrapping our existing public API on top of JSC and preserving the semantics to a very large (read: full) extent. So far it's going pretty well, but I'm stumbling upon some issues. If there's another channel I should use for reporting and discussing these things (e.g. the bug tracker), let me know, thanks. Two more issues I've found with host functions: 1) CallFrame::thisValue() crashes for host functions because it goes via codeBlock, which is 0 for host functions. Why does codeBlock need to be used to get the thisRegister, can't it be stored/deduced in the callFrame itself (other places there are calculations like thisRegister = callFrame->registers() - RegisterFile::CallFrameHeaderSize - argCount)? It would be nice if you could get the thisValue() for host function frames too. 2) CallData::call() doesn't create a new frame to execute the host function in. So for instance if you have a host setter function, the setter is going to be executed in the current frame (see JSObject::put()); whereas if a host function is called like a normal function (through the interpreter), it _will_ be executed in a new frame. I've attached a patch where I try to create a new frame in CallData::call(), but it only "almost" seems to work. I copied some code from Interpreter::execute() and the call opcode handler and tried to adapt it. I suspect there's a problem with my calculation of the registerOffset, because the arguments are fetched incorrectly later. Anyone care to comment if I'm on the right track? On a related note, I think there's a bug in op_call for host functions. The scope chain of the caller frame is used as the scope chain of the new frame. This is not consistent with ECMA 10.2.3: "The scope chain is initialised to contain the activation object followed by the objects in the scope chain stored in the [[Scope]] property of the Function object." It's done correctly for JS functions, but not for host functions. It would be nice if host functions could have a [[Scope]] too (i.e. move scope() and setScope() to InternalFunction). Best regards, Kent

2 2

CallFrame::callee() for host functions
by Kent Hansen 07 Jan '09

07 Jan '09

Hi, Since CallFrame::callee() returns a JSFunction, callee() always returns 0 for host functions. This means that if you want to know the callee at some later point (e.g. walking up the stack), you'd have to store the callee argument that the host function received, and maintain some mapping from the callFrame to callee. And that only works if you implemented the host function yourself. It would be nice if CallFrame::callee() always gave you the callee, regardless of whether it's a host call or JS call; i.e. change the type of callee() to be InternalFunction. Is there anything that prohibits such a change? (If not, I can try to work out a patch.) Regards, Kent

2 1