Vm-dev October 2015

vm-dev@lists.squeakfoundation.org

29 participants
86 discussions

VM Maker: VMMaker.oscog-EstebanLorenzano.1503.mcz
by commits＠source.squeak.org 30 Oct '15

30 Oct '15

Esteban Lorenzano uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-EstebanLorenzano.1503.mcz ==================== Summary ==================== Name: VMMaker.oscog-EstebanLorenzano.1503 Author: EstebanLorenzano Time: 30 October 2015, 2:37:07.998518 pm UUID: 2cc25296-b733-4968-bd6b-8f3db5277693 Ancestors: VMMaker.oscog-eem.1502 general: - applying correct cast so compiler does not fails. ThreadedFFIPlugin: - ExternalAddress now are taken correctly (nor as ByteArray or Alien, because they are different beasts). - allow pushing of pointers to any type (into a ByteArray, an ExternalAddress or an Alien), to allow passing parameters style int*, float*, etc. - allow reading in memory to allow read chunks of ByteArray =============== Diff against VMMaker.oscog-eem.1502 =============== Item was changed: ----- Method: Cogit>>addressIsInInstructions: (in category 'testing') ----- addressIsInInstructions: address <var: #address type: #'AbstractInstruction *'> + ^self cCode: '!!((((unsigned)address) & BytesPerWord-1)) \ - ^self cCode: '!!((unsigned)((address) & BytesPerWord-1)) \ && (address) >= &abstractOpcodes[0] \ && (address) < &abstractOpcodes[opcodeIndex]' inSmalltalk: [(abstractOpcodes object identityIndexOf: address) between: 1 and: opcodeIndex]! Item was changed: ----- Method: ThreadedFFIPlugin>>ffiAddressOf:startingAt:size: (in category 'primitive support') ----- ffiAddressOf: rcvr startingAt: byteOffset size: byteSize "return an int of the address of the byteSize slot (byte, short, int, whatever) at byteOffset in rcvr. Nominally intended for use with ExternalAddress objects, this code will work (for obscure historical reasons) with plain Byte or Word Arrays as well. " | rcvrClass rcvrSize addr | (interpreterProxy isBytes: rcvr) ifFalse:[^interpreterProxy primitiveFail]. (byteOffset > 0) ifFalse:[^interpreterProxy primitiveFail]. rcvrClass := interpreterProxy fetchClassOf: rcvr. rcvrSize := interpreterProxy byteSizeOf: rcvr. rcvrClass = interpreterProxy classExternalAddress ifTrue:[ (rcvrSize = 4) ifFalse:[^interpreterProxy primitiveFail]. addr := interpreterProxy fetchPointer: 0 ofObject: rcvr. "don't you dare to read from object memory!!" + (addr == 0 "or:[interpreterProxy isInMemory: addr]") - (addr == 0 or:[interpreterProxy isInMemory: addr]) ifTrue:[^interpreterProxy primitiveFail]. ] ifFalse:[ (byteOffset+byteSize-1 <= rcvrSize) ifFalse:[^interpreterProxy primitiveFail]. addr := self cCoerce: (interpreterProxy firstIndexableField: rcvr) to: 'int'. ]. addr := addr + byteOffset - 1. ^addr! Item was changed: ----- Method: ThreadedFFIPlugin>>ffiAtomicArgByReference:Class:in: (in category 'callout support') ----- ffiAtomicArgByReference: oop Class: oopClass in: calloutState <var: #calloutState type: #'CalloutState *'> "Support for generic callout. Prepare a pointer reference to an atomic type for callout. + Note: + for type 'void*' we allow ByteArray/String/Symbol, wordVariableSubclass, Alien or ExternalAddress. + for other types we allow ByteArray, wordVariableSubclass, Alien or ExternalAddress." - Note: for type 'void*' we allow ByteArray/String/Symbol, wordVariableSubclass or Alien." | atomicType isString isAlien | <inline: true> atomicType := self atomicTypeOf: calloutState ffiArgHeader. (atomicType = FFITypeBool) ifTrue: "No bools on input" [^FFIErrorCoercionFailed]. isAlien := (isString := interpreterProxy includesBehavior: oopClass ThatOf: interpreterProxy classString) ifTrue: [false] ifFalse: [interpreterProxy includesBehavior: oopClass ThatOf: interpreterProxy classAlien]. ((atomicType >> 1) = (FFITypeSignedChar >> 1)) ifTrue:"string value (char*)" "note: the only types allowed for passing into char* types are ByteArray, String, Symbol, Alien and *no* other byte indexed objects (e.g., CompiledMethod, LargeInteger). We only check for strings here and fall through to the byte* check otherwise." [isString ifTrue:"String/Symbol" "Strings must be allocated by the ffi support code" [^self ffiPushString: (interpreterProxy firstIndexableField: oop) OfLength: (interpreterProxy byteSizeOf: oop) in: calloutState]. "Fall through to byte* test" atomicType := FFITypeUnsignedByte]. self cppIf: COGMTVM ifTrue: ["Since all the following pass the address of the first indexable field we need to fail the call if it is threaded and the object is young, since it may move during the call." ((calloutState callFlags anyMask: FFICallFlagThreaded) and: [(isAlien not or: [self isDirectAlien: oop]) and: [interpreterProxy isYoung: oop]]) ifTrue: [^PrimErrObjectMayMove negated]]. (atomicType = FFITypeVoid or:[(atomicType >> 1) = (FFITypeSignedByte >> 1)]) ifTrue: "byte* -- see comment on string above" + [(isString or: [oopClass = interpreterProxy classByteArray]) ifTrue: "String/Symbol/ByteArray" - [(isString - or: [oopClass = interpreterProxy classByteArray]) ifTrue:"String/Symbol/ByteArray" [^self ffiPushPointer: (interpreterProxy firstIndexableField: oop) in: calloutState]. + (oopClass = interpreterProxy classExternalAddress) ifTrue: + [^self ffiPushPointer: (self longAt: oop + interpreterProxy baseHeaderSize) in: calloutState]. + isAlien ifTrue: - (isAlien or: [oopClass = interpreterProxy classExternalAddress]) ifTrue: [^self ffiPushPointer: (self pointerForOop: (self startOfData: oop)) in: calloutState]. atomicType = FFITypeVoid ifFalse: [^FFIErrorCoercionFailed]]. "note: type void falls through" + "I can push pointers to any type (take for instance calls who receive int* output arguments, etc.) + but I need to store them into a ByteArray, ExternalAddress or Alien" + (atomicType <= FFITypeDoubleFloat) ifTrue: + [((interpreterProxy isWords: oop) or: [oopClass = interpreterProxy classByteArray]) ifTrue: + [^self ffiPushPointer: (interpreterProxy firstIndexableField: oop) in: calloutState]. + (oopClass = interpreterProxy classExternalAddress) ifTrue: + [^self ffiPushPointer: (self longAt: oop + interpreterProxy baseHeaderSize) in: calloutState]. + isAlien ifTrue: + [^self ffiPushPointer: (self pointerForOop: (self startOfData: oop)) in: calloutState]]. - (atomicType <= FFITypeSignedInt "void/short/int" - or:[atomicType = FFITypeSingleFloat]) ifTrue: - ["require a word subclass to work" - (interpreterProxy isWords: oop) ifTrue: - [^self ffiPushPointer: (interpreterProxy firstIndexableField: oop) in: calloutState]]. ^FFIErrorCoercionFailed!

1 0

My first OSX 64 bits stack spur VM is running
by Nicolas Cellier 29 Oct '15

29 Oct '15

Hello, just FYI, it's now possible :)

3 6

clang analyzer detects potential usage of a null pointer
by Nicolas Cellier 29 Oct '15

29 Oct '15

Hi again, here is another simple patch for platforms/iOS/vm/OSX/sqMacUnixExternalPrims.m: a NULL pluginName might be used if for some reason the call to dlopen would fail...

1 0

objc prefix header macros
by Nicolas Cellier 29 Oct '15

29 Oct '15

In platforms/iOS/vm/SqueakPureObjc_Prefix.pch, shouldn't the macros be defined #if ... #define RELEASEOBJ(x) #else #define RELEASEOBJ(x) [x release] #endif same for autorelease and retain because I currently have compilation errors with ARC enabled and clang 7.0 It also seems that #define SUPERDEALLOC is missing in the ARC branch. Thanks

2 3

ARM64
by Ben Coman 28 Oct '15

28 Oct '15

Not sure if this is useful, but just some articles I bumped into... Migrating code from ARM to ARM64 https://www.linuxplumbersconf.org/2014/ocw/system/presentations/2343/origin… ARM64 vs ARM32 -- What's different for Linux programmers http://www.edn.com/design/systems-design/4440662/3/ARM64-vs-ARM32-What-s-di… cheers -ben

1 0

iOS VM problems with dragging a morph and PasteUpMorph>>flashRects:color:
by Javier Diaz-Reinoso 27 Oct '15

27 Oct '15

As i reported in the Squeak list I find a bug with dragging a morph and in PasteUpMorph>>flashRects:color:, I find the problem is with the implementation of DisplayScreen>>forceDisplayUpdate in the VM, specifically in ioForceDisplayUpdateActual in sqSqueakScreenAndWindow.m, the actual implementation call drawTheLayers using a queue: dispatch_async(dispatch_get_main_queue(), ^{ [[self getMainView] drawThelayers]; }); I find this is unreliable, calling directly to drawTheLayers the bug is corrected, I believe that this is the correct operation because what you want is an immediate update otherwise the “timing” of the operation is lost, like in PasteUpMorph>>flashRects:color:, between the first and second call to copyBits are only 15 milliseconds, I suppose is the same with the drag of a morph. I don’t know if this change create new problems, appears to be working OK in an iPad2 (32 bits) and a iPad mini (64 bits). By the way I tested also using [SqueakUIViewOpenGL class] instead of [SqueakUIViewCALayer class] in whatRenderCanWeUse, that works OK in the iPad2 but create a small image in the iPad mini, I think is not taking into account the 2X of the retina display.

2 2

attempting to build BochsX64Plugin on 32-bit Ubuntu
by Robert Withers 26 Oct '15

26 Oct '15

I am not sure it is possible, even. Can this plugin be build 32-bit; does it make any sense to do so? Here is the make portion of this attempt for this plugin. I am unsure how to deal with this C++ issue in the Cross/plugins/BochsX64Plugin/sqUnixBochsX64Plugin.cpp file. /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp: In function 'int resetCPU(void*)': /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:68:47: error: 'struct bx_descriptor_t::<anonymous union>::<anonymous>' has no member named 'l' bx_cpu.sregs[BX_SEG_REG_CS].cache.u.segment.l = 1; // 64-bit seg Thank you, Robert /bin/bash /home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build/libtool --mode=compile gcc -m32 -g -O2 -DNDEBUG -DDEBUGVM=0 -msse2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DCOGMTVM=0 -DLSB_FIRST=1 -m32 -DHAVE_CONFIG_H -I/home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build -I/home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build -I/home/rabbit/warren/oscogvm/platforms/unix/vm -I/home/rabbit/warren/oscogvm/platforms/Cross/vm -I/home/rabbit/warren/oscogvm/spursrc/vm -I/home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin -m32 -DUSE_BOCHS_CONFIG_H=1 -I/home/rabbit/warren/oscogvm/processors/IA32/linuxbochs -I/home/rabbit/warren/oscogvm/processors/IA32/bochs -I/home/rabbit/warren/oscogvm/processors/IA32/bochs/instrument/stubs -c -o BochsX64Plugin.lo /home/rabbit/warren/oscogvm/src/plugins/BochsX64Plugin/BochsX64Plugin.c gcc -m32 -g -O2 -DNDEBUG -DDEBUGVM=0 -msse2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DCOGMTVM=0 -DLSB_FIRST=1 -m32 -DHAVE_CONFIG_H -I/home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build -I/home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build -I/home/rabbit/warren/oscogvm/platforms/unix/vm -I/home/rabbit/warren/oscogvm/platforms/Cross/vm -I/home/rabbit/warren/oscogvm/spursrc/vm -I/home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin -m32 -DUSE_BOCHS_CONFIG_H=1 -I/home/rabbit/warren/oscogvm/processors/IA32/linuxbochs -I/home/rabbit/warren/oscogvm/processors/IA32/bochs -I/home/rabbit/warren/oscogvm/processors/IA32/bochs/instrument/stubs -c /home/rabbit/warren/oscogvm/src/plugins/BochsX64Plugin/BochsX64Plugin.c -fPIC -DPIC -DPIC -o BochsX64Plugin.o mv -f BochsX64Plugin.o BochsX64Plugin.lo /bin/bash /home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build/libtool --mode=compile g++ -g -O2 -DNDEBUG -DDEBUGVM=0 -msse2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DCOGMTVM=0 -DLSB_FIRST=1 -m32 -DHAVE_CONFIG_H -I/home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build -I/home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build -I/home/rabbit/warren/oscogvm/platforms/unix/vm -I/home/rabbit/warren/oscogvm/platforms/Cross/vm -I/home/rabbit/warren/oscogvm/spursrc/vm -I/home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin -m32 -DUSE_BOCHS_CONFIG_H=1 -I/home/rabbit/warren/oscogvm/processors/IA32/linuxbochs -I/home/rabbit/warren/oscogvm/processors/IA32/bochs -I/home/rabbit/warren/oscogvm/processors/IA32/bochs/instrument/stubs -c -o sqBochsX64Plugin.lo /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp g++ -g -O2 -DNDEBUG -DDEBUGVM=0 -msse2 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DCOGMTVM=0 -DLSB_FIRST=1 -m32 -DHAVE_CONFIG_H -I/home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build -I/home/rabbit/warren/oscogvm/build.linux32x86/squeak.cog.spur/build -I/home/rabbit/warren/oscogvm/platforms/unix/vm -I/home/rabbit/warren/oscogvm/platforms/Cross/vm -I/home/rabbit/warren/oscogvm/spursrc/vm -I/home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin -m32 -DUSE_BOCHS_CONFIG_H=1 -I/home/rabbit/warren/oscogvm/processors/IA32/linuxbochs -I/home/rabbit/warren/oscogvm/processors/IA32/bochs -I/home/rabbit/warren/oscogvm/processors/IA32/bochs/instrument/stubs -c /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp -fPIC -DPIC -DPIC -o sqBochsX64Plugin.o /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:17:0: warning: "setjmp" redefined [enabled by default] # define setjmp(jb) _setjmp(jb) ^ In file included from /home/rabbit/warren/oscogvm/processors/IA32/bochs/gui/siminterface.h:1069:0, from /home/rabbit/warren/oscogvm/processors/IA32/bochs/bochs.h:117, from /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:7: /usr/include/setjmp.h:67:0: note: this is the location of the previous definition #define setjmp(env) _setjmp (env) ^ /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp: In function 'int resetCPU(void*)': /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:68:47: error: 'struct bx_descriptor_t::<anonymous union>::<anonymous>' has no member named 'l' bx_cpu.sregs[BX_SEG_REG_CS].cache.u.segment.l = 1; // 64-bit seg ^ /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:74:47: error: 'struct bx_descriptor_t::<anonymous union>::<anonymous>' has no member named 'l' bx_cpu.sregs[BX_SEG_REG_DS].cache.u.segment.l = 1; // 64-bit seg ^ /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:80:47: error: 'struct bx_descriptor_t::<anonymous union>::<anonymous>' has no member named 'l' bx_cpu.sregs[BX_SEG_REG_SS].cache.u.segment.l = 1; // 64-bit seg ^ /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:101:10: error: 'class bx_cpu_c' has no member named 'efer' bx_cpu.efer.set_LMA(1); /* Hack. The old version we use have doesn't support set_EFER */ ^ /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp: In function 'int disassembleForAtInSize(void*, ulong, void*, ulong)': /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:213:52: error: 'struct bx_descriptor_t::<anonymous union>::<anonymous>' has no member named 'l' anx64->sregs[BX_SEG_REG_CS].cache.u.segment.l, ^ /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp: At global scope: /home/rabbit/warren/oscogvm/platforms/Cross/plugins/BochsX64Plugin/sqBochsX64Plugin.cpp:324:39: error: no 'void bx_cpu_c::TLB_flushNonGlobal()' member function declared in class 'bx_cpu_c' void BX_CPU_C::TLB_flushNonGlobal(void) ^ make[1]: *** [sqBochsX64Plugin.lo] Error 1 make: *** [BochsX64Plugin.la] Error 2 rabbit@rabbithole:~/warren/oscogvm/build.linux32x86/squeak.cog.spur/build$ mv -f BochsX64Plugin.o BochsX64Plugin.lo mv: cannot stat ‘BochsX64Plugin.o’: No such file or directory

2 11

[commit][3480] Add the new Bochs plugin build structure to the build. linux32x86 directory.
by commits＠squeakvm.org 26 Oct '15

26 Oct '15

Revision: 3480 Author: eliot Date: 2015-10-25 21:05:32 -0700 (Sun, 25 Oct 2015) Log Message: ----------- Add the new Bochs plugin build structure to the build.linux32x86 directory. Add Clement's ubuntu build config info to the linux HowToBuild. Modified Paths: -------------- branches/Cog/build.linux32x86/HowToBuild branches/Cog/build.linux32x86/bochsx64/conf.COG branches/Cog/build.linux32x86/bochsx64/conf.COG.dbg branches/Cog/build.linux32x86/bochsx86/conf.COG branches/Cog/platforms/unix/plugins/BochsIA32Plugin/Makefile.inc Added Paths: ----------- branches/Cog/build.linux32x86/bochsx64/ branches/Cog/build.linux32x86/bochsx86/ branches/Cog/platforms/unix/plugins/BochsX64Plugin/ branches/Cog/platforms/unix/plugins/BochsX64Plugin/Makefile.inc Modified: branches/Cog/build.linux32x86/HowToBuild =================================================================== --- branches/Cog/build.linux32x86/HowToBuild 2015-10-21 23:18:32 UTC (rev 3479) +++ branches/Cog/build.linux32x86/HowToBuild 2015-10-26 04:05:32 UTC (rev 3480) @@ -244,3 +244,15 @@ sudo apt-get install libpango1.0-dev More advice and examples for other distros gratefully received. + + +Clement Bera's Ubuntu install script +sudo apt-get install cmake zip bash-completion ruby git xz-utils debhelper devscripts +sudo apt-get install libc6-dev:i386 libasound2:i386 libasound2-dev:i386 libasound2-plugins:i386 libssl-dev:i386 libssl0.9.8:i386 libfreetype6-dev:i386 libx11-dev:i386 libsm-dev:i386 libice-dev:i386 +sudo apt-get install build-essential gcc-multilib g++ +sudo apt-get install libgl1-mesa-dev libgl1-mesa-glx:i386 +sudo apt-get install binutils:i386 g++-4.6:i386 gcc-4.6:i386 +sudo ln -s /usr/lib/i386-linux-gnu/mesa/libGL.so /usr/lib/i386-linux-gnu/libGL.so +sudo ln -s /usr/lib/i386-linux-gnu/mesa/libGL.so.1 /usr/lib/i386-linux-gnu/mesa/libGL.so +sudo apt-get install uuid-dev:i386 +sudo apt-get install subversion Modified: branches/Cog/build.linux32x86/bochsx64/conf.COG =================================================================== --- branches/Cog/build.macos32x86/bochsx64/conf.COG 2015-10-21 23:18:32 UTC (rev 3479) +++ branches/Cog/build.linux32x86/bochsx64/conf.COG 2015-10-26 04:05:32 UTC (rev 3480) @@ -1,7 +1,7 @@ #!/bin/sh # Bochs x64 simulator for the BochsX64Alien plugin. -# this sets up the x64 compile for Cog on Mac OS X. Disable as much inessential +# this sets up the x64 compile for Cog on linux. Disable as much inessential # stuff as possible leaving only the cpu/fpu & memory interface # build cpu/libcpu.a disasm/libdisasm.a fpu/libfpu.a using Modified: branches/Cog/build.linux32x86/bochsx64/conf.COG.dbg =================================================================== --- branches/Cog/build.macos32x86/bochsx64/conf.COG.dbg 2015-10-21 23:18:32 UTC (rev 3479) +++ branches/Cog/build.linux32x86/bochsx64/conf.COG.dbg 2015-10-26 04:05:32 UTC (rev 3480) @@ -1,7 +1,7 @@ #!/bin/sh # Bochs x64 simulator for the BochsX64Alien plugin. -# this sets up the x64 compile for Cog on Mac OS X. Disable as much inessential +# this sets up the x64 compile for Cog on linux. Disable as much inessential # stuff as possible leaving only the cpu/fpu & memory interface # build cpu/libcpu.a disasm/libdisasm.a fpu/libfpu.a using Modified: branches/Cog/build.linux32x86/bochsx86/conf.COG =================================================================== --- branches/Cog/build.macos32x86/bochsx86/conf.COG 2015-10-21 23:18:32 UTC (rev 3479) +++ branches/Cog/build.linux32x86/bochsx86/conf.COG 2015-10-26 04:05:32 UTC (rev 3480) @@ -1,7 +1,7 @@ #!/bin/sh # Bochs x86 simulator for the BochsIA32Alien plugin. -# this sets up the x86 compile for Cog on Mac OS X. Disable as much inessential +# this sets up the x86 compile for Cog on linux. Disable as much inessential # stuff as possible leaving only the cpu/fpu & memory interface # build cpu/libcpu.a disasm/libdisasm.a fpu/libfpu.a using Modified: branches/Cog/platforms/unix/plugins/BochsIA32Plugin/Makefile.inc =================================================================== --- branches/Cog/platforms/unix/plugins/BochsIA32Plugin/Makefile.inc 2015-10-21 23:18:32 UTC (rev 3479) +++ branches/Cog/platforms/unix/plugins/BochsIA32Plugin/Makefile.inc 2015-10-26 04:05:32 UTC (rev 3480) @@ -1,11 +1,11 @@ XCFLAGS=-m32 XCPPFLAGS= -m32 -DUSE_BOCHS_CONFIG_H=1 \ - -I$(topdir)/processors/IA32/linuxbochs \ + -I../../../bochsx32 \ -I$(topdir)/processors/IA32/bochs \ -I$(topdir)/processors/IA32/bochs/instrument/stubs -XLDFLAGS= -L$(topdir)/processors/IA32/linuxbochs/cpu \ - -L$(topdir)/processors/IA32/linuxbochs/fpu \ - -L$(topdir)/processors/IA32/linuxbochs/disasm \ +XLDFLAGS= -L../../../bochsx32/cpu \ + -L../../../bochsx32/fpu \ + -L../../../bochsx32/disasm \ -lcpu -lfpu -ldisasm \ -lstdc++ Added: branches/Cog/platforms/unix/plugins/BochsX64Plugin/Makefile.inc =================================================================== --- branches/Cog/platforms/unix/plugins/BochsX64Plugin/Makefile.inc (rev 0) +++ branches/Cog/platforms/unix/plugins/BochsX64Plugin/Makefile.inc 2015-10-26 04:05:32 UTC (rev 3480) @@ -0,0 +1,11 @@ +XCFLAGS=-m32 +XCPPFLAGS= -m32 -DUSE_BOCHS_CONFIG_H=1 \ + -I../../../bochsx64 \ + -I$(topdir)/processors/IA32/bochs \ + -I$(topdir)/processors/IA32/bochs/instrument/stubs + +XLDFLAGS= -L../../../bochsx64/cpu \ + -L../../../bochsx64/fpu \ + -L../../../bochsx64/disasm \ + -lcpu -lfpu -ldisasm \ + -lstdc++

1 0

inlining constant with branch elimination is aggressive
by Nicolas Cellier 25 Oct '15

25 Oct '15

Example in shortPrintContext: (self isMachineCodeFrame: (theFP := self frameOfMarriedContext: aContext)) ifTrue: [self print: ' M ('] ifFalse: [self print: ' I (']. self printHex: theFP asUnsignedInteger; print: ') ' in a stack vm, isMachineCodeFrame: is unconditionnally false, so theFP never gets a chance to be initialized... It's potentially dagerous... Either we change the generator to generate the side effects, or we manually change the code.

2 2

64 bits VM: how sign extension may strike
by Nicolas Cellier 25 Oct '15

25 Oct '15

Hi, one potential problem with 64 bits is for example the (usqInt) cast when shifting >> Example: (((usqInt) sourceWord) >> 24) will correctly extract the 8 high bit is sourceWord is an unsigned int. if source word is an declared as int, then we will get a sign extension (usqInt) 0xF7654321 -> 0xFFFFFFFFF7654321UL If with don't bitAnd: 0xFF the result, it ain't gonna meet our expectations. That rarely happen that we declare int, by default it's rather sqInt. But things like sqInt foo = long32At( oop , index ); will also perform a sign extension... Even if (((usqInt) foo) >>24) doesn't perform the sign extension by itself, we get the same behaviour at the end. If we carefully place a 'unsigned int' type hint in a method, it is thus vital that this type be preserved with inlining. When I analyze the code, most long32At could in fact be replaced with unsignedLong32At. If tthat didn't make much difference in 32 bits, it seems more vital to me with 64 bits sqInt...

1 0

← Newer
1
2
3
4
5
6
7
8
9
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev October 2015