Vm-dev May 2016

vm-dev@lists.squeakfoundation.org

36 participants
150 discussions

[commit][3725] Apply the changes of Mr Hachisuka for multibyte character composition.
by commits＠squeakvm.org 23 May '16

23 May '16

Revision: 3725 Author: lewis Date: 2016-05-22 16:29:12 -0700 (Sun, 22 May 2016) Log Message: ----------- Apply the changes of Mr Hachisuka for multibyte character composition. The diff is in the earlier email from Yoshiki: http://lists.squeakfoundation.org/pipermail/squeak-dev/2016-May/189599.html http://www2.asu.ac.jp/hachi/v3/scratch14ime.html Changes are licenced MIT: http://lists.squeakfoundation.org/pipermail/squeak-dev/2016-May/189646.html Adapt the patch to use the -compositioninput option of the X11 display module. Also fix invocation of that option, which was missing the necessary declaration in sqUnixMain.c (hence did not work). With this change, if the -compositioninput options is on the command line, or if the equivalent SQUEAK_COMPOSITIONINPUT environment variable is set, then the recordPendingKeys() function in the X11 desplay module will use the enhancement of Mr. Hachisuka. Otherwise the original recordPendingKeys() logic is used. Modified Paths: -------------- trunk/platforms/unix/vm/sqUnixMain.c trunk/platforms/unix/vm-display-X11/sqUnixX11.c Modified: trunk/platforms/unix/vm/sqUnixMain.c =================================================================== --- trunk/platforms/unix/vm/sqUnixMain.c 2016-05-20 23:36:03 UTC (rev 3724) +++ trunk/platforms/unix/vm/sqUnixMain.c 2016-05-22 23:29:12 UTC (rev 3725) @@ -1051,6 +1051,7 @@ moduleArg("-headless", display, "X11"); moduleArg("-fullscreen", display, "X11"); moduleArg("-fullscreenDirect", display, "X11"); + moduleArg("-compositioninput", display, "X11"); moduleArg("-quartz", display, "Quartz"); moduleArg("-nosound", sound, "null"); Modified: trunk/platforms/unix/vm-display-X11/sqUnixX11.c =================================================================== --- trunk/platforms/unix/vm-display-X11/sqUnixX11.c 2016-05-20 23:36:03 UTC (rev 3724) +++ trunk/platforms/unix/vm-display-X11/sqUnixX11.c 2016-05-22 23:29:12 UTC (rev 3725) @@ -1709,29 +1709,100 @@ static int recordPendingKeys(void) { - if (inputCount > 0) + if (compositionInput) { - int i= iebOut - iebIn; - for (i= (i > 0 ? i : IEB_SIZE + i) / 4; i > 0; -- i) + if (inputCount <= 0) { + if (inputBuf != inputString) { + free(inputBuf); + inputBuf= inputString; + } + return 0; + } + + int utf32= 0; + while (inputCount > 0) { +# if defined(DEBUG_XIM) + fprintf(stderr, "%3d pending key 0x%02x\n", inputCount, *pendingKey); +# endif + /* 110x xxxx 10xx xxxx */ + if (inputCount >= 2 && + pendingKey[0] >= 0xc0 && pendingKey[0] <= 0xdf && + pendingKey[1] >= 0x80 && pendingKey[1] <= 0xbf) + { + utf32= ((pendingKey[0] & 0x1f) << 6) | + (pendingKey[1] & 0x3f); + recordKeyboardEvent(0, EventKeyDown, modifierState, utf32); + recordKeyboardEvent(0, EventKeyChar, modifierState, utf32); + pendingKey += 2; + inputCount -= 2; + } + /* 1110 xxxx 10xx xxxx 10xx xxxx */ + else if (inputCount >= 3 && + pendingKey[0] >= 0xe0 && pendingKey[0] <= 0xef && + pendingKey[1] >= 0x80 && pendingKey[1] <= 0xbf && + pendingKey[2] >= 0x80 && pendingKey[2] <= 0xbf) + { + utf32= ((pendingKey[0] & 0x0f) << 12) | + ((pendingKey[1] & 0x3f) << 6) | + (pendingKey[2] & 0x3f); + recordKeyboardEvent(0, EventKeyDown, modifierState, utf32); + recordKeyboardEvent(0, EventKeyChar, modifierState, utf32); + pendingKey += 3; + inputCount -= 3; + } + /* 1111 0xxx 10xx xxxx 10xx xxxx 10xx xxxx */ + else if (inputCount >= 4 && + pendingKey[0] >= 0xf0 && pendingKey[0] <= 0xf7 && + pendingKey[1] >= 0x80 && pendingKey[1] <= 0xbf && + pendingKey[2] >= 0x80 && pendingKey[2] <= 0xbf && + pendingKey[3] >= 0x80 && pendingKey[3] <= 0xbf) + { + utf32= ((pendingKey[0] & 0x07) << 18) | + ((pendingKey[1] & 0x3f) << 12) | + ((pendingKey[2] & 0x3f) << 6) | + (pendingKey[3] & 0x3f); + recordKeyboardEvent(0, EventKeyDown, modifierState, utf32); + recordKeyboardEvent(0, EventKeyChar, modifierState, utf32); + pendingKey += 4; + inputCount -= 4; + } + else + { + recordKeyboardEvent(*pendingKey, EventKeyDown, modifierState, 0); + recordKeyboardEvent(*pendingKey, EventKeyChar, modifierState, 0); + recordKeystroke(*pendingKey); /* DEPRECATED */ + pendingKey++; + inputCount--; + } + } + return 0; + } + else + { + if (inputCount > 0) { + int i= iebOut - iebIn; + for (i= (i > 0 ? i : IEB_SIZE + i) / 4; i > 0; -- i) + { # if defined(DEBUG_XIM) - fprintf(stderr, "%3d pending key %2d=0x%02x\n", inputCount, i, *pendingKey); + fprintf(stderr, "%3d pending key %2d=0x%02x\n", inputCount, i, *pendingKey); # endif - recordKeyboardEvent(*pendingKey, EventKeyDown, modifierState, 0); - recordKeyboardEvent(*pendingKey, EventKeyChar, modifierState, 0); - recordKeystroke(*pendingKey); /* DEPRECATED */ - ++pendingKey; - if (--inputCount == 0) break; + recordKeyboardEvent(*pendingKey, EventKeyDown, modifierState, 0); + recordKeyboardEvent(*pendingKey, EventKeyChar, modifierState, 0); + recordKeystroke(*pendingKey); /* DEPRECATED */ + ++pendingKey; + if (--inputCount == 0) break; + } + return 1; } - return 1; + /* inputBuf is allocated by lookupKeys */ + if (inputBuf != inputString) + { + free(inputBuf); + inputBuf= inputString; + } + return 0; } - /* inputBuf is allocated by lookupKeys */ - if (inputBuf != inputString) - { - free(inputBuf); - inputBuf= inputString; - } - return 0; } static int xkeysym2ucs4(KeySym keysym);

1 0

REPL image for simulation (was Re: [Vm-dev] handleCallOrJumpSimulationTrap: curiousity)
by Ben Coman 22 May '16

22 May '16

On Sat, May 21, 2016 at 4:44 AM, Clément Bera <bera.clement(a)gmail.com> wrote: > > On Fri, May 20, 2016 at 7:29 PM, Ben Coman <btc(a)openinworld.com> wrote: >> >> On Fri, May 20, 2016 at 8:44 PM, Clément Bera <bera.clement(a)gmail.com> wrote: > > Normally if you build a cog development image: > $ svn co http://www.squeakvm.org/svn/squeak/branches/Cog/image > $ cd ./image > $ ./buildsqueaktrunkvmmakerimage.sh > > You have multiple scripts available with comments to run the simulator that work out of the box. It should take a couple minutes to get it working. Then the easiest is to simulate a REPL image to easily debug what you want. Hi Clement, Could you point me to such a REPL image? cheers -ben

2 1

buildspurtrunkvmmakerimage StackInterpreterSimulator trunk50 incompatible image format
by Ben Coman 22 May '16

22 May '16

Having run buildspurtrunkvmmakerimage.sh and opened the produced SpurVMMaker.image trying to start the simulator like... (StackInterpreterSimulator new openOn: 'trunk50.image') test. produces an "Error: Incompatible image format" However the following works... (StackInterpreterSimulator new openOn: 'Squeak4.5-13680.image') test How do I get the simulator working with a Spur image? VMMaker.oscog-tfel.1864 is loaded. cheers -ben

1 1

VM Maker: Cog-eem.323.mcz
by commits＠source.squeak.org 21 May '16

21 May '16

Eliot Miranda uploaded a new version of Cog to project VM Maker: http://source.squeak.org/VMMaker/Cog-eem.323.mcz ==================== Summary ==================== Name: Cog-eem.323 Author: eem Time: 21 May 2016, 12:25:21.445269 pm UUID: 4e20a484-a807-44c1-a1e8-b78954ba7ec1 Ancestors: Cog-tpr.322 Complete the Spur bootstrap for Cuis 4.2. 3+4=7. Update some older prototypes to reflect the current state of code in Squeak 5. =============== Diff against Cog-tpr.322 =============== Item was changed: ----- Method: CompiledMethod class>>CompiledMethodclassPROTOTYPEnewBytes:trailerBytes:nArgs:nTemps:nStack:nLits:primitive: (in category '*Cog-method prototypes') ----- CompiledMethodclassPROTOTYPEnewBytes: numberOfBytes trailerBytes: trailer nArgs: nArgs nTemps: nTemps nStack: stackSize nLits: nLits primitive: primitiveIndex "Answer an instance of me. The header is specified by the message arguments. The remaining parts are not as yet determined." | method pc | nArgs > 15 ifTrue: [^self error: 'Cannot compile -- too many arguments']. nTemps > 63 ifTrue: [^self error: 'Cannot compile -- too many temporary variables']. + nLits > 32768 ifTrue: - nLits > 65535 ifTrue: [^self error: 'Cannot compile -- too many literals']. method := trailer createMethod: numberOfBytes class: self header: (nArgs bitShift: 24) + (nTemps bitShift: 18) + ((nTemps + stackSize) > SmallFrame ifTrue: [1 bitShift: 17] ifFalse: [0]) + nLits + (primitiveIndex > 0 ifTrue: [1 bitShift: 16] ifFalse: [0]). primitiveIndex > 0 ifTrue: [pc := method initialPC. method at: pc + 0 put: method encoderClass callPrimitiveCode; at: pc + 1 put: (primitiveIndex bitAnd: 16rFF); at: pc + 2 put: (primitiveIndex bitShift: -8)]. ^method! Item was changed: ----- Method: CompiledMethod class>>CompiledMethodclassPROTOTYPEnewBytes:trailerBytes:nArgs:nTemps:nStack:nLits:primitive:flag: (in category '*Cog-method prototypes') ----- CompiledMethodclassPROTOTYPEnewBytes: numberOfBytes trailerBytes: trailer nArgs: nArgs nTemps: nTemps nStack: stackSize nLits: nLits primitive: primitiveIndex flag: flag "Answer an instance of me. The header is specified by the message arguments. The remaining parts are not as yet determined." | method pc | nArgs > 15 ifTrue: [^self error: 'Cannot compile -- too many arguments']. nTemps > 63 ifTrue: [^self error: 'Cannot compile -- too many temporary variables']. + nLits > 32768 ifTrue: - nLits > 65535 ifTrue: [^self error: 'Cannot compile -- too many literals']. method := trailer createMethod: numberOfBytes class: self header: (nArgs bitShift: 24) + (nTemps bitShift: 18) + ((nTemps + stackSize) > SmallFrame ifTrue: [1 bitShift: 17] ifFalse: [0]) + nLits + (primitiveIndex > 0 ifTrue: [1 bitShift: 16] ifFalse: [0]) + (flag ifTrue: [1 bitShift: 29] ifFalse: [0]). primitiveIndex > 0 ifTrue: [pc := method initialPC. method at: pc + 0 put: method encoderClass callPrimitiveCode; at: pc + 1 put: (primitiveIndex bitAnd: 16rFF); at: pc + 2 put: (primitiveIndex bitShift: -8)]. ^method! Item was added: + ----- Method: MethodNode>>MethodNodePROTOTYPEgenerate:using:ifQuick: (in category '*Cog-method prototypes') ----- + MethodNodePROTOTYPEgenerate: trailer using: aCompiledMethodClass ifQuick: methodBlock + | v | + (primitive = 0 and: [arguments size = 0 and: [block isQuick]]) ifFalse: + [^self]. + v := block code. + v < 0 ifTrue: + [^self]. + v = LdSelf ifTrue: + [^methodBlock value: (aCompiledMethodClass toReturnSelfTrailerBytes: trailer)]. + (v between: LdTrue and: LdMinus1 + 3) ifTrue: + [^methodBlock value: (aCompiledMethodClass toReturnConstant: v - LdSelf trailerBytes: trailer)]. + v < ((CodeBases at: LdInstType) + (CodeLimits at: LdInstType)) ifTrue: + [^methodBlock value: (aCompiledMethodClass toReturnField: v trailerBytes: trailer)]. + v // 256 = 1 ifTrue: + [^methodBlock value: (aCompiledMethodClass toReturnField: v \\ 256 trailerBytes: trailer)]! Item was changed: ----- Method: SpurBootstrap>>rehashImage (in category 'bootstrap image') ----- rehashImage "Rehash all collections in newHeap. Find out which classes implement rehash, entering a 1 against their classIndex in rehashFlags. Enumerate all objects, rehashing those whose class has a bit set in rehashFlags." | n sim rehashFlags dotDate rehashSym sizeSym | rehashSym := map at: (self findSymbol: #rehash). sizeSym := map at: (self findSymbol: #size). sim := StackInterpreterSimulator onObjectMemory: newHeap options: #(ObjectMemory #Spur32BitMemoryManager). sim setImageHeaderFlagsFrom: oldInterpreter getImageHeaderFlags; imageName: 'spur image'; assertValidExecutionPointersAtEachStep: false.. newHeap coInterpreter: sim. sim bootstrapping: true. sim initializeInterpreter: 0. sim instVarNamed: 'methodDictLinearSearchLimit' put: SmallInteger maxVal. + (imageTypes includes: 'cuis') ifTrue: + [newHeap scavenger growRememberedSet]. "Rehashing a 4.2 Cuis image overflows the 768 element high tide." sim redirectTranscriptToHost. newHeap setHashBitsOf: newHeap nilObject to: 1; setHashBitsOf: newHeap falseObject to: 2; setHashBitsOf: newHeap trueObject to: 3. rehashFlags := ByteArray new: newHeap numClassTablePages * newHeap classTablePageSize. n := 0. newHeap classTableObjectsDo: [:class| | classIndex | sim messageSelector: rehashSym. "Lookup rehash but don't be fooled by ProtoObject>>rehash, which is just ^self." ((sim lookupOrdinaryNoMNUEtcInClass: class) = 0 and: [(sim isQuickPrimitiveIndex: (sim primitiveIndexOf: (sim instVarNamed: 'newMethod'))) not]) ifTrue: [n := n + 1. classIndex := newHeap rawHashBitsOf: class. rehashFlags at: classIndex >> 3 + 1 put: ((rehashFlags at: classIndex >> 3 + 1) bitOr: (1 << (classIndex bitAnd: 7)))]]. Transcript cr; print: n; nextPutAll: ' classes understand rehash. rehashing instances...'; flush. dotDate := Time now asSeconds. n := 0. self withExecutableInterpreter: sim do: [sim setBreakSelector: 'error:'. "don't rehash twice (actually without limit), so don't rehash any new objects created." newHeap allExistingOldSpaceObjectsDo: [:o| | classIndex | classIndex := newHeap classIndexOf: o. ((rehashFlags at: classIndex >> 3 + 1) anyMask: 1 << (classIndex bitAnd: 7)) ifTrue: [Time now asSeconds > dotDate ifTrue: [Transcript nextPut: $.; flush. dotDate := Time now asSeconds]. "2845 = n ifTrue: [self halt]." "Rehash an object if its size is > 0. Symbol implements rehash, but let's not waste time rehashing it; in Squeak up to 2013 symbols are kept in a set which will get reashed anyway.. Don't rehash empty collections; they may be large for a reason and rehashing will shrink them." ((sim addressCouldBeClassObj: o) or: [(self interpreter: sim object: o perform: sizeSym withArguments: #()) = (newHeap integerObjectOf: 0)]) ifFalse: [self interpreter: sim object: o perform: rehashSym withArguments: #()]]]]! Item was added: + ----- Method: SpurBootstrapCuisPrototypes>>ArrayPROTOTYPEcreateMethod:class:header: (in category 'method prototypes') ----- + ArrayPROTOTYPEcreateMethod: numberOfBytesForAllButTrailer class: aCompiledMethodClass header: headerWord + | meth delta | + meth := aCompiledMethodClass newMethod: numberOfBytesForAllButTrailer + self size header: headerWord. + "copy the trailer data" + delta := meth size - self size. + 1 to: self size do: + [:i | meth at: delta + i put: (self at: i)]. + ^meth! Item was added: + ----- Method: SpurBootstrapCuisPrototypes>>MethodNodePROTOTYPEgenerate:using:ifQuick: (in category 'method prototypes') ----- + MethodNodePROTOTYPEgenerate: trailer using: aCompiledMethodClass ifQuick: methodBlock + <indirect>! Item was changed: ----- Method: SpurBootstrapPrototypes>>ArrayPROTOTYPEelementsForwardIdentityTo: (in category 'method prototypes') ----- ArrayPROTOTYPEelementsForwardIdentityTo: otherArray "This primitive performs a bulk mutation, causing all pointers to the elements of the receiver to be replaced by pointers to the corresponding elements of otherArray. The identityHashes remain with the pointers rather than with the objects so that the objects in this array should still be properly indexed in any existing hashed structures after the mutation." <primitive: 72 error: ec> + ec == #'bad receiver' ifTrue: + [^self error: 'receiver must be of class Array']. + ec == #'bad argument' ifTrue: + [^self error: (otherArray class == Array + ifTrue: ['arg must be of class Array'] + ifFalse: ['receiver and argument must have the same size'])]. + ec == #'inappropriate operation' ifTrue: + [^self error: 'can''t become immediates such as SmallIntegers or Characters']. + ec == #'no modification' ifTrue: + [^self error: 'can''t become immutable objects']. + ec == #'object is pinned' ifTrue: + [^self error: 'can''t become pinned objects']. + ec == #'insufficient object memory' ifTrue: + [Smalltalk garbageCollect < 1048576 ifTrue: + [Smalltalk growMemoryByAtLeast: 1048576]. + ^self elementsForwardIdentityTo: otherArray]. self primitiveFailed! Item was changed: ----- Method: SpurBootstrapPrototypes>>ArrayPROTOTYPEelementsForwardIdentityTo:copyHash: (in category 'method prototypes') ----- ArrayPROTOTYPEelementsForwardIdentityTo: otherArray copyHash: copyHash "This primitive performs a bulk mutation, causing all pointers to the elements of the receiver to be replaced by pointers to the corresponding elements of otherArray. If copyHash is true, the identityHashes remain with the pointers rather than with the objects so that the objects in the receiver should still be properly indexed in any existing hashed structures after the mutation. If copyHash is false, then the hashes of the objects in otherArray remain unchanged. If you know what you're doing this may indeed be what you want." <primitive: 249 error: ec> + ec == #'bad receiver' ifTrue: + [^self error: 'receiver must be of class Array']. + ec == #'bad argument' ifTrue: + [^self error: (otherArray class == Array + ifTrue: ['arg must be of class Array'] + ifFalse: ['receiver and argument must have the same size'])]. + ec == #'inappropriate operation' ifTrue: + [^self error: 'can''t become immediates such as SmallIntegers or Characters']. + ec == #'no modification' ifTrue: + [^self error: 'can''t become immutable objects']. + ec == #'object is pinned' ifTrue: + [^self error: 'can''t become pinned objects']. + ec == #'insufficient object memory' ifTrue: + [Smalltalk garbageCollect < 1048576 ifTrue: + [Smalltalk growMemoryByAtLeast: 1048576]. + ^self elementsForwardIdentityTo: otherArray copyHash: copyHash]. self primitiveFailed! Item was changed: ----- Method: SpurBootstrapSqueakFamilyPrototypes>>InstructionPrinterPROTOTYPEcallPrimitive: (in category 'method prototypes') ----- InstructionPrinterPROTOTYPEcallPrimitive: index + "Print the callPrimitive bytecode." - "Print the callPrimitive." + self print: 'callPrimitive: ' , index printString! - self print: 'callPrimtive: ' , index printString!

1 0

Interpreter versus StackInterpreter hierarchy
by Ben Coman 21 May '16

21 May '16

The hierarchy goes VMClass InterpreterPrimitives StackIntepreter StackIntepreterPrimitives CoInterpreter CoInterpreterPrimitives from which a tick / tock (XXInterpreter / XXInterpreterPrimitives) pattern is apparent, and I wonder why Interpreter is missing above InterpreterPrimitives sitting off to the side. I guess it is something to do with InterpreterPrimitives composing objectMemory as an instance variable rather than inheriting from ObjectMemory as Interpeter does? The reason I ask is that I'm trying a new approach to the OwnedLock primitives where if waitAcquire primitive sleeps, when the process wakes up it retries the primitive. So effectively the process sleeps at the top of the waitAquire rather than the bottom, and the process truly cannot proceed past that point until it gains the lock. I have this working if StackInterpreterPrimitives holds the primitive, but if I move the primitive to InterpreterPrimitives, instructionPointer is unknown. P.S. For the curious, here is the proof of concept I tried. A single call to the primitive counts up to four by backing up the instructionPointer such that the primitive is executed again, until the exit condition of four is reached. (Note the use of Semaphore and ExcessSignalsIndex is not significant, just an expedient template I was familiar with.) # VM SIDE... StackInterpreterPrimitives >> primitiveRetryExperiment | excessSignals stackTop | stackTop := self stackTop. excessSignals := objectMemory fetchInteger: ExcessSignalsIndex ofObject: stackTop. excessSignals := excessSignals + 1. objectMemory storeInteger: ExcessSignalsIndex ofObject: stackTop withValue: excessSignals. [ excessSignals > 3 ] ifFalse: [ instructionPointer := instructionPointer - 1 ]. StackInterpreter class >> initializePrimitiveTable (234 primitiveRetryExperiment) # IMAGE SIDE... Semaphore subclass: #PrimExp instanceVariableNames: '' classVariableNames: '' package: '0PrimitiveRetryExperiment' PrimExp >> initialize excessSignals := 1. PrimExp >> primRetryExperiment <primitive: 234> PrimExp >> excessSignals ^ excessSignals # TEST CASE... PrimExp new primRetryExperiment excessSignals --> 4 I've only done this with the Stack VM so far. I'll report further when I try it with Cog. cheers -ben

3 5

VM Maker: Cog-tpr.322.mcz
by commits＠source.squeak.org 21 May '16

21 May '16

tim Rowledge uploaded a new version of Cog to project VM Maker: http://source.squeak.org/VMMaker/Cog-tpr.322.mcz ==================== Summary ==================== Name: Cog-tpr.322 Author: tpr Time: 21 May 2016, 10:54:50.890085 am UUID: de0414b6-8d89-4319-9243-f7e4c0db3206 Ancestors: Cog-eem.321 Update ARM simulation of FP code; not complete as yet but mostly there. =============== Diff against Cog-eem.321 =============== Item was added: + ----- Method: GdbARMAlien>>fpCPSR (in category 'accessing') ----- + fpCPSR + "The VFP cpsr flags are kept as a single word in the Alien structure. Return just the top 4 bits, the actual flags" + ^(self unsignedLongAt: 1785) >>28! Item was added: + ----- Method: GdbARMAlien>>handleBasicDoubleArithmetic:at: (in category 'floating-point emulation') ----- + handleBasicDoubleArithmetic: instr at: pc + "Emulate a basic math - add/sub/mul/div - VFP instruction." + | rn rd rm vn vm | + rn := instr >> 16 bitAnd: 15. + rd := instr >> 12 bitAnd: 15. + rm := instr bitAnd: 15. + vn := Float fromIEEE64BitWord: (self perform: (self registerStateGetters at: rn + 18)). "Assume accesses fp regs" + vm := Float fromIEEE64BitWord: (self perform: (self registerStateGetters at: rm + 18)). "Assume accesses fp regs" + + "simplest to match the entire instruction pattern rather than mess around shifting and masking and merging" + (instr bitAnd: 16rFF00FF0) + caseOf: { + [16rE200B00 "FMULD"] -> + [| r | + r := vn * vm. + self perform: (self registerStateSetters at: rd + 18) with: r asIEEE64BitWord]. + [16rE300B00 "FADDD"] -> + [| r | + r := vn + vm. + self perform: (self registerStateSetters at: rd + 18) with: r asIEEE64BitWord]. + [16rE300B40 "FSUBD"] -> + [| r | + r := vn - vm. + self perform: (self registerStateSetters at: rd + 18) with: r asIEEE64BitWord]. + [16rE800B00"FDIVD"] -> + [| r | + r := vn / vm. + self perform: (self registerStateSetters at: rd + 18) with: r asIEEE64BitWord].} + otherwise: [self reportPrimitiveFailure]. + self pc: pc + 4! Item was added: + ----- Method: GdbARMAlien>>handleExtendedDoubleArithmetic:at: (in category 'floating-point emulation') ----- + handleExtendedDoubleArithmetic: instr at: pc + "Emulate an extended math - cmp/sqrt/sitod - VFP instruction." + | rn rd rm vn vm vd | + rn := instr >> 16 bitAnd: 15. + rd := instr >> 12 bitAnd: 15. + rm := instr bitAnd: 15. + vn := Float fromIEEE64BitWord: (self perform: (self registerStateGetters at: rn + 18)). "Assume accesses fp regs" + vm := Float fromIEEE64BitWord: (self perform: (self registerStateGetters at: rm + 18)). "Assume accesses fp regs" + + "simplest to match the entire instruction pattern rather than mess around shifting and masking and merging" + (instr bitAnd: 16rFF00FF0) + caseOf: { + [16rEB80B80 "FCMPD"] -> + ["read rd, compare with rm (ignore rn) and set FPSCR NZCV flags. Sigh" + vd := Float fromIEEE64BitWord: (self perform: (self registerStateGetters at: rd + 18)). + self break]. + [16rEB80BC0 "FSITOD"] -> + [| r | + r := vm asFloat. + self perform: (self registerStateSetters at: rd + 18) with: r asIEEE64BitWord]. + [16rEB10BC0 "FSQRTD"] -> + [| r | + r := vm sqrt. + self perform: (self registerStateSetters at: rd + 18) with: r asIEEE64BitWord]. + } + otherwise: [self reportPrimitiveFailure]. + self pc: pc + 4! Item was added: + ----- Method: GdbARMAlien>>handleFPLoadStore:at: (in category 'floating-point emulation') ----- + handleFPLoadStore: instr at: pc + "Emulate a VFP load/store instruction." + | rn rd offset | + rn := instr >> 16 bitAnd: 15. + rd := instr >> 12 bitAnd: 15. + offset := instr bitAnd: 16rFF. + + "examine the U and Lbits" + (instr >>20 bitAnd: 16rF) + caseOf: { + [0"Store neg"] -> + [| r addr| + addr := (self register: rn) - (offset<<2). + r := self perform: (self registerStateGetters at: rd + 18). + self unsignedLongLongAt: addr put: r]. + [1"Load neg"] -> + [| r addr| + addr := (self register: rn) - (offset<<2). + r := self unsignedLongLongAt: addr. + self perform: (self registerStateSetters at: rd + 18) with: r ]. + [8"Store pos"] -> + [| r addr| + addr := (self register: rn) + (offset<<2). + r := self perform: (self registerStateGetters at: rd + 18). + self unsignedLongLongAt: addr put: r]. + [9"Load pos"] -> + [| r addr| + addr := (self register: rn) + (offset<<2). + r := self unsignedLongLongAt: addr. + self perform: (self registerStateSetters at: rd + 18) with: r ].} + otherwise: [self reportPrimitiveFailure]. + self pc: pc + 4! Item was added: + ----- Method: GdbARMAlien>>handleFPStatus:at: (in category 'floating-point emulation') ----- + handleFPStatus: instr at: pc + "Emulate transferring the FP status to the ARM CPSR." + | fpcpsr | + fpcpsr := self fpCPSR. + self vflag: (fpcpsr bitAnd: 1). + self cflag: ((fpcpsr >>1) bitAnd: 1). + self zflag: ((fpcpsr >>2) bitAnd: 1). + self nflag: ((fpcpsr >>3) bitAnd: 1). + self pc: pc + 4! Item was changed: ----- Method: GdbARMAlien>>handleFailingFPArithmetic:at: (in category 'error handling') ----- handleFailingFPArithmetic: instr at: pc + + "basic arithmetic" + (instr bitAnd: 16rF400FB0) = 16rE000B00 ifTrue: + [^self handleBasicDoubleArithmetic: instr at: pc]. + + "extension instructions sqrt/cmp/sitod" + (instr bitAnd: 16rFF00F70) = 16rEB00B40 ifTrue: + [^self handleExtendedDoubleArithmetic: instr at: pc]. + + "move ARM reg to coproc reg. " + (instr bitAnd: 16rFF00FFF) = 16rE000A10 ifTrue: + [^self handleRegToDoubleTransfer: instr at: pc]. + + "move FPSCR reg to ARM CPSR" + (instr bitAnd: 16rFFFFFFF) = 16rEF1FA10 ifTrue: + [^self handleFPStatus: instr at: pc]. + + "load and store ops. All doubles; we only use FLDD & FSTD" + (instr bitAnd: 16rF600F00) = 16rD000B00 ifTrue: + [^self handleFPLoadStore: instr at: pc]. + + - | oneRegTransferMask twoRegTransferMask | - oneRegTransferMask := 16rF000F00. - (instr bitAnd: oneRegTransferMask) = 16rE000B00 ifTrue: - [^self handleOneRegTransferDoubleArithmetic: instr at: pc]. - (instr bitAnd: oneRegTransferMask) = 16rE000A00 ifTrue: - [^self handleOneRegTransferSingleArithmetic: instr at: pc]. - twoRegTransferMask := 16rFE00FC0. - (instr bitAnd: twoRegTransferMask) = 16rC400B00 ifTrue: - [^self handleTwoRegTransferDoubleArithmetic: instr at: pc]. - (instr bitAnd: twoRegTransferMask) = 16rC400A00 ifTrue: - [^self handleTwoRegTransferSingleArithmetic: instr at: pc]. ^self reportPrimitiveFailure! Item was added: + ----- Method: GdbARMAlien>>handleRegToDoubleTransfer:at: (in category 'floating-point emulation') ----- + handleRegToDoubleTransfer: instr at: pc + "Emulate an ARM to VFP instruction." + | rn rd vn | + rn := (instr >> 16 bitAnd: 15) << 1 bitOr: (instr >>6 bitAnd: 1). + rd := instr >> 12 bitAnd: 15. + vn := self register: rn. + + self perform: (self registerStateSetters at: rd + 18) with: vn. + self pc: pc + 4! Item was changed: ----- Method: GdbARMAlien>>instructionIsAnyFPArithmetic: (in category 'testing') ----- instructionIsAnyFPArithmetic: instr + "Identify VFP instructions. + See C3.1 - C3.4 in the ARM ARM v5 DDI01001." + | cp isFP | + + "All FP instructions are coprocessor instructions on coproc 10 or 11" + cp := instr>>8 bitAnd: 16rF. + isFP := cp = 10 or:[cp = 11]. + (isFP and: [(instr>>25 bitAnd: 7) = 6]) ifTrue: [^true]. + (isFP and: [(instr>>24 bitAnd: 16rF) = 16rE]) ifTrue: [^true]. + + "nope" - "Identify single register transfer and double register transfer VFP instructions. - See C3.3 & C3.4 in the ARM ARM." - | oneRegTransferMask twoRegTransferMask | - oneRegTransferMask := 16rF000F00. - (instr bitAnd: oneRegTransferMask) = 16rE000A00 ifTrue: - [^true]. - (instr bitAnd: oneRegTransferMask) = 16rE000B00 ifTrue: - [^true]. - twoRegTransferMask := 16rFE00FC0. - (instr bitAnd: twoRegTransferMask) = 16rC400A00 ifTrue: - [^true]. - (instr bitAnd: twoRegTransferMask) = 16rC400B00 ifTrue: - [^true]. ^false!

1 0

[commit][3724] Add realWindowHandle() macro to conform to image convention of treating window handle 1 as a reference to the main display window .
by commits＠squeakvm.org 21 May '16

21 May '16

Revision: 3724 Author: lewis Date: 2016-05-20 16:36:03 -0700 (Fri, 20 May 2016) Log Message: ----------- Add realWindowHandle() macro to conform to image convention of treating window handle 1 as a reference to the main display window. Note: For X11, handles for HostWindowPlugin are currently *Window, but a handle registry could be added such that the handles are 1, 2, 3, ... as per original design intent. This would eliminate the need for special handling of handle 1. Reformat additions to conform with README.Contributing guidelines (matches update r3723). Modified Paths: -------------- branches/Cog/platforms/unix/vm-display-X11/sqUnixX11.c Modified: branches/Cog/platforms/unix/vm-display-X11/sqUnixX11.c =================================================================== --- branches/Cog/platforms/unix/vm-display-X11/sqUnixX11.c 2016-05-18 01:20:49 UTC (rev 3723) +++ branches/Cog/platforms/unix/vm-display-X11/sqUnixX11.c 2016-05-20 23:36:03 UTC (rev 3724) @@ -7048,14 +7048,18 @@ int affectedL, int affectedR, int affectedT, int affectedB, int windowIndex) { return 0; } -#define isWindowHandle(winIdx) ((winIdx) >= 65536) +/* By convention for HostWindowPlugin, handle 1 refers to the display window */ +#define realWindowHandle(handleFromImage) (handleFromImage == 1 ? stParent : handleFromImage) +/* Window struct addresses are not small integers */ +#define isWindowHandle(winIdx) ((realWindowHandle(winIdx)) >= 65536) + static long display_ioSizeOfNativeWindow(void *windowHandle); static long display_hostWindowGetSize(long windowIndex) { - return isWindowHandle(windowIndex) - ? display_ioSizeOfNativeWindow((void *)windowIndex) - : -1; + return isWindowHandle(windowIndex) + ? display_ioSizeOfNativeWindow((void *)realWindowHandle(windowIndex)) + : -1; } /* ioSizeOfWindowSetxy: args are int windowIndex, int w & h for the @@ -7063,31 +7067,31 @@ * produced in (width<<16 | height) format or -1 for failure as above. */ static long display_hostWindowSetSize(long windowIndex, long w, long h) { - XWindowAttributes attrs; - int real_border_width; + XWindowAttributes attrs; + int real_border_width; - if (!isWindowHandle(windowIndex) - || !XGetWindowAttributes(stDisplay, (Window)windowIndex, &attrs)) - return -1; + if (!isWindowHandle(windowIndex) + || !XGetWindowAttributes(stDisplay, (Window)realWindowHandle(windowIndex), &attrs)) + return -1; - /* At least under Gnome a window's border width in its attributes is zero - * but the relative position of the left-hand edge is the actual border - * width. - */ - real_border_width = attrs.border_width ? attrs.border_width : attrs.x; - return XResizeWindow(stDisplay, (Window)windowIndex, - w - 2 * real_border_width, - h - attrs.y - real_border_width) - ? display_ioSizeOfNativeWindow((void *)windowIndex) - : -1; + /* At least under Gnome a window's border width in its attributes is zero + * but the relative position of the left-hand edge is the actual border + * width. + */ + real_border_width= attrs.border_width ? attrs.border_width : attrs.x; + return XResizeWindow(stDisplay, (Window)realWindowHandle(windowIndex), + w - 2 * real_border_width, + h - attrs.y - real_border_width) + ? display_ioSizeOfNativeWindow((void *)realWindowHandle(windowIndex)) + : -1; } static long display_ioPositionOfNativeWindow(void *windowHandle); static long display_hostWindowGetPosition(long windowIndex) { - return isWindowHandle(windowIndex) - ? display_ioPositionOfNativeWindow((void *)windowIndex) - : -1; + return isWindowHandle(windowIndex) + ? display_ioPositionOfNativeWindow((void *)realWindowHandle(windowIndex)) + : -1; } /* ioPositionOfWindowSetxy: args are int windowIndex, int x & y for the @@ -7095,11 +7099,11 @@ * produced in (left<<16 | top) format or -1 for failure, as above */ static long display_hostWindowSetPosition(long windowIndex, long x, long y) { - if (!isWindowHandle(windowIndex)) - return -1; - return XMoveWindow(stDisplay, (Window)windowIndex, x, y) - ? display_ioPositionOfNativeWindow((void *)windowIndex) - : -1; + if (!isWindowHandle(windowIndex)) + return -1; + return XMoveWindow(stDisplay, (Window)realWindowHandle(windowIndex), x, y) + ? display_ioPositionOfNativeWindow((void *)windowIndex) + : -1; } @@ -7116,6 +7120,38 @@ return 0; } +static long display_ioSizeOfNativeWindow(void *windowHandle) +{ + XWindowAttributes attrs; + int real_border_width; + + if (!XGetWindowAttributes(stDisplay, (Window)windowHandle, &attrs)) + return -1; + + /* At least under Gnome a window's border width in its attributes is zero + * but the relative position of the left-hand edge is the actual border + * width. + */ + real_border_width= attrs.border_width ? attrs.border_width : attrs.x; + return (attrs.width + 2 * real_border_width << 16) + | (attrs.height + attrs.y + real_border_width); +} + +static long display_ioPositionOfNativeWindow(void *windowHandle) +{ + XWindowAttributes attrs; + Window neglected_child; + int rootx, rooty; + + if (!XGetWindowAttributes(stDisplay, (Window)windowHandle, &attrs) + || !XTranslateCoordinates(stDisplay, (Window)windowHandle, attrs.root, + -attrs.border_width, -attrs.border_width, + &rootx, &rooty, &neglected_child)) + return -1; + + return (rootx - attrs.x << 16) | (rooty - attrs.y); +} + #endif /* (SqDisplayVersionMajor >= 1 && SqDisplayVersionMinor >= 2) */ @@ -7252,39 +7288,6 @@ return (attrs.width << 16) | attrs.height; } -static long -display_ioPositionOfNativeWindow(void *windowHandle) -{ - XWindowAttributes attrs; - Window neglected_child; - int rootx, rooty; - - if (!XGetWindowAttributes(stDisplay, (Window)windowHandle, &attrs) - || !XTranslateCoordinates(stDisplay, (Window)windowHandle, attrs.root, - -attrs.border_width, -attrs.border_width, - &rootx, &rooty, &neglected_child)) - return -1; - - return (rootx - attrs.x << 16) | (rooty - attrs.y); -} - -static long -display_ioSizeOfNativeWindow(void *windowHandle) -{ - XWindowAttributes attrs; - int real_border_width; - - if (!XGetWindowAttributes(stDisplay, (Window)windowHandle, &attrs)) - return -1; - - /* At least under Gnome a window's border width in its attributes is zero - * but the relative position of the left-hand edge is the actual border - * width. - */ - real_border_width = attrs.border_width ? attrs.border_width : attrs.x; - return (attrs.width + 2 * real_border_width << 16) - | (attrs.height + attrs.y + real_border_width); -} #endif /* SqDisplayVersionMajor >= 1 && SqDisplayVersionMinor >= 3 */

2 1

handleCallOrJumpSimulationTrap: curiousity
by Ben Coman 20 May '16

20 May '16

Minor curiosity & clueless speculation... in passing I came across the following in #handleCallOrJumpSimulationTrap and given the similarity of critical section and ownedlock primitives to alter the running process like primitiveWait, I wondered if they should also be in this list ... "Verify the stack layout assumption compileInterpreterPrimitive: makes, provided we've not called something that has built a frame, such as closure value or evaluate method, **or switched frames**, such as primitiveSignal, primitiveWait, primitiveResume, primitiveSuspend et al." ifTrue: [(#( primitiveClosureValue primitiveClosureValueWithArgs primitiveClosureValueNoContextSwitch primitiveFullClosureValue primitiveFullClosureValueWithArgs primitiveFullClosureValueNoContextSwitch primitiveSignal primitiveWait primitiveResume primitiveSuspend primitiveYield primitiveExecuteMethodArgsArray primitiveExecuteMethod primitivePerform primitivePerformWithArgs primitivePerformInSuperclass primitiveTerminateTo primitiveStoreStackp primitiveDoPrimitiveWithArgs) cheers -ben

2 3

NULL pointer deref in fetching events?
by Eliot Miranda 20 May '16

20 May '16

Process: Squeak [2023] Path: /Users/USER/*/CocoaFast.app/Contents/MacOS/Squeak Identifier: org.squeak.Squeak$(VM_MONIKER) Version: 5.0.3721 (5.0.3721) Code Type: X86 (Native) Parent Process: launchd [186] Responsible: Squeak [2023] User ID: 594 Date/Time: 2016-05-19 16:37:13.900 -0700 OS Version: Mac OS X 10.9.5 (13F1808) Report Version: 11 Anonymous UUID: F54BABC9-4764-81AE-0375-EA1A9A4A38C2 Crashed Thread: 0 Dispatch queue: com.apple.main-thread Exception Type: EXC_BAD_ACCESS (SIGBUS) Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000000 VM Regions Near 0: --> __PAGEZERO 0000000000000000-0000000000001000 [ 4K] ---/--- SM=NUL /Users/USER/*/CocoaFast.app/Contents/MacOS/Squeak VM_ALLOCATE 0000000000001000-00000000000cf000 [ 824K] ---/--- SM=NUL Thread 0 Crashed:: Dispatch queue: com.apple.main-thread 0 com.apple.CoreFoundation 0x9821a95d CFRunLoopRunSpecific + 253 1 com.apple.CoreFoundation 0x9821a84b CFRunLoopRunInMode + 123 2 com.apple.Foundation 0x91c48559 -[NSRunLoop(NSRunLoop) runMode:beforeDate:] + 277 3 org.squeak.Squeak$(VM_MONIKER) 0x00145f70 -[sqSqueakMainApplication(events) pumpRunLoop] + 99 (sqSqueakMainApplication+events.m:69) 4 org.squeak.Squeak$(VM_MONIKER) 0x0013beb8 -[sqSqueakOSXApplication(events) pumpRunLoop] + 50 (sqSqueakOSXApplication+events.m:127) 5 org.squeak.Squeak$(VM_MONIKER) 0x00143c56 nativeIoProcessEvents + 208 (sqSqueakEventsAPI.m:70) 6 org.squeak.Squeak$(VM_MONIKER) 0x00143caa ioProcessEvents + 35 (sqSqueakEventsAPI.m:84) 7 org.squeak.Squeak$(VM_MONIKER) 0x000dc82a checkForEventsMayContextSwitch + 884 (gcc3x-cointerp.c:61595) 8 org.squeak.Squeak$(VM_MONIKER) 0x000dbaec handleStackOverflowOrEventAllowContextSwitch + 33 (gcc3x-cointerp.c:64780) 9 org.squeak.Squeak$(VM_MONIKER) 0x000dfb2e ceStackOverflow + 108 (gcc3x-cointerp.c:11701) 10 ??? 0x055262c6 0 + 89285318 _,,,^..^,,,_ best, Eliot

2 1

Pointers to why PGP signing, what is git, version management best practices & other thoughts
by John McIntosh 20 May '16

20 May '16

Signing https://github.com/blog/2144-gpg-signature-verification https://git-scm.com/book/en/v2/Git-Tools-Signing-Your-Work https://mikegerwitz.com/papers/git-horror-story http://stackoverflow.com/questions/10161198/is-there-a-way-to-autosign-comm… I think we'll have to sort out if one signs the commits or the tags. ------------- Some best practices for using GIT. https://www.atlassian.com/git/tutorials/what-is-git http://nvie.com/posts/a-successful-git-branching-model Also cmd line is fine, but for the rest of us consider a GUI, such as https://www.atlassian.com/software/sourcetree -- =========================================================================== John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk ===========================================================================

1 0

← Newer
1
...
4
5
6
7
8
9
10
...
15
Older →

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 May 2016