Vm-dev December 2016

vm-dev@lists.squeakfoundation.org

34 participants
102 discussions

RSqueak/VM 0.2 release
by Tim Felgentreff 24 Dec '16

24 Dec '16

Hi, just in time for Christmas, we're releasing RSqueak/VM version 0.2. This is another alpha-quality release for the brave. Bundles with prepared images can be found at https://www.hpi.uni-potsdam.de/swa/artefacts/rsqueak/bundle/RSqueak.tar.gz <https://www.hpi.uni-potsdam.de/hirschfeld/artefacts/rsqueak/bundle/RSqueak.…> or https://www.hpi.uni-potsdam.de/swa/artefacts/rsqueak/bundle/RSqueak.zip <https://www.hpi.uni-potsdam.de/hirschfeld/artefacts/rsqueak/bundle/RSqueak.…>. These start with a Help browser to explain a bit about RSqueak/VM and some of the features we have. This release includes many bug fixes, performance improvements, and new features. - We are now closer to using RSqueak/VM for development, but our main objective is still to make code fast that we think is likely to be run in production, rather than at development time. We run benchmarks every day on http://speed.squeak.org/. - We have added and are currently working on optional, non-standard plugins that enable tail-call optimization, JIT profiling, and integration of a Ruby and a SQLite runtime directly with the VM and its JIT (so code across the language boundaries is also JITed). - We have worked to improve RSqueak/VM's test coverage with unit tests from previously ~50% to over 80%. This helps us give students a more reliable base to work against in our teaching with RSqueak/VM and is also meant to make it easier for other new contributors to approach the code base (over 30% of the code is now in tests). - Since the first alpha release, we have fixed many bugs and crashes, and are now able to run all but two tests in the default Squeak image. Of the two tests that do not work, one segfaults (TestValueWithinFix) and one hangs (SUnit>>testTimeoutLoop). Of the remaining tests (~4500) about 300 produce errors (many due to incomplete features, such as Sockets) and about 200 fail (some of those failures assert things that we believe should not be part of the language definition, but that's another story). As before, RSqueak/VM is still being tested on a regular basis to work with images starting with Squeak 2 (the mini.image) up to the most recent Spur images. We still do not include BitBlt or Balloon as plugins, so we rely on the Slang code or some other fallback code for those plugins to be loaded into the image. As of now, running those plugins in Smalltalk rather than compiling them as a plugin can be between 5-100x slower compared to Cog, with 32-bit operations being on the faster end of that spectrum and blitting between different color depths (some bitmapped fonts are 16-bit) being on the slower end. All RSqueak/VMs by default work with 32-bit images, so even the 64-bit VMs for Linux and macOS run off a 32-bit image. Image saving for older image versions is still in its early stages, so make sure to have backups of any image you test. The bundle only includes 64-bit Linux, 64-bit macOS, and 32-bit Windows VMs. 32-bit x86 and ARM VMs for Linux are available separately from our Github page: https://github.com/HPI-SWA-Lab/RSqueak Season's Greetings and a Happy New Year from the people at HPI's Software Architecture Group

1 0

[commit][3753] Do not use -O3 optimization, -O2 is safer and works well.
by commits＠squeakvm.org 24 Dec '16

24 Dec '16

Revision: 3753 Author: lewis Date: 2016-12-20 17:03:04 -0800 (Tue, 20 Dec 2016) Log Message: ----------- Do not use -O3 optimization, -O2 is safer and works well. gcc 4.9.2 gives crashes and heisenbugs with OSPP compiled with -O3. Symptoms: (1 to: 10) collect: [ :e | (CommandShell new pipeline: 'ps > /dev/null | cat') output] ==> segfaults in some of the spawned child processes This appears to be an actual compiler bug, although it goes away when print statements are added, so I cannot say for sure (and it is difficult to attach gdb to the newly spawned child process before it crashes). The bad behavior happens only with -O3 and there are no real performance benefits compared to -O2 (bytecodes faster, sends slower). Modified Paths: -------------- trunk/platforms/unix/cmake/Makefile.example Modified: trunk/platforms/unix/cmake/Makefile.example =================================================================== --- trunk/platforms/unix/cmake/Makefile.example 2016-12-11 16:49:35 UTC (rev 3752) +++ trunk/platforms/unix/cmake/Makefile.example 2016-12-21 01:03:04 UTC (rev 3753) @@ -9,7 +9,7 @@ # Assume platforms is ../platforms and src is ../src # CFLAGS setting to pass to cmake configure. If undefined, use compiler defaults. -CFLAGS_PARAM="--CFLAGS='-O3 -D_FILE_OFFSET_BITS=64'" +CFLAGS_PARAM="--CFLAGS='-O2 -D_FILE_OFFSET_BITS=64'" #CFLAGS_PARAM="--CFLAGS='-O0 -g'" squeakvm: build/squeakvm build64/squeakvm64

6 11

VM Maker: VMMaker-dtl.389.mcz
by commits＠source.squeak.org 22 Dec '16

22 Dec '16

David T. Lewis uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker-dtl.389.mcz ==================== Summary ==================== Name: VMMaker-dtl.389 Author: dtl Time: 21 December 2016, 6:31:30.057 pm UUID: 3cce2df0-2d06-46f6-bed2-8946fb48c878 Ancestors: VMMaker-dtl.388 VMMaker 4.15.11 Fix the interpreter for running images of the Squeak 3.6 and 3.8 era. Several obsolete primitives were removed in VMMaker-dtl.387, but these are still required for older images. For this update, the primitive table initialiation is unchanged, but a run time update is done for the ContextInterpreter to restore the old primitive assignments. It is not clear if this will be a good strategy but for now add InterpreterPrimitives>>installPrimitive:at: and use it to update the table on first entry to the interpret() loop. Possible future updates could support run time fixup to the primitive table for Squeak 3.2 and earlier images. =============== Diff against VMMaker-dtl.388 =============== Item was changed: ----- Method: ContextInterpreter>>interpret (in category 'interpreter shell') ----- interpret "This is the main interpreter loop. It normally loops forever, fetching and executing bytecodes. When running in the context of a browser plugin VM, however, it must return control to the browser periodically. This should done only when the state of the currently running Squeak thread is safely stored in the object heap. Since this is the case at the moment that a check for interrupts is performed, that is when we return to the browser if it is time to do so. Interrupt checks happen quite frequently." <inline: false> "should not be inlined into any senders" + "Update the primitive table as required for the current running image." + self updatePrimitiveTable.. "record entry time when running as a browser plug-in" self browserPluginInitialiseIfNeeded. self initializeImageFormatVersionIfNeeded. self internalizeIPandSP. self fetchNextBytecode. [true] whileTrue: [self dispatchOn: currentBytecode in: BytecodeTable]. localIP := localIP - 1. "undo the pre-increment of IP before returning" self externalizeIPandSP. ! Item was removed: - ----- Method: ContextInterpreter>>primIndex (in category 'primitive support') ----- - primIndex - ^ primitiveIndex! Item was added: + ----- Method: ContextInterpreter>>primitiveValueUninterruptably (in category 'control primitives') ----- + primitiveValueUninterruptably + "The only purpose of this primitive is to indicate that the new EH mechanisms are supported." + <inline: false> + ^self primitiveValue! Item was added: + ----- Method: ContextInterpreter>>primitiveValueWithArgs (in category 'control primitives') ----- + primitiveValueWithArgs + | argumentArray blockContext blockArgumentCount arrayArgumentCount initialIP | + argumentArray := self popStack. + blockContext := self popStack. + blockArgumentCount := self argumentCountOfBlock: blockContext. + "If the argArray isnt actually an Array we ahve to unpop the above two" + (objectMemory isArray: argumentArray) ifFalse: [self unPop:2. ^self primitiveFail]. + + self successful ifTrue: [arrayArgumentCount := objectMemory fetchWordLengthOf: argumentArray. + self success: (arrayArgumentCount = blockArgumentCount + and: [(objectMemory fetchPointer: CallerIndex ofObject: blockContext) = objectMemory getNilObj])]. + self successful + ifTrue: [self + transfer: arrayArgumentCount + fromIndex: 0 + ofObject: argumentArray + toIndex: TempFrameStart + ofObject: blockContext. + "Assume: The call to transfer:... makes blockContext a root if necessary, + allowing use to use unchecked stored in the following code. " + initialIP := objectMemory fetchPointer: InitialIPIndex ofObject: blockContext. + objectMemory + storePointerUnchecked: InstructionPointerIndex + ofObject: blockContext + withValue: initialIP. + self storeStackPointerValue: arrayArgumentCount inContext: blockContext. + objectMemory + storePointerUnchecked: CallerIndex + ofObject: blockContext + withValue: activeContext. + self newActiveContext: blockContext] + ifFalse: [self unPop: 2]! Item was changed: ----- Method: Interpreter class>>initializePrimitiveTable (in category 'initialization') ----- (excessive size, no diff calculated) Item was changed: VMClass subclass: #InterpreterPrimitives + instanceVariableNames: 'objectMemory primitiveTable primitiveTableUpdated primFailCode argumentCount interruptKeycode newMethod preemptionYields' - instanceVariableNames: 'objectMemory primitiveTable primFailCode argumentCount interruptKeycode newMethod preemptionYields' classVariableNames: 'CrossedX EndOfRun InterpreterSourceVersion MillisecondClockMask PrimitiveExternalCallIndex PrimitiveTable' poolDictionaries: 'VMObjectIndices VMSqueakV3ObjectRepresentationConstants' category: 'VMMaker-Interpreter'! !InterpreterPrimitives commentStamp: 'dtl 4/14/2013 23:16' prior: 0! InterpreterPrimitives implements most of the VM's core primitives. It is the root of the interpreter hierarchy so as to share the core primitives amongst the various interpreters. Instance Variables argumentCount: <Integer> messageSelector: <Integer> newMethod: <Integer> nextProfileTick: <Integer> objectMemory: <ObjectMemory> (simulation only) preemptionYields: <Boolean> primFailCode: <Integer> profileMethod: <Integer> profileProcess: <Integer> profileSemaphore: <Integer> argumentCount - the number of arguments of the current message messageSelector - the oop of the selector of the current message newMethod - the oop of the result of looking up the current message nextProfileTick - the millisecond clock value of the next profile tick (if profiling is in effect) objectMemory - the memory manager and garbage collector that manages the heap preemptionYields - a boolean controlling the process primitives. If true (old, incorrect, blue-book semantics) a preempted process is sent to the back of its run-queue. If false, a process preempted by a higher-priority process is put back at the head of its run queue, hence preserving cooperative scheduling within priorities. primFailCode - primtiive success/failure flag, 0 for success, otherwise the reason code for failure profileMethod - the oop of the method at the time nextProfileTick was reached profileProcess - the oop of the activeProcess at the time nextProfileTick was reached profileSemaphore - the oop of the semaphore to signal when nextProfileTick is reached ! Item was added: + ----- Method: InterpreterPrimitives>>installPrimitive:at: (in category 'primitive table') ----- + installPrimitive: selector at: anInteger + "The primitive table is normally initialized at compile time. This allows an entry + in the table to be istalled at runtime." + + <inline: false> "for readability" + primitiveTable at: anInteger put: selector. + ! Item was added: + ----- Method: InterpreterPrimitives>>updatePrimitiveTable (in category 'primitive table') ----- + updatePrimitiveTable + "Make any run time updates to the primitive table that may be required for supporting + the current running image. May be called by an interpreter that determines at run time + that it neeces to provide support for an older image." + + "As of VMMaker-dtl.387 8 November 2016 (VMMaker 4.15.9), these primitives were + removed from the primitive table for consistency with recent Cog/Spur. They are still + required for running many older Squeak images. Install them here to support those + images." + + "TODO Need a general way to set up the table at runtime for a range of images" + + <inline: false> + + primitiveTableUpdated = true ifFalse: [ + primitiveTableUpdated := true. + self installPrimitive: #primitiveBlockCopy at: 80. + self installPrimitive: #primitiveValue at: 81. + self installPrimitive: #primitiveValueWithArgs at: 82. + self installPrimitive: #primitiveValueUninterruptably at: 123. + ] + ! Item was changed: ----- Method: VMMaker class>>versionString (in category 'version testing') ----- versionString "VMMaker versionString" + ^'4.15.11'! - ^'4.15.10'!

1 0

VM Maker: VMMaker.oscog-eem.2047.mcz
by commits＠source.squeak.org 21 Dec '16

21 Dec '16

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-eem.2047.mcz ==================== Summary ==================== Name: VMMaker.oscog-eem.2047 Author: eem Time: 21 December 2016, 11:04:16.555177 am UUID: d0d330ab-fb50-43c8-9d48-075414e51a42 Ancestors: VMMaker.oscog-cb.2046 Refactor ensureMethodIsCogged: & ensureBlockIsCogged:numCopied: into ensureMethodIsCogged:maybeClosure: to reduce duplication. At the same time streamline ensureMethodIsCogged:maybeClosure: to avoid multiple calls on the Cogit. Fux tow speeling rorres. =============== Diff against VMMaker.oscog-cb.2046 =============== Item was removed: - ----- Method: CoInterpreter>>ensureBlockIsCogged:numCopied: (in category 'frame access') ----- - ensureBlockIsCogged: methodObj numCopied: numCopied - <returnTypeC: #'CogMethod *'> - | rawHeader cogMethod | - <inline: true> - <var: #cogMethod type: #'CogMethod *'> - rawHeader := self rawHeaderOf: methodObj. - (self isCogMethodReference: rawHeader) ifTrue: - [^self cCoerceSimple: rawHeader to: #'CogMethod *']. - cogMethod := cogit cogFullBlockMethod: methodObj numCopied: numCopied. - (cogMethod = nil - and: [cogCompiledCodeCompactionCalledFor]) ifTrue: - [self commenceCogCompiledCodeCompaction. - cogMethod := cogMethod := cogit cogFullBlockMethod: methodObj numCopied: numCopied]. - (self asserta: cogMethod ~= nil) ifFalse: - [self error: 'could not compile method that should have been compiled']. - ^cogMethod! Item was removed: - ----- Method: CoInterpreter>>ensureMethodIsCogged: (in category 'frame access') ----- - ensureMethodIsCogged: methodObj - <returnTypeC: #'CogMethod *'> - | rawHeader cogMethod | - <inline: true> - <var: #cogMethod type: #'CogMethod *'> - rawHeader := self rawHeaderOf: methodObj. - (self isCogMethodReference: rawHeader) ifTrue: - [^self cCoerceSimple: rawHeader to: #'CogMethod *']. - cogMethod := cogit cog: methodObj selector: objectMemory nilObject. - (cogMethod = nil - and: [cogCompiledCodeCompactionCalledFor]) ifTrue: - [self commenceCogCompiledCodeCompaction. - cogMethod := cogit cog: methodObj selector: objectMemory nilObject]. - (self asserta: cogMethod ~= nil) ifFalse: - [self error: 'could not compile method that should have been compiled']. - ^cogMethod! Item was added: + ----- Method: CoInterpreter>>ensureMethodIsCogged:maybeClosure: (in category 'frame access') ----- + ensureMethodIsCogged: methodObj maybeClosure: maybeClosure + "Ensure that methodObj has been cogged. It may be a FullBlockMethod if maybeClosure is a FullBlockClosure." + <returnTypeC: #'CogMethod *'> + | rawHeader cogMethod yetToCompact | + <inline: true> + <var: #cogMethod type: #'CogMethod *'> + rawHeader := self rawHeaderOf: methodObj. + (self isCogMethodReference: rawHeader) ifTrue: + [^self cCoerceSimple: rawHeader to: #'CogMethod *']. + yetToCompact := true. + [(maybeClosure ~= objectMemory nilObject and: [(self isVanillaBlockClosure: maybeClosure) not]) + ifTrue: [cogMethod := cogit cogFullBlockMethod: methodObj numCopied: (self copiedValueCountOfFullClosure: maybeClosure)] + ifFalse: [cogMethod := cogit cog: methodObj selector: objectMemory nilObject]. + cogMethod == nil + and: [cogCompiledCodeCompactionCalledFor + and: [yetToCompact]]] whileTrue: + [yetToCompact := false. + self commenceCogCompiledCodeCompaction]. + (self asserta: cogMethod ~~ nil) ifFalse: + [self error: 'could not compile method that should have been compiled']. + ^cogMethod! Item was changed: ----- Method: CoInterpreter>>makeBaseFrameFor: (in category 'frame access') ----- makeBaseFrameFor: aContext "<Integer>" "Marry aContext with the base frame of a new stack page. Build the base frame to reflect the context's state. Answer the new page. Override to hold the caller context in a different place, In the StackInterpreter we use the caller saved ip, but in the Cog VM caller saved ip is the ceBaseReturn: trampoline. Simply hold the caller context in the first word of the stack." <returnTypeC: #'StackPage *'> | page pointer theMethod theIP numArgs stackPtrIndex maybeClosure rcvr | <inline: false> <var: #page type: #'StackPage *'> <var: #pointer type: #'char *'> <var: #cogMethod type: #'CogMethod *'> "theIP must be typed as signed because it is assigned ceCannotResumePC and so maybe implicitly typed as unsigned." <var: #theIP type: #sqInt> self assert: (objectMemory isContext: aContext). self assert: (self isSingleContext: aContext). self assert: (objectMemory goodContextSize: aContext). theIP := objectMemory fetchPointer: InstructionPointerIndex ofObject: aContext. self assert: HasBeenReturnedFromMCPC < 0. theIP := (objectMemory isIntegerObject: theIP) ifTrue: [objectMemory integerValueOf: theIP] ifFalse: [HasBeenReturnedFromMCPC]. theMethod := objectMemory followObjField: MethodIndex ofObject: aContext. page := stackPages newStackPage. "first word on stack is caller context of base frame" stackPages longAt: (pointer := page baseAddress) put: (objectMemory followObjField: SenderIndex ofObject: aContext). "second word is the context itself; needed for cannotReturn processing; see ceBaseReturn:." stackPages longAt: (pointer := pointer - objectMemory wordSize) put: aContext. rcvr := objectMemory followField: ReceiverIndex ofObject: aContext. "If the frame is a closure activation then the closure should be on the stack in the pushed receiver position (closures receive the value[:value:] messages). Otherwise it should be the receiver proper." maybeClosure := objectMemory fetchPointer: ClosureIndex ofObject: aContext. maybeClosure ~= objectMemory nilObject ifTrue: [(objectMemory isForwarded: maybeClosure) ifTrue: [maybeClosure := objectMemory fixFollowedField: ClosureIndex ofObject: aContext withInitialValue: maybeClosure]. numArgs := self argumentCountOfClosure: maybeClosure. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: maybeClosure] ifFalse: [| header | header := objectMemory methodHeaderOf: theMethod. numArgs := self argumentCountOfMethodHeader: header. "If this is a synthetic context its IP could be pointing at the CallPrimitive opcode. If so, skip it." ((self methodHeaderHasPrimitive: header) and: [theIP = (1 + (self startPCOfMethodHeader: header))]) ifTrue: [theIP := theIP + (self sizeOfCallPrimitiveBytecode: header)]. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: rcvr]. "Put the arguments on the stack" 1 to: numArgs do: [:i| stackPages longAt: (pointer := pointer - objectMemory wordSize) put: (objectMemory fetchPointer: ReceiverIndex + i ofObject: aContext)]. "saved caller ip is base return trampoline" stackPages longAt: (pointer := pointer - objectMemory wordSize) put: cogit ceBaseFrameReturnPC. "base frame's saved fp is null" stackPages longAt: (pointer := pointer - objectMemory wordSize) put: 0. "N.B. Don't set the baseFP, which marks the page as in use, until after ensureMethodIsCogged: and/or instructionPointer:forContext:frame:. These can cause a compiled code compaction which, if marked as in use, will examine this partially initialized page and crash." page headFP: pointer. "Create either a machine code frame or an interpreter frame based on the pc. If the pc is -ve it is a machine code pc and so we produce a machine code frame. If +ve an interpreter frame. N.B. Do *not* change this to try and map from a bytecode pc to a machine code frame under any circumstances. See ensureContextIsExecutionSafeAfterAssignToStackPointer:" theIP < 0 ifTrue: [| cogMethod | "Since we would have to generate a machine-code method to be able to map the native pc anyway we should create a native method and native frame." + cogMethod := self ensureMethodIsCogged: theMethod maybeClosure: maybeClosure. - (maybeClosure ~= objectMemory nilObject and: [(self isVanillaBlockClosure: maybeClosure) not]) - ifTrue: [cogMethod := self ensureBlockIsCogged: theMethod numCopied: (self copiedValueCountOfFullClosure: maybeClosure)] - ifFalse: [cogMethod := self ensureMethodIsCogged: theMethod]. theMethod := cogMethod asInteger. maybeClosure ~= objectMemory nilObject ifTrue: [(self isVanillaBlockClosure: maybeClosure) ifTrue: ["If the pc is the special HasBeenReturnedFromMCPC pc set the pc appropriately so that the frame stays in the cannotReturn: state." theIP = HasBeenReturnedFromMCPC ifTrue: [theMethod := (cogit findMethodForStartBcpc: (self startPCOfClosure: maybeClosure) inHomeMethod: (self cCoerceSimple: theMethod to: #'CogMethod *')) asInteger. theMethod = 0 ifTrue: [self error: 'cannot find machine code block matching closure''s startpc']. theIP := cogit ceCannotResumePC] ifFalse: [self assert: (theIP signedBitShift: -16) < -1. "See contextInstructionPointer:frame:" theMethod := theMethod - ((theIP signedBitShift: -16) * cogit blockAlignment). theIP := theMethod - theIP signedIntFromShort]] ifFalse: [self assert: (theIP signedBitShift: -16) >= -1. "If the pc is the special HasBeenReturnedFromMCPC pc set the pc appropriately so that the frame stays in the cannotReturn: state." theIP := theIP = HasBeenReturnedFromMCPC ifTrue: [cogit ceCannotResumePC] ifFalse: [theMethod asInteger - theIP]]. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: theMethod + MFMethodFlagHasContextFlag + MFMethodFlagIsBlockFlag] ifFalse: [self assert: (theIP signedBitShift: -16) >= -1. "If the pc is the special HasBeenReturnedFromMCPC pc set the pc appropriately so that the frame stays in the cannotReturn: state." theIP := theIP = HasBeenReturnedFromMCPC ifTrue: [cogit ceCannotResumePC] ifFalse: [theMethod asInteger - theIP]. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: theMethod + MFMethodFlagHasContextFlag]. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: aContext] ifFalse: [stackPages longAt: (pointer := pointer - objectMemory wordSize) put: theMethod. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: aContext. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: (self encodeFrameFieldHasContext: true isBlock: maybeClosure ~= objectMemory nilObject numArgs: numArgs). stackPages longAt: (pointer := pointer - objectMemory wordSize) put: 0. "FoxIFSavedIP" theIP := self iframeInstructionPointerForIndex: theIP method: theMethod]. page baseFP: page headFP. self assert: (self frameHasContext: page baseFP). self assert: (self frameNumArgs: page baseFP) == numArgs. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: rcvr. stackPtrIndex := self quickFetchInteger: StackPointerIndex ofObject: aContext. self assert: ReceiverIndex + stackPtrIndex < (objectMemory lengthOf: aContext). numArgs + 1 to: stackPtrIndex do: [:i| stackPages longAt: (pointer := pointer - objectMemory wordSize) put: (objectMemory fetchPointer: ReceiverIndex + i ofObject: aContext)]. "top of stack is the instruction pointer" stackPages longAt: (pointer := pointer - objectMemory wordSize) put: theIP. page headSP: pointer. self assert: (self context: aContext hasValidInversePCMappingOf: theIP in: page baseFP). "Mark context as married by setting its sender to the frame pointer plus SmallInteger tags and the InstructionPointer to the saved fp (which ensures correct alignment w.r.t. the frame when we check for validity) plus SmallInteger tags." objectMemory storePointerUnchecked: SenderIndex ofObject: aContext withValue: (self withSmallIntegerTags: page baseFP). objectMemory storePointerUnchecked: InstructionPointerIndex ofObject: aContext withValue: (self withSmallIntegerTags: 0). self assert: (objectMemory isIntegerObject: (objectMemory fetchPointer: SenderIndex ofObject: aContext)). self assert: (self frameOfMarriedContext: aContext) = page baseFP. self assert: (self validStackPageBaseFrame: page). ^page! Item was changed: ----- Method: CoInterpreter>>mustMapMachineCodePC:context: (in category 'frame access') ----- mustMapMachineCodePC: theIP context: aOnceMarriedContext "Map the native pc theIP into a bytecode pc integer object and answer it. See contextInstructionPointer:frame: for the explanation." | maybeClosure methodObj cogMethod startBcpc bcpc | <inline: false> <var: #cogMethod type: #'CogMethod *'> theIP = HasBeenReturnedFromMCPC ifTrue: [^objectMemory nilObject]. maybeClosure := objectMemory fetchPointer: ClosureIndex ofObject: aOnceMarriedContext. methodObj := objectMemory fetchPointer: MethodIndex ofObject: aOnceMarriedContext. (maybeClosure ~= objectMemory nilObject and: [self isVanillaBlockClosure: maybeClosure]) ifTrue: [self assert: (theIP signedBitShift: -16) < -1. startBcpc := self startPCOfClosure: maybeClosure] ifFalse: [self assert: (theIP signedBitShift: -16) = -1. startBcpc := self startPCOfMethod: methodObj]. + cogMethod := self ensureMethodIsCogged: methodObj maybeClosure: maybeClosure. - (maybeClosure ~= objectMemory nilObject and: [(self isVanillaBlockClosure: maybeClosure) not]) - ifTrue: [cogMethod := self ensureBlockIsCogged: methodObj numCopied: (self copiedValueCountOfFullClosure: maybeClosure)] - ifFalse: [cogMethod := self ensureMethodIsCogged: methodObj]. bcpc := self bytecodePCFor: theIP cogMethod: cogMethod startBcpc: startBcpc. self assert: bcpc >= (self startPCOfMethod: methodObj). "If there's a CallPrimitive we need to skip it." (bcpc = startBcpc and: [maybeClosure = objectMemory nilObject and: [self methodHeaderHasPrimitive: cogMethod methodHeader]]) ifTrue: [bcpc := bcpc + (self sizeOfCallPrimitiveBytecode: cogMethod methodHeader)]. ^objectMemory integerObjectOf: bcpc + 1! Item was removed: - ----- Method: CogVMSimulator>>ensureMethodIsCogged: (in category 'frame access') ----- - ensureMethodIsCogged: methodObj - "Uncomment this to compact frequently and hence test if clients are ready for the shock." - "[self commenceCogCompiledCodeCompaction] - on: Halt - do: [:ex| ex resume: nil]." - ^super ensureMethodIsCogged: methodObj! Item was added: + ----- Method: CogVMSimulator>>ensureMethodIsCogged:maybeClosure: (in category 'frame access') ----- + ensureMethodIsCogged: methodObj maybeClosure: maybeClosure + "Uncomment this to compact frequently and hence test if clients are ready for the shock." + "[self commenceCogCompiledCodeCompaction] + on: Halt + do: [:ex| ex resume: nil]." + ^super ensureMethodIsCogged: methodObj maybeClosure: maybeClosure! Item was changed: ----- Method: Spur32BitMemoryManager>>objectAfter:limit: (in category 'object enumeration') ----- objectAfter: objOop limit: limit "Object parsing. 1. all objects have at least a word following the header, for a forwarding pointer. + 2. objects with an overflow size have a preceding word with a saturated numSlots. If the word - 2. objects with an overflow size have a preceeing word with a saturated numSlots. If the word following an object doesn't have a saturated numSlots field it must be a single-header object. If the word following does have a saturated numSlots it must be the overflow size word." <inline: true> | followingWordAddress followingWord | followingWordAddress := self addressAfter: objOop. (self oop: followingWordAddress isGreaterThanOrEqualTo: limit) ifTrue: [^limit]. self flag: #endianness. followingWord := self longAt: followingWordAddress + 4. ^followingWord >> self numSlotsHalfShift = self numSlotsMask ifTrue: [followingWordAddress + self baseHeaderSize] ifFalse: [followingWordAddress]! Item was changed: ----- Method: Spur64BitMemoryManager>>objectAfter:limit: (in category 'object enumeration') ----- objectAfter: objOop limit: limit "Object parsing. 1. all objects have at least a word following the header, for a forwarding pointer. + 2. objects with an overflow size have a preceding word with a saturated numSlots. If the word - 2. objects with an overflow size have a preceeing word with a saturated numSlots. If the word following an object doesn't have a saturated numSlots field it must be a single-header object. If the word following does have a saturated numSlots it must be the overflow size word." <inline: true> | followingWordAddress followingWord | followingWordAddress := self addressAfter: objOop. (self oop: followingWordAddress isGreaterThanOrEqualTo: limit) ifTrue: [^limit]. self flag: #endianness. followingWord := self longAt: followingWordAddress. ^followingWord >> self numSlotsFullShift = self numSlotsMask ifTrue: [followingWordAddress + self baseHeaderSize] ifFalse: [followingWordAddress]!

1 0

VM Maker: VMMaker.oscog-cb.2046.mcz
by commits＠source.squeak.org 21 Dec '16

21 Dec '16

ClementBera uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-cb.2046.mcz ==================== Summary ==================== Name: VMMaker.oscog-cb.2046 Author: cb Time: 21 December 2016, 5:17:46.303077 pm UUID: cb74888d-8648-4ed4-87ce-d7fd1d9353f1 Ancestors: VMMaker.oscog-cb.2045 Fixed a bug where Cog would compiled compiledBlock like a compiledMethod, leading to very strange crashes. =============== Diff against VMMaker.oscog-cb.2045 =============== Item was added: + ----- Method: CoInterpreter>>ensureBlockIsCogged:numCopied: (in category 'frame access') ----- + ensureBlockIsCogged: methodObj numCopied: numCopied + <returnTypeC: #'CogMethod *'> + | rawHeader cogMethod | + <inline: true> + <var: #cogMethod type: #'CogMethod *'> + rawHeader := self rawHeaderOf: methodObj. + (self isCogMethodReference: rawHeader) ifTrue: + [^self cCoerceSimple: rawHeader to: #'CogMethod *']. + cogMethod := cogit cogFullBlockMethod: methodObj numCopied: numCopied. + (cogMethod = nil + and: [cogCompiledCodeCompactionCalledFor]) ifTrue: + [self commenceCogCompiledCodeCompaction. + cogMethod := cogMethod := cogit cogFullBlockMethod: methodObj numCopied: numCopied]. + (self asserta: cogMethod ~= nil) ifFalse: + [self error: 'could not compile method that should have been compiled']. + ^cogMethod! Item was changed: ----- Method: CoInterpreter>>makeBaseFrameFor: (in category 'frame access') ----- makeBaseFrameFor: aContext "<Integer>" "Marry aContext with the base frame of a new stack page. Build the base frame to reflect the context's state. Answer the new page. Override to hold the caller context in a different place, In the StackInterpreter we use the caller saved ip, but in the Cog VM caller saved ip is the ceBaseReturn: trampoline. Simply hold the caller context in the first word of the stack." <returnTypeC: #'StackPage *'> | page pointer theMethod theIP numArgs stackPtrIndex maybeClosure rcvr | <inline: false> <var: #page type: #'StackPage *'> <var: #pointer type: #'char *'> <var: #cogMethod type: #'CogMethod *'> "theIP must be typed as signed because it is assigned ceCannotResumePC and so maybe implicitly typed as unsigned." <var: #theIP type: #sqInt> self assert: (objectMemory isContext: aContext). self assert: (self isSingleContext: aContext). self assert: (objectMemory goodContextSize: aContext). theIP := objectMemory fetchPointer: InstructionPointerIndex ofObject: aContext. self assert: HasBeenReturnedFromMCPC < 0. theIP := (objectMemory isIntegerObject: theIP) ifTrue: [objectMemory integerValueOf: theIP] ifFalse: [HasBeenReturnedFromMCPC]. theMethod := objectMemory followObjField: MethodIndex ofObject: aContext. page := stackPages newStackPage. "first word on stack is caller context of base frame" stackPages longAt: (pointer := page baseAddress) put: (objectMemory followObjField: SenderIndex ofObject: aContext). "second word is the context itself; needed for cannotReturn processing; see ceBaseReturn:." stackPages longAt: (pointer := pointer - objectMemory wordSize) put: aContext. rcvr := objectMemory followField: ReceiverIndex ofObject: aContext. "If the frame is a closure activation then the closure should be on the stack in the pushed receiver position (closures receive the value[:value:] messages). Otherwise it should be the receiver proper." maybeClosure := objectMemory fetchPointer: ClosureIndex ofObject: aContext. maybeClosure ~= objectMemory nilObject ifTrue: [(objectMemory isForwarded: maybeClosure) ifTrue: [maybeClosure := objectMemory fixFollowedField: ClosureIndex ofObject: aContext withInitialValue: maybeClosure]. numArgs := self argumentCountOfClosure: maybeClosure. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: maybeClosure] ifFalse: [| header | header := objectMemory methodHeaderOf: theMethod. numArgs := self argumentCountOfMethodHeader: header. "If this is a synthetic context its IP could be pointing at the CallPrimitive opcode. If so, skip it." ((self methodHeaderHasPrimitive: header) and: [theIP = (1 + (self startPCOfMethodHeader: header))]) ifTrue: [theIP := theIP + (self sizeOfCallPrimitiveBytecode: header)]. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: rcvr]. "Put the arguments on the stack" 1 to: numArgs do: [:i| stackPages longAt: (pointer := pointer - objectMemory wordSize) put: (objectMemory fetchPointer: ReceiverIndex + i ofObject: aContext)]. "saved caller ip is base return trampoline" stackPages longAt: (pointer := pointer - objectMemory wordSize) put: cogit ceBaseFrameReturnPC. "base frame's saved fp is null" stackPages longAt: (pointer := pointer - objectMemory wordSize) put: 0. "N.B. Don't set the baseFP, which marks the page as in use, until after ensureMethodIsCogged: and/or instructionPointer:forContext:frame:. These can cause a compiled code compaction which, if marked as in use, will examine this partially initialized page and crash." page headFP: pointer. "Create either a machine code frame or an interpreter frame based on the pc. If the pc is -ve it is a machine code pc and so we produce a machine code frame. If +ve an interpreter frame. N.B. Do *not* change this to try and map from a bytecode pc to a machine code frame under any circumstances. See ensureContextIsExecutionSafeAfterAssignToStackPointer:" theIP < 0 ifTrue: [| cogMethod | "Since we would have to generate a machine-code method to be able to map the native pc anyway we should create a native method and native frame." + (maybeClosure ~= objectMemory nilObject and: [(self isVanillaBlockClosure: maybeClosure) not]) + ifTrue: [cogMethod := self ensureBlockIsCogged: theMethod numCopied: (self copiedValueCountOfFullClosure: maybeClosure)] + ifFalse: [cogMethod := self ensureMethodIsCogged: theMethod]. - cogMethod := self ensureMethodIsCogged: theMethod. theMethod := cogMethod asInteger. maybeClosure ~= objectMemory nilObject ifTrue: [(self isVanillaBlockClosure: maybeClosure) ifTrue: ["If the pc is the special HasBeenReturnedFromMCPC pc set the pc appropriately so that the frame stays in the cannotReturn: state." theIP = HasBeenReturnedFromMCPC ifTrue: [theMethod := (cogit findMethodForStartBcpc: (self startPCOfClosure: maybeClosure) inHomeMethod: (self cCoerceSimple: theMethod to: #'CogMethod *')) asInteger. theMethod = 0 ifTrue: [self error: 'cannot find machine code block matching closure''s startpc']. theIP := cogit ceCannotResumePC] ifFalse: [self assert: (theIP signedBitShift: -16) < -1. "See contextInstructionPointer:frame:" theMethod := theMethod - ((theIP signedBitShift: -16) * cogit blockAlignment). theIP := theMethod - theIP signedIntFromShort]] ifFalse: [self assert: (theIP signedBitShift: -16) >= -1. "If the pc is the special HasBeenReturnedFromMCPC pc set the pc appropriately so that the frame stays in the cannotReturn: state." theIP := theIP = HasBeenReturnedFromMCPC ifTrue: [cogit ceCannotResumePC] ifFalse: [theMethod asInteger - theIP]]. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: theMethod + MFMethodFlagHasContextFlag + MFMethodFlagIsBlockFlag] ifFalse: [self assert: (theIP signedBitShift: -16) >= -1. "If the pc is the special HasBeenReturnedFromMCPC pc set the pc appropriately so that the frame stays in the cannotReturn: state." theIP := theIP = HasBeenReturnedFromMCPC ifTrue: [cogit ceCannotResumePC] ifFalse: [theMethod asInteger - theIP]. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: theMethod + MFMethodFlagHasContextFlag]. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: aContext] ifFalse: [stackPages longAt: (pointer := pointer - objectMemory wordSize) put: theMethod. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: aContext. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: (self encodeFrameFieldHasContext: true isBlock: maybeClosure ~= objectMemory nilObject numArgs: numArgs). stackPages longAt: (pointer := pointer - objectMemory wordSize) put: 0. "FoxIFSavedIP" theIP := self iframeInstructionPointerForIndex: theIP method: theMethod]. page baseFP: page headFP. self assert: (self frameHasContext: page baseFP). self assert: (self frameNumArgs: page baseFP) == numArgs. stackPages longAt: (pointer := pointer - objectMemory wordSize) put: rcvr. stackPtrIndex := self quickFetchInteger: StackPointerIndex ofObject: aContext. self assert: ReceiverIndex + stackPtrIndex < (objectMemory lengthOf: aContext). numArgs + 1 to: stackPtrIndex do: [:i| stackPages longAt: (pointer := pointer - objectMemory wordSize) put: (objectMemory fetchPointer: ReceiverIndex + i ofObject: aContext)]. "top of stack is the instruction pointer" stackPages longAt: (pointer := pointer - objectMemory wordSize) put: theIP. page headSP: pointer. self assert: (self context: aContext hasValidInversePCMappingOf: theIP in: page baseFP). "Mark context as married by setting its sender to the frame pointer plus SmallInteger tags and the InstructionPointer to the saved fp (which ensures correct alignment w.r.t. the frame when we check for validity) plus SmallInteger tags." objectMemory storePointerUnchecked: SenderIndex ofObject: aContext withValue: (self withSmallIntegerTags: page baseFP). objectMemory storePointerUnchecked: InstructionPointerIndex ofObject: aContext withValue: (self withSmallIntegerTags: 0). self assert: (objectMemory isIntegerObject: (objectMemory fetchPointer: SenderIndex ofObject: aContext)). self assert: (self frameOfMarriedContext: aContext) = page baseFP. self assert: (self validStackPageBaseFrame: page). ^page! Item was changed: ----- Method: CoInterpreter>>mustMapMachineCodePC:context: (in category 'frame access') ----- mustMapMachineCodePC: theIP context: aOnceMarriedContext "Map the native pc theIP into a bytecode pc integer object and answer it. See contextInstructionPointer:frame: for the explanation." | maybeClosure methodObj cogMethod startBcpc bcpc | <inline: false> <var: #cogMethod type: #'CogMethod *'> theIP = HasBeenReturnedFromMCPC ifTrue: [^objectMemory nilObject]. maybeClosure := objectMemory fetchPointer: ClosureIndex ofObject: aOnceMarriedContext. methodObj := objectMemory fetchPointer: MethodIndex ofObject: aOnceMarriedContext. (maybeClosure ~= objectMemory nilObject and: [self isVanillaBlockClosure: maybeClosure]) ifTrue: [self assert: (theIP signedBitShift: -16) < -1. startBcpc := self startPCOfClosure: maybeClosure] ifFalse: [self assert: (theIP signedBitShift: -16) = -1. startBcpc := self startPCOfMethod: methodObj]. + (maybeClosure ~= objectMemory nilObject and: [(self isVanillaBlockClosure: maybeClosure) not]) + ifTrue: [cogMethod := self ensureBlockIsCogged: methodObj numCopied: (self copiedValueCountOfFullClosure: maybeClosure)] + ifFalse: [cogMethod := self ensureMethodIsCogged: methodObj]. - cogMethod := self ensureMethodIsCogged: methodObj. bcpc := self bytecodePCFor: theIP cogMethod: cogMethod startBcpc: startBcpc. self assert: bcpc >= (self startPCOfMethod: methodObj). "If there's a CallPrimitive we need to skip it." (bcpc = startBcpc and: [maybeClosure = objectMemory nilObject and: [self methodHeaderHasPrimitive: cogMethod methodHeader]]) ifTrue: [bcpc := bcpc + (self sizeOfCallPrimitiveBytecode: cogMethod methodHeader)]. ^objectMemory integerObjectOf: bcpc + 1!

1 0

VM Maker: VMMaker.oscog-cb.2045.mcz
by commits＠source.squeak.org 21 Dec '16

21 Dec '16

ClementBera uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-cb.2045.mcz ==================== Summary ==================== Name: VMMaker.oscog-cb.2045 Author: cb Time: 21 December 2016, 2:59:29.957723 pm UUID: c23c0e96-df64-4990-b40a-d34c40e5c1fa Ancestors: VMMaker.oscog-nice.2044 use cmacro instead of hand-written constant. =============== Diff against VMMaker.oscog-nice.2044 =============== Item was changed: ----- Method: SimpleStackBasedCogit>>genExtPushRemoteTempOrInstVarLongBytecode (in category 'bytecode generators') ----- genExtPushRemoteTempOrInstVarLongBytecode | index | + ^ (byte2 noMask: coInterpreter remoteIsInstVarAccess) - ^ (byte2 noMask: 1 << 7) ifTrue: [ self genPushRemoteTempLongBytecode ] ifFalse: [ index := byte1 + (extA << 8). extA := 0. extB := 0. "don't use flags in the simple cogit" (coInterpreter isReadMediatedContextInstVarIndex: index) + ifTrue: [self + genPushMaybeContextRemoteInstVar: index + inObjectAt: byte2 - coInterpreter remoteIsInstVarAccess] + ifFalse: [self + genPushRemoteInstVar: index + inObjectAt: byte2 - coInterpreter remoteIsInstVarAccess]]! - ifTrue: [ self genPushMaybeContextRemoteInstVar: index inObjectAt: byte2 - (1 << 7) ] - ifFalse: [ self genPushRemoteInstVar: index inObjectAt: byte2 - (1 << 7) ] ]! Item was changed: ----- Method: SimpleStackBasedCogit>>genExtStorePopRemoteTempOrInstVarLongBytecodePopBoolean: (in category 'bytecode generators') ----- genExtStorePopRemoteTempOrInstVarLongBytecodePopBoolean: popBoolean | index | extB := 0. "simple cogit don't use the extra flag" + (byte2 noMask: coInterpreter remoteIsInstVarAccess) - (byte2 noMask: 1 << 7) ifTrue: [ self genStorePop: popBoolean RemoteTemp: byte1 At: byte2. self cppIf: IMMUTABILITY ifTrue: [ self annotateBytecode: self Label ] ] ifFalse: [ index := byte1 + (extA << 8). extA := 0. (coInterpreter isWriteMediatedContextInstVarIndex: index) ifTrue: [ self genStorePop: popBoolean MaybeContextRemoteInstVar: index + ofObjectAt: byte2 - coInterpreter remoteIsInstVarAccess ] - ofObjectAt: byte2 - (1 << 7) ] ifFalse: [ self genStorePop: popBoolean RemoteInstVar: index + ofObjectAt: byte2 - coInterpreter remoteIsInstVarAccess ] ]. - ofObjectAt: byte2 - (1 << 7) ] ]. ^ 0!

1 0

VM Maker: VMMaker.oscog-nice.2044.mcz
by commits＠source.squeak.org 21 Dec '16

21 Dec '16

Nicolas Cellier uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-nice.2044.mcz ==================== Summary ==================== Name: VMMaker.oscog-nice.2044 Author: nice Time: 21 December 2016, 2:43:59.503877 pm UUID: bf51cf5c-9e79-499e-80e4-2e7d40c359aa Ancestors: VMMaker.oscog-cb.2043 Correct generateCogMethod/FullBlock which might write one byte too far. Indeed, (padIfPossibleWithStopsFrom: startAddress to: endAddress) is inclusive (it will write a byte into endAddress). So it must be called with (self padIfPossibleWithStopsFrom: startAddress to: startAddress + size - 1). Maybe we were saved by roundUpLength:, or maybe the methodZone further allocated size was not yet used, but we don't have to live that dangerously. =============== Diff against VMMaker.oscog-cb.2043 =============== Item was changed: ----- Method: Cogit>>generateCogFullBlock (in category 'compile abstract instructions') ----- generateCogFullBlock "We handle jump sizing simply. First we make a pass that asks each instruction to compute its maximum size. Then we make a pass that sizes jumps based on the maxmimum sizes. Then we make a pass that fixes up jumps. When fixing up a jump the jump is not allowed to choose a smaller offset but must stick to the size set in the second pass." <returnTypeC: #'CogMethod *'> <option: #SistaV1BytecodeSet> | codeSize headerSize mapSize totalSize startAddress result method | <var: #method type: #'CogMethod *'> headerSize := self sizeof: CogMethod. methodLabel address: methodZone freeStart. self computeMaximumSizes. methodLabel concretizeAt: methodZone freeStart. codeSize := self generateInstructionsAt: methodLabel address + headerSize. mapSize := self generateMapAt: nil start: methodLabel address + cbNoSwitchEntryOffset. . totalSize := methodZone roundUpLength: headerSize + codeSize + mapSize. totalSize > MaxMethodSize ifTrue: [^self cCoerceSimple: MethodTooBig to: #'CogMethod *']. startAddress := methodZone allocate: totalSize. startAddress = 0 ifTrue: [^self cCoerceSimple: InsufficientCodeSpace to: #'CogMethod *']. self assert: startAddress + cbEntryOffset = fullBlockEntry address. self assert: startAddress + cbNoSwitchEntryOffset = fullBlockNoContextSwitchEntry address. result := self outputInstructionsAt: startAddress + headerSize. self assert: startAddress + headerSize + codeSize = result. + backEnd padIfPossibleWithStopsFrom: result to: startAddress + totalSize - mapSize - 1. - backEnd padIfPossibleWithStopsFrom: result to: startAddress + totalSize - mapSize. self generateMapAt: startAddress + totalSize - 1 start: startAddress + cbNoSwitchEntryOffset. self flag: #TOCHECK. "It's not clear we want the same header than regular methods. It could be of the same size, but maybe the cmType could be different and the selector could be ignored." method := self fillInMethodHeader: (self cCoerceSimple: startAddress to: #'CogMethod *') size: totalSize selector: objectMemory nilObject. method cpicHasMNUCaseOrCMIsFullBlock: true. postCompileHook ifNotNil: [self perform: postCompileHook with: method. postCompileHook := nil]. ^method! Item was changed: ----- Method: Cogit>>generateCogMethod: (in category 'generate machine code') ----- generateCogMethod: selector "We handle jump sizing simply. First we make a pass that asks each instruction to compute its maximum size. Then we make a pass that sizes jumps based on the maxmimum sizes. Then we make a pass that fixes up jumps. When fixing up a jump the jump is not allowed to choose a smaller offset but must stick to the size set in the second pass." <returnTypeC: #'CogMethod *'> | codeSize headerSize mapSize totalSize startAddress result method | <var: #method type: #'CogMethod *'> headerSize := self sizeof: CogMethod. methodLabel address: methodZone freeStart. self computeMaximumSizes. methodLabel concretizeAt: methodZone freeStart. codeSize := self generateInstructionsAt: methodLabel address + headerSize. mapSize := self generateMapAt: nil start: methodLabel address + cmNoCheckEntryOffset. totalSize := methodZone roundUpLength: headerSize + codeSize + mapSize. totalSize > MaxMethodSize ifTrue: [^self cCoerceSimple: MethodTooBig to: #'CogMethod *']. startAddress := methodZone allocate: totalSize. startAddress = 0 ifTrue: [^self cCoerceSimple: InsufficientCodeSpace to: #'CogMethod *']. self assert: startAddress + cmEntryOffset = entry address. self assert: startAddress + cmNoCheckEntryOffset = noCheckEntry address. result := self outputInstructionsAt: startAddress + headerSize. self assert: startAddress + headerSize + codeSize = result. + backEnd padIfPossibleWithStopsFrom: result to: startAddress + totalSize - mapSize - 1. - backEnd padIfPossibleWithStopsFrom: result to: startAddress + totalSize - mapSize. self generateMapAt: startAddress + totalSize - 1 start: startAddress + cmNoCheckEntryOffset. self fillInBlockHeadersAt: startAddress. method := self fillInMethodHeader: (self cCoerceSimple: startAddress to: #'CogMethod *') size: totalSize selector: selector. postCompileHook ifNotNil: [self perform: postCompileHook with: method. postCompileHook := nil]. ^method!

1 0

Problem with OSSubprocess / signals / heartbeat ?
by Mariano Martinez Peck 20 Dec '16

20 Dec '16

Hi guys, There is a bug happening in OSSubprocess since quite some time which we were never able to fix. What happens is that at certain point (after having used OSSubprocess for a while), a simple command like the following would "hang" Pharo: OSSUnixSubprocess new command: 'ls'; arguments: #(); redirectStdout; runAndWaitOnExitDo: [ :process :outString | outString ] Once the image does that, it keeps doing it in a 80% of the cases (so it is easy enough to debug). Guille Polito kept one of these images if someone can give us a hand. He also proposed the great idea of using `strace` to see what was going on. He (together with Pable Tesone) suspected that the heartbeat could be interrupting the `clone()` function which is (I think) called internally by the 'posix_spawn()' which is the one used by OSSubprocess. When these images are "hung" they found at a infinitive loop like this: [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 593498}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 600126}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 606282}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 612528}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 618689}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 624922}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 631145}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 637241}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 643578}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 17477] gettimeofday({1482152630, 649974}, NULL) = 0 [pid 17477] sigreturn() (mask []) = 120 [pid 17477] clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf7578768) = ? ERESTARTNOINTR (To be restarted) [pid 17477] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- As you can see, there is a SIGALARM involved. It also looks like the `gettimeofday` is used by the heartbeat ? Could it be that someone the heartbeat is interrupting the `clone()` ? Guille also showed me the `strace` output with a regular / working image: [pid 18647] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- [pid 18647] gettimeofday({1482152829, 481014}, NULL) = 0 [pid 18647] sigreturn() (mask []) = -1 EINTR (Interrupted system call) [pid 18647] getitimer(ITIMER_REAL, {it_interval={0, 2000}, it_value={0, 1917}}) = 0 [pid 18647] recvmsg(3, 0xff7b0734, 0) = -1 EAGAIN (Resource temporarily unavailable) [pid 18647] select(4, [3], [], [3], {0, 1000}) = 0 (Timeout) [pid 18647] getitimer(ITIMER_REAL, {it_interval={0, 2000}, it_value={0, 797}}) = 0 [pid 18647] recvmsg(3, 0xff7b0734, 0) = -1 EAGAIN (Resource temporarily unavailable) [pid 18647] select(4, [3], [], [3], {0, 1000}) = ? ERESTARTNOHAND (To be restarted if no handler) [pid 18647] --- SIGALRM {si_signo=SIGALRM, si_code=SI_KERNEL} --- Does anyone have any hint here? Thanks in advance, -- Mariano http://marianopeck.wordpress.com

3 2

Integrated decentralized localized social networks (was Re: [Pharo-users] ptsd rage)
by Charlie Robbats 20 Dec '16

20 Dec '16

Which I have done over the years, used other venues for peaceful warfare. Sorry to disturb you. The 20th century model of society is stale and dated. Look at all the damage done to people for no good reason. We can effect change through technology. What kind of change? Something that leverages modern computing, including the blockchain for trusted interactions, lots of hand-held IoT devices combined into a massive decentralized computational network, leaving REST services behind and presenting functional semantics with robust data marshaling. With these capabilities, transform the way physical interactions occur. I like to term what I am seeing as virtualizing physicality & physicalizing virtuality. Is Pharo just aiming to be a better WebApp? That market is totally saturated. Pharo's/Squeak's strengtsh are robust dynamic presentation of information in graphical form and interconnectivity. Let us leverage that. Smalltalk is perfectly positioned to define the next transformation, because it is the only language system designed to be online 24/7, exactly what is needed for a massive decentralized computer with dynamic un/loading. I merely made a few comments regarding some of the practical functional requirements around the security facet of the new system. My apologies if you felt it was inappropriate. I think you were concerned it may get further out of hand and I do not disagree with you. Let's collect functional requirements for how this new social network will interact with personalized locality. This is my proposition. Why, we could eliminate prisons, massively reducing mental illness and eliminate collective confrontation. I believe completely, with all my heart, that this is in the original spirit of Smalltalk, but I cannot speak for the founders, this is my belief. Thank you for your thoughts, I did not mean to offend or go so far astray. Charlie On 12/19/2016 7:17 PM, Brad wrote: > Don't know about anyone else, but I would appreciate it if this kind > of commentary never made it to this email list. This topic is so > inappropriate. Let's keep the discussion about Pharo. If you feel you > must go down the political route, then please be gracious enough to > find another venue. > > Brad Selfridge > 913-269-2385 > > On Dec 19, 2016, at 6:48 PM, Charlie Robbats > <charlie.robbats(a)gmail.com <mailto:charlie.robbats@gmail.com>> wrote: > >> Alright, one more sad observation. Of all the Law Enforcement >> violence, either perpetrated by LEOs against citizens or by citizens >> against LEOs, in all cases I am aware or recall having heard, all of >> these involve police officers. No violence, or very little, is >> perpetrated against or by Sheriff Deputies. As counties are a good >> and proper and consistent scope of crime administration, regardless >> of the density, police ought to report to the Sheriffs and let's put >> and end to broken windows policy and discriminatory enforcement. >> Police are always in our shit, and have their egos to boot. Sheriff >> Deputies, by comparison, are laid back and easy to talk to. They >> engage the public in respect, without attitude. >> >> Bob didn't shoot the deputy. >> >> On 12/17/2016 10:11 PM, Charlie Robbats wrote: >>> >>> Last bit, a mantra: Crime & Punishment creates massive mental >>> illness in our society. Ask the Russians. >>> >>> >>> On 12/17/2016 10:05 PM, Charlie Robbats wrote: >>>> >>>> >>>> >>>> Well, my paranoia is kicking. I spoke too broadly of police, I have >>>> interacted with many and they do a lot of good. A bad apple spoils >>>> the whatever and their code of silence allows for that sort of rot. >>>> I tell it straight, because they monitor the list. I pissed them >>>> off! Do you think it passed the sanity test that cops run all >>>> street drugs with their felon database in hand? Damn! All those >>>> Fellas are going to turn and bite their masters' hand, you watch! >>>> Stacking these grown men like Twinkies in overcrowded prisons. >>>> There is a debt that needs to be paid and the righteous are those >>>> in prison on a bunch of bullshit paper. Vive la revolucion! Ok, >>>> sorry I went political. I am done. >>>> >>>> >>>> On 12/17/2016 9:49 PM, Charlie Robbats wrote: >>>>> >>>>> I am sad with you in your diagnosis. I get that too, the blues. I >>>>> get high if I am holding and listen to "Kind of Blue" or deep >>>>> house or reggae. Some X would be good, but herbage is my mainstay >>>>> to drop my constant threat mode (hypervigilence). I do not want a >>>>> repeat of what happened! Unfortunately I drive people away to make >>>>> sure of that. I appreciate your alls kindness. I wanted to share a >>>>> solid explanation of why I am such a dick at times. I flip to >>>>> active threat mode, I struggle to control this. It doesn't help >>>>> that the police have been stalking me for 2 years, I am all caught >>>>> up in their noise. Don't believe a word of their truthfulness. The >>>>> blue wall of silence breaks God's commandment thou shalt not bear >>>>> false witness. Police violate that daily, protecting their >>>>> brothers who shoot unarmed black men in the back and claim they >>>>> are fearing for their lives. Really? No, really? Well, that is a >>>>> different story. That just freaking triggers me. And my opinion >>>>> makes me their target. At least in my C-PTSD befuddled mind. >>>>> >>>>> Be well all! >>>>> >>

1 2

ptsd rage
by Robert Withers 20 Dec '16

20 Dec '16

I need to apologize for all my dread behavior. Here is the reason, and I am on service-connected disability. Current diagnosis is Complex PTSD. Sorry about all my sturm und drang. http://www.ptsd.va.gov/public/problems/anger-and-trauma.asp Sincerely, Robert

3 5

← Newer
1
2
3
4
5
6
7
...
11
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev December 2016