Vm-dev August 2016

vm-dev@lists.squeakfoundation.org

33 participants
98 discussions

[OpenSmalltalk/opensmalltalk-vm] 02232a: Add HostWindowPlugin to all plugins.int for Linux
by GitHub 20 Aug '16

20 Aug '16

1 0

[OpenSmalltalk/opensmalltalk-vm] edeea4: Add HostWindowPlugin to all plugins.int for Linux
by GitHub 19 Aug '16

19 Aug '16

1 0

HostWindowPlugin for Linux
by Fabio Niephaus 19 Aug '16

19 Aug '16

Tim mentioned that the HostWindowPlugin is missing in the plugins.int for all Linux builds. Is there a particular reason why? Should we add it? Best, Fabio -- ---------- Forwarded message --------- From: tim Rowledge <tim(a)rowledge.org> Date: Thu, Aug 18, 2016 at 1:00 AM Subject: Re: [Vm-dev] Pi vm builds need a priority checker applet To: Squeak Virtual Machine Development Discussion < vm-dev(a)lists.squeakfoundation.org> > On 17-08-2016, at 3:34 PM, Fabio Niephaus <lists(a)fniephaus.com> wrote: > > > On Thu, Aug 18, 2016 at 12:17 AM tim Rowledge <tim(a)rowledge.org> wrote: > > A persistent problem with the PI vm is that users of xrdp need to wrap the start-up in ‘sudo -E’ because.. well, never mind, unix is just plain stupid. Whatever the issue, we can test for the problem of needing to use sudo by running a trivial test program in the shell script; all it does is try changing a thread priority and returns whether it worked or not. > > Just to make sure: it is not enough to create a squeak.conf as describe at [1] and reboot? Unfortunately that doesn’t work when using xrdp for your remote display. It’s not even a strictly squeak issue - some googling will reveal a depressingly long list of problems to do with the same error. The Raspbian kernel; has no problems at all with squeak when using a direct display. It doesn’t have this particular problem with vnc - but it *does* have a fatal problem with trying to use the composition input window calls. Again, not a specifically squeak, raspbian nor ARM issue but a generic bit of nastiness in linux. > Could you please give [2] a try on a raspi? Feedback is more than welcome as we are about to release Squeak 5.1 very soon. As long as it’s made from the very latest source Eliot put out today it will be ok. *Except* I note that the HostWindowPlugin is missing from the plugins/int file in build.linuxARMv6/squeak.cog.spur and in fact after a quick look it is missing form all the linux related build directories that I checked. Whilst the plugin is not complete it does provide support for title and size aspects and so I’d recommend including it. 5.1 will be the basis of the next Pi Scratch release so I have several millions of reasons to help getting it right... tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Don't sweat petty things....or pet sweaty things.

2 1

[OpenSmalltalk/opensmalltalk-vm] 614fad: Make the pc offset longer in sqWin32Backtrace, jus...
by GitHub 19 Aug '16

19 Aug '16

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 614fadf2fdb1270428e0763ca6bc4fa1cc1dfc84 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/614fadf2fdb1270428… Author: Nicolas Cellier <nicolas.cellier(a)sirehna.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/win32/vm/sqWin32Backtrace.c M platforms/win32/vm/sqWin32Backtrace.h Log Message: ----------- Make the pc offset longer in sqWin32Backtrace, just in case Commit: 57fa38ee5fde3dd87fb83c628f6b09c6b46d2957 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/57fa38ee5fde3dd87f… Author: Nicolas Cellier <nicolas.cellier(a)sirehna.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/win32/plugins/IA32ABI/Makefile M platforms/win32/plugins/IA32ABI/Makefile.plugin Log Message: ----------- Don't hardcode the IA32 ABI on windows (well, despite the directory name...) Commit: 546cd55c8f05e228973d1b0e7ee0513e979ba1b5 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/546cd55c8f05e22897… Author: Nicolas Cellier <nicolas.cellier(a)sirehna.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M build.win64x64/squeak.stack.spur/plugins.ext M build.win64x64/squeak.stack.spur/plugins.int A platforms/Cross/plugins/IA32ABI/dax64win64business.h M platforms/Cross/plugins/IA32ABI/ia32abi.h A platforms/Cross/plugins/IA32ABI/x64win64ia32abicc.c M platforms/Cross/plugins/IA32ABI/xabicc.c Log Message: ----------- Attempt a first x64 win64 FFI plugin Warning: MOSTLY UNTESTED Implementation note: do not use the loadFloatRegs(...) fake call for loading floating point registers xmm0...xmm3 Instead, simply and dumbly switch on the 16 possible combinations of int/float register parameters The reasons for this choice are: -1) I don't understand how loadFloatRegs would work with x64-win64 ABI I believe the caller should save/restore the registers before/after loadFloatRegs call - 2) be more robust versus aggressive compiler optimizations and eventually link time optimizations (loadFloatRegs needs to be in a separate file, but LTO might spoil this) - 3) it's simple enough to encode (even if a bit copy/paste error prone) For thunkEntry, use few asm instructions (REMINDER: have to replace asm with some intrinsics for MSVC - see https://software.intel.com/sites/landingpage/IntrinsicsGuide/#cats=Load&exp… ) Conflicts: src/plugins/SqueakFFIPrims/X64Win64FFIPlugin.c Commit: c2d0de1fe75adf823026a3643c2fe9f6012e59a3 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/c2d0de1fe75adf8230… Author: Nicolas Cellier <nicolas.cellier(a)sirehna.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/Cross/plugins/IA32ABI/x64win64ia32abicc.c Log Message: ----------- Fix win64 x64 compilation 2 problems here: - erroneous usage of asm directives - it was necessary to republish SqueakFFIPrims for correction of WIN64 #ifdef Commit: 0af004c7a34df823f37ef178a37dcc07a4c7cdc9 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/0af004c7a34df823f3… Author: Nicolas Cellier <nicolas.cellier(a)sirehna.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/Cross/plugins/IA32ABI/arm32ia32abicc.c M platforms/Cross/plugins/IA32ABI/ia32abi.h M platforms/Cross/plugins/IA32ABI/ia32abicc.c M platforms/Cross/plugins/IA32ABI/ppc32abicc.c M platforms/Cross/plugins/IA32ABI/x64ia32abicc.c M platforms/Cross/plugins/IA32ABI/x64win64ia32abicc.c Log Message: ----------- Make sure that the stack pointer points to large enough words in thunkEntry This is a fix for LLP64 (x64 win64) Words are large enough when == sizeof(void *) Indeed, we want to skip other return address and things like that... Commit: 5f7a7505f6d14a77eed5788f2aa61baf119fa0e9 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/5f7a7505f6d14a77ee… Author: Nicolas Cellier <nicolas.cellier(a)sirehna.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M build.win64x64/squeak.stack.spur/plugins.ext Log Message: ----------- Temporarily revert compilation of FFI for win64 Changes of sources generated by VMMaker.oscog are necessary for compilation to succeed. Compare: https://github.com/OpenSmalltalk/opensmalltalk-vm/compare/5160d1803ca8...5f…

1 0

Pi vm builds need a priority checker applet
by tim Rowledge 19 Aug '16

19 Aug '16

A persistent problem with the PI vm is that users of xrdp need to wrap the start-up in ‘sudo -E’ because.. well, never mind, unix is just plain stupid. Whatever the issue, we can test for the problem of needing to use sudo by running a trivial test program in the shell script; all it does is try changing a thread priority and returns whether it worked or not. So far as I know this is only an issue on the Pi but then I don’t use an x86 linux machine. I’d imagine any machine with an older kernel might need the check too but others will have to say. If it is of use to general linux machines then it is pretty obvious that we should put the source in platforms/unix/vm and add the build to the general make process. It if is really only a Pi issue then maybe someone has a strong view on having the source/build elsewhere. Any thoughts? tim -- tim Rowledge; tim(a)rowledge.org; http://www.rowledge.org/tim Strange OpCodes: SDP: Search and Destroy Pointer

3 16

[OpenSmalltalk/opensmalltalk-vm] a11109: Fix generation of left shift in FFI plugin
by GitHub 18 Aug '16

18 Aug '16

Branch: refs/heads/LLP64 Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: a1110961aebca3e43b1241e372b0982f15029493 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/a1110961aebca3e43b… Author: nicolas-cellier-aka-nice <nicolas.cellier.aka.nice(a)gmail.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M src/plugins/SqueakFFIPrims/ARM32FFIPlugin.c M src/plugins/SqueakFFIPrims/IA32FFIPlugin.c M src/plugins/SqueakFFIPrims/X64SysVFFIPlugin.c M src/plugins/SqueakFFIPrims/X64Win64FFIPlugin.c Log Message: ----------- Fix generation of left shift in FFI plugin If applied on 32 bits constants in a 64bits VM, left shift might overflow (or have undefined behavior)

1 0

64bits Alien testSignedLong failing
by Nicolas Cellier 18 Aug '16

18 Aug '16

The test fails with -2^31 and 2^31-1 It's not so amazing, because when I debug, I find this incorrect version: 64769 sqInt 64770 signed32BitIntegerFor(sqInt integerValue) 64771 { -> 64772 return (((((int) integerValue)) << 3) | 1); 64773 64774 } 64775 of course, the shift will overflow for large positive or large negative 32bits int. But I don't understand how it's been generated... I would expect return ((((usqInt)((int) integerValue)) << 3) | 1); Indeed: signed32BitIntegerFor: integerValue "Answer a full 32 bit integer object for the given integer value. N.B. Returning in each arm separately enables Slang inlining. /Don't/ return the ifTrue:ifFalse: unless Slang inlining of conditionals is fixed." <inline: true> objectMemory hasSixtyFourBitImmediates ifTrue: [^objectMemory integerObjectOf: (self cCode: [self cCoerceSimple: integerValue to: #int] inSmalltalk: [(integerValue bitAnd: 16r7FFFFFFF) - ((integerValue >> 31 anyMask: 1) ifTrue: [-16r100000000] ifFalse: [0])])] ifFalse: [^self noInlineSigned32BitIntegerFor: integerValue] has the cast to int. Then there is another cast here: integerObjectOf: value "Convert the integer value, assumed to be in SmallInteger range, into a tagged SmallInteger object. In C, use a shift and an add to set the tag bit. In Smalltalk we have to work harder because the simulator works with strictly positive bit patterns." <returnTypeC: #sqInt> ^self cCode: [value asUnsignedInteger << self numTagBits + 1] inSmalltalk: [value << self numTagBits + (value >= 0 ifTrue: [1] ifFalse: [16r10000000000000001])] and asUnsignedInteger is translated like this by CCodeGenerator: generateAsUnsignedInteger: msgNode on: aStream indent: level "Generate the C code for this message onto the given stream." aStream nextPutAll:'((usqInt)'. self emitCExpression: msgNode receiver on: aStream. aStream nextPut: $) How the hell could the hardcoded (usqInt) be not generated?

1 1

[OpenSmalltalk/opensmalltalk-vm] 5160d1: Fix a regression in FFI (can't pass single float o...
by GitHub 18 Aug '16

18 Aug '16

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 5160d1803ca8a4b60914ed193a0f89dc6629a7b9 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/5160d1803ca8a4b609… Author: Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/Cross/vm/sqMemoryAccess.h Log Message: ----------- Fix a regression in FFI (can't pass single float on the stack) The macros storeSingleFloatAtPointerfrom and fetchSingleFloatAtPointerinto are used with double precision values. We cannot naively memcpy without first copying into a single float...

1 0

[OpenSmalltalk/opensmalltalk-vm] 4122e5: Fix minor slip in register printing (crash.dmp)
by GitHub 18 Aug '16

18 Aug '16

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 4122e5904c40b59d977cd8947859625d9b4c6fbe https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/4122e5904c40b59d97… Author: Nicolas Cellier <nicolas.cellier(a)sirehna.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/win32/vm/sqWin32Main.c Log Message: ----------- Fix minor slip in register printing (crash.dmp) Commit: c48471433388082da052a607e539583aedf65717 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/c48471433388082da0… Author: nicolas-cellier-aka-nice <nicolas.cellier.aka.nice(a)gmail.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M build.win64x64/makeall M build.win64x64/makeallinstall M build.win64x64/makeproduct M build.win64x64/makeproductinstall Log Message: ----------- Add the -e flags to build.win64x64 bash files Commit: 9efdcb734d06caba1a2de31f98869eb7887e14b3 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/9efdcb734d06caba1a… Author: Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/Cross/plugins/IA32ABI/arm32ia32abicc.c M platforms/Cross/plugins/IA32ABI/x64ia32abicc.c Log Message: ----------- Authorize usage of Squeak types in vmCallback For example, vmCallback might rely on sqIntptr_t for LLP64 compatibility. The order of include files must be revised in order to do so This change has been applied to ia32abicc.c cf. https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/336c3d5ba394eb2812… (336c3d5) It must also be applied to other supported ABI (arm32, x64) Note sure what is needed for ppc32abicc... Commit: cd8ac277e8a98a35348f7575ef2c4c778c59632c https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/cd8ac277e8a98a3534… Author: Nicolas Cellier <nicolas.cellier(a)sirehna.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/win32/plugins/SocketPlugin/sqWin32NewNet.c M platforms/win32/vm/sqWin32Window.c Log Message: ----------- Cosmetic: remove some compiler warning Commit: 4c5dc3d051e976fea8c388e093d0561a6a832305 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/4c5dc3d051e976fea8… Author: nicolas-cellier-aka-nice <nicolas.cellier.aka.nice(a)gmail.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/win32/plugins/SocketPlugin/sqWin32NewNet.c Log Message: ----------- cosmetic: remove a few warnings Commit: 332e52a3cbdffad551532989e8afa1eaf5669560 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/332e52a3cbdffad551… Author: nicolas-cellier-aka-nice <nicolas.cellier.aka.nice(a)gmail.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/win32/plugins/FilePlugin/sqWin32File.h Log Message: ----------- remove duplicated alloca definition Commit: ca215bc1fcd5c42fb76e0eee67cbaac199fe20e2 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/ca215bc1fcd5c42fb7… Author: nicolas-cellier-aka-nice <nicolas.cellier.aka.nice(a)gmail.com> Date: 2016-08-18 (Thu, 18 Aug 2016) Changed paths: M platforms/win32/vm/sqWin32Main.c Log Message: ----------- correct return type of findEmbeddedImage Compare: https://github.com/OpenSmalltalk/opensmalltalk-vm/compare/2491466dc8ba...ca…

1 0

[OpenSmalltalk/opensmalltalk-vm] 249146: Provide a ioScreenScaleFactor() for the legacy 'Ma...
by GitHub 17 Aug '16

17 Aug '16

Branch: refs/heads/Cog Home: https://github.com/OpenSmalltalk/opensmalltalk-vm Commit: 2491466dc8ba513f224dcc451177b097f7f37c85 https://github.com/OpenSmalltalk/opensmalltalk-vm/commit/2491466dc8ba513f22… Author: Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com> Date: 2016-08-17 (Wed, 17 Aug 2016) Changed paths: M platforms/Mac OS/vm/sqMacWindow.h M platforms/Mac OS/vm/sqMacWindowUniversal.c Log Message: ----------- Provide a ioScreenScaleFactor() for the legacy 'Mac OS' platform

1 0

← Newer
1
2
3
4
5
6
7
8
9
10
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev August 2016