Vm-dev February 2018

vm-dev@lists.squeakfoundation.org

27 participants
131 discussions

FFI Struct Argument Pass By Value Fails on Mac 64 bit
by Todd Blanchard 06 Mar '18

06 Mar '18

I've been trying to track this down for a couple weeks now. I have concluded that structs passed by value to functions on the 64 bit VM are not properly populated. The struct's memory is all zero'd. I found this while trying to work with LibClang and found that functions that fetched code locations from code ranges always returned invalid zero'd locations. After spending some time with lldb I have traced the problem into the native code and found that the argument is not correct. I've carved out the wee bit of clang to reproduce this in a tiny library. The gist of it is below and the entire file is included. Basically the struct passed to the function clang_getRangeStart is zero'd memory regardless of the data I send from the image side. The build command I used on sierra is clang -shared -undefined dynamic_lookup -o microclang.dylib microclang.c I'm stuck. The FFI64 plugin is way beyond my comprehension. // microclang.c typedef struct { const void *ptr_data[2]; unsigned int_data; } CXSourceLocation; /** * \brief Identifies a half-open character range in the source code. * * Use clang_getRangeStart() and clang_getRangeEnd() to retrieve the * starting and end locations from a source range, respectively. */ typedef struct { const void *ptr_data[2]; unsigned begin_int_data; unsigned end_int_data; } CXSourceRange; const char* first = "first_pointer"; const char* second = "second_pointer"; // return a fake range with non zero data CXSourceRange clang_getArbitraryRange() { CXSourceRange range = {0}; range.ptr_data[0] = (void*)first; range.ptr_data[1] = (void*)second; range.begin_int_data = 17; range.end_int_data = 24; return range; } // Actual clang function - range is always zero'd here despite having values in the image CXSourceLocation clang_getRangeStart(CXSourceRange range) { // Special decoding for CXSourceLocations for CXLoadedDiagnostics. if ((uintptr_t)range.ptr_data[0] & 0x1) { CXSourceLocation Result = { { range.ptr_data[0], nullptr }, 0 }; return Result; } CXSourceLocation Result = { { range.ptr_data[0], range.ptr_data[1] }, range.begin_int_data }; return Result; }

3 12

Fwd: [Pharo-dev] Pharo 7 Repeatable crash: XIO: fatal IO error 14 (Bad address) on X server ":0"
by Martin McClure 02 Mar '18

02 Mar '18

Anyone here interested in this crash? Is there a newer VM I should test with? Thanks, -Martin -------- Forwarded Message -------- Subject: [Pharo-dev] Pharo 7 Repeatable crash: XIO: fatal IO error 14 (Bad address) on X server ":0" Date: Wed, 21 Feb 2018 10:37:18 -0800 From: Martin McClure <martin(a)hand2mouse.com> Reply-To: Pharo Development List <pharo-dev(a)lists.pharo.org> To: Pharo Development List <pharo-dev(a)lists.pharo.org> On the current Linux32 Pharo 7.0, (https://get.pharo.org/70+vm as of 30 minutes ago) the VM crashes with the message XIO: fatal IO error 14 (Bad address) on X server ":0" whenever I try to enlarge the Pharo window by dragging the corner, and get to a size of about 2Mpixels. This is with a virgin image with no windows open inside the main window. A workaround seems to be to open a System Browser before resizing the main window. This does not reproduce on Pharo 6.1 (https://get.pharo.org/). This *does* reproduce when running the Pharo7 image on the Pharo 6.1 VM, so it may not be entirely a VM problem. Most times, no crash dump is created, and nothing is logged to PharoDebug.log. Occasionally, especially when rapidly changing the size of the window, it segfaults and logs a dump. This may or may not be the same problem. This is less annoying since I found the workaround, but still seems to be something that should be fixed. Does anyone want further information from me in order to fix it? Regards, -Martin

3 7

[64 bits] Object pointers in jitted code
by Javier Pimás 02 Mar '18

02 Mar '18

Hi! This time I'm investigating how cog jit handles pointers to objects in native code. In x86-32 its easier because you have immediates of the size of a pointer, but in x64 the immediates are restricted to 32bits (and I think less in arm). So I wonder how people works around that, if using a movabs instruction every time you need a pointer or if doing something else. I found a mail in the list dated from 2011 (titled "questions about cog internals") where you (Eliot) said that pointers were inlined in jit code, but I don't know if that's still the case. Looking at the slang code I found CogOutOfLineLiteralsX64Compiler, but it seems it is not used (yet?). Cheers! Pocho -- Javier Pimás Ciudad de Buenos Aires

4 10

[OpenSmalltalk/opensmalltalk-vm] 4b58ed: deploy needs to use filter-exec.sh
by GitHub 28 Feb '18

28 Feb '18

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 4b58edcae0b8da8eda977815ed12e190ea0df616 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/4b58edcae0b8da8eda… Author: Esteban Lorenzano <estebanlm(a)gmail.com> Date: 2018-02-28 (Wed, 28 Feb 2018) Changed paths: M .travis.yml Log Message: ----------- deploy needs to use filter-exec.sh

1 0

[OpenSmalltalk/opensmalltalk-vm] b41228: try to fix pharo deploy by correctly verifying FLA...
by GitHub 28 Feb '18

28 Feb '18

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: b412289650944404712b68440133e7b297f91e71 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/b41228965094440471… Author: Esteban Lorenzano <estebanlm(a)gmail.com> Date: 2018-02-28 (Wed, 28 Feb 2018) Changed paths: M deploy/pharo/filter-exec.sh Log Message: ----------- try to fix pharo deploy by correctly verifying FLAVOR

1 0

VM Maker: VMMaker.oscog-eem.2339.mcz
by commits＠source.squeak.org 27 Feb '18

27 Feb '18

Eliot Miranda uploaded a new version of VMMaker to project VM Maker: http://source.squeak.org/VMMaker/VMMaker.oscog-eem.2339.mcz ==================== Summary ==================== Name: VMMaker.oscog-eem.2339 Author: eem Time: 26 February 2018, 4:41:13.643544 pm UUID: 1f823336-4229-4929-8a75-545797124625 Ancestors: VMMaker.oscog-eem.2338 Immutability: Fix primitiveSlotAtPut (& InterpreterPrimitives>>primitiveInstVarAtPut) to fail with PrimErrorNoModification for read-only objects. =============== Diff against VMMaker.oscog-eem.2338 =============== Item was changed: ----- Method: InterpreterPrimitives>>primitiveInstVarAtPut (in category 'object access primitives') ----- primitiveInstVarAtPut | newValue index rcvr hdr fmt totalLength fixedFields | newValue := self stackTop. index := self stackValue: 1. rcvr := self stackValue: 2. ((objectMemory isNonIntegerObject: index) or: [argumentCount > 2 "e.g. object:instVarAt:put:" and: [objectMemory isOopForwarded: rcvr]]) ifTrue: [^self primitiveFailFor: PrimErrBadArgument]. + (objectMemory isImmediate: rcvr) ifTrue: + [^self primitiveFailFor: PrimErrBadReceiver]. + (objectMemory isObjImmutable: rcvr) ifTrue: + [^self primitiveFailFor: PrimErrNoModification]. index := objectMemory integerValueOf: index. hdr := objectMemory baseHeader: rcvr. fmt := objectMemory formatOfHeader: hdr. totalLength := objectMemory lengthOf: rcvr baseHeader: hdr format: fmt. fixedFields := objectMemory fixedFieldsOf: rcvr format: fmt length: totalLength. (index >= 1 and: [index <= fixedFields]) ifFalse: [^self primitiveFailFor: PrimErrBadIndex]. self subscript: rcvr with: index storing: newValue format: fmt. self pop: argumentCount + 1 thenPush: newValue! Item was changed: ----- Method: InterpreterPrimitives>>primitiveSlotAtPut (in category 'object access primitives') ----- primitiveSlotAtPut "Assign a slot in an object. This numbers all slots from 1, ignoring the distinction between named and indexed inst vars. In objects with both named and indexed inst vars, the named inst vars precede the indexed ones. In non-object indexed objects (objects that contain bits, not object references) this primitive assigns a raw integral value at each slot." | newValue index rcvr fmt numSlots value | newValue := self stackTop. index := self stackValue: 1. rcvr := self stackValue: 2. (objectMemory isIntegerObject: index) ifFalse: [^self primitiveFailFor: PrimErrBadArgument]. (objectMemory isImmediate: rcvr) ifTrue: [^self primitiveFailFor: PrimErrBadReceiver]. + (objectMemory isObjImmutable: rcvr) ifTrue: + [^self primitiveFailFor: PrimErrNoModification]. fmt := objectMemory formatOf: rcvr. index := (objectMemory integerValueOf: index) - 1. fmt <= objectMemory lastPointerFormat ifTrue: [numSlots := objectMemory numSlotsOf: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storePointer: index ofObject: rcvr withValue: newValue. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. value := self positiveMachineIntegerValueOf: newValue. self failed ifTrue: [primFailCode := PrimErrBadArgument. ^0]. fmt >= objectMemory firstByteFormat ifTrue: [fmt >= objectMemory firstCompiledMethodFormat ifTrue: [^self primitiveFailFor: PrimErrUnsupported]. (self asUnsigned: value) > 16rFF ifTrue: [^self primitiveFailFor: PrimErrBadArgument]. numSlots := objectMemory numBytesOfBytes: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storeByte: index ofObject: rcvr withValue: value. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. (objectMemory hasSpurMemoryManagerAPI and: [fmt >= objectMemory firstShortFormat]) ifTrue: [(self asUnsigned: value) > 16rFFFF ifTrue: [^self primitiveFailFor: PrimErrBadArgument]. numSlots := objectMemory num16BitUnitsOf: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storeShort16: index ofObject: rcvr withValue: value. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. (objectMemory bytesPerOop = 8 and: [fmt = objectMemory sixtyFourBitIndexableFormat]) ifTrue: [numSlots := objectMemory num64BitUnitsOf: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storeLong64: index ofObject: rcvr withValue: value. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. fmt >= objectMemory firstLongFormat ifTrue: [(objectMemory wordSize > 4 and: [(self asUnsigned: value) > 16rFFFFFFFF]) ifTrue: [^self primitiveFailFor: PrimErrBadArgument]. numSlots := objectMemory num32BitUnitsOf: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storeLong32: index ofObject: rcvr withValue: value. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. ^self primitiveFailFor: PrimErrBadReceiver! Item was changed: ----- Method: StackInterpreterPrimitives>>primitiveSlotAtPut (in category 'object access primitives') ----- primitiveSlotAtPut "Assign a slot in an object. This numbers all slots from 1, ignoring the distinction between named and indexed inst vars. In objects with both named and indexed inst vars, the named inst vars precede the indexed ones. In non-object indexed objects (objects that contain bits, not object references) this primitive assigns a raw integral value at each slot." + | newValue index rcvr fmt numSlots value | - | newValue index rcvr fmt numSlots value badRcvr | newValue := self stackTop. index := self stackValue: 1. rcvr := self stackValue: 2. (objectMemory isIntegerObject: index) ifFalse: [^self primitiveFailFor: PrimErrBadArgument]. + (objectMemory isImmediate: rcvr) ifTrue: - self cppIf: IMMUTABILITY - ifTrue: [ badRcvr := objectMemory isOopImmutable: rcvr ] - ifFalse: [ badRcvr := objectMemory isImmediate: rcvr ]. - badRcvr ifTrue: [^self primitiveFailFor: PrimErrBadReceiver]. + (objectMemory isObjImmutable: rcvr) ifTrue: + [^self primitiveFailFor: PrimErrNoModification]. + fmt := objectMemory formatOf: rcvr. index := (objectMemory integerValueOf: index) - 1. fmt <= objectMemory lastPointerFormat ifTrue: [numSlots := objectMemory numSlotsOf: rcvr. (self asUnsigned: index) < numSlots ifTrue: [(objectMemory isContextNonImm: rcvr) ifTrue: [self externalInstVar: index ofContext: rcvr put: newValue] ifFalse: [objectMemory storePointer: index ofObject: rcvr withValue: newValue]. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. value := self positiveMachineIntegerValueOf: newValue. self failed ifTrue: [primFailCode := PrimErrBadArgument. ^0]. fmt >= objectMemory firstByteFormat ifTrue: [fmt >= objectMemory firstCompiledMethodFormat ifTrue: [^self primitiveFailFor: PrimErrUnsupported]. (self asUnsigned: value) > 16rFF ifTrue: [^self primitiveFailFor: PrimErrBadArgument]. numSlots := objectMemory numBytesOfBytes: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storeByte: index ofObject: rcvr withValue: value. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. (objectMemory hasSpurMemoryManagerAPI and: [fmt >= objectMemory firstShortFormat]) ifTrue: [(self asUnsigned: value) > 16rFFFF ifTrue: [^self primitiveFailFor: PrimErrBadArgument]. numSlots := objectMemory num16BitUnitsOf: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storeShort16: index ofObject: rcvr withValue: value. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. (objectMemory bytesPerOop = 8 and: [fmt = objectMemory sixtyFourBitIndexableFormat]) ifTrue: [numSlots := objectMemory num64BitUnitsOf: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storeLong64: index ofObject: rcvr withValue: value. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. fmt >= objectMemory firstLongFormat ifTrue: [(objectMemory wordSize > 4 and: [(self asUnsigned: value) > 16rFFFFFFFF]) ifTrue: [^self primitiveFailFor: PrimErrBadArgument]. numSlots := objectMemory num32BitUnitsOf: rcvr. (self asUnsigned: index) < numSlots ifTrue: [objectMemory storeLong32: index ofObject: rcvr withValue: value. self pop: argumentCount + 1 thenPush: newValue. ^0]. ^self primitiveFailFor: PrimErrBadIndex]. ^self primitiveFailFor: PrimErrBadReceiver!

1 0

[OpenSmalltalk/opensmalltalk-vm] 443be0: Update filter-exec.sh [ci skip]
by GitHub 26 Feb '18

26 Feb '18

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 443be0c044f88e2400f7e9289f420cebfac347c8 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/443be0c044f88e2400… Author: Fabio Niephaus <code(a)fniephaus.com> Date: 2018-02-26 (Mon, 26 Feb 2018) Changed paths: M deploy/pharo/filter-exec.sh Log Message: ----------- Update filter-exec.sh [ci skip]

1 0

[OpenSmalltalk/opensmalltalk-vm] 9debee: deploy to files.pharo.org just if flavour is pharo...
by GitHub 26 Feb '18

26 Feb '18

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 9debee039700a2c243442475cc04bc3221261cba https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/9debee039700a2c243… Author: Esteban Lorenzano <estebanlm(a)gmail.com> Date: 2018-02-26 (Mon, 26 Feb 2018) Changed paths: M deploy/pharo/filter-exec.sh Log Message: ----------- deploy to files.pharo.org just if flavour is pharo.cog.spur (no sub-flavours for now)

1 0

Re: [Vm-dev] [Pharo-dev] OSSubprocess 0.2.5 fails in Pharo 6.1 64 bits (macOS)
by Eliot Miranda 26 Feb '18

26 Feb '18

Hi Alistair, On Wed, Feb 21, 2018 at 9:31 AM, Alistair Grant <akgrant0710(a)gmail.com> wrote: > Yep, OSSubprocess doesn't work on 64 bit VMs at the moment. > > Eliot, Mariano has indicated he's waiting on > https://github.com/pharo-project/pharo-vm/pull/142 > Do you know of a reason why it shouldn't be merged? > Well, the changes to FilePlugin.class belong in VMMaker.oscog. And cfileRecordSize doesn't make sense , since all pointers are of the same size in C: cfileRecordSize + "Return the size of a stdio FILE* handle" + <option: #PharoVM> + <static: false> + ^self sizeof: #'FILE*' I would simple use (self sizeof: #'void *') or (self sizeof: #'FILE *'). Other than that I can't see any problems. > Thanks, > Alistair > > > > On 21 February 2018 at 18:22, phideaux <jcasler(a)cdix.org> wrote: > > When trying the simple example (ls -la /Users) in Pharo 6.1 64bit you > get a > > talkback because > > ExternalAddress integerAt:put:size:signed: fails using > > 'primitiveFFIIntegerAtPut' in module 'SqueakFFIPrims' > > > > Something to do with OSSubprocess not accommodating 64 bit pointers > perhaps? > > > > Same code works fine on 32bit version. > > > > VM is CoInterpreter VMMaker.oscog-eem.2254 Jul 20 2017 for macOS > downloaded > > from pharo.org/downloaded page > > > > Regards, > > Jay+ > > > > > > > > -- > > Sent from: http://forum.world.st/Pharo-Smalltalk-Developers-f1294837. > html > > > > -- _,,,^..^,,,_ best, Eliot

4 5

Re: [Vm-dev] Segfaults from #copyStack
by Max Leske 25 Feb '18

25 Feb '18

Thanks Eliot. On 24 February 2018 at 22:15:21, vm-dev-request(a)lists.squeakfoundation.org ( vm-dev-request(a)lists.squeakfoundation.org) wrote: Hi Max, On Sat, Feb 24, 2018 at 1:07 PM <http://airmail.calendar/2018-02-24 13:07:00 CET>, Max Leske <maxleske(a)gmail.com> wrote: > > Hi, > > I'm seeing many segfaults (around 20 to 30% of the time) when running Fuel > builds on Linux. Everytime I see it it's caused by #copyStack (e.g. > https://travis-ci.org/theseion/Fuel/jobs/345741310). AFAICT it happens > more often (exclusively?) on 32 bit with Pharo 7 (latest VM + latest image). > Try and construct a test case that crashes without user interaction. i.e. something that starts up with a known image, supplying a script to load via Fuel. Once this crashes reliably debugging is easy. Otherwise its very hard. As soon as you have the crashing test case let me know and I'll take a look. Ideally it would run on Mac OS X. I've tried for the last two hours to come up with a reliable way of triggering the segfault but haven't had any luck. I can only reproduce it on Linux, not macOS and I've only been able to force the segfault with using smalltalkCI: git clone https://github.com/hpi-swa/smalltalkCI.git && git clone https://github.com/theseion/Fuel.git && smalltalkCI/run.sh -s Pharo-alpha Fuel/.default.ston I used Ubuntu 17.04 (32-bit), TravisCI uses 14.04.5 (64-bit). Maybe there's a way I can extract some information you can use? And then I'll just play mole whacking until I trigger it... :) Cheers, Max

2 1

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev February 2018