Vm-dev November 2014

vm-dev@lists.squeakfoundation.org

36 participants
165 discussions

Re: goto instruction with Cog VM
by Eliot Miranda 07 Nov '14

07 Nov '14

Hi Ralph, On Tue, Nov 4, 2014 at 9:09 PM, Ralph Boland <rpboland(a)gmail.com> wrote: > > > Hi Ralph, > > > > On Sun, Nov 2, 2014 at 9:48 PM, Ralph Boland <rpboland(a)gmail.com> wrote: > > > > > > > > >> > > > >> I am working on a parser generator tool (a replacement for SmaCC) > and > > > >> one of the things I a m interested in is the direct translation of > > > >> language specifications into the virtual machine code (SmaCC and my > > > >> current version of it use Squeak as the target language). > > > >> > > > > > > > First, a different approach than compiling to Smalltalk is to > compile to > > > a > > > > parse tree. We do this in the pseudo JavaScript compiler we've > > > implemented > > > > at Cadence, translating an AST into a Squeak compiler parse tree for > code > > > > generation. Targeting a parse tree gives you much more freedom; you > can > > > > express things that aren't expressible in Smalltalk. And if you > target > > > > bytecodes you can do even more. > > > > > > > > > I never considered using a parse tree as the target. An interesting > idea > > > which in many instances may be the best approach. But for my regular > > > expression > > > example I would still want to generate byte codes. In any case I > wouldn't > > > want to > > > restrict users of my parser generator tool to any one of the three > options > > > (Smalltalk code, > > > parse tree, byte code). It is my responsibility to make all three > options > > > as easy and > > > efficient as reasonably possible for users of the parser generator > tool. > > > Haven't put > > > much thought into this yet though. So far Smalltalk (Squeak) is the > only > > > option. > > > > > > > > > >> One of the problems I have is that, for some languages, the natural > > > >> translation > > > >> into VM code uses computed gotos. > > > >> There are two scenarios here: > > > >> > > > >> 1) goto X where X is a variable. > > > >> 2) goto (coll at: y) where coll is a Collection. > > > >> > > > > > > > There are several ways of implementing this without computed > bytecodes in > > > > the instruction set, but there is also the possibility of > implementing it > > > > directly in the instruction set. > > > > > > > Off the top of my head one could > > > > > > > - map to perform: using some mangled selector. Yes this is > problematic > > > > because one has to split the scope between the two methods, so in > general > > > > it's not a solution > > > > > > Doesn't appeal to me. > > > > > > > - map to a case statement, which is what Squeak does. Just map it to > a > > > > sequence of compare and branches. Or better still, to a binary tree. > > > > Coincidentally this is used by the JIT to implement block dispatch in > > > > methods that contain more than one block. I know of other VM > > > > implementations using it for PIC dispatch with really good > performance. > > > > > > I don't know what you mean my Squeak mapping to a case statement since > > > there is no case statement in Squeak/Smalltalk and I can't off hand > think > > > of > > > where one is needed (some Squeak users might feel they need one but > that > > > is a > > > different matter). > > > > > > > But there is. And it is very convenient when one doesn't want to spread > a > > case over different classes, or when the cases distribute over values of > > the same class (e.g. integer values). People often claim that a case > > statement isn't necessary because one has polymorphism, but unless the > > language supports singletons (which of course Smalltalk does not) a case > > statement is much more readable than e.g. an if-then-else tree or a > > "Dictionary mapping values to blocks or selectors" solution. > > > > Since you're unaware of the case statement I suggest you browse senders > of > > caseOf: and caseOf:otherwise: in Squeak trunk (which includes selectors > of > > optimized selectors that end up not being sent) or caseError, which is > sent > > when there's no otherwise clause. Here's an example: > > > > menuHook: aMenu named: aSymbol shifted: aBool > > "Enhance aMenu with registered services." > > aSymbol > > caseOf: > > { [ #classListMenu ] -> [ ServiceGui browser: self classMenu: aMenu ]. > > [ #codePaneMenu ] -> [ ServiceGui browser: self codePaneMenu: aMenu ]. > > [ #messageCategoryMenu] -> [ ServiceGui browser: self > messageCategoryMenu: > > aMenu ]. > > [ #messageListMenu ] -> [ ServiceGui browser: self messageListMenu: > aMenu ]. > > [ #systemCategoryMenu ] -> [ ServiceGui browser: self classCategoryMenu: > > aMenu ] } > > otherwise: [ "do nothing" ] > > > > This compiles down to a sequence of comparisons. > > I was aware of caseOf: in Squeak. I always found it awkward to use and > felt a true case statement would be simpler. Alas, it's impossible to > have a true case statement added to Smalltalk now I think. > So what's a "true" case statement? For me, at least, the Squeak one *is*, and is more general than one limited to purely integer keys, as for example is C's switch statement. A number of languages provide case statements that are like Squeak's. What do you consider a "true" case statement? > caseOf: is user code and thus, in principle, outside the scope of the VM. > For example, if I didn't want the sequential search of caseOf: I could > implement a method doBlockAt: method in class Dictionary that assumes > the values stored in the dictionary are blocks and invokes the block > found by performing at: on the arg of doBlockAt:. > But the code for both caseOf: and doBlockAt: are written by users so the > VM can't optimize it, is forced to have the overhead of using blocks, > and must be locked in to the search algorithm the user has chosen (for > better or for worse). > > > > The use of compare and branches might be OK in some cases > > > but a mess for the finite state machines generated from regular > > > expressions. > > > Actually, even with computed gotos FSMs are somewhat messy but without > > > them > > > it's worse. I don't know what 'PIC dispatch' is. > > > > > > > Polymorphic inline cache dispatch. In JIT VMs such as Cog polymorphic > send > > sites are optimized using jump tables, one jump for each class > encountered > > at the send site, up tio some small limit such as 6 cases. Right now in > > Cog these jump tables are sequential comparisons. But in some VMs, > where > > the degree of polymorphism may be much higher (think prototype languages > > such as JavaScript) a binary tree may be much miore efficient, if harder > to > > organize. > > > > To use a binary tree don't I need some kind of computed goto for when I > > > reach > > > a leaf of the tree???? > > > > > > > No. Once you're at the leaf you just include the bytecodes. So for > > example take the following case statement: > > > > quickPrimitiveGeneratorFor: aQuickPrimitiveIndex > > <api> > > <returnTypeC: 'int (*quickPrimitiveGeneratorFor(sqInt > > aQuickPrimitiveIndex))(void)'> > > ^aQuickPrimitiveIndex > > caseOf: { > > [256] -> [#genQuickReturnSelf]. > > [257] -> [#genQuickReturnConstNil]. > > [258] -> [#genQuickReturnConstTrue]. > > [259] -> [#genQuickReturnConstFalse]. > > [260] -> [#genQuickReturnConstMinusOne]. > > [261] -> [#genQuickReturnConstZero]. > > [262] -> [#genQuickReturnConstOne]. > > [263] -> [#genQuickReturnConstTwo] } > > otherwise: [#genQuickReturnInstVar] > > > > This is compiled in the current Squeak compiler as > > > > 61 <10> pushTemp: 0 > > 62 <88> dup > > 63 <2A> pushConstant: 256 > > 64 <B6> send: = > > 65 <9B> jumpFalse: 70 > > 66 <87> pop > > 67 <29> pushConstant: #genQuickReturnSelf > > 68 <A4 35> jumpTo: 123 > > 70 <88> dup > > 71 <28> pushConstant: 257 > > 72 <B6> send: = > > 73 <9B> jumpFalse: 78 > > 74 <87> pop > > 75 <21> pushConstant: #genQuickReturnConstNil > > 76 <A4 2D> jumpTo: 123 > > ... > > 117 <22> pushConstant: 263 > > 118 <B6> send: = > > 119 <99> jumpFalse: 122 > > 120 <21> pushConstant: #genQuickReturnConstTwo > > 121 <90> jumpTo: 123 > > 122 <20> pushConstant: #genQuickReturnInstVar > > 123 <7C> returnTop > > > > but it could be more efficient if the code were > > > > pushTemp: 0 > > dup > > pushConstant: 256 > > send: < > > jumpFalse: L1 > > dup > > pushConstant: 260 > > send: < > > jumpFalse: L2 > > dup > > pushConstant: 258 > > send: < > > jumpFalse: L3 > > dup > > pushConstant: 257 > > send: < > > jumpFalse: L4 > > pushConstant: #genQuickReturnSelf > > jumpTo: L0 > > pushConstant: #genQuickReturnConstNil > > jumpTo: L0 > > ... > > L1: > > pushConstant: #genQuickReturnInstVar > > L0: > > returnTop > > I have no problem with this code per say; if Squeak, the language, > had a case statement it could be implemented this way. > You keep on saying it doesn't have one. I disagree. > But I wouldn't want to be forced to implement my FSMs this way. > It might be acceptable for small FSMs. > I want to avoid sequential search and > even binary search might be rather expensive. > I look at computed gotos as the solution but, > as you pointed out, computed gotos pose problems for JIT. > Admittedly, for large FSM's, it might be best or necessary to > use a FSM simulator anyway, as I do now. > Nah. One should always be able to map it down somehow. Tis will be easier with the Spur instruction set which lifts number of literals and length of branches limits. > > > > - use thisContext pc: value. > > > > > > This would be a possibility for me to experiment with for now. When I > > > have a working > > > parser generator tool I could campaign for my computed goto > instructions > > > to be added > > > to the VM. > > > > > > > This /should/ be fine in the stack VM, but > > > > slooooow in the JIT because internally mapping bytecode pcs to > machine > > > code > > > > pcs is slow, and currently slower still because the frame will be > > > converted > > > > to a pure context and then converted back into a frame on the return > from > > > > pc:. But this solution isn't to be rejected out-of-hand. It can be > > > > optimized to avoid the frame conversion and the JIT might be able to > > > > optimize it. > > > > > > I assume that if computed gotos were used the translation to machine > code > > > would require a direct > > > mapping of (virtually labeled) bytecode locations to machine code > > > locations. I think this can be done > > > in a reasonable amount of time but others such as yourself clearly > > > understand the issues far better than > > > I do. The dirty solution to start would be to simply not JIT the code > > > that uses computed gotos. > > > > > > > Yes, it could be. The JIT would generate a jump table from the literal > > containing the bytecoded pcs, and there would be no conversion; only > > indexing the jump table and jumping. Again, organizing that table as a > > binary switch may well be faster on modern architectures. Indirect jumps > > typically involve pipeline stalls, whereas binary jump trees don't. > > > > > The main problem is the compiler has no support for labels so > > > > there would be work here. > > > > > > I don't mind doing the work but to my way of thinking "goto X" is > pretty > > > basic > > > and is thus best handled at the VM/byte code level. Anything else is > doing > > > in a complicated way something that is fairly simple. Of course > changing > > > the VM/byte codes by even a single byte code is a major deal unless > done > > > when the VM/byte codes are initially created. Alas I must deal with > what > > > already exists. Even so, my preference is to work with the VM if at > all > > > possible. > > > > > > > > > >> For example, one such language is that of regular expressions, > which I > > > >> wish to translate into finite state machines implemented in VM code. > > > >> In this case I need case 2) gotos where coll is a collection of > > > >> associations, possibly a > > > >> Dictionary. I also plan to write a debugger for this (and other > > > languages) > > > >> but that is another story. > > > >> > > > >> I realize that the Cog VM is being built for Smalltalk (Squeak? > Pharo?) > > > >> for which the goto instructions are not needed and thus I assume > > > >> unavailable. But there is something to > > > >> viewing a virtual machine as general purpose and thus the target of > > > >> multiple languages as is > > > >> the case for the Java virtual machine. > > > >> If the Cog VM is viewed this way then I argue there is a need for my > > > goto > > > >> instructions > > > >> because some languages have need for them. > > > >> For example, many languages have case statements. (I am all for > object > > > >> oriented > > > >> but I would be willing to accept a case statement in Smalltalk > too; the > > > >> Squeak code > > > >> implemented one in Squeak doesn't cut it). > > > > > > > > > > > I've occasionally thought about this for many years. A computed jump > > > might > > > > be nice. Eg index an Array literal of pcs with the integer on top of > > > > stack, falling through on bad type or out of range. > > > > > > This is the way I am thinking. If there are other reasons for a > computed > > > jumpTo > > > as well all the better. > > > > > > >> Anyway, I am not arguing to Change Squeak or Smalltalk but I am > arguing > > > >> to have my goto instructions in Cog VM. Is there any chance of > this????? > > > >> > > > > > > > There's no chance of me spending time implementing this any time > soon. I > > > > have too much high-priority tasks to tackle this. But I want to > > > encourage > > > > you or others to have a go implementing it. It's fun! > > > > > > I understand and am willing to be the one to add one or more computed > jump > > > instructions, including working on the JIT code generator if needed. > > > As you say it should be fun (and also educational). But > > > 1) I am pretty busy now too and probably won't get to this for a > year. > > > 2) If I am to do this it would be great if someone can write a > > > specification as to > > > what is to be done. If someone can write this now that would > be > > > great but > > > if they write it when I post that I am ready to do the work > that > > > would also > > > be fine. > > > 3) I don't want to just have my own private VM/byte codes. I want > > > users of my > > > parser generator tool to be able to load it into a standard > > > version of Squeak > > > and run it there including the possible generation of compilers > > > for compiling > > > their domain specific language programs into byte codes if > desired. > > > > > > > Sure. Get back to me when you're ready to work on it and I'll write the > > specification, and help you with whatever info you need. There should be > > room in the bytecode sets we'll likely be using next year. > > > WILL DO. SHOULD BE FUN. > WILL DO. SHOULD BE FUN. > WILL DO. SHOULD BE FUN. > Great, thanks!! > > > > > > > >> I don't know the Squeak VM or the Cog VM either but I assume these > > > >> instructions don't exist because I see no need of them when the > source > > > >> language is > > > >> Squeak or any version of Smalltalk for that matter. I also assume > that > > > >> there is already > > > >> a full list of 256 instructions in the Cog VM and thus no room for > my > > > goto > > > >> instructions > > > >> unless some instructions are removed. > > > >> > > > >> Are there Cog VM instructions that are so rarely used that they > could be > > > >> removed without > > > >> unreasonably slowing down the Cog VM interpretation of byte codes > > > >> generated from Squeak source code????? > > > >> > > > > > > > The current set has 3 unused bytecodes, one of which Spur uses, so > > > > effectively there are two unused bytecodes. > > > > > > Levente Uzonyi in his posting pointed out that only one instruction is > > > needed. > > > I don't like having to push the address to jump to onto the stack, > > > preferring a byte > > > code with an argument, but I could live with his solution if that is > what > > > is decided. > > > In the case of goto coll at: X the address is likely to end up on > top > > > of the stack > > > anyway so Levente's jumpToTop instruction looks good in any case. > > > > > > > If the bytecode is one that takes an integer on top of stack, and an > Array > > literal containing bytecode pcs, falling through on out of range, then > > nothing other than the index need be pushed on top of stack. That would > be > > my preference. > > > > Again, for my FSM, case this would often be considered to be good. > But if the state transition tables are sparse then Dictionaries > might be preferable to Arrays. > Yes, but getting to the limit of what the VM can reasonably interpret. Better would be an Array of value. pc pairs, where the keys are the values the switch bytecode compares top of stack against, and the pcs are where to jump to on a match. The JIT can therefore implement the table as it sees fit, whereas the interpreter can just do a linear search through the Array. My expection is that at: be sent to the collection object > to get the address to go to. Knowing that the collection > is an array though makes it easier for the compiler/VM to > ensure that the addresses stored in the collection are valid. > Actually, the compiler will be generating the addresses. > Does the VM have absolute trust in the compiler to generate valid > addresses? > Yes. Generate bad bytecode and the VM crashes. > What if the compiler is generated by my parser generator tool? > Same. Once the tool is generating correct bytecode the VM should stop crashing ;-) > > > > > The Cog VMs support multiple bytecode sets. If you look at the > > > > BytecodeSets package on VMMaker you can read the class comments of > the > > > > BytecodeEncoder subclasses such as EncoderForSistaV1. These bytecode > > > sets > > > > have a few more unused bytecodes. This multiple bytecode set > support is > > > > better implemented in Spur where there is only one compiled method > header > > > > format and support for 64k literals. So let me encourage you to > move to > > > > Spur and to look at the Sista set. The class comment of each encoder > > > class > > > > specifies the instruction set it targets. > > > > > > I am prepared to work with Spur and the Sista set. I am looking for > > > someone to > > > say that if I do this work that incorporating the work into Spur will > be > > > seriously > > > considered. > > > > > > > I can say that, yes. I will seriously consider adding it. I think its a > > useful thing in the bytecode set, precisely to allow compiling other > > languages to the VM. > > > GLAD TO HEAR THIS. > You're welcome. > > > > > > > > > > > > > > Ralph Boland > > > > > > > > > > > > > > > -- > > best, > > Eliot > > > Ralph > > On 3 November 2014 12:34, <vm-dev-request(a)lists.squeakfoundation.org> > wrote: > >> Send Vm-dev mailing list submissions to >> vm-dev(a)lists.squeakfoundation.org >> >> To subscribe or unsubscribe via the World Wide Web, visit >> http://lists.squeakfoundation.org/mailman/listinfo/vm-dev >> or, via email, send a message with subject or body 'help' to >> vm-dev-request(a)lists.squeakfoundation.org >> >> You can reach the person managing the list at >> vm-dev-owner(a)lists.squeakfoundation.org >> >> When replying, please edit your Subject line so it is more specific >> than "Re: Contents of Vm-dev digest..." >> >> >> Today's Topics: >> >> 1. Re: goto instruction with Cog VM (Eliot Miranda) >> >> >> ---------------------------------------------------------------------- >> >> Message: 1 >> Date: Mon, 3 Nov 2014 11:34:48 -0800 >> From: Eliot Miranda <eliot.miranda(a)gmail.com> >> Subject: Re: [Vm-dev] goto instruction with Cog VM >> To: Squeak Virtual Machine Development Discussion >> <vm-dev(a)lists.squeakfoundation.org> >> Message-ID: >> < >> CAC20JE1RduKQW_EdMsfah156QnTbR8USt9dEb4mj1rb15rndqA(a)mail.gmail.com> >> Content-Type: text/plain; charset="utf-8" >> >> Hi Ralph, >> >> On Sun, Nov 2, 2014 at 9:48 PM, Ralph Boland <rpboland(a)gmail.com> wrote: >> >> > >> > >> >> > >> I am working on a parser generator tool (a replacement for SmaCC) and >> > >> one of the things I a m interested in is the direct translation of >> > >> language specifications into the virtual machine code (SmaCC and my >> > >> current version of it use Squeak as the target language). >> > >> >> > >> > > First, a different approach than compiling to Smalltalk is to compile >> to >> > a >> > > parse tree. We do this in the pseudo JavaScript compiler we've >> > implemented >> > > at Cadence, translating an AST into a Squeak compiler parse tree for >> code >> > > generation. Targeting a parse tree gives you much more freedom; you >> can >> > > express things that aren't expressible in Smalltalk. And if you >> target >> > > bytecodes you can do even more. >> > >> > >> > I never considered using a parse tree as the target. An interesting >> idea >> > which in many instances may be the best approach. But for my regular >> > expression >> > example I would still want to generate byte codes. In any case I >> wouldn't >> > want to >> > restrict users of my parser generator tool to any one of the three >> options >> > (Smalltalk code, >> > parse tree, byte code). It is my responsibility to make all three >> options >> > as easy and >> > efficient as reasonably possible for users of the parser generator tool. >> > Haven't put >> > much thought into this yet though. So far Smalltalk (Squeak) is the >> only >> > option. >> > >> > >> > >> One of the problems I have is that, for some languages, the natural >> > >> translation >> > >> into VM code uses computed gotos. >> > >> There are two scenarios here: >> > >> >> > >> 1) goto X where X is a variable. >> > >> 2) goto (coll at: y) where coll is a Collection. >> > >> >> > >> > > There are several ways of implementing this without computed >> bytecodes in >> > > the instruction set, but there is also the possibility of >> implementing it >> > > directly in the instruction set. >> > >> > > Off the top of my head one could >> > >> > > - map to perform: using some mangled selector. Yes this is >> problematic >> > > because one has to split the scope between the two methods, so in >> general >> > > it's not a solution >> > >> > Doesn't appeal to me. >> > >> > > - map to a case statement, which is what Squeak does. Just map it to a >> > > sequence of compare and branches. Or better still, to a binary tree. >> > > Coincidentally this is used by the JIT to implement block dispatch in >> > > methods that contain more than one block. I know of other VM >> > > implementations using it for PIC dispatch with really good >> performance. >> > >> > I don't know what you mean my Squeak mapping to a case statement since >> > there is no case statement in Squeak/Smalltalk and I can't off hand >> think >> > of >> > where one is needed (some Squeak users might feel they need one but that >> > is a >> > different matter). >> > >> >> But there is. And it is very convenient when one doesn't want to spread a >> case over different classes, or when the cases distribute over values of >> the same class (e.g. integer values). People often claim that a case >> statement isn't necessary because one has polymorphism, but unless the >> language supports singletons (which of course Smalltalk does not) a case >> statement is much more readable than e.g. an if-then-else tree or a >> "Dictionary mapping values to blocks or selectors" solution. >> >> Since you're unaware of the case statement I suggest you browse senders of >> caseOf: and caseOf:otherwise: in Squeak trunk (which includes selectors of >> optimized selectors that end up not being sent) or caseError, which is >> sent >> when there's no otherwise clause. Here's an example: >> >> menuHook: aMenu named: aSymbol shifted: aBool >> "Enhance aMenu with registered services." >> aSymbol >> caseOf: >> { [ #classListMenu ] -> [ ServiceGui browser: self classMenu: aMenu ]. >> [ #codePaneMenu ] -> [ ServiceGui browser: self codePaneMenu: aMenu ]. >> [ #messageCategoryMenu] -> [ ServiceGui browser: self messageCategoryMenu: >> aMenu ]. >> [ #messageListMenu ] -> [ ServiceGui browser: self messageListMenu: aMenu >> ]. >> [ #systemCategoryMenu ] -> [ ServiceGui browser: self classCategoryMenu: >> aMenu ] } >> otherwise: [ "do nothing" ] >> >> This compiles down to a sequence of comparisons. >> >> The use of compare and branches might be OK in some cases >> > but a mess for the finite state machines generated from regular >> > expressions. >> > Actually, even with computed gotos FSMs are somewhat messy but without >> > them >> > it's worse. I don't know what 'PIC dispatch' is. >> > >> >> Polymorphic inline cache dispatch. In JIT VMs such as Cog polymorphic >> send >> sites are optimized using jump tables, one jump for each class encountered >> at the send site, up tio some small limit such as 6 cases. Right now in >> Cog these jump tables are sequential comparisons. But in some VMs, where >> the degree of polymorphism may be much higher (think prototype languages >> such as JavaScript) a binary tree may be much miore efficient, if harder >> to >> organize. >> >> To use a binary tree don't I need some kind of computed goto for when I >> > reach >> > a leaf of the tree???? >> > >> >> No. Once you're at the leaf you just include the bytecodes. So for >> example take the following case statement: >> >> quickPrimitiveGeneratorFor: aQuickPrimitiveIndex >> <api> >> <returnTypeC: 'int (*quickPrimitiveGeneratorFor(sqInt >> aQuickPrimitiveIndex))(void)'> >> ^aQuickPrimitiveIndex >> caseOf: { >> [256] -> [#genQuickReturnSelf]. >> [257] -> [#genQuickReturnConstNil]. >> [258] -> [#genQuickReturnConstTrue]. >> [259] -> [#genQuickReturnConstFalse]. >> [260] -> [#genQuickReturnConstMinusOne]. >> [261] -> [#genQuickReturnConstZero]. >> [262] -> [#genQuickReturnConstOne]. >> [263] -> [#genQuickReturnConstTwo] } >> otherwise: [#genQuickReturnInstVar] >> >> This is compiled in the current Squeak compiler as >> >> 61 <10> pushTemp: 0 >> 62 <88> dup >> 63 <2A> pushConstant: 256 >> 64 <B6> send: = >> 65 <9B> jumpFalse: 70 >> 66 <87> pop >> 67 <29> pushConstant: #genQuickReturnSelf >> 68 <A4 35> jumpTo: 123 >> 70 <88> dup >> 71 <28> pushConstant: 257 >> 72 <B6> send: = >> 73 <9B> jumpFalse: 78 >> 74 <87> pop >> 75 <21> pushConstant: #genQuickReturnConstNil >> 76 <A4 2D> jumpTo: 123 >> ... >> 117 <22> pushConstant: 263 >> 118 <B6> send: = >> 119 <99> jumpFalse: 122 >> 120 <21> pushConstant: #genQuickReturnConstTwo >> 121 <90> jumpTo: 123 >> 122 <20> pushConstant: #genQuickReturnInstVar >> 123 <7C> returnTop >> >> but it could be more efficient if the code were >> >> pushTemp: 0 >> dup >> pushConstant: 256 >> send: < >> jumpFalse: L1 >> dup >> pushConstant: 260 >> send: < >> jumpFalse: L2 >> dup >> pushConstant: 258 >> send: < >> jumpFalse: L3 >> dup >> pushConstant: 257 >> send: < >> jumpFalse: L4 >> pushConstant: #genQuickReturnSelf >> jumpTo: L0 >> pushConstant: #genQuickReturnConstNil >> jumpTo: L0 >> ... >> L1: >> pushConstant: #genQuickReturnInstVar >> L0: >> returnTop >> >> > - use thisContext pc: value. >> > >> > This would be a possibility for me to experiment with for now. When I >> > have a working >> > parser generator tool I could campaign for my computed goto instructions >> > to be added >> > to the VM. >> > >> > > This /should/ be fine in the stack VM, but >> > > slooooow in the JIT because internally mapping bytecode pcs to machine >> > code >> > > pcs is slow, and currently slower still because the frame will be >> > converted >> > > to a pure context and then converted back into a frame on the return >> from >> > > pc:. But this solution isn't to be rejected out-of-hand. It can be >> > > optimized to avoid the frame conversion and the JIT might be able to >> > > optimize it. >> > >> > I assume that if computed gotos were used the translation to machine >> code >> > would require a direct >> > mapping of (virtually labeled) bytecode locations to machine code >> > locations. I think this can be done >> > in a reasonable amount of time but others such as yourself clearly >> > understand the issues far better than >> > I do. The dirty solution to start would be to simply not JIT the code >> > that uses computed gotos. >> > >> >> Yes, it could be. The JIT would generate a jump table from the literal >> containing the bytecoded pcs, and there would be no conversion; only >> indexing the jump table and jumping. Again, organizing that table as a >> binary switch may well be faster on modern architectures. Indirect jumps >> typically involve pipeline stalls, whereas binary jump trees don't. >> >> > The main problem is the compiler has no support for labels so >> > > there would be work here. >> > >> > I don't mind doing the work but to my way of thinking "goto X" is >> pretty >> > basic >> > and is thus best handled at the VM/byte code level. Anything else is >> doing >> > in a complicated way something that is fairly simple. Of course >> changing >> > the VM/byte codes by even a single byte code is a major deal unless done >> > when the VM/byte codes are initially created. Alas I must deal with >> what >> > already exists. Even so, my preference is to work with the VM if at all >> > possible. >> > >> > >> > >> For example, one such language is that of regular expressions, which >> I >> > >> wish to translate into finite state machines implemented in VM code. >> > >> In this case I need case 2) gotos where coll is a collection of >> > >> associations, possibly a >> > >> Dictionary. I also plan to write a debugger for this (and other >> > languages) >> > >> but that is another story. >> > >> >> > >> I realize that the Cog VM is being built for Smalltalk (Squeak? >> Pharo?) >> > >> for which the goto instructions are not needed and thus I assume >> > >> unavailable. But there is something to >> > >> viewing a virtual machine as general purpose and thus the target of >> > >> multiple languages as is >> > >> the case for the Java virtual machine. >> > >> If the Cog VM is viewed this way then I argue there is a need for my >> > goto >> > >> instructions >> > >> because some languages have need for them. >> > >> For example, many languages have case statements. (I am all for >> object >> > >> oriented >> > >> but I would be willing to accept a case statement in Smalltalk too; >> the >> > >> Squeak code >> > >> implemented one in Squeak doesn't cut it). >> > > >> > >> > > I've occasionally thought about this for many years. A computed jump >> > might >> > > be nice. Eg index an Array literal of pcs with the integer on top of >> > > stack, falling through on bad type or out of range. >> > >> > This is the way I am thinking. If there are other reasons for a >> computed >> > jumpTo >> > as well all the better. >> > >> > >> Anyway, I am not arguing to Change Squeak or Smalltalk but I am >> arguing >> > >> to have my goto instructions in Cog VM. Is there any chance of >> this????? >> > >> >> > >> > > There's no chance of me spending time implementing this any time >> soon. I >> > > have too much high-priority tasks to tackle this. But I want to >> > encourage >> > > you or others to have a go implementing it. It's fun! >> > >> > I understand and am willing to be the one to add one or more computed >> jump >> > instructions, including working on the JIT code generator if needed. >> > As you say it should be fun (and also educational). But >> > 1) I am pretty busy now too and probably won't get to this for a >> year. >> > 2) If I am to do this it would be great if someone can write a >> > specification as to >> > what is to be done. If someone can write this now that would be >> > great but >> > if they write it when I post that I am ready to do the work that >> > would also >> > be fine. >> > 3) I don't want to just have my own private VM/byte codes. I want >> > users of my >> > parser generator tool to be able to load it into a standard >> > version of Squeak >> > and run it there including the possible generation of compilers >> > for compiling >> > their domain specific language programs into byte codes if >> desired. >> > >> >> Sure. Get back to me when you're ready to work on it and I'll write the >> specification, and help you with whatever info you need. There should be >> room in the bytecode sets we'll likely be using next year. >> >> >> > >> > >> I don't know the Squeak VM or the Cog VM either but I assume these >> > >> instructions don't exist because I see no need of them when the >> source >> > >> language is >> > >> Squeak or any version of Smalltalk for that matter. I also assume >> that >> > >> there is already >> > >> a full list of 256 instructions in the Cog VM and thus no room for my >> > goto >> > >> instructions >> > >> unless some instructions are removed. >> > >> >> > >> Are there Cog VM instructions that are so rarely used that they >> could be >> > >> removed without >> > >> unreasonably slowing down the Cog VM interpretation of byte codes >> > >> generated from Squeak source code????? >> > >> >> > >> > > The current set has 3 unused bytecodes, one of which Spur uses, so >> > > effectively there are two unused bytecodes. >> > >> > Levente Uzonyi in his posting pointed out that only one instruction is >> > needed. >> > I don't like having to push the address to jump to onto the stack, >> > preferring a byte >> > code with an argument, but I could live with his solution if that is >> what >> > is decided. >> > In the case of goto coll at: X the address is likely to end up on top >> > of the stack >> > anyway so Levente's jumpToTop instruction looks good in any case. >> > >> >> If the bytecode is one that takes an integer on top of stack, and an Array >> literal containing bytecode pcs, falling through on out of range, then >> nothing other than the index need be pushed on top of stack. That would >> be >> my preference. >> >> > The Cog VMs support multiple bytecode sets. If you look at the >> > > BytecodeSets package on VMMaker you can read the class comments of the >> > > BytecodeEncoder subclasses such as EncoderForSistaV1. These bytecode >> > sets >> > > have a few more unused bytecodes. This multiple bytecode set support >> is >> > > better implemented in Spur where there is only one compiled method >> header >> > > format and support for 64k literals. So let me encourage you to move >> to >> > > Spur and to look at the Sista set. The class comment of each encoder >> > class >> > > specifies the instruction set it targets. >> > >> > I am prepared to work with Spur and the Sista set. I am looking for >> > someone to >> > say that if I do this work that incorporating the work into Spur will be >> > seriously >> > considered. >> > >> >> I can say that, yes. I will seriously consider adding it. I think its a >> useful thing in the bytecode set, precisely to allow compiling other >> languages to the VM. >> >> >> >> > >> > Ralph Boland >> > >> > >> > >> >> >> -- >> best, >> Eliot >>

1 0

PharoVM - Build issues on CentOS6.5
by phil＠highoctane.be 07 Nov '14

07 Nov '14

I am trying to build a new PharoVM So, I took the master branch and as I have a 3.0 vm that works on my box, I commented out the +vm in get newImage.sh. So, I get the new image and code loads in there. Then I get this: external plugin UnixOSProcessPlugin generated as UnixOSProcessPlugin 7 November 2014 10:14:53 am + cd /extradisk/philippeback/workspace/PharoVM2/pharo-vm/scripts/.. + /extradisk/philippeback/workspace/PharoVM2/pharo-vm/scripts/extract-commit-info.sh + cd /extradisk/philippeback/workspace/PharoVM2/pharo-vm/scripts/../build + bash ./build.sh -- The C compiler identification is GNU -- The CXX compiler identification is GNU -- Check for working C compiler: /usr/bin/gcc -- Check for working C compiler: /usr/bin/gcc -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Check for working CXX compiler: /usr/bin/c++ -- Check for working CXX compiler: /usr/bin/c++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done Configuring thirdparty package: libssh2 Downloading http://files.pharo.org/vm/src/lib/libssh2-1.4.3.tar.gz CMake Error at CMakeLists.txt:25 (add_dependencies): add_dependencies Adding dependency to non-existent target: /extradisk/philippeback/workspace/PharoVM2/pharo-vm/build/..//results/libssh2.so.1.0.1 Configuring thirdparty package: libgit2 Downloading http://files.pharo.org/vm/src/lib/v0.21.1.tar.gz CMake Error at CMakeLists.txt:32 (add_dependencies): add_dependencies Adding dependency to non-existent target: /extradisk/philippeback/workspace/PharoVM2/pharo-vm/build/..//results/libgit2.so.0.21.0 Configuring thirdparty package: libsdl2 Downloading http://files.pharo.org/vm/src/lib/SDL2-2.0.3.tar.gz CMake Error at CMakeLists.txt:39 (add_dependencies): add_dependencies Adding dependency to non-existent target: /extradisk/philippeback/workspace/PharoVM2/pharo-vm/build/..//results/libSDL2-2.0.so.0.2.1 Adding internal plugin: ADPCMCodecPlugin Adding internal plugin: AsynchFilePlugin Adding internal plugin: B2D Looks like something changed in way CMake gets generated. Any idea why it fails? I had a look at https://travis-ci.org/pharo-project/pharo-vm to see how things are normally going on. TIA Phil

2 2

SpurBootstrap is writing into a block argument
by Nicolas Cellier 06 Nov '14

06 Nov '14

#prototypeClassNameMetaSelectorMethodDo: is trying to write in method block arg... Is it well behaved ?

1 0

[commit][3126] Add build directories linux32ARM squeak.stack.spur.
by commits＠squeakvm.org 05 Nov '14

05 Nov '14

Revision: 3126 Author: eliot Date: 2014-11-05 12:48:12 -0800 (Wed, 05 Nov 2014) Log Message: ----------- Add build directories linux32ARM squeak.stack.spur. Update configure to build the UnicodePlugin on linux. Include the Unicode plugin in linux Squeak VMs. Update the linux HowToBuild instructions with info on loading support packages. Thanks to Doug McPherson for getting all this to work. Modified Paths: -------------- branches/Cog/build.linux32ARM/HowToBuild branches/Cog/build.linux32x86/HowToBuild branches/Cog/build.linux32x86/squeak.cog.spur/build/plugins.ext branches/Cog/build.linux32x86/squeak.cog.spur/build.assert/plugins.ext branches/Cog/build.linux32x86/squeak.cog.spur/build.assert.itimerheartbeat/plugins.ext branches/Cog/build.linux32x86/squeak.cog.spur/build.debug/plugins.ext branches/Cog/build.linux32x86/squeak.cog.spur/build.debug.itimerheartbeat/plugins.ext branches/Cog/build.linux32x86/squeak.cog.spur/build.itimerheartbeat/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build.assert/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build.assert.itimerheartbeat/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build.debug/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build.debug.itimerheartbeat/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build.itimerheartbeat/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded.assert/plugins.ext branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded.debug/plugins.ext branches/Cog/build.linux32x86/squeak.stack.spur/build/plugins.ext branches/Cog/build.linux32x86/squeak.stack.spur/build.assert/plugins.ext branches/Cog/build.linux32x86/squeak.stack.spur/build.debug/plugins.ext branches/Cog/build.linux32x86/squeak.stack.v3/build/plugins.ext branches/Cog/build.linux32x86/squeak.stack.v3/build.assert/plugins.ext branches/Cog/build.linux32x86/squeak.stack.v3/build.debug/plugins.ext branches/Cog/platforms/unix/config/configure branches/Cog/platforms/unix/plugins/UnicodePlugin/Makefile.inc branches/Cog/platforms/unix/plugins/UnicodePlugin/acinclude.m4 Added Paths: ----------- branches/Cog/build.linux32ARM/squeak.stack.spur/ branches/Cog/build.linux32ARM/squeak.stack.spur/build/ branches/Cog/build.linux32ARM/squeak.stack.spur/build/mvm branches/Cog/build.linux32ARM/squeak.stack.spur/build/plugins.ext branches/Cog/build.linux32ARM/squeak.stack.spur/build/plugins.int branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/ branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/mvm branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/plugins.ext branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/plugins.int branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/ branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/mvm branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/plugins.ext branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/plugins.int branches/Cog/build.linux32ARM/squeak.stack.spur/makeallclean branches/Cog/build.linux32ARM/squeak.stack.spur/makealldirty Property Changed: ---------------- branches/Cog/platforms/Cross/vm/sqSCCSVersion.h Modified: branches/Cog/build.linux32ARM/HowToBuild =================================================================== --- branches/Cog/build.linux32ARM/HowToBuild 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32ARM/HowToBuild 2014-11-05 20:48:12 UTC (rev 3126) @@ -11,7 +11,8 @@ - How to configure and build a VM on Unix - Testing an external plugin has completely linked - Optimization level and gcc version (please read!) - - BitBlt optimizations on ARM + - BitBlt optimizations on ARM + - Installing support libraries Overview @@ -80,7 +81,10 @@ Building out of the box ----------------------- -Install the tools (gcc, X11-devel, etc (e.g. libpng, libX11 & libxt source)). +Install the tools (gcc, X11-devel, etc, e.g. libpng, libuuid libX11 & libxt +source). See "Installing support libraries" below. If the configure step fails +when "checking for C compiler default output file name", you have yet to install +all the necessary support packages (e.g. libuuid). Then cd to the build directory of your choice, e.g. build.linux32x86/squeak.cog.spur/build Then either remove any BochsIA32Plugin line from plugins.ext or build the @@ -88,6 +92,9 @@ Then execute ./mvm answering "y" to perform a clean build or "n" to rebuild without recionfiguring. +Again, if the configure step fails when "checking for C compiler default output +file name", you have yet to install all the necessary support (e.g. lubuuid). + The subdirectories confrm to the production/assert/debug x itimer vs threaded heartbeat x single vs multi-threaded parts of the matrix described above. For example, build.linux32x86/squeak.cog.v3 includes @@ -145,7 +152,7 @@ You'll also have to add the -m32 switch to all gcc & g++ invocations. The easiest way to do this is to add CC="gcc -m32" & CXX="g++ -m32" to the configure script: - ../../platforms/unix/config/configure CC="gcc -m32" CXX="g++ -m32" --without-npsqueak CFLAGS="-g -O2 -msse2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DNDEBUG -DITIMER_HEARTBEAT=1 -DNO_VM_PROFILE=1 -DCOGMTVM=0 -DDEBUGVM=0" LIBS=-lpthread + ../../platforms/unix/config/configure CC="gcc -m32" CXX="g++ -m32" --without-npsqueak CFLAGS="-g -O2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DNDEBUG -DITIMER_HEARTBEAT=1 -DNO_VM_PROFILE=1 -DCOGMTVM=0 -DDEBUGVM=0" LIBS=-lpthread To run a 32-bit VM on a 64-bit OS, you'll also need the 32-bit libraries provided by the ia32-libs package. @@ -179,16 +186,18 @@ directory. e.g. - ../../platforms/unix/config/configure --without-npsqueak CFLAGS="-g -O2 -msse2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DNDEBUG -DITIMER_HEARTBEAT=1 -DNO_VM_PROFILE=1 -DCOGMTVM=0 -DDEBUGVM=0" LIBS=-lpthread + ../../platforms/unix/config/configure --without-npsqueak CFLAGS="-g -O2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DNDEBUG -DITIMER_HEARTBEAT=1 -DNO_VM_PROFILE=1 -DCOGMTVM=0 -DDEBUGVM=0" LIBS=-lpthread make install prefix=WhereYouWantTheVmToGo N.B. If you're on a 64-bit linux read 3e below!! N.B. On Ubuntu *do not* supply "LIBS=-lpthread -luuid", i.e. use - ../platforms/unix/config/configure --without-npsqueak CFLAGS="-g -O2 -msse2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DNDEBUG -DITIMER_HEARTBEAT=1 -DNO_VM_PROFILE=1 -DCOGMTVM=0 -DDEBUGVM=0" + ../platforms/unix/config/configure --without-npsqueak CFLAGS="-g -O2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DNDEBUG -DITIMER_HEARTBEAT=1 -DNO_VM_PROFILE=1 -DCOGMTVM=0 -DDEBUGVM=0" N.B. The plugin set is defined by plugins.ext and plugins.int in the build dir. +Be prepared to install libuuid support. e.g. on CentOS 6.5 use +sudo yum -y install libuuid-devel Testing an external plugin has completely linked @@ -208,7 +217,7 @@ -fcaller-saves So turn them off. e.g. - ../../platforms/unix/config/configure --without-npsqueak CFLAGS="-g -O2 -msse2 -fno-caller-saves -fno-partial-inlining -fno-tree-pre -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DNDEBUG -DNO_VM_PROFILE=1 -DCOGMTVM=0 -DDEBUGVM=0" LIBS="-lpthread -luuid" + ../../platforms/unix/config/configure --without-npsqueak CFLAGS="-g -O2 -fno-caller-saves -fno-partial-inlining -fno-tree-pre -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DNDEBUG -DNO_VM_PROFILE=1 -DCOGMTVM=0 -DDEBUGVM=0" LIBS="-lpthread -luuid" See http://smallissimo.blogspot.fr/2013/02/compiling-squeak-cog-virtual-machine… There appear to be issues with 3.4.x gcc version on RedHat. In particular @@ -227,4 +236,20 @@ asasm is included in Cog/build.linux32ARM/ as a pre-built binary. The Makefiles for BitBltPlugin and vm-display-X11 reference it directly. +Installing support libraries +---------------------------- +Different linux distributions have different mechanisms for loading packages. +Here are some examples: +CentOS +sudo yum install cairo-devel +sudo yum install pango-devel +sudo yum install libuuid-devel +sudo yum install libX11-devel +sudo yum install mesa-libGL-devel + +Ubuntu +sudo apt-get install libcairo2-dev +sudo apt-get install libpango1.0-dev + +More advice and examples for other distros gratefully received. Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build/mvm =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build/mvm (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build/mvm 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,33 @@ +#!/bin/bash +# Stack Spur VM and threaded heartbeat +INSTALLDIR=stkspurlinuxhtARM +# Some gcc versions create a broken VM using -O2 +case `gcc -v 2>&1 | grep version | sed 's/gcc version *//'` in +3.4.*) OPT="-g -O1 -DNDEBUG -DDEBUGVM=0 -DNO_VM_PROFILE=1";; +*) OPT="-g -O3 -DNDEBUG -DDEBUGVM=0 -DNO_VM_PROFILE=1";; +esac + +if [ $# -ge 1 ]; then + INSTALLDIR="$1"; shift +fi + +echo -n "clean? " +read a +case $a in +n|no|N|NO) echo "ok but this isn't safe!!";; +*) rm config.h; test -f Makefile && make reallyclean +esac +test -f config.h || ../../../platforms/unix/config/configure \ + --with-vmversion=5.0 \ + --with-src=spurstacksrc --with-plugins=src/plugins --disable-cogit \ + --without-vm-display-fbdev --without-npsqueak --enable-fast-bitblt \ + CC="gcc " \ + CXX="g++ " \ + CFLAGS="$OPT -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64" \ + LIBS="-lpthread -luuid" \ + LDFLAGS=-Wl,-z,now +rm -f vm/sqUnixMain.o # nuke version info +rm -rf ../../../products/$INSTALLDIR +# prefer make install prefix=`readlink -f \`pwd\`/../../../products/$INSTALLDIR` +# but older linux readlinks lack the -f flag and Raspbian lacks `readlinks` +make install prefix=`(cd ../../../;pwd)`/products/$INSTALLDIR 2>&1 | tee LOG Property changes on: branches/Cog/build.linux32ARM/squeak.stack.spur/build/mvm ___________________________________________________________________ Added: svn:executable + * Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build/plugins.ext =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build/plugins.ext (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,9 @@ +# Copied, perhaps edited, from ../../src/examplePlugins.ext +EXTERNAL_PLUGINS = \ +LocalePlugin \ +SqueakFFIPrims \ +SqueakSSL \ +UnicodePlugin \ +UnixOSProcessPlugin \ +UUIDPlugin \ +XDisplayControlPlugin Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build/plugins.int =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build/plugins.int (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build/plugins.int 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,37 @@ +# Copied, perhaps edited, from ../../src/examplePlugins.int +INTERNAL_PLUGINS = \ +ADPCMCodecPlugin \ +AioPlugin \ +AsynchFilePlugin \ +B2DPlugin \ +BitBltPlugin \ +BMPReadWriterPlugin \ +CameraPlugin \ +ZipPlugin \ +DropPlugin \ +DSAPrims \ +FFTPlugin \ +FileCopyPlugin \ +FilePlugin \ +FloatArrayPlugin \ +FloatMathPlugin \ +JoystickTabletPlugin \ +JPEGReaderPlugin \ +JPEGReadWriter2Plugin \ +Klatt \ +LargeIntegers \ +Matrix2x3Plugin \ +MIDIPlugin \ +MiscPrimitivePlugin \ +Mpeg3Plugin \ +RePlugin \ +ScratchPlugin \ +SecurityPlugin \ +SerialPlugin \ +SocketPlugin \ +SoundCodecPrims \ +SoundGenerationPlugin \ +SoundPlugin \ +StarSqueakPlugin \ +SurfacePlugin \ +VMProfileLinuxSupportPlugin Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/mvm =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/mvm (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/mvm 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,29 @@ +#!/bin/bash +# assert Stack Spur VM with VM profiler and threaded heartbeat +INSTALLDIR=assert/stkspurlinuxhtARM +OPT="-g3 -O1 -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -DDEBUGVM=0" + +if [ $# -ge 1 ]; then + INSTALLDIR="$1"; shift +fi + +echo -n "clean? " +read a +case $a in +n|no|N|NO) echo "ok but this isn't safe!!";; +*) rm config.h; test -f Makefile && make reallyclean +esac +test -f config.h || ../../../platforms/unix/config/configure \ + --with-vmversion=5.0 \ + --with-src=spurstacksrc --with-plugins=src/plugins --disable-cogit \ + --without-vm-display-fbdev --without-npsqueak --enable-fast-bitblt \ + CC="gcc " \ + CXX="g++ " \ + CFLAGS="$OPT -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64" \ + LIBS="-lpthread -luuid" \ + LDFLAGS=-Wl,-z,now +rm -f vm/sqUnixMain.o # nuke version info +rm -rf ../../../products/$INSTALLDIR +# prefer make install prefix=`readlink -f \`pwd\`/../../../products/$INSTALLDIR` +# but older linux readlinks lack the -f flag and Raspbian lacks `readlinks` +make install prefix=`(cd ../../../;pwd)`/products/$INSTALLDIR 2>&1 | tee LOG Property changes on: branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/mvm ___________________________________________________________________ Added: svn:executable + * Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/plugins.ext =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/plugins.ext (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,8 @@ +# Copied, perhaps edited, from ../../src/examplePlugins.ext +EXTERNAL_PLUGINS = \ +LocalePlugin \ +SqueakFFIPrims \ +SqueakSSL \ +UnixOSProcessPlugin \ +UUIDPlugin \ +XDisplayControlPlugin Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/plugins.int =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/plugins.int (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build.assert/plugins.int 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,36 @@ +# Copied, perhaps edited, from ../../src/examplePlugins.int +INTERNAL_PLUGINS = \ +ADPCMCodecPlugin \ +AioPlugin \ +AsynchFilePlugin \ +B2DPlugin \ +BitBltPlugin \ +BMPReadWriterPlugin \ +CroquetPlugin \ +ZipPlugin \ +DropPlugin \ +DSAPrims \ +FFTPlugin \ +FileCopyPlugin \ +FilePlugin \ +FloatArrayPlugin \ +FloatMathPlugin \ +JoystickTabletPlugin \ +JPEGReaderPlugin \ +JPEGReadWriter2Plugin \ +Klatt \ +LargeIntegers \ +Matrix2x3Plugin \ +MIDIPlugin \ +MiscPrimitivePlugin \ +Mpeg3Plugin \ +RePlugin \ +SecurityPlugin \ +SerialPlugin \ +SocketPlugin \ +SoundCodecPrims \ +SoundGenerationPlugin \ +SoundPlugin \ +StarSqueakPlugin \ +SurfacePlugin \ +VMProfileLinuxSupportPlugin Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/mvm =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/mvm (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/mvm 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,29 @@ +#!/bin/bash +# debug Stack Spur VM with VM profiler and threaded heartbeat +INSTALLDIR=debug/stkspurlinuxhtARM +OPT="-g3 -O0 -DDEBUGVM=1" + +if [ $# -ge 1 ]; then + INSTALLDIR="$1"; shift +fi + +echo -n "clean? " +read a +case $a in +n|no|N|NO) echo "ok but this isn't safe!!";; +*) rm config.h; test -f Makefile && make reallyclean +esac +test -f config.h || ../../../platforms/unix/config/configure \ + --with-vmversion=5.0 \ + --with-src=spurstacksrc --with-plugins=src/plugins --disable-cogit \ + --without-vm-display-fbdev --without-npsqueak --enable-fast-bitblt \ + CC="gcc " \ + CXX="g++ " \ + CFLAGS="$OPT -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64" \ + LIBS="-lpthread -luuid" \ + LDFLAGS=-Wl,-z,now +rm -f vm/sqUnixMain.o # nuke version info +rm -rf ../../../products/$INSTALLDIR +# prefer make install prefix=`readlink -f \`pwd\`/../../../products/$INSTALLDIR` +# but older linux readlinks lack the -f flag and Raspbian lacks `readlinks` +make install prefix=`(cd ../../../;pwd)`/products/$INSTALLDIR 2>&1 | tee LOG Property changes on: branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/mvm ___________________________________________________________________ Added: svn:executable + * Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/plugins.ext =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/plugins.ext (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,9 @@ +# Copied, perhaps edited, from ../../src/examplePlugins.ext +EXTERNAL_PLUGINS = \ +LocalePlugin \ +SqueakFFIPrims \ +SqueakSSL \ +UnixOSProcessPlugin \ +UUIDPlugin \ +UnicodePlugin \ +XDisplayControlPlugin Added: branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/plugins.int =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/plugins.int (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/build.debug/plugins.int 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,36 @@ +# Copied, perhaps edited, from ../../src/examplePlugins.int +INTERNAL_PLUGINS = \ +ADPCMCodecPlugin \ +AioPlugin \ +AsynchFilePlugin \ +B2DPlugin \ +BitBltPlugin \ +BMPReadWriterPlugin \ +CroquetPlugin \ +ZipPlugin \ +DropPlugin \ +DSAPrims \ +FFTPlugin \ +FileCopyPlugin \ +FilePlugin \ +FloatArrayPlugin \ +FloatMathPlugin \ +JoystickTabletPlugin \ +JPEGReaderPlugin \ +JPEGReadWriter2Plugin \ +Klatt \ +LargeIntegers \ +Matrix2x3Plugin \ +MIDIPlugin \ +MiscPrimitivePlugin \ +Mpeg3Plugin \ +RePlugin \ +SecurityPlugin \ +SerialPlugin \ +SocketPlugin \ +SoundCodecPrims \ +SoundGenerationPlugin \ +SoundPlugin \ +StarSqueakPlugin \ +SurfacePlugin \ +VMProfileLinuxSupportPlugin Added: branches/Cog/build.linux32ARM/squeak.stack.spur/makeallclean =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/makeallclean (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/makeallclean 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,6 @@ +#!/bin/sh +trap 'exit 2' HUP INT PIPE TERM +for d in `dirname $0`/build*; do + (cd $d + echo y | ./mvm "$@") +done Property changes on: branches/Cog/build.linux32ARM/squeak.stack.spur/makeallclean ___________________________________________________________________ Added: svn:executable + * Added: branches/Cog/build.linux32ARM/squeak.stack.spur/makealldirty =================================================================== --- branches/Cog/build.linux32ARM/squeak.stack.spur/makealldirty (rev 0) +++ branches/Cog/build.linux32ARM/squeak.stack.spur/makealldirty 2014-11-05 20:48:12 UTC (rev 3126) @@ -0,0 +1,6 @@ +#!/bin/sh +trap 'exit 2' HUP INT PIPE TERM +for d in `dirname $0`/build*; do + (cd $d + echo n | ./mvm "$@") +done Property changes on: branches/Cog/build.linux32ARM/squeak.stack.spur/makealldirty ___________________________________________________________________ Added: svn:executable + * Modified: branches/Cog/build.linux32x86/HowToBuild =================================================================== --- branches/Cog/build.linux32x86/HowToBuild 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/HowToBuild 2014-11-05 20:48:12 UTC (rev 3126) @@ -11,6 +11,7 @@ - How to configure and build a VM on Unix - Testing an external plugin has completely linked - Optimization level and gcc version (please read!) + - Installing support libraries Overview @@ -79,7 +80,10 @@ Building out of the box ----------------------- -Install the tools (gcc, X11-devel, etc (e.g. libpng, libX11 & libxt source)). +Install the tools (gcc, X11-devel, etc, e.g. libpng, libuuid libX11 & libxt +source). See "Installing support libraries" below. If the configure step fails +when "checking for C compiler default output file name", you have yet to install +all the necessary support packages (e.g. libuuid). Then cd to the build directory of your choice, e.g. build.linux32x86/squeak.cog.spur/build Then either remove any BochsIA32Plugin line from plugins.ext or build the @@ -87,6 +91,9 @@ Then execute ./mvm answering "y" to perform a clean build or "n" to rebuild without recionfiguring. +Again, if the configure step fails when "checking for C compiler default output +file name", you have yet to install all the necessary support (e.g. lubuuid). + The subdirectories confrm to the production/assert/debug x itimer vs threaded heartbeat x single vs multi-threaded parts of the matrix described above. For example, build.linux32x86/squeak.cog.v3 includes @@ -218,3 +225,22 @@ gcc version 3.4.6 20060404 (Red Hat 3.4.6-10) using -O2 results in a VM that segfaults early in startup. For these compilers it is probably wise to use -O1, even though -O3 seems to work. + + +Installing support libraries +---------------------------- +Different linux distributions have different mechanisms for loading packages. +Here are some examples: + +CentOS +sudo yum install cairo-devel +sudo yum install pango-devel +sudo yum install libuuid-devel +sudo yum install libX11-devel +sudo yum install mesa-libGL-devel + +Ubuntu +sudo apt-get install libcairo2-dev +sudo apt-get install libpango1.0-dev + +More advice and examples for other distros gratefully received. Modified: branches/Cog/build.linux32x86/squeak.cog.spur/build/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.spur/build/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.spur/build/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.spur/build.assert/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.spur/build.assert/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.spur/build.assert/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.spur/build.assert.itimerheartbeat/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.spur/build.assert.itimerheartbeat/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.spur/build.assert.itimerheartbeat/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.spur/build.debug/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.spur/build.debug/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.spur/build.debug/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.spur/build.debug.itimerheartbeat/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.spur/build.debug.itimerheartbeat/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.spur/build.debug.itimerheartbeat/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.spur/build.itimerheartbeat/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.spur/build.itimerheartbeat/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.spur/build.itimerheartbeat/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build.assert/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build.assert/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build.assert/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build.assert.itimerheartbeat/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build.assert.itimerheartbeat/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build.assert.itimerheartbeat/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build.debug/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build.debug/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build.debug/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build.debug.itimerheartbeat/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build.debug.itimerheartbeat/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build.debug.itimerheartbeat/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build.itimerheartbeat/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build.itimerheartbeat/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build.itimerheartbeat/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded.assert/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded.assert/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded.assert/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded.debug/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded.debug/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.cog.v3/build.multithreaded.debug/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -5,6 +5,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.stack.spur/build/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.stack.spur/build/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.stack.spur/build/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -4,6 +4,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.stack.spur/build.assert/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.stack.spur/build.assert/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.stack.spur/build.assert/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -4,6 +4,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.stack.spur/build.debug/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.stack.spur/build.debug/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.stack.spur/build.debug/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -4,6 +4,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.stack.v3/build/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.stack.v3/build/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.stack.v3/build/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -4,6 +4,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.stack.v3/build.assert/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.stack.v3/build.assert/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.stack.v3/build.assert/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -4,6 +4,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Modified: branches/Cog/build.linux32x86/squeak.stack.v3/build.debug/plugins.ext =================================================================== --- branches/Cog/build.linux32x86/squeak.stack.v3/build.debug/plugins.ext 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/build.linux32x86/squeak.stack.v3/build.debug/plugins.ext 2014-11-05 20:48:12 UTC (rev 3126) @@ -4,6 +4,7 @@ SqueakFFIPrims \ SqueakSSL \ LocalePlugin \ +UnicodePlugin \ UnixOSProcessPlugin \ UUIDPlugin \ XDisplayControlPlugin Property changes on: branches/Cog/platforms/Cross/vm/sqSCCSVersion.h ___________________________________________________________________ Modified: checkindate - Tue Nov 4 17:11:21 PST 2014 + Wed Nov 5 12:46:58 PST 2014 Modified: branches/Cog/platforms/unix/config/configure =================================================================== --- branches/Cog/platforms/unix/config/configure 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/platforms/unix/config/configure 2014-11-05 20:48:12 UTC (rev 3126) @@ -28311,6 +28311,8 @@ echo "$as_me:$LINENO: checking for PangoCairo libraries" >&5 echo $ECHO_N "checking for PangoCairo libraries... $ECHO_C" >&6 +save_CPPFLAGS="$CPPFLAGS" +CPPFLAGS="-I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/lib/arm-linux-gnueabihf/glib-2.0/include" cat >conftest.$ac_ext <<_ACEOF /* confdefs.h. */ _ACEOF @@ -28367,6 +28369,7 @@ fi rm -f conftest.err conftest.$ac_objext conftest.$ac_ext +CPPFLAGS="$save_CPPFLAGS" if test "${plibs}"; then llibs="${LIBS}" for l in ${plibs}; do Modified: branches/Cog/platforms/unix/plugins/UnicodePlugin/Makefile.inc =================================================================== --- branches/Cog/platforms/unix/plugins/UnicodePlugin/Makefile.inc 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/platforms/unix/plugins/UnicodePlugin/Makefile.inc 2014-11-05 20:48:12 UTC (rev 3126) @@ -1,2 +1,2 @@ -XCPPFLAGS= -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabihf/glib-2.0/include -XLDFLAGS= -lpangocairo-1.0 -lcairo -lpango-1.0 -lgobject-2.0 -lglib-2.0 +XCPPFLAGS= -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/lib/arm-linux-gnueabihf/glib-2.0/include +XLDFLAGS= -lpangocairo-1.0 -lcairo -lpango-1.0 -lgobject-2.0 -lglib-2.0 Modified: branches/Cog/platforms/unix/plugins/UnicodePlugin/acinclude.m4 =================================================================== --- branches/Cog/platforms/unix/plugins/UnicodePlugin/acinclude.m4 2014-11-05 01:12:31 UTC (rev 3125) +++ branches/Cog/platforms/unix/plugins/UnicodePlugin/acinclude.m4 2014-11-05 20:48:12 UTC (rev 3126) @@ -1,9 +1,12 @@ -# -*- sh -*- - -AC_MSG_CHECKING([for PangoCairo libraries]) -AC_TRY_COMPILE([#include <pango/pangocairo.h>],[{}],[ - AC_MSG_RESULT(yes) -],[ - AC_MSG_RESULT(no) - AC_PLUGIN_DISABLE -]) +# -*- sh -*- + +AC_MSG_CHECKING([for PangoCairo libraries]) +save_CPPFLAGS="$CPPFLAGS" +CPPFLAGS="-I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/lib/arm-linux-gnueabihf/glib-2.0/include" +AC_TRY_COMPILE([#include <pango/pangocairo.h>],[{}],[ + AC_MSG_RESULT(yes) +],[ + AC_MSG_RESULT(no) + AC_PLUGIN_DISABLE +]) +CPPFLAGS="$save_CPPFLAGS"

1 0

Re: Vm-dev Digest, Vol 101, Issue 7
by Ralph Boland 05 Nov '14

05 Nov '14

> Hi Ralph, > > On Sun, Nov 2, 2014 at 9:48 PM, Ralph Boland <rpboland(a)gmail.com> wrote: > > > > > >> > > >> I am working on a parser generator tool (a replacement for SmaCC) and > > >> one of the things I a m interested in is the direct translation of > > >> language specifications into the virtual machine code (SmaCC and my > > >> current version of it use Squeak as the target language). > > >> > > > > > First, a different approach than compiling to Smalltalk is to compile to > > a > > > parse tree. We do this in the pseudo JavaScript compiler we've > > implemented > > > at Cadence, translating an AST into a Squeak compiler parse tree for code > > > generation. Targeting a parse tree gives you much more freedom; you can > > > express things that aren't expressible in Smalltalk. And if you target > > > bytecodes you can do even more. > > > > > > I never considered using a parse tree as the target. An interesting idea > > which in many instances may be the best approach. But for my regular > > expression > > example I would still want to generate byte codes. In any case I wouldn't > > want to > > restrict users of my parser generator tool to any one of the three options > > (Smalltalk code, > > parse tree, byte code). It is my responsibility to make all three options > > as easy and > > efficient as reasonably possible for users of the parser generator tool. > > Haven't put > > much thought into this yet though. So far Smalltalk (Squeak) is the only > > option. > > > > > > >> One of the problems I have is that, for some languages, the natural > > >> translation > > >> into VM code uses computed gotos. > > >> There are two scenarios here: > > >> > > >> 1) goto X where X is a variable. > > >> 2) goto (coll at: y) where coll is a Collection. > > >> > > > > > There are several ways of implementing this without computed bytecodes in > > > the instruction set, but there is also the possibility of implementing it > > > directly in the instruction set. > > > > > Off the top of my head one could > > > > > - map to perform: using some mangled selector. Yes this is problematic > > > because one has to split the scope between the two methods, so in general > > > it's not a solution > > > > Doesn't appeal to me. > > > > > - map to a case statement, which is what Squeak does. Just map it to a > > > sequence of compare and branches. Or better still, to a binary tree. > > > Coincidentally this is used by the JIT to implement block dispatch in > > > methods that contain more than one block. I know of other VM > > > implementations using it for PIC dispatch with really good performance. > > > > I don't know what you mean my Squeak mapping to a case statement since > > there is no case statement in Squeak/Smalltalk and I can't off hand think > > of > > where one is needed (some Squeak users might feel they need one but that > > is a > > different matter). > > > > But there is. And it is very convenient when one doesn't want to spread a > case over different classes, or when the cases distribute over values of > the same class (e.g. integer values). People often claim that a case > statement isn't necessary because one has polymorphism, but unless the > language supports singletons (which of course Smalltalk does not) a case > statement is much more readable than e.g. an if-then-else tree or a > "Dictionary mapping values to blocks or selectors" solution. > > Since you're unaware of the case statement I suggest you browse senders of > caseOf: and caseOf:otherwise: in Squeak trunk (which includes selectors of > optimized selectors that end up not being sent) or caseError, which is sent > when there's no otherwise clause. Here's an example: > > menuHook: aMenu named: aSymbol shifted: aBool > "Enhance aMenu with registered services." > aSymbol > caseOf: > { [ #classListMenu ] -> [ ServiceGui browser: self classMenu: aMenu ]. > [ #codePaneMenu ] -> [ ServiceGui browser: self codePaneMenu: aMenu ]. > [ #messageCategoryMenu] -> [ ServiceGui browser: self messageCategoryMenu: > aMenu ]. > [ #messageListMenu ] -> [ ServiceGui browser: self messageListMenu: aMenu ]. > [ #systemCategoryMenu ] -> [ ServiceGui browser: self classCategoryMenu: > aMenu ] } > otherwise: [ "do nothing" ] > > This compiles down to a sequence of comparisons. I was aware of caseOf: in Squeak. I always found it awkward to use and felt a true case statement would be simpler. Alas, it's impossible to have a true case statement added to Smalltalk now I think. caseOf: is user code and thus, in principle, outside the scope of the VM. For example, if I didn't want the sequential search of caseOf: I could implement a method doBlockAt: method in class Dictionary that assumes the values stored in the dictionary are blocks and invokes the block found by performing at: on the arg of doBlockAt:. But the code for both caseOf: and doBlockAt: are written by users so the VM can't optimize it, is forced to have the overhead of using blocks, and must be locked in to the search algorithm the user has chosen (for better or for worse). > > The use of compare and branches might be OK in some cases > > but a mess for the finite state machines generated from regular > > expressions. > > Actually, even with computed gotos FSMs are somewhat messy but without > > them > > it's worse. I don't know what 'PIC dispatch' is. > > > > Polymorphic inline cache dispatch. In JIT VMs such as Cog polymorphic send > sites are optimized using jump tables, one jump for each class encountered > at the send site, up tio some small limit such as 6 cases. Right now in > Cog these jump tables are sequential comparisons. But in some VMs, where > the degree of polymorphism may be much higher (think prototype languages > such as JavaScript) a binary tree may be much miore efficient, if harder to > organize. > > To use a binary tree don't I need some kind of computed goto for when I > > reach > > a leaf of the tree???? > > > > No. Once you're at the leaf you just include the bytecodes. So for > example take the following case statement: > > quickPrimitiveGeneratorFor: aQuickPrimitiveIndex > <api> > <returnTypeC: 'int (*quickPrimitiveGeneratorFor(sqInt > aQuickPrimitiveIndex))(void)'> > ^aQuickPrimitiveIndex > caseOf: { > [256] -> [#genQuickReturnSelf]. > [257] -> [#genQuickReturnConstNil]. > [258] -> [#genQuickReturnConstTrue]. > [259] -> [#genQuickReturnConstFalse]. > [260] -> [#genQuickReturnConstMinusOne]. > [261] -> [#genQuickReturnConstZero]. > [262] -> [#genQuickReturnConstOne]. > [263] -> [#genQuickReturnConstTwo] } > otherwise: [#genQuickReturnInstVar] > > This is compiled in the current Squeak compiler as > > 61 <10> pushTemp: 0 > 62 <88> dup > 63 <2A> pushConstant: 256 > 64 <B6> send: = > 65 <9B> jumpFalse: 70 > 66 <87> pop > 67 <29> pushConstant: #genQuickReturnSelf > 68 <A4 35> jumpTo: 123 > 70 <88> dup > 71 <28> pushConstant: 257 > 72 <B6> send: = > 73 <9B> jumpFalse: 78 > 74 <87> pop > 75 <21> pushConstant: #genQuickReturnConstNil > 76 <A4 2D> jumpTo: 123 > ... > 117 <22> pushConstant: 263 > 118 <B6> send: = > 119 <99> jumpFalse: 122 > 120 <21> pushConstant: #genQuickReturnConstTwo > 121 <90> jumpTo: 123 > 122 <20> pushConstant: #genQuickReturnInstVar > 123 <7C> returnTop > > but it could be more efficient if the code were > > pushTemp: 0 > dup > pushConstant: 256 > send: < > jumpFalse: L1 > dup > pushConstant: 260 > send: < > jumpFalse: L2 > dup > pushConstant: 258 > send: < > jumpFalse: L3 > dup > pushConstant: 257 > send: < > jumpFalse: L4 > pushConstant: #genQuickReturnSelf > jumpTo: L0 > pushConstant: #genQuickReturnConstNil > jumpTo: L0 > ... > L1: > pushConstant: #genQuickReturnInstVar > L0: > returnTop I have no problem with this code per say; if Squeak, the language, had a case statement it could be implemented this way. But I wouldn't want to be forced to implement my FSMs this way. It might be acceptable for small FSMs. I want to avoid sequential search and even binary search might be rather expensive. I look at computed gotos as the solution but, as you pointed out, computed gotos pose problems for JIT. Admittedly, for large FSM's, it might be best or necessary to use a FSM simulator anyway, as I do now. > > - use thisContext pc: value. > > > > This would be a possibility for me to experiment with for now. When I > > have a working > > parser generator tool I could campaign for my computed goto instructions > > to be added > > to the VM. > > > > > This /should/ be fine in the stack VM, but > > > slooooow in the JIT because internally mapping bytecode pcs to machine > > code > > > pcs is slow, and currently slower still because the frame will be > > converted > > > to a pure context and then converted back into a frame on the return from > > > pc:. But this solution isn't to be rejected out-of-hand. It can be > > > optimized to avoid the frame conversion and the JIT might be able to > > > optimize it. > > > > I assume that if computed gotos were used the translation to machine code > > would require a direct > > mapping of (virtually labeled) bytecode locations to machine code > > locations. I think this can be done > > in a reasonable amount of time but others such as yourself clearly > > understand the issues far better than > > I do. The dirty solution to start would be to simply not JIT the code > > that uses computed gotos. > > > > Yes, it could be. The JIT would generate a jump table from the literal > containing the bytecoded pcs, and there would be no conversion; only > indexing the jump table and jumping. Again, organizing that table as a > binary switch may well be faster on modern architectures. Indirect jumps > typically involve pipeline stalls, whereas binary jump trees don't. > > > The main problem is the compiler has no support for labels so > > > there would be work here. > > > > I don't mind doing the work but to my way of thinking "goto X" is pretty > > basic > > and is thus best handled at the VM/byte code level. Anything else is doing > > in a complicated way something that is fairly simple. Of course changing > > the VM/byte codes by even a single byte code is a major deal unless done > > when the VM/byte codes are initially created. Alas I must deal with what > > already exists. Even so, my preference is to work with the VM if at all > > possible. > > > > > > >> For example, one such language is that of regular expressions, which I > > >> wish to translate into finite state machines implemented in VM code. > > >> In this case I need case 2) gotos where coll is a collection of > > >> associations, possibly a > > >> Dictionary. I also plan to write a debugger for this (and other > > languages) > > >> but that is another story. > > >> > > >> I realize that the Cog VM is being built for Smalltalk (Squeak? Pharo?) > > >> for which the goto instructions are not needed and thus I assume > > >> unavailable. But there is something to > > >> viewing a virtual machine as general purpose and thus the target of > > >> multiple languages as is > > >> the case for the Java virtual machine. > > >> If the Cog VM is viewed this way then I argue there is a need for my > > goto > > >> instructions > > >> because some languages have need for them. > > >> For example, many languages have case statements. (I am all for object > > >> oriented > > >> but I would be willing to accept a case statement in Smalltalk too; the > > >> Squeak code > > >> implemented one in Squeak doesn't cut it). > > > > > > > > I've occasionally thought about this for many years. A computed jump > > might > > > be nice. Eg index an Array literal of pcs with the integer on top of > > > stack, falling through on bad type or out of range. > > > > This is the way I am thinking. If there are other reasons for a computed > > jumpTo > > as well all the better. > > > > >> Anyway, I am not arguing to Change Squeak or Smalltalk but I am arguing > > >> to have my goto instructions in Cog VM. Is there any chance of this????? > > >> > > > > > There's no chance of me spending time implementing this any time soon. I > > > have too much high-priority tasks to tackle this. But I want to > > encourage > > > you or others to have a go implementing it. It's fun! > > > > I understand and am willing to be the one to add one or more computed jump > > instructions, including working on the JIT code generator if needed. > > As you say it should be fun (and also educational). But > > 1) I am pretty busy now too and probably won't get to this for a year. > > 2) If I am to do this it would be great if someone can write a > > specification as to > > what is to be done. If someone can write this now that would be > > great but > > if they write it when I post that I am ready to do the work that > > would also > > be fine. > > 3) I don't want to just have my own private VM/byte codes. I want > > users of my > > parser generator tool to be able to load it into a standard > > version of Squeak > > and run it there including the possible generation of compilers > > for compiling > > their domain specific language programs into byte codes if desired. > > > > Sure. Get back to me when you're ready to work on it and I'll write the > specification, and help you with whatever info you need. There should be > room in the bytecode sets we'll likely be using next year. WILL DO. SHOULD BE FUN. WILL DO. SHOULD BE FUN. WILL DO. SHOULD BE FUN. > > > > > >> I don't know the Squeak VM or the Cog VM either but I assume these > > >> instructions don't exist because I see no need of them when the source > > >> language is > > >> Squeak or any version of Smalltalk for that matter. I also assume that > > >> there is already > > >> a full list of 256 instructions in the Cog VM and thus no room for my > > goto > > >> instructions > > >> unless some instructions are removed. > > >> > > >> Are there Cog VM instructions that are so rarely used that they could be > > >> removed without > > >> unreasonably slowing down the Cog VM interpretation of byte codes > > >> generated from Squeak source code????? > > >> > > > > > The current set has 3 unused bytecodes, one of which Spur uses, so > > > effectively there are two unused bytecodes. > > > > Levente Uzonyi in his posting pointed out that only one instruction is > > needed. > > I don't like having to push the address to jump to onto the stack, > > preferring a byte > > code with an argument, but I could live with his solution if that is what > > is decided. > > In the case of goto coll at: X the address is likely to end up on top > > of the stack > > anyway so Levente's jumpToTop instruction looks good in any case. > > > > If the bytecode is one that takes an integer on top of stack, and an Array > literal containing bytecode pcs, falling through on out of range, then > nothing other than the index need be pushed on top of stack. That would be > my preference. > Again, for my FSM, case this would often be considered to be good. But if the state transition tables are sparse then Dictionaries might be preferable to Arrays. My expection is that at: be sent to the collection object to get the address to go to. Knowing that the collection is an array though makes it easier for the compiler/VM to ensure that the addresses stored in the collection are valid. Actually, the compiler will be generating the addresses. Does the VM have absolute trust in the compiler to generate valid addresses? What if the compiler is generated by my parser generator tool? > > The Cog VMs support multiple bytecode sets. If you look at the > > > BytecodeSets package on VMMaker you can read the class comments of the > > > BytecodeEncoder subclasses such as EncoderForSistaV1. These bytecode > > sets > > > have a few more unused bytecodes. This multiple bytecode set support is > > > better implemented in Spur where there is only one compiled method header > > > format and support for 64k literals. So let me encourage you to move to > > > Spur and to look at the Sista set. The class comment of each encoder > > class > > > specifies the instruction set it targets. > > > > I am prepared to work with Spur and the Sista set. I am looking for > > someone to > > say that if I do this work that incorporating the work into Spur will be > > seriously > > considered. > > > > I can say that, yes. I will seriously consider adding it. I think its a > useful thing in the bytecode set, precisely to allow compiling other > languages to the VM. GLAD TO HEAR THIS. > > > > > > > Ralph Boland > > > > > > > > > -- > best, > Eliot Ralph On 3 November 2014 12:34, <vm-dev-request(a)lists.squeakfoundation.org> wrote: > Send Vm-dev mailing list submissions to > vm-dev(a)lists.squeakfoundation.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://lists.squeakfoundation.org/mailman/listinfo/vm-dev > or, via email, send a message with subject or body 'help' to > vm-dev-request(a)lists.squeakfoundation.org > > You can reach the person managing the list at > vm-dev-owner(a)lists.squeakfoundation.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Vm-dev digest..." > > > Today's Topics: > > 1. Re: goto instruction with Cog VM (Eliot Miranda) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Mon, 3 Nov 2014 11:34:48 -0800 > From: Eliot Miranda <eliot.miranda(a)gmail.com> > Subject: Re: [Vm-dev] goto instruction with Cog VM > To: Squeak Virtual Machine Development Discussion > <vm-dev(a)lists.squeakfoundation.org> > Message-ID: > < > CAC20JE1RduKQW_EdMsfah156QnTbR8USt9dEb4mj1rb15rndqA(a)mail.gmail.com> > Content-Type: text/plain; charset="utf-8" > > Hi Ralph, > > On Sun, Nov 2, 2014 at 9:48 PM, Ralph Boland <rpboland(a)gmail.com> wrote: > > > > > >> > > >> I am working on a parser generator tool (a replacement for SmaCC) and > > >> one of the things I a m interested in is the direct translation of > > >> language specifications into the virtual machine code (SmaCC and my > > >> current version of it use Squeak as the target language). > > >> > > > > > First, a different approach than compiling to Smalltalk is to compile > to > > a > > > parse tree. We do this in the pseudo JavaScript compiler we've > > implemented > > > at Cadence, translating an AST into a Squeak compiler parse tree for > code > > > generation. Targeting a parse tree gives you much more freedom; you > can > > > express things that aren't expressible in Smalltalk. And if you target > > > bytecodes you can do even more. > > > > > > I never considered using a parse tree as the target. An interesting idea > > which in many instances may be the best approach. But for my regular > > expression > > example I would still want to generate byte codes. In any case I > wouldn't > > want to > > restrict users of my parser generator tool to any one of the three > options > > (Smalltalk code, > > parse tree, byte code). It is my responsibility to make all three > options > > as easy and > > efficient as reasonably possible for users of the parser generator tool. > > Haven't put > > much thought into this yet though. So far Smalltalk (Squeak) is the > only > > option. > > > > > > >> One of the problems I have is that, for some languages, the natural > > >> translation > > >> into VM code uses computed gotos. > > >> There are two scenarios here: > > >> > > >> 1) goto X where X is a variable. > > >> 2) goto (coll at: y) where coll is a Collection. > > >> > > > > > There are several ways of implementing this without computed bytecodes > in > > > the instruction set, but there is also the possibility of implementing > it > > > directly in the instruction set. > > > > > Off the top of my head one could > > > > > - map to perform: using some mangled selector. Yes this is problematic > > > because one has to split the scope between the two methods, so in > general > > > it's not a solution > > > > Doesn't appeal to me. > > > > > - map to a case statement, which is what Squeak does. Just map it to a > > > sequence of compare and branches. Or better still, to a binary tree. > > > Coincidentally this is used by the JIT to implement block dispatch in > > > methods that contain more than one block. I know of other VM > > > implementations using it for PIC dispatch with really good performance. > > > > I don't know what you mean my Squeak mapping to a case statement since > > there is no case statement in Squeak/Smalltalk and I can't off hand > think > > of > > where one is needed (some Squeak users might feel they need one but that > > is a > > different matter). > > > > But there is. And it is very convenient when one doesn't want to spread a > case over different classes, or when the cases distribute over values of > the same class (e.g. integer values). People often claim that a case > statement isn't necessary because one has polymorphism, but unless the > language supports singletons (which of course Smalltalk does not) a case > statement is much more readable than e.g. an if-then-else tree or a > "Dictionary mapping values to blocks or selectors" solution. > > Since you're unaware of the case statement I suggest you browse senders of > caseOf: and caseOf:otherwise: in Squeak trunk (which includes selectors of > optimized selectors that end up not being sent) or caseError, which is sent > when there's no otherwise clause. Here's an example: > > menuHook: aMenu named: aSymbol shifted: aBool > "Enhance aMenu with registered services." > aSymbol > caseOf: > { [ #classListMenu ] -> [ ServiceGui browser: self classMenu: aMenu ]. > [ #codePaneMenu ] -> [ ServiceGui browser: self codePaneMenu: aMenu ]. > [ #messageCategoryMenu] -> [ ServiceGui browser: self messageCategoryMenu: > aMenu ]. > [ #messageListMenu ] -> [ ServiceGui browser: self messageListMenu: aMenu > ]. > [ #systemCategoryMenu ] -> [ ServiceGui browser: self classCategoryMenu: > aMenu ] } > otherwise: [ "do nothing" ] > > This compiles down to a sequence of comparisons. > > The use of compare and branches might be OK in some cases > > but a mess for the finite state machines generated from regular > > expressions. > > Actually, even with computed gotos FSMs are somewhat messy but without > > them > > it's worse. I don't know what 'PIC dispatch' is. > > > > Polymorphic inline cache dispatch. In JIT VMs such as Cog polymorphic send > sites are optimized using jump tables, one jump for each class encountered > at the send site, up tio some small limit such as 6 cases. Right now in > Cog these jump tables are sequential comparisons. But in some VMs, where > the degree of polymorphism may be much higher (think prototype languages > such as JavaScript) a binary tree may be much miore efficient, if harder to > organize. > > To use a binary tree don't I need some kind of computed goto for when I > > reach > > a leaf of the tree???? > > > > No. Once you're at the leaf you just include the bytecodes. So for > example take the following case statement: > > quickPrimitiveGeneratorFor: aQuickPrimitiveIndex > <api> > <returnTypeC: 'int (*quickPrimitiveGeneratorFor(sqInt > aQuickPrimitiveIndex))(void)'> > ^aQuickPrimitiveIndex > caseOf: { > [256] -> [#genQuickReturnSelf]. > [257] -> [#genQuickReturnConstNil]. > [258] -> [#genQuickReturnConstTrue]. > [259] -> [#genQuickReturnConstFalse]. > [260] -> [#genQuickReturnConstMinusOne]. > [261] -> [#genQuickReturnConstZero]. > [262] -> [#genQuickReturnConstOne]. > [263] -> [#genQuickReturnConstTwo] } > otherwise: [#genQuickReturnInstVar] > > This is compiled in the current Squeak compiler as > > 61 <10> pushTemp: 0 > 62 <88> dup > 63 <2A> pushConstant: 256 > 64 <B6> send: = > 65 <9B> jumpFalse: 70 > 66 <87> pop > 67 <29> pushConstant: #genQuickReturnSelf > 68 <A4 35> jumpTo: 123 > 70 <88> dup > 71 <28> pushConstant: 257 > 72 <B6> send: = > 73 <9B> jumpFalse: 78 > 74 <87> pop > 75 <21> pushConstant: #genQuickReturnConstNil > 76 <A4 2D> jumpTo: 123 > ... > 117 <22> pushConstant: 263 > 118 <B6> send: = > 119 <99> jumpFalse: 122 > 120 <21> pushConstant: #genQuickReturnConstTwo > 121 <90> jumpTo: 123 > 122 <20> pushConstant: #genQuickReturnInstVar > 123 <7C> returnTop > > but it could be more efficient if the code were > > pushTemp: 0 > dup > pushConstant: 256 > send: < > jumpFalse: L1 > dup > pushConstant: 260 > send: < > jumpFalse: L2 > dup > pushConstant: 258 > send: < > jumpFalse: L3 > dup > pushConstant: 257 > send: < > jumpFalse: L4 > pushConstant: #genQuickReturnSelf > jumpTo: L0 > pushConstant: #genQuickReturnConstNil > jumpTo: L0 > ... > L1: > pushConstant: #genQuickReturnInstVar > L0: > returnTop > > > - use thisContext pc: value. > > > > This would be a possibility for me to experiment with for now. When I > > have a working > > parser generator tool I could campaign for my computed goto instructions > > to be added > > to the VM. > > > > > This /should/ be fine in the stack VM, but > > > slooooow in the JIT because internally mapping bytecode pcs to machine > > code > > > pcs is slow, and currently slower still because the frame will be > > converted > > > to a pure context and then converted back into a frame on the return > from > > > pc:. But this solution isn't to be rejected out-of-hand. It can be > > > optimized to avoid the frame conversion and the JIT might be able to > > > optimize it. > > > > I assume that if computed gotos were used the translation to machine code > > would require a direct > > mapping of (virtually labeled) bytecode locations to machine code > > locations. I think this can be done > > in a reasonable amount of time but others such as yourself clearly > > understand the issues far better than > > I do. The dirty solution to start would be to simply not JIT the code > > that uses computed gotos. > > > > Yes, it could be. The JIT would generate a jump table from the literal > containing the bytecoded pcs, and there would be no conversion; only > indexing the jump table and jumping. Again, organizing that table as a > binary switch may well be faster on modern architectures. Indirect jumps > typically involve pipeline stalls, whereas binary jump trees don't. > > > The main problem is the compiler has no support for labels so > > > there would be work here. > > > > I don't mind doing the work but to my way of thinking "goto X" is pretty > > basic > > and is thus best handled at the VM/byte code level. Anything else is > doing > > in a complicated way something that is fairly simple. Of course changing > > the VM/byte codes by even a single byte code is a major deal unless done > > when the VM/byte codes are initially created. Alas I must deal with what > > already exists. Even so, my preference is to work with the VM if at all > > possible. > > > > > > >> For example, one such language is that of regular expressions, which I > > >> wish to translate into finite state machines implemented in VM code. > > >> In this case I need case 2) gotos where coll is a collection of > > >> associations, possibly a > > >> Dictionary. I also plan to write a debugger for this (and other > > languages) > > >> but that is another story. > > >> > > >> I realize that the Cog VM is being built for Smalltalk (Squeak? > Pharo?) > > >> for which the goto instructions are not needed and thus I assume > > >> unavailable. But there is something to > > >> viewing a virtual machine as general purpose and thus the target of > > >> multiple languages as is > > >> the case for the Java virtual machine. > > >> If the Cog VM is viewed this way then I argue there is a need for my > > goto > > >> instructions > > >> because some languages have need for them. > > >> For example, many languages have case statements. (I am all for > object > > >> oriented > > >> but I would be willing to accept a case statement in Smalltalk too; > the > > >> Squeak code > > >> implemented one in Squeak doesn't cut it). > > > > > > > > I've occasionally thought about this for many years. A computed jump > > might > > > be nice. Eg index an Array literal of pcs with the integer on top of > > > stack, falling through on bad type or out of range. > > > > This is the way I am thinking. If there are other reasons for a computed > > jumpTo > > as well all the better. > > > > >> Anyway, I am not arguing to Change Squeak or Smalltalk but I am > arguing > > >> to have my goto instructions in Cog VM. Is there any chance of > this????? > > >> > > > > > There's no chance of me spending time implementing this any time > soon. I > > > have too much high-priority tasks to tackle this. But I want to > > encourage > > > you or others to have a go implementing it. It's fun! > > > > I understand and am willing to be the one to add one or more computed > jump > > instructions, including working on the JIT code generator if needed. > > As you say it should be fun (and also educational). But > > 1) I am pretty busy now too and probably won't get to this for a > year. > > 2) If I am to do this it would be great if someone can write a > > specification as to > > what is to be done. If someone can write this now that would be > > great but > > if they write it when I post that I am ready to do the work that > > would also > > be fine. > > 3) I don't want to just have my own private VM/byte codes. I want > > users of my > > parser generator tool to be able to load it into a standard > > version of Squeak > > and run it there including the possible generation of compilers > > for compiling > > their domain specific language programs into byte codes if > desired. > > > > Sure. Get back to me when you're ready to work on it and I'll write the > specification, and help you with whatever info you need. There should be > room in the bytecode sets we'll likely be using next year. > > > > > > >> I don't know the Squeak VM or the Cog VM either but I assume these > > >> instructions don't exist because I see no need of them when the source > > >> language is > > >> Squeak or any version of Smalltalk for that matter. I also assume that > > >> there is already > > >> a full list of 256 instructions in the Cog VM and thus no room for my > > goto > > >> instructions > > >> unless some instructions are removed. > > >> > > >> Are there Cog VM instructions that are so rarely used that they could > be > > >> removed without > > >> unreasonably slowing down the Cog VM interpretation of byte codes > > >> generated from Squeak source code????? > > >> > > > > > The current set has 3 unused bytecodes, one of which Spur uses, so > > > effectively there are two unused bytecodes. > > > > Levente Uzonyi in his posting pointed out that only one instruction is > > needed. > > I don't like having to push the address to jump to onto the stack, > > preferring a byte > > code with an argument, but I could live with his solution if that is what > > is decided. > > In the case of goto coll at: X the address is likely to end up on top > > of the stack > > anyway so Levente's jumpToTop instruction looks good in any case. > > > > If the bytecode is one that takes an integer on top of stack, and an Array > literal containing bytecode pcs, falling through on out of range, then > nothing other than the index need be pushed on top of stack. That would be > my preference. > > > The Cog VMs support multiple bytecode sets. If you look at the > > > BytecodeSets package on VMMaker you can read the class comments of the > > > BytecodeEncoder subclasses such as EncoderForSistaV1. These bytecode > > sets > > > have a few more unused bytecodes. This multiple bytecode set support > is > > > better implemented in Spur where there is only one compiled method > header > > > format and support for 64k literals. So let me encourage you to move > to > > > Spur and to look at the Sista set. The class comment of each encoder > > class > > > specifies the instruction set it targets. > > > > I am prepared to work with Spur and the Sista set. I am looking for > > someone to > > say that if I do this work that incorporating the work into Spur will be > > seriously > > considered. > > > > I can say that, yes. I will seriously consider adding it. I think its a > useful thing in the bytecode set, precisely to allow compiling other > languages to the VM. > > > > > > > Ralph Boland > > > > > > > > > -- > best, > Eliot >

1 0

[commit][3125] CogVM source as per VMMaker.oscog-eem.923
by commits＠squeakvm.org 05 Nov '14

05 Nov '14

Revision: 3125 Author: eliot Date: 2014-11-04 17:12:31 -0800 (Tue, 04 Nov 2014) Log Message: ----------- CogVM source as per VMMaker.oscog-eem.923 Back out of Ryan's recent absent receiver changes, while accumulating the other recent fixes. There's some incompatibility with the working classes machinery at Cadence and Ryan's recent changes which we don't have time to resolve now. The kept changes are: Fix asserts in addFreeSubTree: & addToFreeTree:bytes:. Fix possible forwarding of the receiver in implicit receiver sends. Eliminate some compiler warnings in the refactored instVar:ofContext: changes. Modified Paths: -------------- branches/Cog/nscogsrc/vm/cogit.c branches/Cog/nscogsrc/vm/cogit.h branches/Cog/nscogsrc/vm/cointerp.c branches/Cog/nscogsrc/vm/cointerp.h branches/Cog/nscogsrc/vm/gcc3x-cointerp.c branches/Cog/nsspursrc/vm/cogit.c branches/Cog/nsspursrc/vm/cogit.h branches/Cog/nsspursrc/vm/cointerp.c branches/Cog/nsspursrc/vm/cointerp.h branches/Cog/nsspursrc/vm/gcc3x-cointerp.c branches/Cog/nsspurstacksrc/vm/gcc3x-interp.c branches/Cog/nsspurstacksrc/vm/interp.c branches/Cog/sistasrc/vm/cogit.c branches/Cog/sistasrc/vm/cogit.h branches/Cog/sistasrc/vm/cointerp.c branches/Cog/sistasrc/vm/cointerp.h branches/Cog/sistasrc/vm/gcc3x-cointerp.c branches/Cog/spursistasrc/vm/cogit.c branches/Cog/spursistasrc/vm/cogit.h branches/Cog/spursistasrc/vm/cointerp.c branches/Cog/spursistasrc/vm/cointerp.h branches/Cog/spursistasrc/vm/gcc3x-cointerp.c branches/Cog/spursrc/vm/cogit.c branches/Cog/spursrc/vm/cogit.h branches/Cog/spursrc/vm/cointerp.c branches/Cog/spursrc/vm/cointerp.h branches/Cog/spursrc/vm/gcc3x-cointerp.c branches/Cog/spurstacksrc/vm/gcc3x-interp.c branches/Cog/spurstacksrc/vm/interp.c branches/Cog/src/vm/cogit.c branches/Cog/src/vm/cogit.h branches/Cog/src/vm/cointerp.c branches/Cog/src/vm/cointerp.h branches/Cog/src/vm/cointerpmt.c branches/Cog/src/vm/cointerpmt.h branches/Cog/src/vm/gcc3x-cointerp.c branches/Cog/src/vm/gcc3x-cointerpmt.c branches/Cog/stacksrc/vm/gcc3x-interp.c branches/Cog/stacksrc/vm/interp.c Property Changed: ---------------- branches/Cog/platforms/Cross/vm/sqSCCSVersion.h Modified: branches/Cog/nscogsrc/vm/cogit.c =================================================================== --- branches/Cog/nscogsrc/vm/cogit.c 2014-11-04 22:36:20 UTC (rev 3124) +++ branches/Cog/nscogsrc/vm/cogit.c 2014-11-05 01:12:31 UTC (rev 3125) @@ -1,9 +1,9 @@ /* Automatically generated by - CCodeGenerator VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGenerator VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb from - StackToRegisterMappingCogit VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + StackToRegisterMappingCogit VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb */ -static char __buildInfo[] = "StackToRegisterMappingCogit VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 " __DATE__ ; +static char __buildInfo[] = "StackToRegisterMappingCogit VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb " __DATE__ ; char *__cogitBuildInfo = __buildInfo; @@ -924,8 +924,6 @@ static sqInt genExtReturnTopFromBlock(void); static sqInt genExtSendAbsentDynamicSuperBytecode(void); static sqInt genExtSendAbsentImplicitBytecode(void); -static sqInt genExtSendAbsentOuterBytecode(void); -static sqInt genExtSendAbsentSelfBytecode(void); static sqInt genExtSendBytecode(void); static sqInt genExtSendSuperBytecode(void); static sqInt genExtStoreAndPopLiteralVariableBytecode(void); @@ -1090,8 +1088,6 @@ static sqInt genReturnTopFromMethod(void); static sqInt genSendAbsentDynamicSupernumArgs(sqInt selector, sqInt numArgs) NoDbgRegParms; static sqInt genSendAbsentImplicitnumArgs(sqInt selector, sqInt numArgs) NoDbgRegParms; -static sqInt genSendAbsentOuternumArgsdepth(sqInt selector, sqInt numArgs, sqInt depth) NoDbgRegParms; -static sqInt genSendAbsentSelfnumArgs(sqInt selector, sqInt numArgs) NoDbgRegParms; static sqInt genSendDynamicSupernumArgs(sqInt selector, sqInt numArgs) NoDbgRegParms; static sqInt genSendSupernumArgs(sqInt selector, sqInt numArgs) NoDbgRegParms; static sqInt genSendTrampolineFornumArgscalledargargarg(void *aRoutine, sqInt numArgs, char *aString, sqInt regOrConst0, sqInt regOrConst1, sqInt regOrConst2) NoDbgRegParms; @@ -1114,7 +1110,7 @@ static void initSimStackForFramelessBlock(sqInt startpc) NoDbgRegParms; static void initSimStackForFramelessMethod(sqInt startpc) NoDbgRegParms; static sqInt liveRegisters(void); -static void marshallAbsentReceiverSendArguments(sqInt numArgs) NoDbgRegParms; +static void marshallImplicitReceiverSendArguments(sqInt numArgs) NoDbgRegParms; static void marshallSendArguments(sqInt numArgs) NoDbgRegParms; static void mergeafterContinuation(BytecodeFixup *fixup, sqInt mergeWithContinuation) NoDbgRegParms; static sqInt methodAbortTrampolineFor(sqInt numArgs) NoDbgRegParms; @@ -1753,16 +1749,16 @@ { genExtUnconditionalJump, v4LongBranchDistance, 0, 0, 0, 2, 0, 0, 0, 0, 1, 0, 0, 0 }, { genExtJumpIfTrue, v4LongBranchDistance, 0, 0, 0, 2, 1, 0, 0, 0, 1, 0, 0, 0 }, { genExtJumpIfFalse, v4LongBranchDistance, 0, 0, 0, 2, 0, 1, 0, 0, 1, 0, 0, 0 }, - { genExtSendAbsentSelfBytecode, 0, 0, 0, 0, 2, 0, 0, 0, 0, 1, 0, 0, 0 }, { unknownBytecode, 0, 0, 0, Nop, 2, 0, 0, 0, 0, 0, 0, 1, 0 }, { unknownBytecode, 0, 0, 0, Nop, 2, 0, 0, 0, 0, 0, 0, 1, 0 }, { unknownBytecode, 0, 0, 0, Nop, 2, 0, 0, 0, 0, 0, 0, 1, 0 }, + { unknownBytecode, 0, 0, 0, Nop, 2, 0, 0, 0, 0, 0, 0, 1, 0 }, { callPrimitiveBytecode, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0 }, { genPushRemoteTempLongBytecode, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0 }, { genStoreRemoteTempLongBytecode, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0 }, { genStoreAndPopRemoteTempLongBytecode, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0 }, { genExtPushClosureBytecode, v4BlockCodeSize, 0, 0, 0, 3, 0, 0, 0, 1, 0, 0, 0, 0 }, - { genExtSendAbsentOuterBytecode, 0, 0, 0, 0, 3, 0, 0, 0, 0, 1, 0, 0, 0 }, + { unknownBytecode, 0, 0, 0, Nop, 3, 0, 0, 0, 0, 0, 0, 1, 0 }, { unknownBytecode, 0, 0, 0, Nop, 3, 0, 0, 0, 0, 0, 0, 1, 0 } }; static sqInt guardPageSize; @@ -17448,45 +17444,6 @@ } -/* 254 11111110 i i i i i j j j kkkkkkkk Send To Absent Outer Literal - Selector #iiiii (+ Extend A * 32) with jjj (+ Extend B * 8) Arguments at - Depth kkkkkkkk - */ - -static sqInt -genExtSendAbsentOuterBytecode(void) -{ - sqInt depth; - sqInt litIndex; - sqInt nArgs; - - litIndex = (((usqInt) byte1) >> 3) + (extA << 5); - extA = 0; - nArgs = (byte1 & 7) + (extB << 3); - extB = 0; - depth = byte2; - return genSendAbsentOuternumArgsdepth(getLiteral(litIndex), nArgs, depth); -} - - -/* 245 11110101 i i i i i j j j Send To Absent Self Literal Selector #iiiii - (+ Extend A * 32) with jjj (+ Extend B * 8) Arguments - */ - -static sqInt -genExtSendAbsentSelfBytecode(void) -{ - sqInt litIndex; - sqInt nArgs; - - litIndex = (((usqInt) byte1) >> 3) + (extA << 5); - extA = 0; - nArgs = (byte1 & 7) + (extB << 3); - extB = 0; - return genSendAbsentSelfnumArgs(getLiteral(litIndex), nArgs); -} - - /* 238 11101110 i i i i i j j j Send Literal Selector #iiiii (+ Extend A * 32) with jjj (+ Extend B * 8) Arguments */ @@ -20977,7 +20934,7 @@ ssAllocateCallReg(ReceiverResultReg); /* begin MoveMw:r:R: */ genoperandoperandoperand(MoveMwrR, FoxMFReceiver, FPReg, ReceiverResultReg); - marshallAbsentReceiverSendArguments(numArgs); + marshallImplicitReceiverSendArguments(numArgs); return genMarshalledSendDynamicSupernumArgs(selector, numArgs); } @@ -20995,53 +20952,11 @@ if (result != 0) { return result; } - marshallAbsentReceiverSendArguments(numArgs); + marshallImplicitReceiverSendArguments(numArgs); return genMarshalledSendnumArgs(selector, numArgs); } - -/* OK, we could do better and avoid spilling ReceiverResultReg if we - refactored marshallAbsentReceiverSendArguments: to take a flag saying - whether the - receiver was in ReceiverResultReg (absent receiver send) or on the stack - (absent dynamic super send) and in the latter case loading - ReceiverResultReg from the stack after marshalling. But this is a rare - bytecode so for the moment - don't bother. */ - static sqInt -genSendAbsentOuternumArgsdepth(sqInt selector, sqInt numArgs, sqInt depth) -{ - (optStatus.isReceiverResultRegLive = 0); - ssAllocateCallReg(SendNumArgsReg); - /* begin MoveCq:R: */ - genoperandoperand(MoveCqR, depth, SendNumArgsReg); - CallRT(ceEnclosingObjectTrampoline); - marshallAbsentReceiverSendArguments(numArgs); - return genMarshalledSendnumArgs(selector, numArgs); -} - - -/* OK, we could do better and avoid spilling ReceiverResultReg if we - refactored marshallAbsentReceiverSendArguments: to take a flag saying - whether the - receiver was in ReceiverResultReg (absent receiver send) or on the stack - (absent dynamic super send) and in the latter case loading - ReceiverResultReg from the stack after marshalling. But this is a rare - bytecode so for the moment - don't bother. */ - -static sqInt -genSendAbsentSelfnumArgs(sqInt selector, sqInt numArgs) -{ - ssAllocateCallReg(ReceiverResultReg); - /* begin MoveMw:r:R: */ - genoperandoperandoperand(MoveMwrR, FoxMFReceiver, FPReg, ReceiverResultReg); - marshallAbsentReceiverSendArguments(numArgs); - return genMarshalledSendnumArgs(selector, numArgs); -} - -static sqInt genSendDynamicSupernumArgs(sqInt selector, sqInt numArgs) { marshallSendArguments(numArgs); @@ -22029,7 +21944,7 @@ necessary. */ static void -marshallAbsentReceiverSendArguments(sqInt numArgs) +marshallImplicitReceiverSendArguments(sqInt numArgs) { sqInt index; sqInt numSpilled; Modified: branches/Cog/nscogsrc/vm/cogit.h =================================================================== --- branches/Cog/nscogsrc/vm/cogit.h 2014-11-04 22:36:20 UTC (rev 3124) +++ branches/Cog/nscogsrc/vm/cogit.h 2014-11-05 01:12:31 UTC (rev 3125) @@ -1,5 +1,5 @@ /* Automatically generated by - CCodeGenerator VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGenerator VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb */ Modified: branches/Cog/nscogsrc/vm/cointerp.c =================================================================== --- branches/Cog/nscogsrc/vm/cointerp.c 2014-11-04 22:36:20 UTC (rev 3124) +++ branches/Cog/nscogsrc/vm/cointerp.c 2014-11-05 01:12:31 UTC (rev 3125) @@ -1,9 +1,9 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb from - CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e + CoInterpreter VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb */ -static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e " __DATE__ ; +static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb " __DATE__ ; char *__interpBuildInfo = __buildInfo; @@ -20,8 +20,8 @@ #endif #include <stddef.h> /* for e.g. alloca */ -#include <wchar.h> #include <setjmp.h> +#include <wchar.h> /* for wint_t */ #include "sq.h" #include "vmCallback.h" #include "sqMemoryFence.h" @@ -463,7 +463,7 @@ static sqInt iframeReceiver(char *theFP) NoDbgRegParms; static sqInt iframeSavedIP(char *theFP) NoDbgRegParms; void ifValidWriteBackStackPointersSaveTo(void *theCFP, void *theCSP, char **savedFPP, char **savedSPP); -sqInt implicitReceiverFormixinimplementing(sqInt methodReceiver, sqInt methodMixin, sqInt selector); +sqInt implicitReceiverFormixinimplementing(sqInt rcvr, sqInt mixin, sqInt selector); usqInt instructionPointerAddress(void); static sqInt instructionPointerForFramecurrentFPcurrentIP(char *spouseFP, char *currentFP, sqInt instrPtr) NoDbgRegParms; usqInt interpretAddress(void); @@ -853,7 +853,7 @@ void printWronglySizedContexts(sqInt printContexts); static void restoreHeadersFromtofromandtofrom(sqInt firstIn, sqInt lastIn, sqInt hdrBaseIn, sqInt firstOut, sqInt lastOut, sqInt hdrBaseOut) NoDbgRegParms; static void runLeakCheckerForFullGC(sqInt fullGCFlag) NoDbgRegParms; -static usqInt safeObjectAfter(sqInt oop) NoDbgRegParms; +static sqInt safeObjectAfter(sqInt oop) NoDbgRegParms; static sqInt safePrintStringOf(sqInt oop) NoDbgRegParms; static sqInt shortentoIndexableSize(sqInt obj, sqInt nSlots) NoDbgRegParms; static sqInt sizeBitsOfSafe(sqInt oop) NoDbgRegParms; @@ -1055,7 +1055,7 @@ sqInt doSignalSemaphoreWithIndex(sqInt index); static void dummyReferToProxy(void); EXPORT(sqInt) dumpImage(sqInt fileName); -static sqInt enclosingObjectAtwithObjectwithMixin(sqInt depth, sqInt methodReceiver, sqInt methodMixin) NoDbgRegParms; +static sqInt enclosingObjectAtwithObjectwithMixin(sqInt n, sqInt anObject, sqInt mixin) NoDbgRegParms; static sqInt ensureCallerContext(char *theFP) NoDbgRegParms; static void ensureImageFormatIsUpToDate(sqInt swapBytes) NoDbgRegParms; static char * establishFrameForContextToReturnTo(sqInt contextToReturnTo) NoDbgRegParms; @@ -1069,6 +1069,8 @@ sqInt fetchIntegerofObject(sqInt fieldIndex, sqInt objectPointer); static sqInt fetchPointerofMarriedContext(sqInt offset, sqInt aContext) NoDbgRegParms; static sqInt fetchStackPointerOf(sqInt aContext) NoDbgRegParms; +static sqInt findApplicationOfTargetMixinstartingAtBehavior(sqInt targetMixin, sqInt aBehavior) NoDbgRegParms; +static sqInt findApplicationOfTargetMixinstartingAtNonMetaClass(sqInt targetMixin, sqInt aClass) NoDbgRegParms; static sqInt findClassContainingMethodstartingAt(sqInt meth, sqInt classObj) NoDbgRegParms; sqInt findClassOfMethodforReceiver(sqInt meth, sqInt rcvr); static char * findFrameAboveinPage(char *theFP, StackPage *thePage) NoDbgRegParms; @@ -1401,6 +1403,7 @@ _iss sqInt tenuringThreshold; _iss sqInt forceTenureFlag; _iss sqInt longRunningPrimitiveCheckSequenceNumber; +_iss sqInt methodDictLinearSearchLimit; _iss sqInt signalLowSpace; _iss usqLong statFullGCUsecs; _iss sqInt statFullGCs; @@ -1416,7 +1419,6 @@ _iss sqInt interruptKeycode; _iss sqInt interruptPending; _iss usqInt memory; -_iss sqInt methodDictLinearSearchLimit; _iss usqLong nextPollUsecs; _iss usqLong statForceInterruptCheck; _iss usqLong statIGCDeltaUsecs; @@ -2086,7 +2088,7 @@ /* 574 */ (void (*)(void))0, /* 575 */ (void (*)(void))0, 0 }; -const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.922"; +const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.923"; sqInt minBackwardJumpCountForCompile = MinBackwardJumpCountForCompile /* 40 */; volatile int sendTrace; @@ -4951,12 +4953,9 @@ case 126: /* dynamicSuperSendBytecode */ { - sqInt aBehavior; sqInt ccIndex; sqInt mClassMixin; - sqInt mixin; sqInt mixinApplication; - sqInt mixinApplication1; sqInt rcvr; VM_LABEL(0dynamicSuperSendBytecode); @@ -4966,19 +4965,11 @@ GIV(messageSelector) = longAt((GIV(method) + BaseHeaderSize) + (((byteAtPointer(++localIP)) + LiteralStart) << ShiftForWord)); rcvr = longAtPointer(localSP + (GIV(argumentCount) * BytesPerOop)); mClassMixin = methodClassOf(GIV(method)); - /* begin findApplicationOfTargetMixin:startingAtBehavior: */ - aBehavior = ((rcvr & 1) + mixinApplication = findApplicationOfTargetMixinstartingAtBehavior(mClassMixin, ((rcvr & 1) ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) : (((ccIndex = (((usqInt) (longAt(rcvr))) >> (compactClassFieldLSB())) & 0x1F)) == 0 ? (longAt(rcvr - BaseHeaderSize)) & AllButTypeMask - : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))); - mixinApplication1 = aBehavior; - while (!((mixinApplication1 == GIV(nilObj)) - || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << ShiftForWord)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); - } - mixinApplication = mixinApplication1; + : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord))))); lkupClassTag = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); goto commonSend; } @@ -12767,14 +12758,11 @@ case 497: /*241*/ /* extSendAbsentDynamicSuperBytecode */ { - sqInt aBehavior; sqInt byte; sqInt ccIndex; sqInt i; sqInt mClassMixin; - sqInt mixin; sqInt mixinApplication; - sqInt mixinApplication1; sqInt oop; VM_LABEL(0extSendAbsentDynamicSuperBytecode); @@ -12796,19 +12784,11 @@ } longAtPointerput(localSP + (GIV(argumentCount) * BytesPerOop), longAt(localFP + FoxIFReceiver)); mClassMixin = methodClassOf(GIV(method)); - /* begin findApplicationOfTargetMixin:startingAtBehavior: */ - aBehavior = (((longAt(localFP + FoxIFReceiver)) & 1) + mixinApplication = findApplicationOfTargetMixinstartingAtBehavior(mClassMixin, (((longAt(localFP + FoxIFReceiver)) & 1) ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) : (((ccIndex = (((usqInt) (longAt(longAt(localFP + FoxIFReceiver)))) >> (compactClassFieldLSB())) & 0x1F)) == 0 ? (longAt((longAt(localFP + FoxIFReceiver)) - BaseHeaderSize)) & AllButTypeMask - : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))); - mixinApplication1 = aBehavior; - while (!((mixinApplication1 == GIV(nilObj)) - || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << ShiftForWord)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); - } - mixinApplication = mixinApplication1; + : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord))))); lkupClassTag = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); goto commonSend; } @@ -12957,45 +12937,10 @@ } break; case 501: /*245*/ - /* extSendAbsentSelfBytecode */ - { - sqInt byte; - sqInt ccIndex; - sqInt classObj; - sqInt i; - sqInt oop; - - VM_LABEL(0extSendAbsentSelfBytecode); - byte = byteAtPointer(++localIP); - /* begin literal: */ - assert(GIV(method) == (iframeMethod(localFP))); - GIV(messageSelector) = longAt((GIV(method) + BaseHeaderSize) + ((((((usqInt) byte) >> 3) + (extA << 5)) + LiteralStart) << ShiftForWord)); - extA = 0; - GIV(argumentCount) = (byte & 7) + (extB << 3); - extB = 0; - /* begin shuffleArgumentsAndStoreAbsentReceiver: */ - - /* a.k.a. self internalPush: anything */ - - localSP -= BytesPerOop; - for (i = 1; i <= GIV(argumentCount); i += 1) { - oop = longAtPointer(localSP + (i * BytesPerOop)); - longAtPointerput(localSP + ((i - 1) * BytesPerOop), oop); - } - longAtPointerput(localSP + (GIV(argumentCount) * BytesPerOop), longAt(localFP + FoxIFReceiver)); - /* begin classTagForClass: */ - classObj = (((longAt(localFP + FoxIFReceiver)) & 1) - ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) - : (((ccIndex = (((usqInt) (longAt(longAt(localFP + FoxIFReceiver)))) >> (compactClassFieldLSB())) & 0x1F)) == 0 - ? (longAt((longAt(localFP + FoxIFReceiver)) - BaseHeaderSize)) & AllButTypeMask - : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))); - lkupClassTag = classObj; - goto commonSend; - } - break; case 502: /*246*/ case 503: /*247*/ case 504: /*248*/ + case 510: /*254*/ case 511: /*255*/ /* unknownBytecode */ { @@ -13113,47 +13058,6 @@ longAtPointerput((localSP -= BytesPerOop), newClosure); } break; - case 510: /*254*/ - /* extSendAbsentOuterBytecode */ - { - sqInt absentReceiver; - sqInt byte; - sqInt ccIndex; - sqInt classObj; - sqInt depth; - sqInt i; - sqInt oop; - - VM_LABEL(0extSendAbsentOuterBytecode); - byte = byteAtPointer(++localIP); - /* begin literal: */ - assert(GIV(method) == (iframeMethod(localFP))); - GIV(messageSelector) = longAt((GIV(method) + BaseHeaderSize) + ((((((usqInt) byte) >> 3) + (extA << 5)) + LiteralStart) << ShiftForWord)); - extA = 0; - GIV(argumentCount) = (byte & 7) + (extB << 3); - extB = 0; - depth = byteAtPointer(++localIP); - absentReceiver = enclosingObjectAtwithObjectwithMixin(depth, longAt(localFP + FoxIFReceiver), methodClassOf(GIV(method))); - /* begin shuffleArgumentsAndStoreAbsentReceiver: */ - - /* a.k.a. self internalPush: anything */ - - localSP -= BytesPerOop; - for (i = 1; i <= GIV(argumentCount); i += 1) { - oop = longAtPointer(localSP + (i * BytesPerOop)); - longAtPointerput(localSP + ((i - 1) * BytesPerOop), oop); - } - longAtPointerput(localSP + (GIV(argumentCount) * BytesPerOop), absentReceiver); - /* begin classTagForClass: */ - classObj = ((absentReceiver & 1) - ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) - : (((ccIndex = (((usqInt) (longAt(absentReceiver))) >> (compactClassFieldLSB())) & 0x1F)) == 0 - ? (longAt(absentReceiver - BaseHeaderSize)) & AllButTypeMask - : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))); - lkupClassTag = classObj; - goto commonSend; - } - break; } } @@ -14286,7 +14190,6 @@ sqInt ceDynamicSuperSendtonumArgs(sqInt selector, sqInt rcvr, sqInt numArgs) { DECL_MAYBE_SQ_GLOBAL_STRUCT - sqInt aBehavior; sqInt ccIndex; sqInt classTag; CogMethod *cogMethod; @@ -14295,9 +14198,7 @@ sqInt mClassMixin; sqInt methodHeader; sqInt methodHeader1; - sqInt mixin; sqInt mixinApplication; - sqInt mixinApplication1; CogMethod *newCogMethod; sqInt top; sqInt top1; @@ -14311,19 +14212,11 @@ /* begin sendBreakpoint:receiver: */ sendBreakpointreceiver(firstFixedFieldOfMaybeImmediate(selector), lengthOfMaybeImmediate(selector), rcvr); mClassMixin = methodClassOf(((mframeHomeMethod(GIV(framePointer)))->methodObject)); - /* begin findApplicationOfTargetMixin:startingAtBehavior: */ - aBehavior = ((rcvr & 1) + mixinApplication = findApplicationOfTargetMixinstartingAtBehavior(mClassMixin, ((rcvr & 1) ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) : (((ccIndex = (((usqInt) (longAt(rcvr))) >> (compactClassFieldLSB())) & 0x1F)) == 0 ? (longAt(rcvr - BaseHeaderSize)) & AllButTypeMask - : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))); - mixinApplication1 = aBehavior; - while (!((mixinApplication1 == GIV(nilObj)) - || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << ShiftForWord)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); - } - mixinApplication = mixinApplication1; + : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord))))); assert((lengthOf(mixinApplication)) > (InstanceSpecificationIndex + 1)); classTag = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); GIV(argumentCount) = numArgs; @@ -16726,145 +16619,245 @@ } -/* This is used to implement implicit receiver sends in Newspeak. Find the - nearest lexically-enclosing implementation of selector by searching up the - static chain of the - method's receiver, starting at mixin's application. This implementation is +/* This is used to implement the innards of the pushImplicitReceiverBytecode, + used for implicit receiver sends in NS2/NS3. Find the nearest + lexically-enclosing implementation of selector by searching up the static + chain of anObject, + starting at mixin's application. This is an iterative implementation derived from - <ContextPart> implicitReceiverFor: methodReceiver <Object> - in: methodMixin <Mixin> + <ContextPart> implicitReceiverFor: obj <Object> + withMixin: mixin <Mixin> implementing: selector <Symbol> ^<Object> */ -/* This is used to implement implicit receiver sends in Newspeak. Find the - nearest lexically-enclosing implementation of selector by searching up the - static chain of the - method's receiver, starting at mixin's application. This implementation is +/* This is used to implement the innards of the pushImplicitReceiverBytecode, + used for implicit receiver sends in NS2/NS3. Find the nearest + lexically-enclosing implementation of selector by searching up the static + chain of anObject, + starting at mixin's application. This is an iterative implementation derived from - <ContextPart> implicitReceiverFor: methodReceiver <Object> - in: methodMixin <Mixin> + <ContextPart> implicitReceiverFor: obj <Object> + withMixin: mixin <Mixin> implementing: selector <Symbol> ^<Object> */ sqInt -implicitReceiverFormixinimplementing(sqInt methodReceiver, sqInt methodMixin, sqInt selector) +implicitReceiverFormixinimplementing(sqInt rcvr, sqInt mixin, sqInt selector) { DECL_MAYBE_SQ_GLOBAL_STRUCT - sqInt aBehavior; - sqInt candidateMixin; - sqInt candidateMixinApplication; - sqInt candidateReceiver; sqInt ccIndex; + sqInt ccIndex1; sqInt dictionary; + sqInt dictionary1; sqInt found; + sqInt found1; sqInt header; + sqInt header1; + sqInt implicitReceiver; + sqInt implicitReceiverOrNil; sqInt index; + sqInt index1; sqInt length; + sqInt length1; sqInt mask; + sqInt mask1; sqInt methodArray; - sqInt mixin; + sqInt methodArray1; sqInt mixinApplication; + sqInt mixinApplication1; sqInt nextSelector; + sqInt nextSelector1; sqInt sz; + sqInt sz1; + sqInt targetMixin; + sqInt theMixin; + sqInt theMixin1; sqInt wrapAround; + sqInt wrapAround1; if (breakOnImplicitReceiver()) { /* begin sendBreakpoint:receiver: */ sendBreakpointreceiver(firstFixedFieldOfMaybeImmediate(selector), lengthOfMaybeImmediate(selector), null); } - assert(!((isOopForwarded(methodReceiver)))); - assert(!((isForwarded(methodMixin)))); + assert(!((isOopForwarded(rcvr)))); + assert(!((isForwarded(mixin)))); GIV(messageSelector) = selector; - candidateReceiver = methodReceiver; - candidateMixin = methodMixin; - while (1) { - /* begin findApplicationOfTargetMixin:startingAtBehavior: */ - aBehavior = ((candidateReceiver & 1) - ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) - : (((ccIndex = (((usqInt) (longAt(candidateReceiver))) >> (compactClassFieldLSB())) & 0x1F)) == 0 - ? (longAt(candidateReceiver - BaseHeaderSize)) & AllButTypeMask - : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))); - mixinApplication = aBehavior; - while (!((mixinApplication == GIV(nilObj)) - || ((mixinApplication == candidateMixin) - || (((mixin = longAt((mixinApplication + BaseHeaderSize) + (MixinIndex << ShiftForWord)))) == candidateMixin)))) { - mixinApplication = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); - } - candidateMixinApplication = mixinApplication; - if (candidateMixinApplication == GIV(nilObj)) { - return methodReceiver; - } - dictionary = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << ShiftForWord)); - /* begin lookupMethodInDictionary: */ - /* begin numSlotsOf: */ - header = longAt(dictionary); - sz = ((header & TypeMask) == HeaderTypeSizeAndClass - ? (longAt(dictionary - (BytesPerWord * 2))) & AllButTypeMask - : header & SizeMask); - length = ((usqInt) (sz - BaseHeaderSize)) >> ShiftForWord; + mixinApplication = findApplicationOfTargetMixinstartingAtBehavior(mixin, ((rcvr & 1) + ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) + : (((ccIndex = (((usqInt) (longAt(rcvr))) >> (compactClassFieldLSB())) & 0x1F)) == 0 + ? (longAt(rcvr - BaseHeaderSize)) & AllButTypeMask + : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord))))); + if (mixinApplication == GIV(nilObj)) { + return rcvr; + } + dictionary = longAt((mixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << ShiftForWord)); + /* begin lookupMethodInDictionary: */ + /* begin numSlotsOf: */ + header = longAt(dictionary); + sz = ((header & TypeMask) == HeaderTypeSizeAndClass + ? (longAt(dictionary - (BytesPerWord * 2))) & AllButTypeMask + : header & SizeMask); + length = ((usqInt) (sz - BaseHeaderSize)) >> ShiftForWord; - /* Use linear search on small dictionaries; its cheaper. - Also the limit can be set to force linear search of all dictionaries, which supports the - booting of images that need rehashing (e.g. because a tracer has generated an image - with different hashes but hasn't rehashed it yet.) */ + /* Use linear search on small dictionaries; its cheaper. + Also the limit can be set to force linear search of all dictionaries, which supports the + booting of images that need rehashing (e.g. because a tracer has generated an image + with different hashes but hasn't rehashed it yet.) */ - mask = (length - SelectorStart) - 1; - if (mask <= GIV(methodDictLinearSearchLimit)) { - index = 0; - while (index <= mask) { - nextSelector = longAt((dictionary + BaseHeaderSize) + ((index + SelectorStart) << ShiftForWord)); - - if (nextSelector == GIV(messageSelector)) { - methodArray = longAt((dictionary + BaseHeaderSize) + (MethodArrayIndex << ShiftForWord)); - GIV(newMethod) = longAt((methodArray + BaseHeaderSize) + (index << ShiftForWord)); - found = 1; - goto l1; - } - index += 1; + mask = (length - SelectorStart) - 1; + if (mask <= GIV(methodDictLinearSearchLimit)) { + index = 0; + while (index <= mask) { + nextSelector = longAt((dictionary + BaseHeaderSize) + ((index + SelectorStart) << ShiftForWord)); + + if (nextSelector == GIV(messageSelector)) { + methodArray = longAt((dictionary + BaseHeaderSize) + (MethodArrayIndex << ShiftForWord)); + GIV(newMethod) = longAt((methodArray + BaseHeaderSize) + (index << ShiftForWord)); + found = 1; + goto l1; } - found = 0; - goto l1; + index += 1; } + found = 0; + goto l1; + } - /* It is assumed that there are some nils in this dictionary, and search will - stop when one is encountered. However, if there are no nils, then wrapAround - will be detected the second time the loop gets to the end of the table. */ + /* It is assumed that there are some nils in this dictionary, and search will + stop when one is encountered. However, if there are no nils, then wrapAround + will be detected the second time the loop gets to the end of the table. */ - index = SelectorStart + (mask & (((GIV(messageSelector) & 1) + index = SelectorStart + (mask & (((GIV(messageSelector) & 1) ? (GIV(messageSelector) >> 1) : (((usqInt) (longAt(GIV(messageSelector)))) >> HashBitsOffset) & HashMaskUnshifted))); - wrapAround = 0; - while (1) { - nextSelector = longAt((dictionary + BaseHeaderSize) + (index << ShiftForWord)); - if (nextSelector == GIV(nilObj)) { + wrapAround = 0; + while (1) { + nextSelector = longAt((dictionary + BaseHeaderSize) + (index << ShiftForWord)); + if (nextSelector == GIV(nilObj)) { + found = 0; + goto l1; + } + + if (nextSelector == GIV(messageSelector)) { + methodArray = longAt((dictionary + BaseHeaderSize) + (MethodArrayIndex << ShiftForWord)); + GIV(newMethod) = longAt((methodArray + BaseHeaderSize) + ((index - SelectorStart) << ShiftForWord)); + found = 1; + goto l1; + } + index += 1; + if (index == length) { + if (wrapAround) { found = 0; goto l1; } - - if (nextSelector == GIV(messageSelector)) { - methodArray = longAt((dictionary + BaseHeaderSize) + (MethodArrayIndex << ShiftForWord)); - GIV(newMethod) = longAt((methodArray + BaseHeaderSize) + ((index - SelectorStart) << ShiftForWord)); - found = 1; - goto l1; + wrapAround = 1; + index = SelectorStart; + } + } + found = 0; +l1: /* end lookupMethodInDictionary: */; + if (!found) { + theMixin = longAt((mixinApplication + BaseHeaderSize) + (MixinIndex << ShiftForWord)); + /* begin nextImplicitReceiverFor:withMixin: */ + assert(!((isOopForwarded(longAt((mixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << ShiftForWord)))))); + assert(!((isForwarded(longAt((theMixin + BaseHeaderSize) + (EnclosingMixinIndex << ShiftForWord)))))); + implicitReceiver = longAt((mixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << ShiftForWord)); + targetMixin = longAt((theMixin + BaseHeaderSize) + (EnclosingMixinIndex << ShiftForWord)); + while (1) { + if (targetMixin == GIV(nilObj)) { + + /* or: [implicitReceiver = objectMemory nilObject] */ + + implicitReceiverOrNil = GIV(nilObj); + goto l3; } - index += 1; - if (index == length) { - if (wrapAround) { - found = 0; - goto l1; + mixinApplication1 = findApplicationOfTargetMixinstartingAtNonMetaClass(targetMixin, ((implicitReceiver & 1) + ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) + : (((ccIndex1 = (((usqInt) (longAt(implicitReceiver))) >> (compactClassFieldLSB())) & 0x1F)) == 0 + ? (longAt(implicitReceiver - BaseHeaderSize)) & AllButTypeMask + : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex1 - 1) << ShiftForWord))))); + if (mixinApplication1 == GIV(nilObj)) { + implicitReceiverOrNil = GIV(nilObj); + goto l3; + } + dictionary1 = longAt((mixinApplication1 + BaseHeaderSize) + (MethodDictionaryIndex << ShiftForWord)); + /* begin lookupMethodInDictionary: */ + /* begin numSlotsOf: */ + header1 = longAt(dictionary1); + sz1 = ((header1 & TypeMask) == HeaderTypeSizeAndClass + ? (longAt(dictionary1 - (BytesPerWord * 2))) & AllButTypeMask + : header1 & SizeMask); + length1 = ((usqInt) (sz1 - BaseHeaderSize)) >> ShiftForWord; + + /* Use linear search on small dictionaries; its cheaper. + Also the limit can be set to force linear search of all dictionaries, which supports the + booting of images that need rehashing (e.g. because a tracer has generated an image + with different hashes but hasn't rehashed it yet.) */ + + mask1 = (length1 - SelectorStart) - 1; + if (mask1 <= GIV(methodDictLinearSearchLimit)) { + index1 = 0; + while (index1 <= mask1) { + nextSelector1 = longAt((dictionary1 + BaseHeaderSize) + ((index1 + SelectorStart) << ShiftForWord)); + + if (nextSelector1 == GIV(messageSelector)) { + methodArray1 = longAt((dictionary1 + BaseHeaderSize) + (MethodArrayIndex << ShiftForWord)); + GIV(newMethod) = longAt((methodArray1 + BaseHeaderSize) + (index1 << ShiftForWord)); + found1 = 1; + goto l2; + } + index1 += 1; } - wrapAround = 1; - index = SelectorStart; + found1 = 0; + goto l2; } + + /* It is assumed that there are some nils in this dictionary, and search will + stop when one is encountered. However, if there are no nils, then wrapAround + will be detected the second time the loop gets to the end of the table. */ + + index1 = SelectorStart + (mask1 & (((GIV(messageSelector) & 1) + ? (GIV(messageSelector) >> 1) + : (((usqInt) (longAt(GIV(messageSelector)))) >> HashBitsOffset) & HashMaskUnshifted))); + wrapAround1 = 0; + while (1) { + nextSelector1 = longAt((dictionary1 + BaseHeaderSize) + (index1 << ShiftForWord)); + if (nextSelector1 == GIV(nilObj)) { + found1 = 0; + goto l2; + } + + if (nextSelector1 == GIV(messageSelector)) { + methodArray1 = longAt((dictionary1 + BaseHeaderSize) + (MethodArrayIndex << ShiftForWord)); + GIV(newMethod) = longAt((methodArray1 + BaseHeaderSize) + ((index1 - SelectorStart) << ShiftForWord)); + found1 = 1; + goto l2; + } + index1 += 1; + if (index1 == length1) { + if (wrapAround1) { + found1 = 0; + goto l2; + } + wrapAround1 = 1; + index1 = SelectorStart; + } + } + found1 = 0; + l2: /* end lookupMethodInDictionary: */; + if (found1) break; + implicitReceiver = longAt((mixinApplication1 + BaseHeaderSize) + (EnclosingObjectIndex << ShiftForWord)); + theMixin1 = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << ShiftForWord)); + if (theMixin1 == GIV(nilObj)) { + implicitReceiverOrNil = GIV(nilObj); + goto l3; + } + targetMixin = longAt((theMixin1 + BaseHeaderSize) + (EnclosingMixinIndex << ShiftForWord)); } - found = 0; - l1: /* end lookupMethodInDictionary: */; - if (found) { - return candidateReceiver; + implicitReceiverOrNil = implicitReceiver; + l3: /* end nextImplicitReceiverFor:withMixin: */; + if (implicitReceiverOrNil != GIV(nilObj)) { + return implicitReceiverOrNil; } - candidateMixin = longAt((candidateMixin + BaseHeaderSize) + (EnclosingMixinIndex << ShiftForWord)); - if (candidateMixin == GIV(nilObj)) break; - candidateReceiver = longAt((candidateMixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << ShiftForWord)); } - return methodReceiver; + return rcvr; } usqInt @@ -36647,15 +36640,17 @@ sqInt header1; sqInt header2; sqInt header3; + sqInt header4; usqInt lastWord; sqInt newFreeChunk; sqInt newOop; - usqInt next; + sqInt next; sqInt oop; sqInt realHeader; sqInt sz; sqInt sz1; sqInt sz2; + sqInt sz3; sqInt target; usqInt w; @@ -36666,7 +36661,21 @@ /* begin objectAfterWhileForwarding: */ header2 = longAt(oop); if ((header2 & MarkBit) == 0) { - next = ((sqInt) (objectAfter(oop))); + /* begin objectAfter: */ + if (!(asserta(oopisLessThan(oop, GIV(freeStart))))) { + error("no objects after the end of memory"); + } + if (((longAt(oop)) & TypeMask) == HeaderTypeFree) { + sz2 = (longAt(oop)) & AllButTypeMask; + } + else { + /* begin sizeBitsOf: */ + header3 = longAt(oop); + sz2 = ((header3 & TypeMask) == HeaderTypeSizeAndClass + ? (longAt(oop - (BytesPerWord * 2))) & LongSizeMask + : header3 & SizeMask); + } + next = (oop + sz2) + (headerTypeBytes[(longAt(oop + sz2)) & TypeMask]); goto l1; } fwdBlock1 = (header2 & AllButMarkBitAndTypeMask) << 1; @@ -36681,7 +36690,7 @@ else { sz1 = realHeader & SizeMask; } - next = ((sqInt) ((oop + sz1) + (headerTypeBytes[(longAt(oop + sz1)) & TypeMask]))); + next = (oop + sz1) + (headerTypeBytes[(longAt(oop + sz1)) & TypeMask]); l1: /* end objectAfterWhileForwarding: */; if (!(((longAt(oop)) & TypeMask) == HeaderTypeFree)) { @@ -36735,18 +36744,18 @@ } /* begin safeObjectAfter: */ if (((longAt(newFreeChunk)) & TypeMask) == HeaderTypeFree) { - sz2 = (longAt(newFreeChunk)) & AllButTypeMask; + sz3 = (longAt(newFreeChunk)) & AllButTypeMask; } else { /* begin sizeBitsOf: */ - header3 = longAt(newFreeChunk); - sz2 = ((header3 & TypeMask) == HeaderTypeSizeAndClass + header4 = longAt(newFreeChunk); + sz3 = ((header4 & TypeMask) == HeaderTypeSizeAndClass ? (longAt(newFreeChunk - (BytesPerWord * 2))) & LongSizeMask - : header3 & SizeMask); + : header4 & SizeMask); } - next = ((((usqInt)(newFreeChunk + sz2))) >= GIV(freeStart) + next = ((((usqInt)(newFreeChunk + sz3))) >= GIV(freeStart) ? GIV(freeStart) - : (newFreeChunk + sz2) + (headerTypeBytes[(longAt(newFreeChunk + sz2)) & TypeMask])); + : (newFreeChunk + sz3) + (headerTypeBytes[(longAt(newFreeChunk + sz3)) & TypeMask])); assert((next == GIV(freeStart)) || (next == (oopFromChunk(GIV(compEnd))))); if (next == GIV(freeStart)) { @@ -38622,7 +38631,7 @@ given object or free chunk in memory. Return freeStart when enumeration is complete. This is for assertion checking only. */ -static usqInt +static sqInt safeObjectAfter(sqInt oop) { DECL_MAYBE_SQ_GLOBAL_STRUCT sqInt header; @@ -44572,40 +44581,35 @@ } -/* This is used to implement outer sends and outer expressions in Newspeak. */ +/* This is used to implement the innards of the pushEnclosingObjectBytecode, + used for explicit outer sends in NS2/NS3. */ static sqInt -enclosingObjectAtwithObjectwithMixin(sqInt depth, sqInt methodReceiver, sqInt methodMixin) +enclosingObjectAtwithObjectwithMixin(sqInt n, sqInt anObject, sqInt mixin) { DECL_MAYBE_SQ_GLOBAL_STRUCT - sqInt aBehavior; sqInt ccIndex; sqInt count; sqInt enclosingObject; - sqInt mixin; sqInt mixinApplication; - sqInt mixinApplication1; sqInt targetMixin; - enclosingObject = methodReceiver; - targetMixin = methodMixin; + enclosingObject = anObject; + targetMixin = mixin; count = 0; - while (count < depth) { + while (count < n) { count += 1; - assert(!((targetMixin == (nilObject())))); - /* begin findApplicationOfTargetMixin:startingAtBehavior: */ - aBehavior = ((enclosingObject & 1) + if ((targetMixin == GIV(nilObj)) + || (enclosingObject == GIV(nilObj))) { + return GIV(nilObj); + } + mixinApplication = findApplicationOfTargetMixinstartingAtNonMetaClass(targetMixin, ((enclosingObject & 1) ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) : (((ccIndex = (((usqInt) (longAt(enclosingObject))) >> (compactClassFieldLSB())) & 0x1F)) == 0 ? (longAt(enclosingObject - BaseHeaderSize)) & AllButTypeMask - : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))); - mixinApplication1 = aBehavior; - while (!((mixinApplication1 == GIV(nilObj)) - || ((mixinApplication1 == targetMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << ShiftForWord)))) == targetMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); + : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord))))); + if (mixinApplication == GIV(nilObj)) { + return GIV(nilObj); } - mixinApplication = mixinApplication1; - assert(!((mixinApplication == (nilObject())))); enclosingObject = longAt((mixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << ShiftForWord)); targetMixin = longAt((targetMixin + BaseHeaderSize) + (EnclosingMixinIndex << ShiftForWord)); } @@ -45365,7 +45369,62 @@ return (sp >> 1); } + +/* This is used to implement the innards of the pushImplicitReceiverBytecode, + used for outer sends in NS2/NS3. Find the MixinApplcation of which + aBehavior is a subclass that is an application of targetMixin. This is an + implementation derived from + + <ContextPart> findApplicationOfTargetMixin: targetMixin + startingAtBehavior: aBehavior + */ + static sqInt +findApplicationOfTargetMixinstartingAtBehavior(sqInt targetMixin, sqInt aBehavior) +{ DECL_MAYBE_SQ_GLOBAL_STRUCT + sqInt ccIndex; + sqInt mixin; + sqInt mixinOrMixinApplication; + + mixinOrMixinApplication = aBehavior; + while (!((mixinOrMixinApplication == GIV(nilObj)) + || ((mixinOrMixinApplication == targetMixin) + || ((((mixin = longAt((mixinOrMixinApplication + BaseHeaderSize) + (MixinIndex << ShiftForWord)))) == targetMixin) + || (((((ccIndex = (((usqInt) (longAt(mixin))) >> (compactClassFieldLSB())) & 0x1F)) == 0 + ? (longAt(mixin - BaseHeaderSize)) & AllButTypeMask + : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))) == targetMixin))))) { + mixinOrMixinApplication = longAt((mixinOrMixinApplication + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); + } + return mixinOrMixinApplication; +} + + +/* This is used to implement the innards of the pushImplicitReceiverBytecode, + used for outer sends in NS2/NS3. Find the MixinApplcation of which aClass + is a subclass that is an application of targetMixin. This is an + implementation derived from + + <ContextPart> findApplicationOfTargetMixin: targetMixin + startingAtNonMetaClass: aClass + */ + +static sqInt +findApplicationOfTargetMixinstartingAtNonMetaClass(sqInt targetMixin, sqInt aClass) +{ DECL_MAYBE_SQ_GLOBAL_STRUCT + sqInt mixinOrMixinApplication; + + assert(!((isForwarded(targetMixin)))); + assert(!((isForwarded(aClass)))); + mixinOrMixinApplication = aClass; + while (!((mixinOrMixinApplication == GIV(nilObj)) + || ((mixinOrMixinApplication == targetMixin) + || ((longAt((mixinOrMixinApplication + BaseHeaderSize) + (MixinIndex << ShiftForWord))) == targetMixin)))) { + mixinOrMixinApplication = longAt((mixinOrMixinApplication + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); + } + return mixinOrMixinApplication; +} + +static sqInt findClassContainingMethodstartingAt(sqInt meth, sqInt classObj) { DECL_MAYBE_SQ_GLOBAL_STRUCT sqInt classDict; Modified: branches/Cog/nscogsrc/vm/cointerp.h =================================================================== --- branches/Cog/nscogsrc/vm/cointerp.h 2014-11-04 22:36:20 UTC (rev 3124) +++ branches/Cog/nscogsrc/vm/cointerp.h 2014-11-05 01:12:31 UTC (rev 3125) @@ -1,5 +1,5 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb */ @@ -64,7 +64,7 @@ void (*functionPointerForCompiledMethodprimitiveIndex(sqInt methodObj, sqInt primIndex))(void) ; sqInt getCheckAllocFiller(void); void ifValidWriteBackStackPointersSaveTo(void *theCFP, void *theCSP, char **savedFPP, char **savedSPP); -sqInt implicitReceiverFormixinimplementing(sqInt methodReceiver, sqInt methodMixin, sqInt selector); +sqInt implicitReceiverFormixinimplementing(sqInt rcvr, sqInt mixin, sqInt selector); usqInt instructionPointerAddress(void); usqInt interpretAddress(void); sqInt isCogMethodReference(sqInt methodHeader); Modified: branches/Cog/nscogsrc/vm/gcc3x-cointerp.c =================================================================== --- branches/Cog/nscogsrc/vm/gcc3x-cointerp.c 2014-11-04 22:36:20 UTC (rev 3124) +++ branches/Cog/nscogsrc/vm/gcc3x-cointerp.c 2014-11-05 01:12:31 UTC (rev 3125) @@ -2,11 +2,11 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb from - CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e + CoInterpreter VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb */ -static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e " __DATE__ ; +static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.923 uuid: b4a88a08-47d9-4173-9d8d-a2c11886c5fb " __DATE__ ; char *__interpBuildInfo = __buildInfo; @@ -23,8 +23,8 @@ #endif #include <stddef.h> /* for e.g. alloca */ -#include <wchar.h> #include <setjmp.h> +#include <wchar.h> /* for wint_t */ #include "sqGnu.h" #include "vmCallback.h" #include "sqMemoryFence.h" @@ -466,7 +466,7 @@ static sqInt iframeReceiver(char *theFP) NoDbgRegParms; static sqInt iframeSavedIP(char *theFP) NoDbgRegParms; void ifValidWriteBackStackPointersSaveTo(void *theCFP, void *theCSP, char **savedFPP, char **savedSPP); -sqInt implicitReceiverFormixinimplementing(sqInt methodReceiver, sqInt methodMixin, sqInt selector); +sqInt implicitReceiverFormixinimplementing(sqInt rcvr, sqInt mixin, sqInt selector); usqInt instructionPointerAddress(void); static sqInt instructionPointerForFramecurrentFPcurrentIP(char *spouseFP, char *currentFP, sqInt instrPtr) NoDbgRegParms; usqInt interpretAddress(void); @@ -856,7 +856,7 @@ void printWronglySizedContexts(sqInt printContexts); static void restoreHeadersFromtofromandtofrom(sqInt firstIn, sqInt lastIn, sqInt hdrBaseIn, sqInt firstOut, sqInt lastOut, sqInt hdrBaseOut) NoDbgRegParms; static void runLeakCheckerForFullGC(sqInt fullGCFlag) NoDbgRegParms; -static usqInt safeObjectAfter(sqInt oop) NoDbgRegParms; +static sqInt safeObjectAfter(sqInt oop) NoDbgRegParms; static sqInt safePrintStringOf(sqInt oop) NoDbgRegParms; static sqInt shortentoIndexableSize(sqInt obj, sqInt nSlots) NoDbgRegParms; static sqInt sizeBitsOfSafe(sqInt oop) NoDbgRegParms; @@ -1058,7 +1058,7 @@ sqInt doSignalSemaphoreWithIndex(sqInt index); static void dummyReferToProxy(void); EXPORT(sqInt) dumpImage(sqInt fileName); -static sqInt enclosingObjectAtwithObjectwithMixin(sqInt depth, sqInt methodReceiver, sqInt methodMixin) NoDbgRegParms; +static sqInt enclosingObjectAtwithObjectwithMixin(sqInt n, sqInt anObject, sqInt mixin) NoDbgRegParms; static sqInt ensureCallerContext(char *theFP) NoDbgRegParms; static void ensureImageFormatIsUpToDate(sqInt swapBytes) NoDbgRegParms; static char * establishFrameForContextToReturnTo(sqInt contextToReturnTo) NoDbgRegParms; @@ -1072,6 +1072,8 @@ sqInt fetchIntegerofObject(sqInt fieldIndex, sqInt objectPointer); static sqInt fetchPointerofMarriedContext(sqInt offset, sqInt aContext) NoDbgRegParms; static sqInt fetchStackPointerOf(sqInt aContext) NoDbgRegParms; +static sqInt findApplicationOfTargetMixinstartingAtBehavior(sqInt targetMixin, sqInt aBehavior) NoDbgRegParms; +static sqInt findApplicationOfTargetMixinstartingAtNonMetaClass(sqInt targetMixin, sqInt aClass) NoDbgRegParms; static sqInt findClassContainingMethodstartingAt(sqInt meth, sqInt classObj) NoDbgRegParms; sqInt findClassOfMethodforReceiver(sqInt meth, sqInt rcvr); static char * findFrameAboveinPage(char *theFP, StackPage *thePage) NoDbgRegParms; @@ -1404,6 +1406,7 @@ _iss sqInt tenuringThreshold; _iss sqInt forceTenureFlag; _iss sqInt longRunningPrimitiveCheckSequenceNumber; +_iss sqInt methodDictLinearSearchLimit; _iss sqInt signalLowSpace; _iss usqLong statFullGCUsecs; _iss sqInt statFullGCs; @@ -1419,7 +1422,6 @@ _iss sqInt interruptKeycode; _iss sqInt interruptPending; _iss usqInt memory; -_iss sqInt methodDictLinearSearchLimit; _iss usqLong nextPollUsecs; _iss usqLong statForceInterruptCheck; _iss usqLong statIGCDeltaUsecs; @@ -2089,7 +2091,7 @@ /* 574 */ (void (*)(void))0, /* 575 */ (void (*)(void))0, 0 }; -const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.922"; +const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.923"; sqInt minBackwardJumpCountForCompile = MinBackwardJumpCountForCompile /* 40 */; volatile int sendTrace; @@ -4960,12 +4962,9 @@ CASE(126) /* dynamicSuperSendBytecode */ { - sqInt aBehavior; sqInt ccIndex; sqInt mClassMixin; - sqInt mixin; sqInt mixinApplication; - sqInt mixinApplication1; sqInt rcvr; VM_LABEL(0dynamicSuperSendBytecode); @@ -4975,19 +4974,11 @@ GIV(messageSelector) = longAt((GIV(method) + BaseHeaderSize) + (((byteAtPointer(++localIP)) + LiteralStart) << ShiftForWord)); rcvr = longAtPointer(localSP + (GIV(argumentCount) * BytesPerOop)); mClassMixin = methodClassOf(GIV(method)); - /* begin findApplicationOfTargetMixin:startingAtBehavior: */ - aBehavior = ((rcvr & 1) + mixinApplication = findApplicationOfTargetMixinstartingAtBehavior(mClassMixin, ((rcvr & 1) ? longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (ClassSmallInteger << ShiftForWord)) : (((ccIndex = (((usqInt) (longAt(rcvr))) >> (compactClassFieldLSB())) & 0x1F)) == 0 ? (longAt(rcvr - BaseHeaderSize)) & AllButTypeMask - : longAt(((longAt((GIV(specialObjectsOop) + BaseHeaderSize) + (CompactClasses << ShiftForWord))) + BaseHeaderSize) + ((ccIndex - 1) << ShiftForWord)))); - mixinApplication1 = aBehavior; - while (!((mixinApplication1 == GIV(nilObj)) - || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << ShiftForWord)))) == mClassMixin)))) { @@ Diff output truncated at 50000 characters. @@

1 0

VM Maker: VMMaker.oscog-eem.923.mcz
by commits＠source.squeak.org 05 Nov '14

05 Nov '14

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-eem.923.mcz ==================== Summary ==================== Name: VMMaker.oscog-eem.923 Author: eem Time: 4 November 2014, 4:45:37.508 pm UUID: b4a88a08-47d9-4173-9d8d-a2c11886c5fb Ancestors: VMMaker.oscog-eem.918 Back out of Ryan's recent absent receiver changes, while accumulating the other recent fixes. There's some incompatibility with the working classes machinery at Cadence and Ryan's recent changes which we don't have time to resolve right now. The kept changes are: Fix asserts in addFreeSubTree: & addToFreeTree:bytes:. Fix possible forwarding of the receiver in implicit receiver sends. Eliminate some compiler warnings in the refactored instVar:ofContext: changes. =============== Diff against VMMaker.oscog-eem.918 =============== Item was changed: ----- Method: CoInterpreter>>instVar:ofContext: (in category 'frame access') ----- instVar: offset ofContext: aContext "Fetch an instance avriable from a maybe married context. If the context is still married compute the value of the relevant inst var from the spouse frame's state. If the context is single but has a negative instruction pointer recognise that the instruction pointer is actually into machine code and convert it to the corresponding bytecode pc." | value spouseFP | <var: #spouseFP type: #'char *'> <inline: true> self assert: offset < MethodIndex. self assert: (objectMemory isContext: aContext). self writeBackHeadFramePointers. (self isMarriedOrWidowedContext: aContext) ifFalse: [value := objectMemory fetchPointer: offset ofObject: aContext. (offset = InstructionPointerIndex and: [(objectMemory isIntegerObject: value) and: [value signedIntFromLong < 0]]) ifTrue: [value := self internalMustMapMachineCodePC: (objectMemory integerValueOf: value) context: aContext]. ^value]. (self isWidowedContext: aContext) ifTrue: [^objectMemory fetchPointer: offset ofObject: aContext]. spouseFP := self frameOfMarriedContext: aContext. offset = SenderIndex ifTrue: [^self ensureCallerContext: spouseFP]. offset = StackPointerIndex ifTrue: [self assert: ReceiverIndex + (self stackPointerIndexForFrame: spouseFP) < (objectMemory lengthOf: aContext). ^objectMemory integerObjectOf: (self stackPointerIndexForFrame: spouseFP)]. offset = InstructionPointerIndex ifTrue: + [^self instructionPointerForFrame: spouseFP currentFP: localFP currentIP: (self oopForPointer: localIP)]. - [^self instructionPointerForFrame: spouseFP currentFP: localFP currentIP: localIP]. self error: 'bad index'. ^0! Item was changed: ----- Method: CoInterpreter>>instructionPointerForFrame:currentFP:currentIP: (in category 'frame access') ----- instructionPointerForFrame: spouseFP currentFP: currentFP currentIP: instrPtr "Answer the bytecode pc object (i.e. SmallInteger) for an active frame. The bytecode pc is derived from the frame's pc. If the frame is the top frame on the current stack the frame pc is whatever the current instruction pointer is. If the frame is the top frame on some other stack the frame pc is the value on top of stack. Otherwise the frame pc is the saved pc of the frame above. Once the frame pc is found it must be mapped to a bytecode pc." + <var: #spouseFP type: #'char *'> + <var: #currentFP type: #'char *'> | value theIP thePage theFPAbove | + <var: #thePage type: #'StackPage *'> + <var: #theFPAbove type: #'char *'> spouseFP = currentFP ifTrue: [theIP := self oopForPointer: instrPtr] ifFalse: [thePage := stackPages stackPageFor: spouseFP. theFPAbove := self findFrameAbove: spouseFP inPage: thePage. theIP := theFPAbove == 0 ifTrue: [stackPages longAt: thePage headSP] ifFalse:[self oopForPointer: (self frameCallerSavedIP: theFPAbove)]]. value := self contextInstructionPointer: theIP frame: spouseFP. ^value signedIntFromLong < 0 ifTrue: [self mustMapMachineCodePC: (objectMemory integerValueOf: value) context: (self frameContext: spouseFP)] ifFalse: [value]! Item was added: + ----- Method: CoInterpreter>>internalFollowedReceiver (in category 'internal interpreter access') ----- + internalFollowedReceiver + <inline: true> + | rcvr | + rcvr := stackPages longAt: localFP + FoxIFReceiver. + (objectMemory isOopForwarded: rcvr) ifTrue: + [rcvr := objectMemory followForwarded: rcvr. + stackPages longAt: localFP + FoxIFReceiver put: rcvr]. + ^rcvr! Item was added: + ----- Method: CogObjectRepresentationForSpur>>genEnsureObjInRegNotForwarded:scratchReg:updatingMw:r: (in category 'compile abstract instructions') ----- + genEnsureObjInRegNotForwarded: reg scratchReg: scratch updatingMw: offset r: baseReg + "Make sure that the object in reg is not forwarded, and update the field reg[offset] is + updated if the object in reg is forwarded. + Use the fact that isForwardedObjectClassIndexPun is a power of two to save an instruction." + | loop imm ok | + <var: #ok type: #'AbstractInstruction *'> + <var: #imm type: #'AbstractInstruction *'> + <var: #loop type: #'AbstractInstruction *'> + self assert: reg ~= scratch. + loop := cogit Label. + cogit MoveR: reg R: scratch. + imm := self genJumpImmediateInScratchReg: scratch. + "notionally + self genGetClassIndexOfNonImm: reg into: scratch. + cogit CmpCq: objectMemory isForwardedObjectClassIndexPun R: TempReg. + but the following is an instruction shorter:" + cogit MoveMw: 0 r: reg R: scratch. + cogit + AndCq: objectMemory classIndexMask - objectMemory isForwardedObjectClassIndexPun + R: scratch. + ok := cogit JumpNonZero: 0. + self genLoadSlot: 0 sourceReg: reg destReg: reg. + cogit MoveR: reg Mw: offset r: baseReg. + cogit Jump: loop. + ok jmpTarget: (imm jmpTarget: cogit Label). + ^0! Item was changed: ----- Method: Cogit>>generateNewspeakRuntime (in category 'initialization') ----- generateNewspeakRuntime <option: #NewspeakVM> | jumpMiss jumpItsTheReceiverStupid retpcReg | <var: #jumpMiss type: #'AbstractInstruction *'> <var: #jumpItsTheReceiverStupid type: #'AbstractInstruction *'> "Generate the non-send runtime support for Newspeak, explicit outer and implicit receiver. The dynamic frequency of explicit outer is so low we merely call an interpreter routine." ceEnclosingObjectTrampoline := self genTrampolineFor: #ceEnclosingObjectAt: called: 'ceEnclosingObjectTrampoline' arg: SendNumArgsReg result: ReceiverResultReg. "Cached push implicit receiver implementation. If objectRepresentation doesn't support pinning then caller looks like mov selector, SendNumArgsReg call ceImplicitReceiver br continue Lclass: .word Lmixin:: .word continue: If objectRepresentation supports pinning then caller looks like mov Lclass, Arg1Reg mov selector, SendNumArgsReg call ceImplicitReceiver and Lclass: .word; Lmixin: .word is somewhere on the heap. If class tag matches class of receiver then mixin contains either 0 or the implicit receiver. If 0, answer the actual receiver, otherwise the mixin. Generate the class fetch and cache probe inline for speed. Smashes Arg1Reg, RegClass and caller-saved regs." opcodeIndex := 0. self MoveMw: FoxMFReceiver r: FPReg R: ReceiverResultReg. objectRepresentation genGetInlineCacheClassTagFrom: ReceiverResultReg into: ClassReg forEntry: false. objectRepresentation canPinObjects ifTrue: [self MoveMw: 0 r: Arg1Reg R: TempReg. self CmpR: ClassReg R: TempReg. jumpMiss := self JumpNonZero: 0. self MoveMw: BytesPerOop r: Arg1Reg R: TempReg. self CmpCq: 0 R: TempReg. jumpItsTheReceiverStupid := self JumpZero: 0. self MoveR: TempReg R: ReceiverResultReg. jumpItsTheReceiverStupid jmpTarget: (self RetN: 0). jumpMiss jmpTarget: self Label. + objectRepresentation + genEnsureObjInRegNotForwarded: ReceiverResultReg + scratchReg: TempReg + updatingMw: FoxMFReceiver + r: FPReg. ceImplicitReceiverTrampoline := self genTrampolineFor: #ceImplicitReceiverFor:receiver:cache: called: 'ceImplicitReceiverTrampoline' numArgs: 3 arg: SendNumArgsReg arg: ReceiverResultReg arg: Arg1Reg arg: nil saveRegs: false pushLinkReg: true resultReg: ReceiverResultReg appendOpcodes: true] ifFalse: [backEnd hasLinkRegister ifTrue: [retpcReg := LinkReg] ifFalse: [self MoveMw: 0 r: SPReg R: (retpcReg := TempReg)]. self MoveMw: 0 r: SPReg R: retpcReg. self MoveMw: backEnd jumpShortByteSize r: retpcReg R: Arg1Reg. self CmpR: ClassReg R: Arg1Reg. jumpMiss := self JumpNonZero: 0. self MoveMw: backEnd jumpShortByteSize + BytesPerOop r: retpcReg R: ClassReg. self CmpCq: 0 R: ClassReg. jumpItsTheReceiverStupid := self JumpZero: 0. self MoveR: ClassReg R: ReceiverResultReg. jumpItsTheReceiverStupid jmpTarget: (self RetN: 0). jumpMiss jmpTarget: self Label. ceImplicitReceiverTrampoline := self genTrampolineFor: #ceImplicitReceiverFor:receiver: called: 'ceImplicitReceiverTrampoline' numArgs: 2 arg: SendNumArgsReg arg: ReceiverResultReg arg: nil arg: nil saveRegs: false pushLinkReg: true resultReg: ReceiverResultReg appendOpcodes: true]! Item was changed: ----- Method: SpurMemoryManager>>addFreeSubTree: (in category 'free space') ----- addFreeSubTree: freeTree "Add a freeChunk sub tree back into the large free chunk tree. This is for allocateOldSpaceChunkOf[Exactly]Bytes:[suchThat:]." | bytesInArg treeNode bytesInNode subNode | "N.B. *can't* use numSlotsOfAny: because of rounding up of odd slots and/or step in size at 1032 bytes in 32-bits or 2048 bytes in 64-bits." self assert: (self isFreeObject: freeTree). bytesInArg := self bytesInObject: freeTree. self assert: bytesInArg >= (self numFreeLists * self allocationUnit). treeNode := freeLists at: 0. self assert: treeNode ~= 0. [bytesInNode := self bytesInObject: treeNode. + "check for overlap; could write this as self oop: (self objectAfter: freeChunk) isLessThanOrEqualTo: child... + but that relies on headers being correct, etc. So keep it clumsy..." + self assert: ((self oop: freeTree + bytesInArg - self baseHeaderSize isLessThanOrEqualTo: treeNode) + or: [self oop: freeTree isGreaterThanOrEqualTo: treeNode + bytesInNode - self baseHeaderSize]). - self assert: ((self oop: freeTree + bytesInArg isLessThanOrEqualTo: treeNode) - or: [self oop: freeTree isGreaterThanOrEqualTo: treeNode + bytesInNode]). self assert: bytesInNode >= (self numFreeLists * self allocationUnit). self assert: bytesInArg ~= bytesInNode. bytesInNode > bytesInArg ifTrue: [subNode := self fetchPointer: self freeChunkSmallerIndex ofFreeChunk: treeNode. subNode = 0 ifTrue: [self storePointer: self freeChunkSmallerIndex ofFreeChunk: treeNode withValue: freeTree. self storePointer: self freeChunkParentIndex ofFreeChunk: freeTree withValue: treeNode. ^self]] ifFalse: [subNode := self fetchPointer: self freeChunkLargerIndex ofFreeChunk: treeNode. subNode = 0 ifTrue: [self storePointer: self freeChunkLargerIndex ofFreeChunk: treeNode withValue: freeTree. self storePointer: self freeChunkParentIndex ofFreeChunk: freeTree withValue: treeNode. ^self]]. treeNode := subNode] repeat! Item was changed: ----- Method: SpurMemoryManager>>addToFreeTree:bytes: (in category 'free space') ----- addToFreeTree: freeChunk bytes: chunkBytes "Add freeChunk to the large free chunk tree. For the benefit of sortedFreeObject:, answer the treeNode it is added to, if it is added to the next list of a freeTreeNode, otherwise answer 0." | childBytes parent child | self assert: (self isFreeObject: freeChunk). self assert: chunkBytes = (self bytesInObject: freeChunk). self assert: chunkBytes >= (self numFreeLists * self allocationUnit). self storePointer: self freeChunkNextIndex ofFreeChunk: freeChunk withValue: 0; storePointer: self freeChunkParentIndex ofFreeChunk: freeChunk withValue: 0; storePointer: self freeChunkSmallerIndex ofFreeChunk: freeChunk withValue: 0; storePointer: self freeChunkLargerIndex ofFreeChunk: freeChunk withValue: 0. "Large chunk list organized as a tree, each node of which is a list of chunks of the same size. Beneath the node are smaller and larger blocks." parent := 0. child := freeLists at: 0. [child ~= 0] whileTrue: [childBytes := self bytesInObject: child. + "check for overlap; could write this as self oop: (self objectAfter: freeChunk) isLessThanOrEqualTo: child... + but that relies on headers being correct, etc. So keep it clumsy..." + self assert: ((self oop: freeChunk + chunkBytes - self baseHeaderSize isLessThanOrEqualTo: child) + or: [self oop: freeChunk isGreaterThanOrEqualTo: child + childBytes - self baseHeaderSize]). - self assert: ((self oop: freeChunk + chunkBytes isLessThanOrEqualTo: child) - or: [self oop: freeChunk isGreaterThanOrEqualTo: child + childBytes]). childBytes = chunkBytes ifTrue: "size match; add to list at node." [self storePointer: self freeChunkNextIndex ofFreeChunk: freeChunk withValue: (self fetchPointer: self freeChunkNextIndex ofFreeChunk: child); storePointer: self freeChunkNextIndex ofFreeChunk: child withValue: freeChunk. ^child]. "walk down the tree" parent := child. child := self fetchPointer: (childBytes > chunkBytes ifTrue: [self freeChunkSmallerIndex] ifFalse: [self freeChunkLargerIndex]) ofFreeChunk: child]. parent = 0 ifTrue: [self assert: (freeLists at: 0) = 0. freeLists at: 0 put: freeChunk. freeListsMask := freeListsMask bitOr: 1. ^0]. self assert: (freeListsMask anyMask: 1). "insert in tree" self storePointer: self freeChunkParentIndex ofFreeChunk: freeChunk withValue: parent. self storePointer: (childBytes > chunkBytes ifTrue: [self freeChunkSmallerIndex] ifFalse: [self freeChunkLargerIndex]) ofFreeChunk: parent withValue: freeChunk. ^0! Item was changed: ----- Method: StackInterpreter>>commonSendAbsentImplicit (in category 'send bytecodes') ----- commonSendAbsentImplicit "Send a message to the implicit receiver for that message." "Assume: messageSelector and argumentCount have been set, and that the arguments but not the receiver have been pushed onto the stack," "Note: This method is inlined into the interpreter dispatch loop." "160-175 1010 i i i i Send To Absent Implicit Receiver Literal Selector #iiii With 0 Arguments" "240 11110000 i i i i i j j j Send To Absent Implicit Receiver Literal Selector #iiiii (+ Extend A * 32) with jjj (+ Extend B * 8) Arguments" <sharedCodeInCase: #extSendAbsentImplicitBytecode> + | followedReceiver implicitReceiver | + followedReceiver := self internalFollowedReceiver. - | implicitReceiver | implicitReceiver := self + implicitReceiverFor: followedReceiver - implicitReceiverFor: self receiver mixin: (self methodClassOf: method) implementing: messageSelector. self shuffleArgumentsAndStoreAbsentReceiver: implicitReceiver. lkupClassTag := objectMemory fetchClassTagOf: implicitReceiver. self assert: lkupClassTag ~= objectMemory nilObject. self commonSend! Item was changed: ----- Method: StackInterpreter>>findApplicationOfTargetMixin:startingAtBehavior: (in category 'newspeak bytecode support') ----- findApplicationOfTargetMixin: targetMixin startingAtBehavior: aBehavior "This is used to implement the innards of the pushImplicitReceiverBytecode, used for outer sends in NS2/NS3. Find the MixinApplcation of which aBehavior is a subclass that is an application of targetMixin. This is an implementation derived from <ContextPart> findApplicationOfTargetMixin: targetMixin startingAtBehavior: aBehavior " | mixinOrMixinApplication mixin | mixinOrMixinApplication := aBehavior. [mixinOrMixinApplication = objectMemory nilObject or: [mixinOrMixinApplication = targetMixin + or: [(mixin := objectMemory followObjField: MixinIndex ofObject: mixinOrMixinApplication) = targetMixin - or: [(mixin := objectMemory fetchPointer: MixinIndex ofObject: mixinOrMixinApplication) = targetMixin or: [(objectMemory fetchClassOfNonImm: mixin) = targetMixin]]]] whileFalse: + [mixinOrMixinApplication := objectMemory followObjField: SuperclassIndex ofObject: mixinOrMixinApplication]. - [mixinOrMixinApplication := objectMemory fetchPointer: SuperclassIndex ofObject: mixinOrMixinApplication]. ^mixinOrMixinApplication! Item was changed: ----- Method: StackInterpreter>>instVar:ofContext: (in category 'frame access') ----- instVar: offset ofContext: aContext "Fetch an instance variable from a maybe married context. If the context is still married compute the value of the relevant inst var from the spouse frame's state." | spouseFP | <var: #spouseFP type: #'char *'> <inline: true> self assert: offset < MethodIndex. self assert: (objectMemory isContext: aContext). (self isMarriedOrWidowedContext: aContext) ifFalse: [^objectMemory fetchPointer: offset ofObject: aContext]. self writeBackHeadFramePointers. (self isWidowedContext: aContext) ifTrue: [^objectMemory fetchPointer: offset ofObject: aContext]. spouseFP := self frameOfMarriedContext: aContext. offset = SenderIndex ifTrue: [^self ensureCallerContext: spouseFP]. offset = StackPointerIndex ifTrue: [self assert: ReceiverIndex + (self stackPointerIndexForFrame: spouseFP) < (objectMemory lengthOf: aContext). ^objectMemory integerObjectOf: (self stackPointerIndexForFrame: spouseFP)]. offset = InstructionPointerIndex ifTrue: + [^self instructionPointerForFrame: spouseFP currentFP: localFP currentIP: (self oopForPointer: localIP)]. - [^self instructionPointerForFrame: spouseFP currentFP: localFP currentIP: localIP]. self error: 'bad index'. ^0! Item was changed: ----- Method: StackInterpreter>>instructionPointerForFrame:currentFP:currentIP: (in category 'frame access') ----- instructionPointerForFrame: spouseFP currentFP: currentFP currentIP: instrPtr "Answer the bytecode pc object (i.e. SmallInteger) for an active frame. The bytecode pc is derived from the frame's pc. If the frame is the top frame on the current stack the frame pc is whatever the current instruction pointer is. If the frame is the top frame on some other stack the frame pc is the value on top of stack. Otherwise the frame pc is the saved pc of the frame above. Once the frame pc is found it must be mapped to a bytecode pc." + <var: #spouseFP type: #'char *'> + <var: #currentFP type: #'char *'> | theIP thePage theFPAbove | + <var: #thePage type: #'StackPage *'> + <var: #theFPAbove type: #'char *'> spouseFP = currentFP ifTrue: [theIP := self oopForPointer: instrPtr] ifFalse: [thePage := stackPages stackPageFor: spouseFP. theFPAbove := self findFrameAbove: spouseFP inPage: thePage. theIP := theFPAbove == 0 ifTrue: [stackPages longAt: thePage headSP] ifFalse:[self oopForPointer: (self frameCallerSavedIP: theFPAbove)]]. ^self contextInstructionPointer: theIP frame: spouseFP! Item was added: + ----- Method: StackInterpreter>>internalFollowedReceiver (in category 'internal interpreter access') ----- + internalFollowedReceiver + <inline: true> + | rcvr | + rcvr := stackPages longAt: localFP + FoxReceiver. + (objectMemory isOopForwarded: rcvr) ifTrue: + [rcvr := objectMemory followForwarded: rcvr. + stackPages longAt: localFP + FoxReceiver put: rcvr]. + ^rcvr! Item was changed: ----- Method: StackInterpreter>>pushImplicitReceiverBytecode (in category 'stack bytecodes') ----- pushImplicitReceiverBytecode "This bytecode is used to implement outer sends in NS2/NS3. The bytecode takes as an argument the literal offset of a selector. It effectively finds the nearest lexically-enclosing implementation of that selector by searching up the static chain of the receiver, starting at the current method." + | selector followedReceiver | - | selector | selector := self literal: self fetchByte. self fetchNextBytecode. + followedReceiver := self internalFollowedReceiver. self internalPush: (self + implicitReceiverFor: followedReceiver - implicitReceiverFor: self receiver mixin: (self methodClassOf: method) implementing: selector)! Item was changed: ----- Method: StackInterpreterPrimitives>>pathTo:using:followWeak: (in category 'object access primitives') ----- pathTo: goal using: stack followWeak: followWeak "Trace objects and frames from the root, marking visited objects, pushing the current path on stack, until goal is found. If found, unmark, leaving path in stack, and answer 0. Otherwise answer an error: PrimErrBadArgument if stack is not an Array PrimErrBadIndex if search overflows stack PrimErrNotFound if goal cannot be found" | current index next stackSize stackp freeStartAtStart | (objectMemory isArray: stack) ifFalse: [^PrimErrBadArgument]. self assert: objectMemory allObjectsUnmarked. freeStartAtStart := objectMemory freeStart. "check no allocations during search" objectMemory beRootIfOld: stack. "so no store checks are necessary on stack" stackSize := objectMemory lengthOf: stack. objectMemory mark: stack. "no need. the current context is not reachable from the active process (suspendedContext is nil)" "objectMemory mark: self activeProcess." current := objectMemory specialObjectsOop. objectMemory mark: current. index := objectMemory lengthOf: current. stackp := 0. [[(index := index - 1) >= -1] whileTrue: [(stackPages couldBeFramePointer: current) ifTrue: [next := index >= 0 ifTrue: [self field: index ofFrame: (self cCoerceSimple: current to: #'char *')] ifFalse: [objectMemory nilObject]] ifFalse: [index >= 0 ifTrue: [next := (objectMemory isContextNonImm: current) ifTrue: [self fieldOrSenderFP: index ofContext: current] ifFalse: [objectMemory fetchPointer: index ofObject: current]] ifFalse: [next := objectMemory fetchClassOfNonImm: current]]. (stackPages couldBeFramePointer: next) + ifTrue: [self assert: (self isFrame: (self cCoerceSimple: next to: #'char *') + onPage: (stackPages stackPageFor: (self cCoerceSimple: next to: #'char *')))] - ifTrue: [self assert: (self isFrame: next onPage: (stackPages stackPageFor: next))] ifFalse: [self assert: (self checkOkayOop: next)]. next = goal ifTrue: [self assert: freeStartAtStart = objectMemory freeStart. self unmarkAfterPathTo. objectMemory storePointer: stackp ofObject: stack withValue: current. self pruneStack: stack stackp: stackp. ^0]. ((objectMemory isNonIntegerObject: next) and: [(stackPages couldBeFramePointer: next) ifTrue: [(self frameIsMarked: next) not] ifFalse: [(objectMemory isMarked: next) not and: [((objectMemory isPointers: next) or: [objectMemory isCompiledMethod: next]) and: [followWeak or: [(objectMemory isWeakNonImm: next) not]]]]]) ifTrue: [stackp + 2 > stackSize ifTrue: [self assert: freeStartAtStart = objectMemory freeStart. self unmarkAfterPathTo. objectMemory nilFieldsOf: stack. ^PrimErrBadIndex]. "PrimErrNoMemory ?" objectMemory storePointerUnchecked: stackp ofObject: stack withValue: current; storePointerUnchecked: stackp + 1 ofObject: stack withValue: (objectMemory integerObjectOf: index). stackp := stackp + 2. (stackPages couldBeFramePointer: (self cCoerceSimple: next to: #'char *')) ifTrue: [self markFrame: next. index := self fieldsInFrame: (self cCoerceSimple: next to: #'char *')] ifFalse: [objectMemory mark: next. (objectMemory isCompiledMethod: next) ifTrue: [index := (objectMemory literalCountOf: next) + LiteralStart] ifFalse: [index := objectMemory lengthOf: next]]. current := next]]. current = objectMemory specialObjectsOop ifTrue: [self assert: freeStartAtStart = objectMemory freeStart. self unmarkAfterPathTo. objectMemory nilFieldsOf: stack. ^PrimErrNotFound]. index := objectMemory integerValueOf: (objectMemory fetchPointer: stackp - 1 ofObject: stack). current := objectMemory fetchPointer: stackp - 2 ofObject: stack. stackp := stackp - 2] repeat! Item was changed: ----- Method: VMMaker class>>generateAllNewspeakConfigurationsUnderVersionControl (in category 'configurations') ----- generateAllNewspeakConfigurationsUnderVersionControl self generateNewspeakCogVM; + generateNewspeakSpurStackVM; generateNewspeakSpurCogVM!

1 0

[commit][3124] CogVM source as per VMMaker.oscog-eem.922
by commits＠squeakvm.org 04 Nov '14

04 Nov '14

Revision: 3124 Author: eliot Date: 2014-11-04 14:36:20 -0800 (Tue, 04 Nov 2014) Log Message: ----------- CogVM source as per VMMaker.oscog-eem.922 Ancestors: VMMaker.oscog-eem.921 Fix a transcription slip in the new implicitReceiverFor:mixin:implementing:. Use followObjField: in findApplicationOfTargetMixin:startingAtBehavior:. Fix the assert in addFreeSubTree: similarly to that in addToFreeTree:bytes:. Modified Paths: -------------- branches/Cog/nscogsrc/vm/cointerp.c branches/Cog/nscogsrc/vm/cointerp.h branches/Cog/nscogsrc/vm/gcc3x-cointerp.c branches/Cog/nsspursrc/vm/cointerp.c branches/Cog/nsspursrc/vm/cointerp.h branches/Cog/nsspursrc/vm/gcc3x-cointerp.c branches/Cog/nsspurstacksrc/vm/gcc3x-interp.c branches/Cog/nsspurstacksrc/vm/interp.c branches/Cog/sistasrc/vm/cointerp.c branches/Cog/sistasrc/vm/cointerp.h branches/Cog/sistasrc/vm/gcc3x-cointerp.c branches/Cog/spursistasrc/vm/cointerp.c branches/Cog/spursistasrc/vm/cointerp.h branches/Cog/spursistasrc/vm/gcc3x-cointerp.c branches/Cog/spursrc/vm/cointerp.c branches/Cog/spursrc/vm/cointerp.h branches/Cog/spursrc/vm/gcc3x-cointerp.c branches/Cog/spurstacksrc/vm/gcc3x-interp.c branches/Cog/spurstacksrc/vm/interp.c branches/Cog/src/vm/cointerp.c branches/Cog/src/vm/cointerp.h branches/Cog/src/vm/cointerpmt.c branches/Cog/src/vm/cointerpmt.h branches/Cog/src/vm/gcc3x-cointerp.c branches/Cog/src/vm/gcc3x-cointerpmt.c branches/Cog/stacksrc/vm/gcc3x-interp.c branches/Cog/stacksrc/vm/interp.c Property Changed: ---------------- branches/Cog/platforms/Cross/vm/sqSCCSVersion.h Modified: branches/Cog/nscogsrc/vm/cointerp.c =================================================================== --- branches/Cog/nscogsrc/vm/cointerp.c 2014-11-04 05:35:18 UTC (rev 3123) +++ branches/Cog/nscogsrc/vm/cointerp.c 2014-11-04 22:36:20 UTC (rev 3124) @@ -1,9 +1,9 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e from - CoInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e */ -static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 " __DATE__ ; +static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e " __DATE__ ; char *__interpBuildInfo = __buildInfo; @@ -853,7 +853,7 @@ void printWronglySizedContexts(sqInt printContexts); static void restoreHeadersFromtofromandtofrom(sqInt firstIn, sqInt lastIn, sqInt hdrBaseIn, sqInt firstOut, sqInt lastOut, sqInt hdrBaseOut) NoDbgRegParms; static void runLeakCheckerForFullGC(sqInt fullGCFlag) NoDbgRegParms; -static sqInt safeObjectAfter(sqInt oop) NoDbgRegParms; +static usqInt safeObjectAfter(sqInt oop) NoDbgRegParms; static sqInt safePrintStringOf(sqInt oop) NoDbgRegParms; static sqInt shortentoIndexableSize(sqInt obj, sqInt nSlots) NoDbgRegParms; static sqInt sizeBitsOfSafe(sqInt oop) NoDbgRegParms; @@ -2086,7 +2086,7 @@ /* 574 */ (void (*)(void))0, /* 575 */ (void (*)(void))0, 0 }; -const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.921"; +const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.922"; sqInt minBackwardJumpCountForCompile = MinBackwardJumpCountForCompile /* 40 */; volatile int sendTrace; @@ -16789,7 +16789,9 @@ mixinApplication = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); } candidateMixinApplication = mixinApplication; - assert(!((candidateMixinApplication == (nilObject())))); + if (candidateMixinApplication == GIV(nilObj)) { + return methodReceiver; + } dictionary = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << ShiftForWord)); /* begin lookupMethodInDictionary: */ /* begin numSlotsOf: */ @@ -36645,17 +36647,15 @@ sqInt header1; sqInt header2; sqInt header3; - sqInt header4; usqInt lastWord; sqInt newFreeChunk; sqInt newOop; - sqInt next; + usqInt next; sqInt oop; sqInt realHeader; sqInt sz; sqInt sz1; sqInt sz2; - sqInt sz3; sqInt target; usqInt w; @@ -36666,21 +36666,7 @@ /* begin objectAfterWhileForwarding: */ header2 = longAt(oop); if ((header2 & MarkBit) == 0) { - /* begin objectAfter: */ - if (!(asserta(oopisLessThan(oop, GIV(freeStart))))) { - error("no objects after the end of memory"); - } - if (((longAt(oop)) & TypeMask) == HeaderTypeFree) { - sz2 = (longAt(oop)) & AllButTypeMask; - } - else { - /* begin sizeBitsOf: */ - header3 = longAt(oop); - sz2 = ((header3 & TypeMask) == HeaderTypeSizeAndClass - ? (longAt(oop - (BytesPerWord * 2))) & LongSizeMask - : header3 & SizeMask); - } - next = (oop + sz2) + (headerTypeBytes[(longAt(oop + sz2)) & TypeMask]); + next = ((sqInt) (objectAfter(oop))); goto l1; } fwdBlock1 = (header2 & AllButMarkBitAndTypeMask) << 1; @@ -36695,7 +36681,7 @@ else { sz1 = realHeader & SizeMask; } - next = (oop + sz1) + (headerTypeBytes[(longAt(oop + sz1)) & TypeMask]); + next = ((sqInt) ((oop + sz1) + (headerTypeBytes[(longAt(oop + sz1)) & TypeMask]))); l1: /* end objectAfterWhileForwarding: */; if (!(((longAt(oop)) & TypeMask) == HeaderTypeFree)) { @@ -36749,18 +36735,18 @@ } /* begin safeObjectAfter: */ if (((longAt(newFreeChunk)) & TypeMask) == HeaderTypeFree) { - sz3 = (longAt(newFreeChunk)) & AllButTypeMask; + sz2 = (longAt(newFreeChunk)) & AllButTypeMask; } else { /* begin sizeBitsOf: */ - header4 = longAt(newFreeChunk); - sz3 = ((header4 & TypeMask) == HeaderTypeSizeAndClass + header3 = longAt(newFreeChunk); + sz2 = ((header3 & TypeMask) == HeaderTypeSizeAndClass ? (longAt(newFreeChunk - (BytesPerWord * 2))) & LongSizeMask - : header4 & SizeMask); + : header3 & SizeMask); } - next = ((((usqInt)(newFreeChunk + sz3))) >= GIV(freeStart) + next = ((((usqInt)(newFreeChunk + sz2))) >= GIV(freeStart) ? GIV(freeStart) - : (newFreeChunk + sz3) + (headerTypeBytes[(longAt(newFreeChunk + sz3)) & TypeMask])); + : (newFreeChunk + sz2) + (headerTypeBytes[(longAt(newFreeChunk + sz2)) & TypeMask])); assert((next == GIV(freeStart)) || (next == (oopFromChunk(GIV(compEnd))))); if (next == GIV(freeStart)) { @@ -38636,7 +38622,7 @@ given object or free chunk in memory. Return freeStart when enumeration is complete. This is for assertion checking only. */ -static sqInt +static usqInt safeObjectAfter(sqInt oop) { DECL_MAYBE_SQ_GLOBAL_STRUCT sqInt header; Modified: branches/Cog/nscogsrc/vm/cointerp.h =================================================================== --- branches/Cog/nscogsrc/vm/cointerp.h 2014-11-04 05:35:18 UTC (rev 3123) +++ branches/Cog/nscogsrc/vm/cointerp.h 2014-11-04 22:36:20 UTC (rev 3124) @@ -1,5 +1,5 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e */ Modified: branches/Cog/nscogsrc/vm/gcc3x-cointerp.c =================================================================== --- branches/Cog/nscogsrc/vm/gcc3x-cointerp.c 2014-11-04 05:35:18 UTC (rev 3123) +++ branches/Cog/nscogsrc/vm/gcc3x-cointerp.c 2014-11-04 22:36:20 UTC (rev 3124) @@ -2,11 +2,11 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e from - CoInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e */ -static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 " __DATE__ ; +static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e " __DATE__ ; char *__interpBuildInfo = __buildInfo; @@ -856,7 +856,7 @@ void printWronglySizedContexts(sqInt printContexts); static void restoreHeadersFromtofromandtofrom(sqInt firstIn, sqInt lastIn, sqInt hdrBaseIn, sqInt firstOut, sqInt lastOut, sqInt hdrBaseOut) NoDbgRegParms; static void runLeakCheckerForFullGC(sqInt fullGCFlag) NoDbgRegParms; -static sqInt safeObjectAfter(sqInt oop) NoDbgRegParms; +static usqInt safeObjectAfter(sqInt oop) NoDbgRegParms; static sqInt safePrintStringOf(sqInt oop) NoDbgRegParms; static sqInt shortentoIndexableSize(sqInt obj, sqInt nSlots) NoDbgRegParms; static sqInt sizeBitsOfSafe(sqInt oop) NoDbgRegParms; @@ -2089,7 +2089,7 @@ /* 574 */ (void (*)(void))0, /* 575 */ (void (*)(void))0, 0 }; -const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.921"; +const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.922"; sqInt minBackwardJumpCountForCompile = MinBackwardJumpCountForCompile /* 40 */; volatile int sendTrace; @@ -16798,7 +16798,9 @@ mixinApplication = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << ShiftForWord)); } candidateMixinApplication = mixinApplication; - assert(!((candidateMixinApplication == (nilObject())))); + if (candidateMixinApplication == GIV(nilObj)) { + return methodReceiver; + } dictionary = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << ShiftForWord)); /* begin lookupMethodInDictionary: */ /* begin numSlotsOf: */ @@ -36654,17 +36656,15 @@ sqInt header1; sqInt header2; sqInt header3; - sqInt header4; usqInt lastWord; sqInt newFreeChunk; sqInt newOop; - sqInt next; + usqInt next; sqInt oop; sqInt realHeader; sqInt sz; sqInt sz1; sqInt sz2; - sqInt sz3; sqInt target; usqInt w; @@ -36675,21 +36675,7 @@ /* begin objectAfterWhileForwarding: */ header2 = longAt(oop); if ((header2 & MarkBit) == 0) { - /* begin objectAfter: */ - if (!(asserta(oopisLessThan(oop, GIV(freeStart))))) { - error("no objects after the end of memory"); - } - if (((longAt(oop)) & TypeMask) == HeaderTypeFree) { - sz2 = (longAt(oop)) & AllButTypeMask; - } - else { - /* begin sizeBitsOf: */ - header3 = longAt(oop); - sz2 = ((header3 & TypeMask) == HeaderTypeSizeAndClass - ? (longAt(oop - (BytesPerWord * 2))) & LongSizeMask - : header3 & SizeMask); - } - next = (oop + sz2) + (headerTypeBytes[(longAt(oop + sz2)) & TypeMask]); + next = ((sqInt) (objectAfter(oop))); goto l1; } fwdBlock1 = (header2 & AllButMarkBitAndTypeMask) << 1; @@ -36704,7 +36690,7 @@ else { sz1 = realHeader & SizeMask; } - next = (oop + sz1) + (headerTypeBytes[(longAt(oop + sz1)) & TypeMask]); + next = ((sqInt) ((oop + sz1) + (headerTypeBytes[(longAt(oop + sz1)) & TypeMask]))); l1: /* end objectAfterWhileForwarding: */; if (!(((longAt(oop)) & TypeMask) == HeaderTypeFree)) { @@ -36758,18 +36744,18 @@ } /* begin safeObjectAfter: */ if (((longAt(newFreeChunk)) & TypeMask) == HeaderTypeFree) { - sz3 = (longAt(newFreeChunk)) & AllButTypeMask; + sz2 = (longAt(newFreeChunk)) & AllButTypeMask; } else { /* begin sizeBitsOf: */ - header4 = longAt(newFreeChunk); - sz3 = ((header4 & TypeMask) == HeaderTypeSizeAndClass + header3 = longAt(newFreeChunk); + sz2 = ((header3 & TypeMask) == HeaderTypeSizeAndClass ? (longAt(newFreeChunk - (BytesPerWord * 2))) & LongSizeMask - : header4 & SizeMask); + : header3 & SizeMask); } - next = ((((usqInt)(newFreeChunk + sz3))) >= GIV(freeStart) + next = ((((usqInt)(newFreeChunk + sz2))) >= GIV(freeStart) ? GIV(freeStart) - : (newFreeChunk + sz3) + (headerTypeBytes[(longAt(newFreeChunk + sz3)) & TypeMask])); + : (newFreeChunk + sz2) + (headerTypeBytes[(longAt(newFreeChunk + sz2)) & TypeMask])); assert((next == GIV(freeStart)) || (next == (oopFromChunk(GIV(compEnd))))); if (next == GIV(freeStart)) { @@ -38645,7 +38631,7 @@ given object or free chunk in memory. Return freeStart when enumeration is complete. This is for assertion checking only. */ -static sqInt +static usqInt safeObjectAfter(sqInt oop) { DECL_MAYBE_SQ_GLOBAL_STRUCT sqInt header; Modified: branches/Cog/nsspursrc/vm/cointerp.c =================================================================== --- branches/Cog/nsspursrc/vm/cointerp.c 2014-11-04 05:35:18 UTC (rev 3123) +++ branches/Cog/nsspursrc/vm/cointerp.c 2014-11-04 22:36:20 UTC (rev 3124) @@ -1,9 +1,9 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e from - CoInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e */ -static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 " __DATE__ ; +static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e " __DATE__ ; char *__interpBuildInfo = __buildInfo; @@ -972,6 +972,7 @@ static sqInt followForwardedInObjStackatIndex(sqInt objStack, sqInt objStackRootIndex) NoDbgRegParms; sqInt followForwardedObjectFieldstoDepth(sqInt objOop, sqInt depth); sqInt followForwarded(sqInt objOop); +static sqInt followObjFieldofObject(sqInt fieldIndex, sqInt anObject) NoDbgRegParms; sqInt formatMask(void); sqInt formatOfClass(sqInt classPointer); static sqInt formatOf(sqInt objOop) NoDbgRegParms; @@ -2335,7 +2336,7 @@ /*540*/ -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, /*560*/ -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, 0, 0, 0,-1,-1 }; -const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.921"; +const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.922"; sqInt minBackwardJumpCountForCompile = MinBackwardJumpCountForCompile /* 40 */; volatile int sendTrace; @@ -5709,6 +5710,7 @@ sqInt mixinApplication1; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt rcvr; sqInt tagBits; @@ -5753,8 +5755,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop1); + } + mixinApplication1 = objOop1; } mixinApplication = mixinApplication1; /* begin classTagForClass: */ @@ -14615,6 +14623,7 @@ sqInt mixinApplication1; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt oop; sqInt tagBits; @@ -14671,8 +14680,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop1); + } + mixinApplication1 = objOop1; } mixinApplication = mixinApplication1; /* begin classTagForClass: */ @@ -16334,6 +16349,7 @@ CogMethod *newCogMethod; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt tagBits; sqInt top; sqInt top1; @@ -16379,19 +16395,25 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop)); + if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { + objOop = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop); + } + mixinApplication1 = objOop; } mixinApplication = mixinApplication1; assert((lengthOf(mixinApplication)) > (InstanceSpecificationIndex + 1)); /* begin classTagForClass: */ /* begin followField:ofObject: */ - objOop = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); - if (((objOop & 3) == 0) - && (((longAt(objOop)) & 0x3FFFFF) == 8)) { - objOop = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication, objOop); + objOop1 = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); + if (((objOop1 & 3) == 0) + && (((longAt(objOop1)) & 0x3FFFFF) == 8)) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication, objOop1); } - classObj = objOop; + classObj = objOop1; assert(addressCouldBeClassObj(classObj)); /* begin ensureBehaviorHash: */ assert(addressCouldBeClassObj(classObj)); @@ -19175,6 +19197,7 @@ sqInt objOop2; sqInt objOop21; sqInt objOop3; + sqInt objOop31; sqInt objOop4; sqInt objOop5; sqInt tagBits; @@ -19225,18 +19248,26 @@ mixinApplication = aBehavior; while (!((mixinApplication == GIV(nilObj)) || ((mixinApplication == candidateMixin) - || (((mixin = longAt((mixinApplication + BaseHeaderSize) + (MixinIndex << 2)))) == candidateMixin)))) { - mixinApplication = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication))) == candidateMixin)))) { + /* begin followObjField:ofObject: */ + objOop = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop)); + if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { + objOop = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication, objOop); + } + mixinApplication = objOop; } candidateMixinApplication = mixinApplication; - assert(!((candidateMixinApplication == (nilObject())))); + if (candidateMixinApplication == GIV(nilObj)) { + return methodReceiver; + } /* begin followObjField:ofObject: */ - objOop = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << 2)); - assert(isNonImmediate(objOop)); - if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { - objOop = fixFollowedFieldofObjectwithInitialValue(MethodDictionaryIndex, candidateMixinApplication, objOop); + objOop1 = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(MethodDictionaryIndex, candidateMixinApplication, objOop1); } - dictionary = objOop; + dictionary = objOop1; /* begin lookupMethodInDictionary: */ /* begin numSlotsOf: */ flag("endianness"); @@ -19323,12 +19354,12 @@ } methodArray = objOop21; /* begin followField:ofObject: */ - objOop3 = longAt((methodArray + BaseHeaderSize) + ((index - SelectorStart) << 2)); - if (((objOop3 & 3) == 0) - && (((longAt(objOop3)) & 0x3FFFFF) == 8)) { - objOop3 = fixFollowedFieldofObjectwithInitialValue(index - SelectorStart, methodArray, objOop3); + objOop31 = longAt((methodArray + BaseHeaderSize) + ((index - SelectorStart) << 2)); + if (((objOop31 & 3) == 0) + && (((longAt(objOop31)) & 0x3FFFFF) == 8)) { + objOop31 = fixFollowedFieldofObjectwithInitialValue(index - SelectorStart, methodArray, objOop31); } - GIV(newMethod) = ((sqInt) objOop3); + GIV(newMethod) = ((sqInt) objOop31); found = 1; goto l2; } @@ -19348,20 +19379,20 @@ return candidateReceiver; } /* begin followObjField:ofObject: */ - objOop1 = longAt((candidateMixin + BaseHeaderSize) + (EnclosingMixinIndex << 2)); - assert(isNonImmediate(objOop1)); - if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { - objOop1 = fixFollowedFieldofObjectwithInitialValue(EnclosingMixinIndex, candidateMixin, objOop1); + objOop2 = longAt((candidateMixin + BaseHeaderSize) + (EnclosingMixinIndex << 2)); + assert(isNonImmediate(objOop2)); + if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { + objOop2 = fixFollowedFieldofObjectwithInitialValue(EnclosingMixinIndex, candidateMixin, objOop2); } - candidateMixin = objOop1; + candidateMixin = objOop2; if (candidateMixin == GIV(nilObj)) break; /* begin followObjField:ofObject: */ - objOop2 = longAt((candidateMixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << 2)); - assert(isNonImmediate(objOop2)); - if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { - objOop2 = fixFollowedFieldofObjectwithInitialValue(EnclosingObjectIndex, candidateMixinApplication, objOop2); + objOop3 = longAt((candidateMixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << 2)); + assert(isNonImmediate(objOop3)); + if (((longAt(objOop3)) & (0x3FFFFF - 8)) == 0) { + objOop3 = fixFollowedFieldofObjectwithInitialValue(EnclosingObjectIndex, candidateMixinApplication, objOop3); } - candidateReceiver = objOop2; + candidateReceiver = objOop3; } return methodReceiver; } @@ -43518,9 +43549,13 @@ treeNode = GIV(freeLists)[0]; assert(treeNode != 0); while(1) { + + /* check for overlap; could write this as self oop: (self objectAfter: freeChunk) isLessThanOrEqualTo: child... + but that relies on headers being correct, etc. So keep it clumsy... */ + bytesInNode = bytesInObject(treeNode); - assert((oopisLessThanOrEqualTo(freeTree + bytesInArg, treeNode)) - || (oopisGreaterThanOrEqualTo(freeTree, treeNode + bytesInNode))); + assert((oopisLessThanOrEqualTo((freeTree + bytesInArg) - BaseHeaderSize, treeNode)) + || (oopisGreaterThanOrEqualTo(freeTree, (treeNode + bytesInNode) - BaseHeaderSize))); assert(bytesInNode >= ((numFreeLists()) * (allocationUnit()))); assert(bytesInArg != bytesInNode); if (bytesInNode > bytesInArg) { @@ -49218,6 +49253,24 @@ } +/* Make sure the obj at fieldIndex in anObject is not forwarded (follow the + forwarder there-in if so). Answer the (possibly followed) obj at + fieldIndex. */ + +static sqInt +followObjFieldofObject(sqInt fieldIndex, sqInt anObject) +{ + sqInt objOop; + + objOop = longAt((anObject + BaseHeaderSize) + (fieldIndex << 2)); + assert(isNonImmediate(objOop)); + if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { + objOop = fixFollowedFieldofObjectwithInitialValue(fieldIndex, anObject, objOop); + } + return objOop; +} + + /* 0 = 0 sized objects (UndefinedObject True False et al) 1 = non-indexable objects with inst vars (Point et al) 2 = indexable objects with no inst vars (Array et al) @@ -64931,6 +64984,7 @@ sqInt mixinApplication1; sqInt objOop; sqInt objOop1; + sqInt objOop2; sqInt tagBits; sqInt targetMixin; @@ -64972,8 +65026,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == targetMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == targetMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == targetMixin)))) { + /* begin followObjField:ofObject: */ + objOop2 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop2)); + if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { + objOop2 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop2); + } + mixinApplication1 = objOop2; } mixinApplication = mixinApplication1; assert(!((mixinApplication == (nilObject())))); Modified: branches/Cog/nsspursrc/vm/cointerp.h =================================================================== --- branches/Cog/nsspursrc/vm/cointerp.h 2014-11-04 05:35:18 UTC (rev 3123) +++ branches/Cog/nsspursrc/vm/cointerp.h 2014-11-04 22:36:20 UTC (rev 3124) @@ -1,5 +1,5 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e */ Modified: branches/Cog/nsspursrc/vm/gcc3x-cointerp.c =================================================================== --- branches/Cog/nsspursrc/vm/gcc3x-cointerp.c 2014-11-04 05:35:18 UTC (rev 3123) +++ branches/Cog/nsspursrc/vm/gcc3x-cointerp.c 2014-11-04 22:36:20 UTC (rev 3124) @@ -2,11 +2,11 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e from - CoInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e */ -static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 " __DATE__ ; +static char __buildInfo[] = "CoInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e " __DATE__ ; char *__interpBuildInfo = __buildInfo; @@ -975,6 +975,7 @@ static sqInt followForwardedInObjStackatIndex(sqInt objStack, sqInt objStackRootIndex) NoDbgRegParms; sqInt followForwardedObjectFieldstoDepth(sqInt objOop, sqInt depth); sqInt followForwarded(sqInt objOop); +static sqInt followObjFieldofObject(sqInt fieldIndex, sqInt anObject) NoDbgRegParms; sqInt formatMask(void); sqInt formatOfClass(sqInt classPointer); static sqInt formatOf(sqInt objOop) NoDbgRegParms; @@ -2338,7 +2339,7 @@ /*540*/ -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, /*560*/ -1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1, 0, 0, 0,-1,-1 }; -const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.921"; +const char *interpreterVersion = "Newspeak Virtual Machine CoInterpreterPrimitives_VMMaker.oscog-eem.922"; sqInt minBackwardJumpCountForCompile = MinBackwardJumpCountForCompile /* 40 */; volatile int sendTrace; @@ -5718,6 +5719,7 @@ sqInt mixinApplication1; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt rcvr; sqInt tagBits; @@ -5762,8 +5764,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop1); + } + mixinApplication1 = objOop1; } mixinApplication = mixinApplication1; /* begin classTagForClass: */ @@ -14624,6 +14632,7 @@ sqInt mixinApplication1; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt oop; sqInt tagBits; @@ -14680,8 +14689,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop1); + } + mixinApplication1 = objOop1; } mixinApplication = mixinApplication1; /* begin classTagForClass: */ @@ -16343,6 +16358,7 @@ CogMethod *newCogMethod; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt tagBits; sqInt top; sqInt top1; @@ -16388,19 +16404,25 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop)); + if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { + objOop = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop); + } + mixinApplication1 = objOop; } mixinApplication = mixinApplication1; assert((lengthOf(mixinApplication)) > (InstanceSpecificationIndex + 1)); /* begin classTagForClass: */ /* begin followField:ofObject: */ - objOop = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); - if (((objOop & 3) == 0) - && (((longAt(objOop)) & 0x3FFFFF) == 8)) { - objOop = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication, objOop); + objOop1 = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); + if (((objOop1 & 3) == 0) + && (((longAt(objOop1)) & 0x3FFFFF) == 8)) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication, objOop1); } - classObj = objOop; + classObj = objOop1; assert(addressCouldBeClassObj(classObj)); /* begin ensureBehaviorHash: */ assert(addressCouldBeClassObj(classObj)); @@ -19184,6 +19206,7 @@ sqInt objOop2; sqInt objOop21; sqInt objOop3; + sqInt objOop31; sqInt objOop4; sqInt objOop5; sqInt tagBits; @@ -19234,18 +19257,26 @@ mixinApplication = aBehavior; while (!((mixinApplication == GIV(nilObj)) || ((mixinApplication == candidateMixin) - || (((mixin = longAt((mixinApplication + BaseHeaderSize) + (MixinIndex << 2)))) == candidateMixin)))) { - mixinApplication = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication))) == candidateMixin)))) { + /* begin followObjField:ofObject: */ + objOop = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop)); + if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { + objOop = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication, objOop); + } + mixinApplication = objOop; } candidateMixinApplication = mixinApplication; - assert(!((candidateMixinApplication == (nilObject())))); + if (candidateMixinApplication == GIV(nilObj)) { + return methodReceiver; + } /* begin followObjField:ofObject: */ - objOop = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << 2)); - assert(isNonImmediate(objOop)); - if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { - objOop = fixFollowedFieldofObjectwithInitialValue(MethodDictionaryIndex, candidateMixinApplication, objOop); + objOop1 = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(MethodDictionaryIndex, candidateMixinApplication, objOop1); } - dictionary = objOop; + dictionary = objOop1; /* begin lookupMethodInDictionary: */ /* begin numSlotsOf: */ flag("endianness"); @@ -19332,12 +19363,12 @@ } methodArray = objOop21; /* begin followField:ofObject: */ - objOop3 = longAt((methodArray + BaseHeaderSize) + ((index - SelectorStart) << 2)); - if (((objOop3 & 3) == 0) - && (((longAt(objOop3)) & 0x3FFFFF) == 8)) { - objOop3 = fixFollowedFieldofObjectwithInitialValue(index - SelectorStart, methodArray, objOop3); + objOop31 = longAt((methodArray + BaseHeaderSize) + ((index - SelectorStart) << 2)); + if (((objOop31 & 3) == 0) + && (((longAt(objOop31)) & 0x3FFFFF) == 8)) { + objOop31 = fixFollowedFieldofObjectwithInitialValue(index - SelectorStart, methodArray, objOop31); } - GIV(newMethod) = ((sqInt) objOop3); + GIV(newMethod) = ((sqInt) objOop31); found = 1; goto l2; } @@ -19357,20 +19388,20 @@ return candidateReceiver; } /* begin followObjField:ofObject: */ - objOop1 = longAt((candidateMixin + BaseHeaderSize) + (EnclosingMixinIndex << 2)); - assert(isNonImmediate(objOop1)); - if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { - objOop1 = fixFollowedFieldofObjectwithInitialValue(EnclosingMixinIndex, candidateMixin, objOop1); + objOop2 = longAt((candidateMixin + BaseHeaderSize) + (EnclosingMixinIndex << 2)); + assert(isNonImmediate(objOop2)); + if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { + objOop2 = fixFollowedFieldofObjectwithInitialValue(EnclosingMixinIndex, candidateMixin, objOop2); } - candidateMixin = objOop1; + candidateMixin = objOop2; if (candidateMixin == GIV(nilObj)) break; /* begin followObjField:ofObject: */ - objOop2 = longAt((candidateMixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << 2)); - assert(isNonImmediate(objOop2)); - if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { - objOop2 = fixFollowedFieldofObjectwithInitialValue(EnclosingObjectIndex, candidateMixinApplication, objOop2); + objOop3 = longAt((candidateMixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << 2)); + assert(isNonImmediate(objOop3)); + if (((longAt(objOop3)) & (0x3FFFFF - 8)) == 0) { + objOop3 = fixFollowedFieldofObjectwithInitialValue(EnclosingObjectIndex, candidateMixinApplication, objOop3); } - candidateReceiver = objOop2; + candidateReceiver = objOop3; } return methodReceiver; } @@ -43527,9 +43558,13 @@ treeNode = GIV(freeLists)[0]; assert(treeNode != 0); while(1) { + + /* check for overlap; could write this as self oop: (self objectAfter: freeChunk) isLessThanOrEqualTo: child... + but that relies on headers being correct, etc. So keep it clumsy... */ + bytesInNode = bytesInObject(treeNode); - assert((oopisLessThanOrEqualTo(freeTree + bytesInArg, treeNode)) - || (oopisGreaterThanOrEqualTo(freeTree, treeNode + bytesInNode))); + assert((oopisLessThanOrEqualTo((freeTree + bytesInArg) - BaseHeaderSize, treeNode)) + || (oopisGreaterThanOrEqualTo(freeTree, (treeNode + bytesInNode) - BaseHeaderSize))); assert(bytesInNode >= ((numFreeLists()) * (allocationUnit()))); assert(bytesInArg != bytesInNode); if (bytesInNode > bytesInArg) { @@ -49227,6 +49262,24 @@ } +/* Make sure the obj at fieldIndex in anObject is not forwarded (follow the + forwarder there-in if so). Answer the (possibly followed) obj at + fieldIndex. */ + +static sqInt +followObjFieldofObject(sqInt fieldIndex, sqInt anObject) +{ + sqInt objOop; + + objOop = longAt((anObject + BaseHeaderSize) + (fieldIndex << 2)); + assert(isNonImmediate(objOop)); + if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { + objOop = fixFollowedFieldofObjectwithInitialValue(fieldIndex, anObject, objOop); + } + return objOop; +} + + /* 0 = 0 sized objects (UndefinedObject True False et al) 1 = non-indexable objects with inst vars (Point et al) 2 = indexable objects with no inst vars (Array et al) @@ -64940,6 +64993,7 @@ sqInt mixinApplication1; sqInt objOop; sqInt objOop1; + sqInt objOop2; sqInt tagBits; sqInt targetMixin; @@ -64981,8 +65035,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == targetMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == targetMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == targetMixin)))) { + /* begin followObjField:ofObject: */ + objOop2 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop2)); + if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { + objOop2 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop2); + } + mixinApplication1 = objOop2; } mixinApplication = mixinApplication1; assert(!((mixinApplication == (nilObject())))); Modified: branches/Cog/nsspurstacksrc/vm/gcc3x-interp.c =================================================================== --- branches/Cog/nsspurstacksrc/vm/gcc3x-interp.c 2014-11-04 05:35:18 UTC (rev 3123) +++ branches/Cog/nsspurstacksrc/vm/gcc3x-interp.c 2014-11-04 22:36:20 UTC (rev 3124) @@ -2,11 +2,11 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e from - StackInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + StackInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e */ -static char __buildInfo[] = "StackInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 " __DATE__ ; +static char __buildInfo[] = "StackInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e " __DATE__ ; char *__interpBuildInfo = __buildInfo; @@ -722,6 +722,7 @@ static sqInt followForwardedInObjStackatIndex(sqInt objStack, sqInt objStackRootIndex) NoDbgRegParms; sqInt followForwardedObjectFieldstoDepth(sqInt objOop, sqInt depth); sqInt followForwarded(sqInt objOop); +static sqInt followObjFieldofObject(sqInt fieldIndex, sqInt anObject) NoDbgRegParms; sqInt formatMask(void); sqInt formatOfClass(sqInt classPointer); static sqInt formatOf(sqInt objOop) NoDbgRegParms; @@ -2084,7 +2085,7 @@ 0 }; char * breakSelector; sqInt breakSelectorLength = -1; -const char *interpreterVersion = "Newspeak Virtual Machine StackInterpreterPrimitives_VMMaker.oscog-eem.921"; +const char *interpreterVersion = "Newspeak Virtual Machine StackInterpreterPrimitives_VMMaker.oscog-eem.922"; volatile int sendTrace; sqInt suppressHeartbeatFlag; @@ -5306,6 +5307,7 @@ sqInt mixinApplication1; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt rcvr; sqInt tagBits; @@ -5350,8 +5352,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop1); + } + mixinApplication1 = objOop1; } mixinApplication = mixinApplication1; /* begin classTagForClass: */ @@ -13898,6 +13906,7 @@ sqInt mixinApplication1; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt oop; sqInt tagBits; @@ -13954,8 +13963,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop1); + } + mixinApplication1 = objOop1; } mixinApplication = mixinApplication1; /* begin classTagForClass: */ @@ -29681,9 +29696,13 @@ treeNode = GIV(freeLists)[0]; assert(treeNode != 0); while(1) { + + /* check for overlap; could write this as self oop: (self objectAfter: freeChunk) isLessThanOrEqualTo: child... + but that relies on headers being correct, etc. So keep it clumsy... */ + bytesInNode = bytesInObject(treeNode); - assert((oopisLessThanOrEqualTo(freeTree + bytesInArg, treeNode)) - || (oopisGreaterThanOrEqualTo(freeTree, treeNode + bytesInNode))); + assert((oopisLessThanOrEqualTo((freeTree + bytesInArg) - BaseHeaderSize, treeNode)) + || (oopisGreaterThanOrEqualTo(freeTree, (treeNode + bytesInNode) - BaseHeaderSize))); assert(bytesInNode >= ((numFreeLists()) * (allocationUnit()))); assert(bytesInArg != bytesInNode); if (bytesInNode > bytesInArg) { @@ -35404,6 +35423,24 @@ } +/* Make sure the obj at fieldIndex in anObject is not forwarded (follow the + forwarder there-in if so). Answer the (possibly followed) obj at + fieldIndex. */ + +static sqInt +followObjFieldofObject(sqInt fieldIndex, sqInt anObject) +{ + sqInt objOop; + + objOop = longAt((anObject + BaseHeaderSize) + (fieldIndex << 2)); + assert(isNonImmediate(objOop)); + if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { + objOop = fixFollowedFieldofObjectwithInitialValue(fieldIndex, anObject, objOop); + } + return objOop; +} + + /* 0 = 0 sized objects (UndefinedObject True False et al) 1 = non-indexable objects with inst vars (Point et al) 2 = indexable objects with no inst vars (Array et al) @@ -51337,6 +51374,7 @@ sqInt mixinApplication1; sqInt objOop; sqInt objOop1; + sqInt objOop2; sqInt tagBits; sqInt targetMixin; @@ -51378,8 +51416,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == targetMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == targetMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == targetMixin)))) { + /* begin followObjField:ofObject: */ + objOop2 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop2)); + if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { + objOop2 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop2); + } + mixinApplication1 = objOop2; } mixinApplication = mixinApplication1; assert(!((mixinApplication == (nilObject())))); @@ -53580,6 +53624,7 @@ sqInt objOop2; sqInt objOop21; sqInt objOop3; + sqInt objOop31; sqInt objOop4; sqInt objOop5; sqInt tagBits; @@ -53626,18 +53671,26 @@ mixinApplication = aBehavior; while (!((mixinApplication == GIV(nilObj)) || ((mixinApplication == candidateMixin) - || (((mixin = longAt((mixinApplication + BaseHeaderSize) + (MixinIndex << 2)))) == candidateMixin)))) { - mixinApplication = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication))) == candidateMixin)))) { + /* begin followObjField:ofObject: */ + objOop = longAt((mixinApplication + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop)); + if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { + objOop = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication, objOop); + } + mixinApplication = objOop; } candidateMixinApplication = mixinApplication; - assert(!((candidateMixinApplication == (nilObject())))); + if (candidateMixinApplication == GIV(nilObj)) { + return methodReceiver; + } /* begin followObjField:ofObject: */ - objOop = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << 2)); - assert(isNonImmediate(objOop)); - if (((longAt(objOop)) & (0x3FFFFF - 8)) == 0) { - objOop = fixFollowedFieldofObjectwithInitialValue(MethodDictionaryIndex, candidateMixinApplication, objOop); + objOop1 = longAt((candidateMixinApplication + BaseHeaderSize) + (MethodDictionaryIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(MethodDictionaryIndex, candidateMixinApplication, objOop1); } - dictionary = objOop; + dictionary = objOop1; /* begin lookupMethodInDictionary: */ /* begin numSlotsOf: */ flag("endianness"); @@ -53724,12 +53777,12 @@ } methodArray = objOop21; /* begin followField:ofObject: */ - objOop3 = longAt((methodArray + BaseHeaderSize) + ((index - SelectorStart) << 2)); - if (((objOop3 & 3) == 0) - && (((longAt(objOop3)) & 0x3FFFFF) == 8)) { - objOop3 = fixFollowedFieldofObjectwithInitialValue(index - SelectorStart, methodArray, objOop3); + objOop31 = longAt((methodArray + BaseHeaderSize) + ((index - SelectorStart) << 2)); + if (((objOop31 & 3) == 0) + && (((longAt(objOop31)) & 0x3FFFFF) == 8)) { + objOop31 = fixFollowedFieldofObjectwithInitialValue(index - SelectorStart, methodArray, objOop31); } - GIV(newMethod) = ((sqInt) objOop3); + GIV(newMethod) = ((sqInt) objOop31); found = 1; goto l2; } @@ -53749,20 +53802,20 @@ return candidateReceiver; } /* begin followObjField:ofObject: */ - objOop1 = longAt((candidateMixin + BaseHeaderSize) + (EnclosingMixinIndex << 2)); - assert(isNonImmediate(objOop1)); - if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { - objOop1 = fixFollowedFieldofObjectwithInitialValue(EnclosingMixinIndex, candidateMixin, objOop1); + objOop2 = longAt((candidateMixin + BaseHeaderSize) + (EnclosingMixinIndex << 2)); + assert(isNonImmediate(objOop2)); + if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { + objOop2 = fixFollowedFieldofObjectwithInitialValue(EnclosingMixinIndex, candidateMixin, objOop2); } - candidateMixin = objOop1; + candidateMixin = objOop2; if (candidateMixin == GIV(nilObj)) break; /* begin followObjField:ofObject: */ - objOop2 = longAt((candidateMixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << 2)); - assert(isNonImmediate(objOop2)); - if (((longAt(objOop2)) & (0x3FFFFF - 8)) == 0) { - objOop2 = fixFollowedFieldofObjectwithInitialValue(EnclosingObjectIndex, candidateMixinApplication, objOop2); + objOop3 = longAt((candidateMixinApplication + BaseHeaderSize) + (EnclosingObjectIndex << 2)); + assert(isNonImmediate(objOop3)); + if (((longAt(objOop3)) & (0x3FFFFF - 8)) == 0) { + objOop3 = fixFollowedFieldofObjectwithInitialValue(EnclosingObjectIndex, candidateMixinApplication, objOop3); } - candidateReceiver = objOop2; + candidateReceiver = objOop3; } return methodReceiver; } Modified: branches/Cog/nsspurstacksrc/vm/interp.c =================================================================== --- branches/Cog/nsspurstacksrc/vm/interp.c 2014-11-04 05:35:18 UTC (rev 3123) +++ branches/Cog/nsspurstacksrc/vm/interp.c 2014-11-04 22:36:20 UTC (rev 3124) @@ -1,9 +1,9 @@ /* Automatically generated by - CCodeGeneratorGlobalStructure VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + CCodeGeneratorGlobalStructure VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e from - StackInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 + StackInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e */ -static char __buildInfo[] = "StackInterpreter VMMaker.oscog-eem.921 uuid: ab06d6ee-a8c8-47db-80cc-7530c9a4b945 " __DATE__ ; +static char __buildInfo[] = "StackInterpreter VMMaker.oscog-eem.922 uuid: 52f1b50e-005d-4035-ad57-953a3790260e " __DATE__ ; char *__interpBuildInfo = __buildInfo; @@ -719,6 +719,7 @@ static sqInt followForwardedInObjStackatIndex(sqInt objStack, sqInt objStackRootIndex) NoDbgRegParms; sqInt followForwardedObjectFieldstoDepth(sqInt objOop, sqInt depth); sqInt followForwarded(sqInt objOop); +static sqInt followObjFieldofObject(sqInt fieldIndex, sqInt anObject) NoDbgRegParms; sqInt formatMask(void); sqInt formatOfClass(sqInt classPointer); static sqInt formatOf(sqInt objOop) NoDbgRegParms; @@ -2081,7 +2082,7 @@ 0 }; char * breakSelector; sqInt breakSelectorLength = -1; -const char *interpreterVersion = "Newspeak Virtual Machine StackInterpreterPrimitives_VMMaker.oscog-eem.921"; +const char *interpreterVersion = "Newspeak Virtual Machine StackInterpreterPrimitives_VMMaker.oscog-eem.922"; volatile int sendTrace; sqInt suppressHeartbeatFlag; @@ -5297,6 +5298,7 @@ sqInt mixinApplication1; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt rcvr; sqInt tagBits; @@ -5341,8 +5343,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop1)); + if (((longAt(objOop1)) & (0x3FFFFF - 8)) == 0) { + objOop1 = fixFollowedFieldofObjectwithInitialValue(SuperclassIndex, mixinApplication1, objOop1); + } + mixinApplication1 = objOop1; } mixinApplication = mixinApplication1; /* begin classTagForClass: */ @@ -13889,6 +13897,7 @@ sqInt mixinApplication1; sqInt newHash; sqInt objOop; + sqInt objOop1; sqInt oop; sqInt tagBits; @@ -13945,8 +13954,14 @@ mixinApplication1 = aBehavior; while (!((mixinApplication1 == GIV(nilObj)) || ((mixinApplication1 == mClassMixin) - || (((mixin = longAt((mixinApplication1 + BaseHeaderSize) + (MixinIndex << 2)))) == mClassMixin)))) { - mixinApplication1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + || (((mixin = followObjFieldofObject(MixinIndex, mixinApplication1))) == mClassMixin)))) { + /* begin followObjField:ofObject: */ + objOop1 = longAt((mixinApplication1 + BaseHeaderSize) + (SuperclassIndex << 2)); + assert(isNonImmediate(objOop1)); @@ Diff output truncated at 50000 characters. @@

1 0

VM Maker: VMMaker.oscog-eem.922.mcz
by commits＠source.squeak.org 04 Nov '14

04 Nov '14

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-eem.922.mcz ==================== Summary ==================== Name: VMMaker.oscog-eem.922 Author: eem Time: 4 November 2014, 2:29:50.312 pm UUID: 52f1b50e-005d-4035-ad57-953a3790260e Ancestors: VMMaker.oscog-eem.921 Fix a transcription slip in the new implicitReceiverFor:mixin:implementing:. Use followObjField: in findApplicationOfTargetMixin:startingAtBehavior:. Fix the assert in addFreeSubTree: similarly to that in addToFreeTree:bytes:. =============== Diff against VMMaker.oscog-eem.921 =============== Item was changed: ----- Method: SpurMemoryManager>>addFreeSubTree: (in category 'free space') ----- addFreeSubTree: freeTree "Add a freeChunk sub tree back into the large free chunk tree. This is for allocateOldSpaceChunkOf[Exactly]Bytes:[suchThat:]." | bytesInArg treeNode bytesInNode subNode | "N.B. *can't* use numSlotsOfAny: because of rounding up of odd slots and/or step in size at 1032 bytes in 32-bits or 2048 bytes in 64-bits." self assert: (self isFreeObject: freeTree). bytesInArg := self bytesInObject: freeTree. self assert: bytesInArg >= (self numFreeLists * self allocationUnit). treeNode := freeLists at: 0. self assert: treeNode ~= 0. [bytesInNode := self bytesInObject: treeNode. + "check for overlap; could write this as self oop: (self objectAfter: freeChunk) isLessThanOrEqualTo: child... + but that relies on headers being correct, etc. So keep it clumsy..." + self assert: ((self oop: freeTree + bytesInArg - self baseHeaderSize isLessThanOrEqualTo: treeNode) + or: [self oop: freeTree isGreaterThanOrEqualTo: treeNode + bytesInNode - self baseHeaderSize]). - self assert: ((self oop: freeTree + bytesInArg isLessThanOrEqualTo: treeNode) - or: [self oop: freeTree isGreaterThanOrEqualTo: treeNode + bytesInNode]). self assert: bytesInNode >= (self numFreeLists * self allocationUnit). self assert: bytesInArg ~= bytesInNode. bytesInNode > bytesInArg ifTrue: [subNode := self fetchPointer: self freeChunkSmallerIndex ofFreeChunk: treeNode. subNode = 0 ifTrue: [self storePointer: self freeChunkSmallerIndex ofFreeChunk: treeNode withValue: freeTree. self storePointer: self freeChunkParentIndex ofFreeChunk: freeTree withValue: treeNode. ^self]] ifFalse: [subNode := self fetchPointer: self freeChunkLargerIndex ofFreeChunk: treeNode. subNode = 0 ifTrue: [self storePointer: self freeChunkLargerIndex ofFreeChunk: treeNode withValue: freeTree. self storePointer: self freeChunkParentIndex ofFreeChunk: freeTree withValue: treeNode. ^self]]. treeNode := subNode] repeat! Item was changed: ----- Method: StackInterpreter>>findApplicationOfTargetMixin:startingAtBehavior: (in category 'newspeak bytecode support') ----- findApplicationOfTargetMixin: targetMixin startingAtBehavior: aBehavior "This is used to implement implicit receiver and enclosing object lookup + for Newspeak. Find the mixin application of which aClass is a subclass that - for Newspeak. Find the mixin applcation of which aClass is a subclass that is an application of targetMixin. This is an implementation derived from <ContextPart> findApplicationOf: targetMixin startingAt: aBehavior " | mixinApplication mixin | mixinApplication := aBehavior. [mixinApplication = objectMemory nilObject or: [mixinApplication = targetMixin + or: [(mixin := objectMemory followObjField: MixinIndex ofObject: mixinApplication) = targetMixin]]] whileFalse: + [mixinApplication := objectMemory followObjField: SuperclassIndex ofObject: mixinApplication]. - or: [(mixin := objectMemory fetchPointer: MixinIndex ofObject: mixinApplication) = targetMixin]]] whileFalse: - [mixinApplication := objectMemory fetchPointer: SuperclassIndex ofObject: mixinApplication]. ^mixinApplication! Item was changed: ----- Method: StackInterpreter>>implicitReceiverFor:mixin:implementing: (in category 'newspeak bytecode support') ----- implicitReceiverFor: methodReceiver mixin: methodMixin implementing: selector "This is used to implement implicit receiver sends in Newspeak. Find the nearest lexically-enclosing implementation of selector by searching up the static chain of the method's receiver, starting at mixin's application. This implementation is derived from <ContextPart> implicitReceiverFor: methodReceiver <Object> in: methodMixin <Mixin> implementing: selector <Symbol> ^<Object>" <api> <option: #NewspeakVM> | candidateReceiver candidateMixin candidateMixinApplication dictionary found | self deny: (objectMemory isOopForwarded: methodReceiver). self deny: (objectMemory isForwarded: methodMixin). "messageSelector is an implicit parameter of lookupMethodInDictionary:" messageSelector := objectMemory followMaybeForwarded: selector. candidateReceiver := methodReceiver. candidateMixin := methodMixin. [candidateMixinApplication := self + findApplicationOfTargetMixin: candidateMixin + startingAtBehavior: (objectMemory fetchClassOf: candidateReceiver). + candidateMixinApplication = objectMemory nilObject ifTrue: + [^methodReceiver]. - findApplicationOfTargetMixin: candidateMixin - startingAtBehavior: (objectMemory fetchClassOf: candidateReceiver). - self deny: (candidateMixinApplication = objectMemory nilObject). dictionary := objectMemory followObjField: MethodDictionaryIndex ofObject: candidateMixinApplication. found := self lookupMethodInDictionary: dictionary. found ifTrue: [^candidateReceiver]. candidateMixin := objectMemory followObjField: EnclosingMixinIndex ofObject: candidateMixin. candidateMixin = objectMemory nilObject] whileFalse: [candidateReceiver := objectMemory followObjField: EnclosingObjectIndex ofObject: candidateMixinApplication]. ^methodReceiver!

1 0

Re: [Vm-dev] VM Maker: VMMakerJS-bf.14.mcz
by Eliot Miranda 04 Nov '14

04 Nov '14

Hi Bert, instead of all these hacks for cCode:inSmalltalk: why not add cCode:jsCode:inSmalltalk: ? On Mon, Nov 3, 2014 at 3:48 AM, <commits(a)source.squeak.org> wrote: > > Bert Freudenberg uploaded a new version of VMMakerJS to project VM Maker: > http://source.squeak.org/VMMaker/VMMakerJS-bf.14.mcz > > ==================== Summary ==================== > > Name: VMMakerJS-bf.14 > Author: bf > Time: 3 November 2014, 12:38:30.528 pm > UUID: 6a767eb3-c538-442f-91d0-c7d61c865651 > Ancestors: VMMakerJS-bf.13 > > Fix exp() math call for Klatt > > =============== Diff against VMMakerJS-bf.13 =============== > > Item was removed: > - ----- Method: JSCodeGenerator>>generateAbs:on:indent: (in category 'JS > translation') ----- > - generateAbs: msgNode on: aStream indent: level > - "Generate the JS code for the negated message onto the given > stream." > - > - aStream nextPutAll: 'Math.abs('. > - self emitJSExpression: msgNode receiver on: aStream. > - aStream nextPutAll: ')'. > - ! > > Item was removed: > - ----- Method: JSCodeGenerator>>generateCos:on:indent: (in category 'JS > translation') ----- > - generateCos: msgNode on: aStream indent: level > - "Generate the JS code for the message onto the given stream." > - > - aStream nextPutAll: 'Math.cos('. > - self emitJSExpression: msgNode receiver on: aStream. > - aStream nextPutAll: ')'. > - ! > > Item was changed: > ----- Method: JSCodeGenerator>>generateJSCodeForCcode:on:indent: (in > category 'JS translation') ----- > generateJSCodeForCcode: cCode on: aStream indent: level > cCode = '' ifTrue: [^self]. > aStream nextPutAll: (cCodeTranslationDict at: cCode ifAbsent: [ > + (#('sqrt' 'sin' 'cos' 'tan' 'atan' 'atan2' 'log' 'exp' > 'min' 'max' 'floor') anySatisfy: [:fn | fn,'(*)' match: cCode]) > - (#('sqrt' 'sin' 'cos' 'tan' 'atan' 'atan2' 'log' 'min' > 'max' 'floor') anySatisfy: [:fn | fn,'(*)' match: cCode]) > ifTrue: ['Math.', cCode] ifFalse: [ > ('opTable[*] = (void *)*' match: cCode) > ifTrue: [cCode copyReplaceAll: '(void *)' with: > ''] ifFalse: [ > ('fn(*Handle, &*Pitch, *)' match: cCode) > ifTrue: [(cCode copyReplaceAll: '&' with: > 'function(p){') copyReplaceAll: 'Pitch' with: 'Pitch = p}'] ifFalse: [ > (' ((*)querySurfaceFn)*(handle, &destWidth, &destHeight, > &destDepth, &destMSB)' match: cCode) > ifTrue: [ 'querySurfaceFn(handle, function(w, h, > d, m){destWidth = w; destHeight = h; destDepth = d; destMSB = m; })'] > ifFalse: [ > (' ((*)querySurfaceFn)*(handle, &sourceWidth, > &sourceHeight, &sourceDepth, &sourceMSB)' match: cCode) > ifTrue: [ 'querySurfaceFn(handle, function(w, h, > d, m){sourceWidth = w; sourceHeight = h; sourceDepth = d; sourceMSB = m; > })'] ifFalse: [ > (#('fn(destHandle, affectedL, affectedT, > affectedR-affectedL, affectedB-affectedT)', > 'fn(sourceHandle, 0, 0, 0, 0)') includes: cCode) > ifTrue: [cCode] ifFalse: [ > "See initializeCTranslationDictionary" > self error: 'C: ' , cCode]]]]]]]) > ! > > Item was added: > + ----- Method: JSCodeGenerator>>generateMath:on:indent: (in category 'JS > translation') ----- > + generateMath: msgNode on: aStream indent: level > + "Generate the JS code for the message onto the given stream." > + aStream nextPutAll: 'Math.'; nextPutAll: msgNode selector; > nextPut: $(. > + self emitJSExpression: msgNode receiver on: aStream. > + aStream nextPut: $). > + ! > > Item was removed: > - ----- Method: JSCodeGenerator>>generateSin:on:indent: (in category 'JS > translation') ----- > - generateSin: msgNode on: aStream indent: level > - "Generate the JS code for the message onto the given stream." > - > - aStream nextPutAll: 'Math.sin('. > - self emitJSExpression: msgNode receiver on: aStream. > - aStream nextPutAll: ')'. > - ! > > Item was removed: > - ----- Method: JSCodeGenerator>>generateSqrt:on:indent: (in category 'JS > translation') ----- > - generateSqrt: msgNode on: aStream indent: level > - "Generate the JS code for the message onto the given stream." > - > - aStream nextPutAll: 'Math.sqrt('. > - self emitJSExpression: msgNode receiver on: aStream. > - aStream nextPutAll: ')'. > - ! > > Item was changed: > ----- Method: JSCodeGenerator>>initializeCTranslationDictionary (in > category 'JS translation') ----- > initializeCTranslationDictionary > "Initialize the dictionary mapping message names to actions for C > code generation." > > | pairs | > translationDict := Dictionary new: 200. > pairs := #( > #& #generateAnd:on:indent: > #| #generateOr:on:indent: > #and: #generateSequentialAnd:on:indent: > #or: #generateSequentialOr:on:indent: > #not #generateNot:on:indent: > > #+ #generatePlus:on:indent: > #- #generateMinus:on:indent: > #negated #generateNegated:on:indent: > + #abs #generateMath:on:indent: > + #sqrt #generateMath:on:indent: > + #sin #generateMath:on:indent: > + #cos #generateMath:on:indent: > + #tan #generateMath:on:indent: > + #atan #generateMath:on:indent: > + #exp #generateMath:on:indent: > + #log #generateMath:on:indent: > - #abs #generateAbs:on:indent: > - #sqrt #generateSqrt:on:indent: > - #sin #generateSin:on:indent: > - #cos #generateCos:on:indent: > #* #generateTimes:on:indent: > #/ #generateDivide:on:indent: > #// #generateDivide:on:indent: > #\\ #generateModulo:on:indent: > #<< #generateShiftLeft:on:indent: > #>> #generateShiftRight:on:indent: > #min: #generateMin:on:indent: > #max: #generateMax:on:indent: > #between:and: #generateBetweenAnd:on:indent: > > #bitAnd: #generateBitAnd:on:indent: > #bitOr: #generateBitOr:on:indent: > #bitXor: #generateBitXor:on:indent: > #bitShift: #generateBitShift:on:indent: > #signedBitShift: #generateSignedBitShift:on:indent: > #bitInvert32 #generateBitInvert32:on:indent: > #bitClear: #generateBitClear:on:indent: > > #< #generateLessThan:on:indent: > #<= #generateLessThanOrEqual:on:indent: > #= #generateEqual:on:indent: > #> #generateGreaterThan:on:indent: > #>= #generateGreaterThanOrEqual:on:indent: > #~= #generateNotEqual:on:indent: > #== #generateEqual:on:indent: > #~~ #generateNotEqual:on:indent: > #isNil #generateIsNil:on:indent: > #notNil #generateNotNil:on:indent: > > #whileTrue: #generateWhileTrue:on:indent: > #whileFalse: #generateWhileFalse:on:indent: > #whileTrue #generateDoWhileTrue:on:indent: > #whileFalse #generateDoWhileFalse:on:indent: > #to:do: #generateToDo:on:indent: > #to:by:do: #generateToByDo:on:indent: > #repeat #generateRepeat:on:indent: > > #ifTrue: #generateIfTrue:on:indent: > #ifFalse: #generateIfFalse:on:indent: > #ifTrue:ifFalse: #generateIfTrueIfFalse:on:indent: > #ifFalse:ifTrue: #generateIfFalseIfTrue:on:indent: > > #at: #generateAt:on:indent: > #at:put: #generateAtPut:on:indent: > #basicAt: #generateAt:on:indent: > #basicAt:put: #generateAtPut:on:indent: > > #integerValueOf: > #generateIntegerValueOf:on:indent: > #integerObjectOf: > #generateIntegerObjectOf:on:indent: > #isIntegerObject: > #generateIsIntegerObject:on:indent: > #cCode: > #generateInlineCCode:on:indent: > #cCode:inSmalltalk: > #generateInlineCCode:on:indent: > #cPreprocessorDirective: > #generateInlineCPreprocessorDirective:on:indent: > #preprocessorExpression: > #generateInlineCppDirective:on:indent: > #isDefined:inSmalltalk:comment:ifTrue: > #generateInlineCppIfDef:on:indent: > #isDefined:inSmalltalk:comment:ifTrue:ifFalse: > #generateInlineCppIfDefElse:on:indent: > #isDefinedTrueExpression:inSmalltalk:comment:ifTrue:ifFalse: > #generateInlineCppIfElse:on:indent: > #cCoerce:to: > #generateCCoercion:on:indent: > #cCoerceSimple:to: > #generateCCoercion:on:indent: > #addressOf: > #generateAddressOf:on:indent: > #signedIntFromLong > #generateSignedIntFromLong:on:indent: > #signedIntToLong > #generateSignedIntToLong:on:indent: > #signedIntFromShort > #generateSignedIntFromShort:on:indent: > #signedIntToShort > #generateSignedIntToShort:on:indent: > #preIncrement > #generatePreIncrement:on:indent: > #preDecrement > #generatePreDecrement:on:indent: > #inline: > #generateInlineDirective:on:indent: > #asFloat > #generateAsFloat:on:indent: > #asInteger > #generateAsInteger:on:indent: > #asUnsignedInteger > #generateAsUnsignedInteger:on:indent: > #asSymbol > #generateAsSymbol:on:indent: > #anyMask: > #generateBitAnd:on:indent: > #raisedTo: > #generateRaisedTo:on:indent: > #touch: > #generateTouch:on:indent: > #bytesPerWord #generateBytesPerWord:on:indent: > #baseHeaderSize #generateBaseHeaderSize:on:indent: > > #sharedCodeNamed:inCase: > #generateSharedCodeDirective:on:indent: > > #perform: > #generatePerform:on:indent: > #perform:with: > #generatePerform:on:indent: > #perform:with:with: > #generatePerform:on:indent: > #perform:with:with:with: > #generatePerform:on:indent: > #perform:with:with:with:with: #generatePerform:on:indent: > #perform:with:with:with:with:with: #generatePerform:on:indent: > > #shouldNotImplement > #generateSmalltalkMetaError:on:indent: > #shouldBeImplemented > #generateSmalltalkMetaError:on:indent: > > "optimized interpreterProxy calls" > #firstIndexableField: > #generateFirstIndexableField:on:indent: > #slotSizeOf: > #generateSlotSizeOf:on:indent: > #stSizeOf: > #generateSlotSizeOf:on:indent: > #byteSizeOfBytes: > #generateByteSizeOf:on:indent: > #byteSizeOf: > #generateByteSizeOf:on:indent: > #fetchClassOf: > #generateFetchClassOf:on:indent: > #superclassOf: > #generateSuperclassOf:on:indent: > #instanceSizeOf: > #generateInstanceSizeOf:on:indent: > #is:KindOf: > #generateIsKindOf:on:indent: > #isFloatObject: > #generateIsFloatObject:on:indent: > #cDigitCopyFrom:to:len: > #generateCDigitCopy:on:indent: > #sizeOfSTArrayFromCPrimitive: > #generateSizeOfSTArrayFromCPrimitive:on:indent: > #asciiValue > #generateIdentityUnary:on:indent: > #primitiveFail > #generateInterpreterProxyCall:on:indent: > #intAtPointer: > #generateIntAtPointer:on:indent: > #byteAtPointer: > #generateByteAtPointer:on:indent: > #oopForPointer: > #generateOopForPointer:on:indent: > #long32At: > #generateLong32At:on:indent: > #srcLongAt: > #generateSrcLongAt:on:indent: > #dstLongAt: > #generateDstLongAt:on:indent: > #dstLongAt:put: > #generateDstLongAtPut:on:indent: > ). > > 1 to: pairs size by: 2 do: [:i | > translationDict at: (pairs at: i) put: (pairs at: i + 1)]. > > pairs := #( > #ifTrue: > #generateIfTrueAsArgument:on:indent: > #ifFalse: > #generateIfFalseAsArgument:on:indent: > #ifTrue:ifFalse: > #generateIfTrueIfFalseAsArgument:on:indent: > #ifFalse:ifTrue: > #generateIfFalseIfTrueAsArgument:on:indent: > #cCode: #generateInlineCCodeAsArgument:on:indent: > #cCode:inSmalltalk: #generateInlineCCodeAsArgument:on:indent: > ). > > asArgumentTranslationDict := Dictionary new: 8. > 1 to: pairs size by: 2 do: [:i | > asArgumentTranslationDict at: (pairs at: i) put: (pairs > at: i + 1)]. > > cCodeTranslationDict := Dictionary new: 8. > pairs := #( > 'fprintf(stderr, "\n%s: %s", moduleName, s)' > 'console.log(moduleName + ": " + s)' > 'interpreterProxy->majorVersion() == VM_PROXY_MAJOR' > 'interpreterProxy.majorVersion() == VM_PROXY_MAJOR' > 'interpreterProxy->minorVersion() >= VM_PROXY_MINOR' > 'interpreterProxy.minorVersion() >= VM_PROXY_MINOR' > 'rand()' > > 'Math.random()' > '*src++' > 'src[_src++]' > '*dst++ = max' > 'dst[_dst++] = max' > 'src++' > '_src++' > '*dst++ = *src; src += 2' > 'dst[_dst++] = src[_src]; _src += 2' > 'tX' > 'tX' > 'tY' > 'tY' > ). > 1 to: pairs size by: 2 do: [:i | > cCodeTranslationDict at: (pairs at: i) put: (pairs at: i + > 1)]. > ! > > -- best, Eliot

2 1

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev November 2014