Vm-dev December 2022

vm-dev@lists.squeakfoundation.org

12 participants
52 discussions

VM Maker: VMMaker.oscog.seperateMarking-eem.3284.mcz
by commits＠source.squeak.org 10 Dec '22

10 Dec '22

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog.seperateMarking-eem.3284.mcz ==================== Summary ==================== Name: VMMaker.oscog.seperateMarking-eem.3284 Author: eem Time: 9 December 2022, 8:13:27.049847 pm UUID: 368a65e4-0689-4992-ac09-c9ed9811ae1e Ancestors: VMMaker.oscog.seperateMarking-eem.3283 Print segments to the standard transcript =============== Diff against VMMaker.oscog.seperateMarking-eem.3283 =============== Item was changed: ----- Method: SpurSegmentManager>>printSegmentAddresses (in category 'simulation only') ----- printSegmentAddresses <doNotGenerate> "Convenient when debugging segment overlap" + manager coInterpreter print: 'Seg addresses: '; cr. - Transcript show: 'Seg addresses: '; cr. 0 to: numSegments - 1 do: [:i| + manager coInterpreter + printHexnp: (segments at: i) segStart; + print: ' -> '; + printHexnp: (segments at: i) segLimit; + cr]! - Transcript - show: (segments at: i) segStart; - show: ' -> '; - show: (segments at: i) segLimit; - show: ' ; ']. - Transcript cr.!

1 0

VM Maker: VMMaker.oscog.seperateMarking-eem.3283.mcz
by commits＠source.squeak.org 10 Dec '22

10 Dec '22

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog.seperateMarking-eem.3283.mcz ==================== Summary ==================== Name: VMMaker.oscog.seperateMarking-eem.3283 Author: eem Time: 9 December 2022, 7:50:07.446363 pm UUID: 538b31ef-81fb-4a2d-9255-f576bc00d202 Ancestors: VMMaker.oscog.seperateMarking-eem.3282, VMMaker.oscog-eem.3281 Merge VMMaker.oscog-eem.3281 =============== Diff against VMMaker.oscog-eem.3281 =============== Item was changed: ----- Method: CogARMv8Compiler>>concretizeMoveCqR (in category 'generate machine code - concretize') ----- concretizeMoveCqR "C3.3.4 Move (immediate) C3-215 The Move (immediate) instructions are aliases for a single MOVZ, MOVN, or ORR (immediate with zero register), instruction to load an immediate value into the destination register. An assembler must permit a signed or unsigned immediate, as long as its binary representation can be generated using one of these instructions, and an assembler error results if the immediate cannot be generated in this way. On disassembly, it is unspecified whether the immediate is output as a signed or an unsigned value. C6.2.191 MOVZ C6-1102 Move wide with zero moves an optionally-shifted 16-bit immediate value to a register. C6.2.190 MOVN C6-1100 Move wide with NOT moves the inverse of an optionally-shifted 16-bit immediate value to a register. C6.2.204 ORR (immediate) C6-1125 Bitwise OR (immediate) performs a bitwise (inclusive) OR of a register value and an immediate register value, and writes the result to the destination register." | constant destReg | constant := operands at: 0. destReg := operands at: 1. destReg ~= SP ifTrue: [| lowBit lowBitMod16 mask | lowBit := constant > 0 ifTrue: [self cCode: [self computeLowBit: constant] inSmalltalk: [constant lowBit - 1]] ifFalse: [0]. lowBitMod16 := lowBit // 16 * 16. mask := 1 << 16 - 1 << lowBitMod16. (constant bitAnd: mask) = constant ifTrue: ["Use MOVZ" machineCode at: 0 put: 2r110100101 << 23 + (lowBitMod16 // 16 << 21) + (constant >> lowBitMod16 << 5) + destReg. ^4]. lowBit := constant signedIntFromLong64 < -1 ifTrue: [self cCode: [self computeLowBit: constant bitInvert64] inSmalltalk: [constant bitInvert64 lowBit - 1]] ifFalse: [0]. (constant bitOr: mask) signedIntFromLong64 = -1 ifTrue: ["Use MOVN" self assert: (constant bitInvert64 >> lowBitMod16) = ((constant bitInvert64 >> lowBitMod16) bitAnd: mask). machineCode at: 0 put: 2r100100101 << 23 + (lowBitMod16 // 16 << 21) + (constant bitInvert64 >> lowBitMod16 << 5) + destReg. ^4]]. ^self isImmNImmSImmREncodableBitmask: constant ifTrue: [:n :imms :immr| "Use ORR" machineCode at: 0 + put: 2r1011001 << 25 + + (n << 22) - put: 2r1011001001 << 22 + (immr << 16) + (imms << 10) + (XZR << 5) + destReg. 4] + ifFalse: [self emitMoveCw: constant intoR: destReg at: 0] + + "cogit processor disassembleInstructionAt: 0 In: machineCode object"! - ifFalse: [self emitMoveCw: constant intoR: destReg at: 0]! Item was removed: - ----- Method: CogARMv8Compiler>>decode64Imms:immr: (in category 'generate machine code - support') ----- - decode64Imms: imms immr: immr - "See aarch64/instrs/integer/bitmasks/DecodeBitMasks J1-7389. - This is a 64-bit version computing the imm mask (wmask) only." - <returnTypeC: #usqInt> - | mask | - <var: 'mask' type: #usqInt> - self assert: ((imms between: 0 and: 63) and: [immr between: 0 and: 63]). - "For logical immediates an all-ones value of S is reserved since it would generate a useless all-ones result (many times)" - imms = 63 ifTrue: - [^self cCode: [0] inSmalltalk: [#undefined]]. - - mask := 1 << (imms + 1) - 1. - ^immr = 0 - ifTrue: [mask] - ifFalse: [(mask << (64 - immr) bitAnd: 1 << 64 - 1) + (mask >> immr)]! Item was added: + ----- Method: CogARMv8Compiler>>decodeN:imms:immr: (in category 'generate machine code - support') ----- + decodeN: n imms: imms immr: immr + "See aarch64/instrs/integer/bitmasks/DecodeBitMasks J1-7389." + <returnTypeC: #usqInt> + | bits immediate mask rotation width | + <var: 'mask' type: #usqInt> + self assert: ((n between: 0 and: 1) and: [(imms between: 0 and: 63) and: [immr between: 0 and: 63]]). + "For logical immediates an all-ones value of S is reserved since it would generate a useless all-ones result (many times)" + imms = 63 ifTrue: + [^self cCode: [0] inSmalltalk: [#undefined]]. + + n = 1 + ifTrue: + [width := 64. + mask := 16rffffffffffffffff. + bits := imms. + rotation := immr] + ifFalse: + [imms < 16r20 + ifTrue: [width := 32. bits := imms] ifFalse: [ + imms < 16r30 + ifTrue: [width := 16. bits := imms bitAnd: 16rF] ifFalse: [ + imms < 16r38 + ifTrue: [width := 8. bits := imms bitAnd: 16r7] ifFalse: [ + imms < 16r3C + ifTrue: [width := 4. bits := imms bitAnd: 16r3] ifFalse: [ + imms < 16r3E + ifTrue: [width := 2. bits := imms bitAnd: 16r1] + ifFalse: [self error: 'invalid logical immediate']]]]]. + mask := 1 << width - 1. + rotation := immr bitAnd: width - 1]. + + width - 1 = bits ifTrue: + [^0]. + + immediate := (1 << (bits + 1)) - 1. + + rotation > 0 ifTrue: + [immediate := immediate << (width - rotation) - 1]. + + (width between: 2 and: 32) ifTrue: + [immediate := (immediate bitShift: width) bitOr: immediate]. + + ^immediate! Item was changed: ----- Method: CogARMv8Compiler>>isImmNImmSImmREncodableBitmask:ifTrue:ifFalse: (in category 'generate machine code - support') ----- isImmNImmSImmREncodableBitmask: constant ifTrue: trinaryBlock "[:n :imms :immr|...]" ifFalse: nullaryBlock "See DecodeBitMasks J1-7389. See https://dinfuehr.github.io/blog/encoding-of-immediate-values-on-aarch64/ This method is adapted from The LLVM Compiler Infrastructure, AArch64AddressingModes.h processLogicalImmediate" <inline: #always> | imm size mask numLeadingOnes numTrailingOnes immr n nImms rotateCount | <var: 'mask' type: #usqInt> <var: 'nImms' type: #usqInt> (constant between: -1 and: 0) ifTrue: [^nullaryBlock value]. imm := self cCode: [constant] inSmalltalk: [constant signedIntToLong64]. "First, determine the element size." size := 32. [mask := 1 << size - 1. (imm bitAnd: mask) ~= (imm >> size) ifTrue: [size := size * 2. false] ifFalse: [size > 2]] whileTrue: [size := size / 2]. "Second, determine the rotation to make the element be: 0^m 1^n." mask := 1 << 64 - 1 >> (64 - size). imm := imm bitAnd: mask. (self isShiftedMask: imm) ifTrue: [rotateCount := self countTrailingZeros: imm. numTrailingOnes := self countTrailingOnes: imm >> rotateCount] ifFalse: [imm := imm bitOr: mask bitInvert64. (self isShiftedMask: imm) ifFalse: [^nullaryBlock value]. numLeadingOnes := self countLeadingOnes: imm. rotateCount := 64 - numLeadingOnes. numTrailingOnes := numLeadingOnes + (self countTrailingOnes: imm) - (64 - size)]. "Encode in Immr the number of RORs it would take to get *from* 0^m 1^n to our target value, where I is the number of RORs to go the opposite direction." self assert: size > rotateCount. "rotateCount should be smaller than element size" immr := size - rotateCount bitAnd: size - 1. "If size has a 1 in the n'th bit, create a value that has zeroes in bits [0, n] and ones above that." + nImms := self cCode: [(size - 1) bitInvert64 << 1] inSmalltalk: [(size - 1) bitInvert64 << 1 bitAnd: 16rFFFFFFFFFFFFFFFF]. - nImms := (size - 1) bitInvert64 << 1. "Or the CTO value into the low bits, which must be below the Nth bit mentioned above." + nImms := nImms bitOr: numTrailingOnes - 1. - nImms := nImms bitOr: numTrailingOnes - 1. "Extract the seventh bit and toggle it to create the N field." n := ((nImms >> 6) bitAnd: 1) bitXor: 1. nImms := nImms bitAnd: 16r3F. + self assert: (self decodeN: n imms: nImms immr: immr) = constant signedIntToLong64. - self assert: (self decode64Imms: nImms immr: immr) = constant signedIntToLong64. ^trinaryBlock value: n value: nImms value: immr ! Item was changed: ----- Method: Spur32BitMMLECoSimulator>>coInterpreter:cogit: (in category 'initialization') ----- coInterpreter: aCoInterpreter cogit: aCogit + coInterpreter := aCoInterpreter. cogit := aCogit. + marker coInterpreter: aCoInterpreter. scavenger coInterpreter: aCoInterpreter. + compactor coInterpreter: aCoInterpreter. + gc coInterpreter: aCoInterpreter + ! - compactor coInterpreter: aCoInterpreter! Item was added: + ----- Method: SpurIncrementalSweepAndCompact>>coInterpreter: (in category 'initialization') ----- + coInterpreter: aVMSimulator + <doNotGenerate> + super coInterpreter: aVMSimulator. + sweeper coInterpreter: aVMSimulator. + compactor coInterpreter: aVMSimulator! Item was changed: ----- Method: StackInterpreter>>objectMemory: (in category 'initialization') ----- objectMemory: anObjectMemory <doNotGenerate> objectMemory ifNotNil: [self halt]. + objectMemory := anObjectMemory. + anObjectMemory coInterpreter: self! - objectMemory := anObjectMemory! Item was changed: ----- Method: VMClass class>>openSpurMultiWindowBrowser (in category 'utilities') ----- openSpurMultiWindowBrowser "Answer a new multi-window browser on the Spur classes, the Cog StackInterpreter classes, and the support classes" "self openSpurMultiWindowBrowser" | b | b := Browser open. #( SpurMemoryManager Spur32BitMemoryManager Spur64BitMemoryManager SpurGenerationScavenger SpurSegmentManager Spur32BitMMLESimulator SpurGenerationScavengerSimulator + SpurMarker SpurAllAtOnceMarker SpurIncrementalMarker + SpurCompactor SpurPlanningCompactor SpurHybridCompactor SpurIncrementalCompactor + SpurIncrementalSweepAndCompact InterpreterPrimitives StackInterpreter StackInterpreterPrimitives VMStructType VMMaker CCodeGenerator TMethod) do: [:className| (Smalltalk classNamed: className) ifNotNil: [:class| b selectCategoryForClass: class; selectClass: class]] separatedBy: [b multiWindowState addNewWindow]. b multiWindowState selectWindowIndex: 1!

1 0

[OpenSmalltalk/opensmalltalk-vm] 1a7f4f: CogVM source as per VMMaker.oscog-eem.3281
by Eliot Miranda 09 Dec '22

09 Dec '22

Branch: refs/heads/virtend Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 1a7f4f3c818fb7a502ac78291cb3a1797ad9a4f1 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/1a7f4f3c818fb7a502… Author: Eliot Miranda <eliot.miranda(a)gmail.com> Date: 2022-12-08 (Thu, 08 Dec 2022) Changed paths: M src/spur64.cog.lowcode/cogitARMv8.c M src/spur64.cog.newspeak/cogitARMv8.c M src/spur64.cog.newspeak/cogitX64SysV.c M src/spur64.cog.newspeak/cogitX64WIN64.c M src/spur64.cog.newspeak/cointerp.c M src/spur64.cog.newspeak/gcc3x-cointerp.c M src/spur64.cog/cogitARMv8.c M src/spur64.sista/cogitARMv8.c Log Message: ----------- CogVM source as per VMMaker.oscog-eem.3281 Cogit: Fix MoveCqR on ARMv8 for certain bit patterns. This one caused 16r20000000 to evaluate to zero on ARMv8. The old code invalidly assumed that N would always be 1 in the immediate ORR logical immediate. Incidentally generate the 64-bit Newspeak VM.

1 0

[OpenSmalltalk/opensmalltalk-vm] aef4d4: CogVM source as per VMMaker.oscog-eem.3281
by Eliot Miranda 09 Dec '22

09 Dec '22

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: aef4d4060a0cf0974fe512295fcedd171d1c3a46 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/aef4d4060a0cf0974f… Author: Eliot Miranda <eliot.miranda(a)gmail.com> Date: 2022-12-08 (Thu, 08 Dec 2022) Changed paths: R src/plugins/MP4MediaFilePlugin M src/spur64.cog.lowcode/cogitARMv8.c M src/spur64.cog.newspeak/cogitARMv8.c M src/spur64.cog.newspeak/cogitX64SysV.c M src/spur64.cog.newspeak/cogitX64WIN64.c M src/spur64.cog.newspeak/cointerp.c M src/spur64.cog.newspeak/gcc3x-cointerp.c M src/spur64.cog/cogitARMv8.c M src/spur64.sista/cogitARMv8.c Log Message: ----------- CogVM source as per VMMaker.oscog-eem.3281 Cogit: Fix MoveCqR on ARMv8 for certain bit patterns. This one caused 16r20000000 to evaluate to zero on ARMv8. The old code invalidly assumed that N would always be 1 in the immediate ORR logical immediate. Incidentally generate the 64-bit Newspeak VM.

1 0

VM Maker: VMMaker.oscog-eem.3281.mcz
by commits＠source.squeak.org 09 Dec '22

09 Dec '22

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-eem.3281.mcz ==================== Summary ==================== Name: VMMaker.oscog-eem.3281 Author: eem Time: 8 December 2022, 4:29:23.109215 pm UUID: 2cf271a5-8c56-4c80-919d-69abbec5f64b Ancestors: VMMaker.oscog-eem.3280 Cogit: Fix MoveCqR on ARMv8 for certain bit patterns. This one caused 16r20000000 to evaluate to zero on ARMv8. The old code invalidly assumed that N would always be 1 in the immediate ORR logocal immediate. =============== Diff against VMMaker.oscog-eem.3280 =============== Item was changed: ----- Method: CogARMv8Compiler>>concretizeMoveCqR (in category 'generate machine code - concretize') ----- concretizeMoveCqR "C3.3.4 Move (immediate) C3-215 The Move (immediate) instructions are aliases for a single MOVZ, MOVN, or ORR (immediate with zero register), instruction to load an immediate value into the destination register. An assembler must permit a signed or unsigned immediate, as long as its binary representation can be generated using one of these instructions, and an assembler error results if the immediate cannot be generated in this way. On disassembly, it is unspecified whether the immediate is output as a signed or an unsigned value. C6.2.191 MOVZ C6-1102 Move wide with zero moves an optionally-shifted 16-bit immediate value to a register. C6.2.190 MOVN C6-1100 Move wide with NOT moves the inverse of an optionally-shifted 16-bit immediate value to a register. C6.2.204 ORR (immediate) C6-1125 Bitwise OR (immediate) performs a bitwise (inclusive) OR of a register value and an immediate register value, and writes the result to the destination register." | constant destReg | constant := operands at: 0. destReg := operands at: 1. destReg ~= SP ifTrue: [| lowBit lowBitMod16 mask | lowBit := constant > 0 ifTrue: [self cCode: [self computeLowBit: constant] inSmalltalk: [constant lowBit - 1]] ifFalse: [0]. lowBitMod16 := lowBit // 16 * 16. mask := 1 << 16 - 1 << lowBitMod16. (constant bitAnd: mask) = constant ifTrue: ["Use MOVZ" machineCode at: 0 put: 2r110100101 << 23 + (lowBitMod16 // 16 << 21) + (constant >> lowBitMod16 << 5) + destReg. ^4]. lowBit := constant signedIntFromLong64 < -1 ifTrue: [self cCode: [self computeLowBit: constant bitInvert64] inSmalltalk: [constant bitInvert64 lowBit - 1]] ifFalse: [0]. (constant bitOr: mask) signedIntFromLong64 = -1 ifTrue: ["Use MOVN" self assert: (constant bitInvert64 >> lowBitMod16) = ((constant bitInvert64 >> lowBitMod16) bitAnd: mask). machineCode at: 0 put: 2r100100101 << 23 + (lowBitMod16 // 16 << 21) + (constant bitInvert64 >> lowBitMod16 << 5) + destReg. ^4]]. ^self isImmNImmSImmREncodableBitmask: constant ifTrue: [:n :imms :immr| "Use ORR" machineCode at: 0 + put: 2r1011001 << 25 + + (n << 22) - put: 2r1011001001 << 22 + (immr << 16) + (imms << 10) + (XZR << 5) + destReg. 4] + ifFalse: [self emitMoveCw: constant intoR: destReg at: 0] + + "cogit processor disassembleInstructionAt: 0 In: machineCode object"! - ifFalse: [self emitMoveCw: constant intoR: destReg at: 0]! Item was removed: - ----- Method: CogARMv8Compiler>>decode64Imms:immr: (in category 'generate machine code - support') ----- - decode64Imms: imms immr: immr - "See aarch64/instrs/integer/bitmasks/DecodeBitMasks J1-7389. - This is a 64-bit version computing the imm mask (wmask) only." - <returnTypeC: #usqInt> - | mask | - <var: 'mask' type: #usqInt> - self assert: ((imms between: 0 and: 63) and: [immr between: 0 and: 63]). - "For logical immediates an all-ones value of S is reserved since it would generate a useless all-ones result (many times)" - imms = 63 ifTrue: - [^self cCode: [0] inSmalltalk: [#undefined]]. - - mask := 1 << (imms + 1) - 1. - ^immr = 0 - ifTrue: [mask] - ifFalse: [(mask << (64 - immr) bitAnd: 1 << 64 - 1) + (mask >> immr)]! Item was added: + ----- Method: CogARMv8Compiler>>decodeN:imms:immr: (in category 'generate machine code - support') ----- + decodeN: n imms: imms immr: immr + "See aarch64/instrs/integer/bitmasks/DecodeBitMasks J1-7389." + <returnTypeC: #usqInt> + | bits immediate mask rotation width | + <var: 'mask' type: #usqInt> + self assert: ((n between: 0 and: 1) and: [(imms between: 0 and: 63) and: [immr between: 0 and: 63]]). + "For logical immediates an all-ones value of S is reserved since it would generate a useless all-ones result (many times)" + imms = 63 ifTrue: + [^self cCode: [0] inSmalltalk: [#undefined]]. + + n = 1 + ifTrue: + [width := 64. + mask := 16rffffffffffffffff. + bits := imms. + rotation := immr] + ifFalse: + [imms < 16r20 + ifTrue: [width := 32. bits := imms] ifFalse: [ + imms < 16r30 + ifTrue: [width := 16. bits := imms bitAnd: 16rF] ifFalse: [ + imms < 16r38 + ifTrue: [width := 8. bits := imms bitAnd: 16r7] ifFalse: [ + imms < 16r3C + ifTrue: [width := 4. bits := imms bitAnd: 16r3] ifFalse: [ + imms < 16r3E + ifTrue: [width := 2. bits := imms bitAnd: 16r1] + ifFalse: [self error: 'invalid logical immediate']]]]]. + mask := 1 << width - 1. + rotation := immr bitAnd: width - 1]. + + width - 1 = bits ifTrue: + [^0]. + + immediate := (1 << (bits + 1)) - 1. + + rotation > 0 ifTrue: + [immediate := immediate << (width - rotation) - 1]. + + (width between: 2 and: 32) ifTrue: + [immediate := (immediate bitShift: width) bitOr: immediate]. + + ^immediate! Item was changed: ----- Method: CogARMv8Compiler>>isImmNImmSImmREncodableBitmask:ifTrue:ifFalse: (in category 'generate machine code - support') ----- isImmNImmSImmREncodableBitmask: constant ifTrue: trinaryBlock "[:n :imms :immr|...]" ifFalse: nullaryBlock "See DecodeBitMasks J1-7389. See https://dinfuehr.github.io/blog/encoding-of-immediate-values-on-aarch64/ This method is adapted from The LLVM Compiler Infrastructure, AArch64AddressingModes.h processLogicalImmediate" <inline: #always> | imm size mask numLeadingOnes numTrailingOnes immr n nImms rotateCount | <var: 'mask' type: #usqInt> <var: 'nImms' type: #usqInt> (constant between: -1 and: 0) ifTrue: [^nullaryBlock value]. imm := self cCode: [constant] inSmalltalk: [constant signedIntToLong64]. "First, determine the element size." size := 32. [mask := 1 << size - 1. (imm bitAnd: mask) ~= (imm >> size) ifTrue: [size := size * 2. false] ifFalse: [size > 2]] whileTrue: [size := size / 2]. "Second, determine the rotation to make the element be: 0^m 1^n." mask := 1 << 64 - 1 >> (64 - size). imm := imm bitAnd: mask. (self isShiftedMask: imm) ifTrue: [rotateCount := self countTrailingZeros: imm. numTrailingOnes := self countTrailingOnes: imm >> rotateCount] ifFalse: [imm := imm bitOr: mask bitInvert64. (self isShiftedMask: imm) ifFalse: [^nullaryBlock value]. numLeadingOnes := self countLeadingOnes: imm. rotateCount := 64 - numLeadingOnes. numTrailingOnes := numLeadingOnes + (self countTrailingOnes: imm) - (64 - size)]. "Encode in Immr the number of RORs it would take to get *from* 0^m 1^n to our target value, where I is the number of RORs to go the opposite direction." self assert: size > rotateCount. "rotateCount should be smaller than element size" immr := size - rotateCount bitAnd: size - 1. "If size has a 1 in the n'th bit, create a value that has zeroes in bits [0, n] and ones above that." + nImms := self cCode: [(size - 1) bitInvert64 << 1] inSmalltalk: [(size - 1) bitInvert64 << 1 bitAnd: 16rFFFFFFFFFFFFFFFF]. - nImms := (size - 1) bitInvert64 << 1. "Or the CTO value into the low bits, which must be below the Nth bit mentioned above." + nImms := nImms bitOr: numTrailingOnes - 1. - nImms := nImms bitOr: numTrailingOnes - 1. "Extract the seventh bit and toggle it to create the N field." n := ((nImms >> 6) bitAnd: 1) bitXor: 1. nImms := nImms bitAnd: 16r3F. + self assert: (self decodeN: n imms: nImms immr: immr) = constant signedIntToLong64. - self assert: (self decode64Imms: nImms immr: immr) = constant signedIntToLong64. ^trinaryBlock value: n value: nImms value: immr !

1 0

VM Maker Inbox: VMMaker.oscog.seperateMarking-eem.3282.mcz
by commits＠source.squeak.org 08 Dec '22

08 Dec '22

Eliot Miranda uploaded a new version of VMMaker to project VM Maker Inbox: http://source.squeak.org/VMMakerInbox/VMMaker.oscog.seperateMarking-eem.328… ==================== Summary ==================== Name: VMMaker.oscog.seperateMarking-eem.3282 Author: eem Time: 8 December 2022, 11:52:37.114851 am UUID: 56db74b3-6af7-486d-985a-3be4f53132f1 Ancestors: VMMaker.oscog.seperateMarking-eem.3281 Get SpurIncrementalSweepAndCompact to set coInterpreter correctly, propagating to the sweeper and compactor. Kick off the propagation in StackInterpreter>>objectMemory:. Add some of the IGC classes to openSpurMultiWindowBrowser =============== Diff against VMMaker.oscog.seperateMarking-eem.3281 =============== Item was changed: ----- Method: Spur32BitMMLECoSimulator>>coInterpreter:cogit: (in category 'initialization') ----- coInterpreter: aCoInterpreter cogit: aCogit + coInterpreter := aCoInterpreter. cogit := aCogit. + marker coInterpreter: aCoInterpreter. scavenger coInterpreter: aCoInterpreter. + compactor coInterpreter: aCoInterpreter. + gc coInterpreter: aCoInterpreter + ! - compactor coInterpreter: aCoInterpreter! Item was added: + ----- Method: SpurIncrementalSweepAndCompact>>coInterpreter: (in category 'initialization') ----- + coInterpreter: aVMSimulator + <doNotGenerate> + super coInterpreter: aVMSimulator. + sweeper coInterpreter: aVMSimulator. + compactor coInterpreter: aVMSimulator! Item was changed: ----- Method: StackInterpreter>>objectMemory: (in category 'initialization') ----- objectMemory: anObjectMemory <doNotGenerate> objectMemory ifNotNil: [self halt]. + objectMemory := anObjectMemory. + anObjectMemory coInterpreter: self! - objectMemory := anObjectMemory! Item was changed: ----- Method: VMClass class>>openSpurMultiWindowBrowser (in category 'utilities') ----- openSpurMultiWindowBrowser "Answer a new multi-window browser on the Spur classes, the Cog StackInterpreter classes, and the support classes" "self openSpurMultiWindowBrowser" | b | b := Browser open. #( SpurMemoryManager Spur32BitMemoryManager Spur64BitMemoryManager SpurGenerationScavenger SpurSegmentManager Spur32BitMMLESimulator SpurGenerationScavengerSimulator + SpurMarker SpurAllAtOnceMarker SpurIncrementalMarker + SpurCompactor SpurPlanningCompactor SpurHybridCompactor SpurIncrementalCompactor + SpurIncrementalSweepAndCompact InterpreterPrimitives StackInterpreter StackInterpreterPrimitives VMStructType VMMaker CCodeGenerator TMethod) do: [:className| (Smalltalk classNamed: className) ifNotNil: [:class| b selectCategoryForClass: class; selectClass: class]] separatedBy: [b multiWindowState addNewWindow]. b multiWindowState selectWindowIndex: 1!

1 0

VM Maker: VMMaker.oscog.seperateMarking-eem.3281.mcz
by commits＠source.squeak.org 08 Dec '22

08 Dec '22

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog.seperateMarking-eem.3281.mcz ==================== Summary ==================== Name: VMMaker.oscog.seperateMarking-eem.3281 Author: eem Time: 8 December 2022, 10:35:27.618989 am UUID: 563eefa5-7e25-41a0-b95c-71a70fec2569 Ancestors: VMMaker.oscog.seperateMarking-eem.3279, VMMaker.oscog-eem.3280 Merge VMMaker.oscog-eem.3280 & VMMaker.oscog-eem.3276 =============== Diff against VMMaker.oscog.seperateMarking-eem.3279 =============== Item was changed: ----- Method: ClipboardExtendedPlugin>>ioAddClipboardData:data:dataFormat: (in category 'io') ----- ioAddClipboardData: clipboard data: data dataFormat: format | dataLength | + self primitive: 'ioAddClipboardData' parameters: #(Unsigned WordsOrBytes Oop). - self primitive: 'ioAddClipboardData' parameters: #(Unsigned ByteArray Oop). dataLength := interpreterProxy byteSizeOf: data cPtrAsOop. (interpreterProxy isIntegerObject: format) ifTrue: [self sqPasteboardPutItemFlavor: clipboard asVoidPointer data: data length: dataLength formatType: (interpreterProxy integerValueOf: format)] ifFalse: [(interpreterProxy isBytes: format) ifTrue: [self sqPasteboardPutItemFlavor: clipboard asVoidPointer data: data length: dataLength formatType: (interpreterProxy firstIndexableField: format) formatLength: (interpreterProxy byteSizeOf: format)] ifFalse: [interpreterProxy primitiveFailFor: PrimErrBadArgument]]! Item was changed: ----- Method: ClipboardExtendedPlugin>>ioGetClipboardFormat:formatNumber: (in category 'io') ----- + ioGetClipboardFormat: clipboard formatNumber: formatNumber - ioGetClipboardFormat: clipboard formatNumber: formatNumber <var: #clipboardAddress type: #'usqIntptr_t'> self primitive: 'ioGetClipboardFormat' parameters: #(Unsigned SmallInteger). ^(self sqPasteboardGetItemCount: clipboard asVoidPointer) > 0 ifTrue: [self sqPasteboardCopyItemFlavors: clipboard asVoidPointer itemNumber: formatNumber] ifFalse: [interpreterProxy nilObject]! Item was added: + ----- Method: InterpreterPrimitives>>primitiveObjectsAccessibleFromRoots (in category 'image segment in/out') ----- + primitiveObjectsAccessibleFromRoots + "This primitive is called from Squeak as... + arrayOfRoots uniquelyAccessibleObjects" + + "This primitive answers an array of the receiver and every object in its proper tree of subParts (ie, that is not refered to from anywhere else outside the tree)." + "This primitive could be used to implement the primitiveStoreImageSegment segment, thanks to a suggestion from Igor Stassenko. Currently it is + used only to debug that primitive." + + | arrayOfRoots result | + arrayOfRoots := self stackTop. + + "Essential type checks" + (objectMemory isArray: arrayOfRoots) "Must be indexable pointers" + ifFalse: [^self primitiveFail]. + + result := objectMemory objectsAccessibleFromRoots: arrayOfRoots. + (objectMemory hasSpurMemoryManagerAPI + and: [(objectMemory isIntegerObject: result) + and: [(objectMemory integerValueOf: result) = PrimErrNoMemory]]) ifTrue: + [objectMemory fullGC. + arrayOfRoots := self stackTop. + result := objectMemory objectsAccessibleFromRoots: arrayOfRoots]. + (objectMemory isIntegerObject: result) + ifTrue: [self primitiveFailFor: (objectMemory integerValueOf: result)] + ifFalse: [self methodReturnValue: result]! Item was added: + ----- Method: ObjectMemory>>objectsAccessibleFromRoots: (in category 'image segment in/out') ----- + objectsAccessibleFromRoots: arrayOfRoots + ^self integerObjectOf: PrimErrUnsupported! Item was added: + ----- Method: SpurMemoryManager>>objectsAccessibleFromRoots: (in category 'image segment in/out') ----- + objectsAccessibleFromRoots: arrayOfRootsArg + "This primitive is called from Squeak as... + arrayOfRoots uniquelyAccessibleObjects + + This primitive answers an array of the receiver and every object in its proper tree of subParts (ie, that is not refered to from anywhere else outside the tree). + + The primitive can fail for the following reasons with the specified failure codes: + PrimErrNoMemory: additional allocations failed" + + <inline: false> + | arrayOfRoots arrayOfObjects | + <var: 'segAddr' type: #usqInt> + + self runLeakCheckerFor: GCCheckImageSegment. + + "First scavenge to collect any new space garbage that refers to the graph." + self scavengingGC. + arrayOfRoots := self updatePostScavenge: arrayOfRootsArg. + + "Now compute the transitive closure, collecting the sequence of objects to be stored in the arrayOfObjects array. + Included in arrayOfObjects are the arrayOfRoots and all its contents. All objects have been unmarked." + arrayOfObjects := self objectsReachableFromRoots: arrayOfRoots. + arrayOfObjects ifNil: + [^self integerObjectOf: PrimErrNoMemory]. + "If objectsReachableFromRoots: answers an integer there is not enough continuous free space in which to allocate the + reachable objects. If there is sufficient free space then answer an error code to prompt a compacting GC and a retry." + (self isIntegerObject: arrayOfObjects) ifTrue: + [totalFreeOldSpace - self allocationUnit >= (self integerValueOf: arrayOfObjects) ifTrue: + [^self integerObjectOf: PrimErrNeedCompaction]. + ^self integerObjectOf: PrimErrNoMemory]. + + self assert: self allObjectsUnmarked. "work to be done when the incremental GC is written" + self deny: (self forwardersIn: arrayOfObjects). + + ^arrayOfObjects! Item was changed: ----- Method: StackInterpreter class>>initializePrimitiveTable (in category 'initialization') ----- (excessive size, no diff calculated)

1 0

[OpenSmalltalk/opensmalltalk-vm] 131bb5: CogVM source as per VMMaker.oscog-eem.3280
by Eliot Miranda 07 Dec '22

07 Dec '22

Branch: refs/heads/virtend Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 131bb5b2e71a790510a23267cf9519861d64151e https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/131bb5b2e71a790510… Author: Eliot Miranda <eliot.miranda(a)gmail.com> Date: 2022-12-06 (Tue, 06 Dec 2022) Changed paths: M src/spur32.cog.lowcode/cointerp.c M src/spur32.cog.lowcode/cointerp.h M src/spur32.cog.lowcode/gcc3x-cointerp.c M src/spur32.cog/cointerp.c M src/spur32.cog/cointerp.h M src/spur32.cog/cointerpmt.c M src/spur32.cog/cointerpmt.h M src/spur32.cog/gcc3x-cointerp.c M src/spur32.cog/gcc3x-cointerpmt.c M src/spur32.sista/cointerp.c M src/spur32.sista/cointerp.h M src/spur32.sista/gcc3x-cointerp.c M src/spur32.stack.lowcode/gcc3x-interp.c M src/spur32.stack.lowcode/interp.c M src/spur32.stack/gcc3x-interp.c M src/spur32.stack/interp.c M src/spur64.cog.lowcode/cointerp.c M src/spur64.cog.lowcode/cointerp.h M src/spur64.cog.lowcode/gcc3x-cointerp.c M src/spur64.cog/cointerp.c M src/spur64.cog/cointerp.h M src/spur64.cog/cointerpmt.c M src/spur64.cog/cointerpmt.h M src/spur64.cog/gcc3x-cointerp.c M src/spur64.cog/gcc3x-cointerpmt.c M src/spur64.sista/cointerp.c M src/spur64.sista/cointerp.h M src/spur64.sista/gcc3x-cointerp.c M src/spur64.stack.lowcode/gcc3x-interp.c M src/spur64.stack.lowcode/interp.c M src/spur64.stack/gcc3x-interp.c M src/spur64.stack/interp.c M src/v3.cog/cointerp.c M src/v3.cog/cointerp.h M src/v3.cog/gcc3x-cointerp.c M src/v3.stack/gcc3x-interp.c M src/v3.stack/interp.c Log Message: ----------- CogVM source as per VMMaker.oscog-eem.3280 Add a primitive that computes the transitive closure of objects accessible from an array of roots. This is the collection half of the storeImageSegment primtiive. This can be used either to debug failures in the storeImageSegment primitive, and may eventually be used to implement it in a more general way. Thanks to Igor Stasenko for suggesting this decomposition a few years ago.

1 0

[OpenSmalltalk/opensmalltalk-vm] 1da73c: CogVM source as per VMMaker.oscog-eem.3280
by Eliot Miranda 07 Dec '22

07 Dec '22

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 1da73cf86e96356dbaa1bf953dea7cc613f7cd22 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/1da73cf86e96356dba… Author: Eliot Miranda <eliot.miranda(a)gmail.com> Date: 2022-12-06 (Tue, 06 Dec 2022) Changed paths: M src/spur32.cog.lowcode/cointerp.c M src/spur32.cog.lowcode/cointerp.h M src/spur32.cog.lowcode/gcc3x-cointerp.c M src/spur32.cog/cointerp.c M src/spur32.cog/cointerp.h M src/spur32.cog/cointerpmt.c M src/spur32.cog/cointerpmt.h M src/spur32.cog/gcc3x-cointerp.c M src/spur32.cog/gcc3x-cointerpmt.c M src/spur32.sista/cointerp.c M src/spur32.sista/cointerp.h M src/spur32.sista/gcc3x-cointerp.c M src/spur32.stack.lowcode/gcc3x-interp.c M src/spur32.stack.lowcode/interp.c M src/spur32.stack/gcc3x-interp.c M src/spur32.stack/interp.c M src/spur64.cog.lowcode/cointerp.c M src/spur64.cog.lowcode/cointerp.h M src/spur64.cog.lowcode/gcc3x-cointerp.c M src/spur64.cog/cointerp.c M src/spur64.cog/cointerp.h M src/spur64.cog/cointerpmt.c M src/spur64.cog/cointerpmt.h M src/spur64.cog/gcc3x-cointerp.c M src/spur64.cog/gcc3x-cointerpmt.c M src/spur64.sista/cointerp.c M src/spur64.sista/cointerp.h M src/spur64.sista/gcc3x-cointerp.c M src/spur64.stack.lowcode/gcc3x-interp.c M src/spur64.stack.lowcode/interp.c M src/spur64.stack/gcc3x-interp.c M src/spur64.stack/interp.c M src/v3.cog/cointerp.c M src/v3.cog/cointerp.h M src/v3.cog/gcc3x-cointerp.c M src/v3.stack/gcc3x-interp.c M src/v3.stack/interp.c Log Message: ----------- CogVM source as per VMMaker.oscog-eem.3280 Add a primitive that computes the transitive closure of objects accessible from an array of roots. This is the collection half of the storeImageSegment primtiive. This can be used either to debug failures in the storeImageSegment primitive, and may eventually be used to implement it in a more general way. Thanks to Igor Stasenko for suggesting this decomposition a few years ago.

1 0

VM Maker: VMMaker.oscog-eem.3280.mcz
by commits＠source.squeak.org 07 Dec '22

07 Dec '22

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-eem.3280.mcz ==================== Summary ==================== Name: VMMaker.oscog-eem.3280 Author: eem Time: 6 December 2022, 8:52:36.917493 pm UUID: d6a800e7-dd9a-44d1-8421-b9a7e664d3bc Ancestors: VMMaker.oscog-eem.3279 Add a primitive that computes the transitive closure of objects accessible from an array of roots. This is the collection half of the storeImageSegment primtiive. This can be used either to debug failures in the storeImageSegment primitive, and may eventually be used to implement it in a more general way. Thanks to Igor Stasenko for suggesting this decomposition a few years ago. =============== Diff against VMMaker.oscog-eem.3279 =============== Item was added: + ----- Method: InterpreterPrimitives>>primitiveObjectsAccessibleFromRoots (in category 'image segment in/out') ----- + primitiveObjectsAccessibleFromRoots + "This primitive is called from Squeak as... + arrayOfRoots uniquelyAccessibleObjects" + + "This primitive answers an array of the receiver and every object in its proper tree of subParts (ie, that is not refered to from anywhere else outside the tree)." + "This primitive could be used to implement the primitiveStoreImageSegment segment, thanks to a suggestion from Igor Stassenko. Currently it is + used only to debug that primitive." + + | arrayOfRoots result | + arrayOfRoots := self stackTop. + + "Essential type checks" + (objectMemory isArray: arrayOfRoots) "Must be indexable pointers" + ifFalse: [^self primitiveFail]. + + result := objectMemory objectsAccessibleFromRoots: arrayOfRoots. + (objectMemory hasSpurMemoryManagerAPI + and: [(objectMemory isIntegerObject: result) + and: [(objectMemory integerValueOf: result) = PrimErrNoMemory]]) ifTrue: + [objectMemory fullGC. + arrayOfRoots := self stackTop. + result := objectMemory objectsAccessibleFromRoots: arrayOfRoots]. + (objectMemory isIntegerObject: result) + ifTrue: [self primitiveFailFor: (objectMemory integerValueOf: result)] + ifFalse: [self methodReturnValue: result]! Item was added: + ----- Method: ObjectMemory>>objectsAccessibleFromRoots: (in category 'image segment in/out') ----- + objectsAccessibleFromRoots: arrayOfRoots + ^self integerObjectOf: PrimErrUnsupported! Item was added: + ----- Method: SpurMemoryManager>>objectsAccessibleFromRoots: (in category 'image segment in/out') ----- + objectsAccessibleFromRoots: arrayOfRootsArg + "This primitive is called from Squeak as... + arrayOfRoots uniquelyAccessibleObjects + + This primitive answers an array of the receiver and every object in its proper tree of subParts (ie, that is not refered to from anywhere else outside the tree). + + The primitive can fail for the following reasons with the specified failure codes: + PrimErrNoMemory: additional allocations failed" + + <inline: false> + | arrayOfRoots arrayOfObjects | + <var: 'segAddr' type: #usqInt> + + self runLeakCheckerFor: GCCheckImageSegment. + + "First scavenge to collect any new space garbage that refers to the graph." + self scavengingGC. + arrayOfRoots := self updatePostScavenge: arrayOfRootsArg. + + "Now compute the transitive closure, collecting the sequence of objects to be stored in the arrayOfObjects array. + Included in arrayOfObjects are the arrayOfRoots and all its contents. All objects have been unmarked." + arrayOfObjects := self objectsReachableFromRoots: arrayOfRoots. + arrayOfObjects ifNil: + [^self integerObjectOf: PrimErrNoMemory]. + "If objectsReachableFromRoots: answers an integer there is not enough continuous free space in which to allocate the + reachable objects. If there is sufficient free space then answer an error code to prompt a compacting GC and a retry." + (self isIntegerObject: arrayOfObjects) ifTrue: + [totalFreeOldSpace - self allocationUnit >= (self integerValueOf: arrayOfObjects) ifTrue: + [^self integerObjectOf: PrimErrNeedCompaction]. + ^self integerObjectOf: PrimErrNoMemory]. + + self assert: self allObjectsUnmarked. "work to be done when the incremental GC is written" + self deny: (self forwardersIn: arrayOfObjects). + + ^arrayOfObjects! Item was changed: ----- Method: StackInterpreter class>>initializePrimitiveTable (in category 'initialization') ----- (excessive size, no diff calculated)

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev December 2022