Quantcast

VM Maker: VMMaker.oscog-rsf.2083.mcz

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

VM Maker: VMMaker.oscog-rsf.2083.mcz

commits-2
 
Ronie Salgado Faila uploaded a new version of VMMaker to project VM Maker:
http://source.squeak.org/VMMaker/VMMaker.oscog-rsf.2083.mcz

==================== Summary ====================

Name: VMMaker.oscog-rsf.2083
Author: rsf
Time: 11 January 2017, 4:42:00.330997 am
UUID: 2debebfc-5008-4ab3-b16d-37ab942d9bc0
Ancestors: VMMaker.oscog-eem.2082

Workaround a GCC crash in Windows when building a Lowcode VM. Too much register allocation pressure for calling a builtin memcpy.

=============== Diff against VMMaker.oscog-eem.2082 ===============

Item was changed:
  ----- Method: StackInterpreter>>internalPushShadowCallStackStructure:size: (in category 'internal interpreter access') -----
  internalPushShadowCallStackStructure: structurePointer size: size
  <option: #LowcodeVM>
  shadowCallStackPointer := shadowCallStackPointer - size.
+ self lowcode_mem: shadowCallStackPointer cp: structurePointer y: size!
- self mem: shadowCallStackPointer cp: structurePointer y: size!

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveInt32ToPointer (in category 'inline primitive generated code') -----
  lowcodePrimitiveInt32ToPointer
  <option: #LowcodeVM> "Lowcode instruction generator"
  | value result |
  <var: #value type: #'sqInt' >
  <var: #result type: #'char*' >
  value := self internalPopStackInt32.
 
+ result := self cCoerce: (self cCoerce: value to: 'uintptr_t') to: 'char*'.
- result := self cCoerce: value to: 'uintptr_t'.
 
  self internalPushPointer: result.
 
  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpy32 (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpy32
  <option: #LowcodeVM> "Lowcode instruction generator"
  | source dest size |
  <var: #source type: #'char*' >
  <var: #dest type: #'char*' >
  <var: #size type: #'sqInt' >
  size := self internalPopStackInt32.
  source := self internalPopStackPointer.
  dest := self internalPopStackPointer.
 
+ self lowcode_mem: dest cp: source y: size.
- self mem: dest cp: source y: size.
 
 
  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpy64 (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpy64
  <option: #LowcodeVM> "Lowcode instruction generator"
  | source dest size |
  <var: #source type: #'char*' >
  <var: #dest type: #'char*' >
  <var: #size type: #'sqLong' >
  size := self internalPopStackInt64.
  source := self internalPopStackPointer.
  dest := self internalPopStackPointer.
 
+ self lowcode_mem: dest cp: source y: size.
- self mem: dest cp: source y: size.
 
 
  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpyFixed (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpyFixed
  <option: #LowcodeVM> "Lowcode instruction generator"
  | source size dest |
  <var: #source type: #'char*' >
  <var: #dest type: #'char*' >
  size := extA.
  source := self internalPopStackPointer.
  dest := self internalPopStackPointer.
 
+ self lowcode_mem: dest cp: source y: size.
- self mem: dest cp: source y: size.
 
  extA := 0.
 
  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitivePerformCallStructure (in category 'inline primitive generated code') -----
  lowcodePrimitivePerformCallStructure
  <option: #LowcodeVM> "Lowcode instruction generator"
  | resultPointer result function structureSize |
  <var: #resultPointer type: #'char*' >
  <var: #result type: #'char*' >
  function := extA.
  structureSize := extB.
  result := self internalPopStackPointer.
 
  self internalPushShadowCallStackPointer: result.
  resultPointer := self lowcodeCalloutPointerResult: (self cCoerce: function to: #'char*').
 
  self internalPushPointer: resultPointer.
  extA := 0.
  extB := 0.
  numExtB := 0.
+
  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitivePointerAddConstantOffset (in category 'inline primitive generated code') -----
  lowcodePrimitivePointerAddConstantOffset
  <option: #LowcodeVM> "Lowcode instruction generator"
  | base offset result |
  <var: #base type: #'char*' >
  <var: #result type: #'char*' >
  offset := extB.
  base := self internalPopStackPointer.
 
  result := base + offset.
 
  self internalPushPointer: result.
  extB := 0.
  numExtB := 0.
 
  !

Item was added:
+ ----- Method: StackInterpreter>>lowcode_mem:cp:y: (in category 'inline primitive support') -----
+ lowcode_mem: destAddress cp: sourceAddress y: bytes
+ "This method is a workaround a GCC bug.
+ In Windows memcpy is putting too much register pressure on GCC when used by Lowcode instructions"
+ <inline: #never>
+ <option: #LowcodeVM>
+ <var: #destAddress type: #'void*'>
+ <var: #sourceAddress type: #'void*'>
+ <var: #bytes type: #'sqInt'>
+
+ "Using memmove instead of memcpy to avoid crashing GCC in Windows."
+ self mem: destAddress mo: sourceAddress ve: bytes!

Item was changed:
  ----- Method: StackToRegisterMappingCogit>>genLowcodePerformCallStructure (in category 'inline primitive generators generated code') -----
  genLowcodePerformCallStructure
  <option: #LowcodeVM> "Lowcode instruction generator"
 
  "Push the result space"
  self ssNativeTop nativeStackPopToReg: TempReg.
  self ssNativePop: 1.
  self PushR: TempReg.
  "Call the function"
  self callSwitchToCStack.
  self MoveCw: extA R: TempReg.
  self CallRT: ceFFICalloutTrampoline.
  "Fetch the result"
  self MoveR: backEnd cResultRegister R: ReceiverResultReg.
  self ssPushNativeRegister: ReceiverResultReg.
  extA := 0.
  extB := 0.
  numExtB := 0.
 
  ^ 0
 
  !

Item was changed:
  ----- Method: StackToRegisterMappingCogit>>genLowcodePointerAddConstantOffset (in category 'inline primitive generators generated code') -----
  genLowcodePointerAddConstantOffset
  <option: #LowcodeVM> "Lowcode instruction generator"
  | base offset |
  offset := extB.
 
  (base := backEnd availableRegisterOrNoneFor: self liveRegisters) = NoReg ifTrue:
  [self ssAllocateRequiredReg:
  (base := optStatus isReceiverResultRegLive
  ifTrue: [Arg0Reg]
  ifFalse: [ReceiverResultReg])].
  base = ReceiverResultReg ifTrue:
  [ optStatus isReceiverResultRegLive: false ].
  self ssNativeTop nativePopToReg: base.
  self ssNativePop: 1.
 
  self AddCq: offset R: base.
  self ssPushNativeRegister: base.
 
  extB := 0.
  numExtB := 0.
  ^ 0
 
  !

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: VM Maker: VMMaker.oscog-rsf.2083.mcz

Eliot Miranda-2
 
Hi Ronie,

    I see this :-)

lowcode_mem: destAddress cp: sourceAddress y: bytes
"This method is a workaround a GCC bug.
In Windows memcpy is putting too much register pressure on GCC when used by Lowcode instructions"
<inline: #never>
<option: #LowcodeVM>
<var: #destAddress type: #'void*'>
<var: #sourceAddress type: #'void*'>
<var: #bytes type: #'sqInt'>
"Using memmove instead of memcpy to avoid crashing GCC in Windows."
self mem: destAddress mo: sourceAddress ve: bytes

Isn't it great when one has to work around compiler bugs??  ;-)

However, let me suggest that this this is perhaps a case where a macro would be better.  If you added the definition of the macro to StackInterpreter class>>#preambleCCode you'd be able to avoid the overhead with compilers that can correctly inline memcpy.  If required, a sqPlatform.h could define a value, say DontInlineMemcpyForLowcode and then in the preamble you could have

#if DontInlineMemcpyForLowcode
# define memcpy(a,b,c) noinline_memcpy(a,b,c)
#endif

?  And then noinline_memcpy could be defined in some platform support file, sqWin32Main.c perhaps?  The simulator's noinline_memcpy would be defined as <doNotGenerate>.

Anyway, some way of making this platform-dependent is nice as you'll get better performance on the other platforms, and on x64.

And yes, feel free to ignore me as this perhaps does count as a premature optimization.

On Tue, Jan 10, 2017 at 11:42 PM, <[hidden email]> wrote:

Ronie Salgado Faila uploaded a new version of VMMaker to project VM Maker:
http://source.squeak.org/VMMaker/VMMaker.oscog-rsf.2083.mcz

==================== Summary ====================

Name: VMMaker.oscog-rsf.2083
Author: rsf
Time: 11 January 2017, 4:42:00.330997 am
UUID: 2debebfc-5008-4ab3-b16d-37ab942d9bc0
Ancestors: VMMaker.oscog-eem.2082

Workaround a GCC crash in Windows when building a Lowcode VM. Too much register allocation pressure for calling a builtin memcpy.

=============== Diff against VMMaker.oscog-eem.2082 ===============

Item was changed:
  ----- Method: StackInterpreter>>internalPushShadowCallStackStructure:size: (in category 'internal interpreter access') -----
  internalPushShadowCallStackStructure: structurePointer size: size
        <option: #LowcodeVM>
        shadowCallStackPointer := shadowCallStackPointer - size.
+       self lowcode_mem: shadowCallStackPointer cp: structurePointer y: size!
-       self mem: shadowCallStackPointer cp: structurePointer y: size!

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveInt32ToPointer (in category 'inline primitive generated code') -----
  lowcodePrimitiveInt32ToPointer
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | value result |
        <var: #value type: #'sqInt' >
        <var: #result type: #'char*' >
        value := self internalPopStackInt32.

+       result := self cCoerce: (self cCoerce: value to: 'uintptr_t') to: 'char*'.
-       result := self cCoerce: value to: 'uintptr_t'.

        self internalPushPointer: result.

  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpy32 (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpy32
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | source dest size |
        <var: #source type: #'char*' >
        <var: #dest type: #'char*' >
        <var: #size type: #'sqInt' >
        size := self internalPopStackInt32.
        source := self internalPopStackPointer.
        dest := self internalPopStackPointer.

+       self lowcode_mem: dest cp: source y: size.
-       self mem: dest cp: source y: size.


  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpy64 (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpy64
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | source dest size |
        <var: #source type: #'char*' >
        <var: #dest type: #'char*' >
        <var: #size type: #'sqLong' >
        size := self internalPopStackInt64.
        source := self internalPopStackPointer.
        dest := self internalPopStackPointer.

+       self lowcode_mem: dest cp: source y: size.
-       self mem: dest cp: source y: size.


  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpyFixed (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpyFixed
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | source size dest |
        <var: #source type: #'char*' >
        <var: #dest type: #'char*' >
        size := extA.
        source := self internalPopStackPointer.
        dest := self internalPopStackPointer.

+       self lowcode_mem: dest cp: source y: size.
-       self mem: dest cp: source y: size.

        extA := 0.

  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitivePerformCallStructure (in category 'inline primitive generated code') -----
  lowcodePrimitivePerformCallStructure
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | resultPointer result function structureSize |
        <var: #resultPointer type: #'char*' >
        <var: #result type: #'char*' >
        function := extA.
        structureSize := extB.
        result := self internalPopStackPointer.

        self internalPushShadowCallStackPointer: result.
        resultPointer := self lowcodeCalloutPointerResult: (self cCoerce: function to: #'char*').

        self internalPushPointer: resultPointer.
        extA := 0.
        extB := 0.
        numExtB := 0.
+
  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitivePointerAddConstantOffset (in category 'inline primitive generated code') -----
  lowcodePrimitivePointerAddConstantOffset
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | base offset result |
        <var: #base type: #'char*' >
        <var: #result type: #'char*' >
        offset := extB.
        base := self internalPopStackPointer.

        result := base + offset.

        self internalPushPointer: result.
        extB := 0.
        numExtB := 0.

  !

Item was added:
+ ----- Method: StackInterpreter>>lowcode_mem:cp:y: (in category 'inline primitive support') -----
+ lowcode_mem: destAddress cp: sourceAddress y: bytes
+       "This method is a workaround a GCC bug.
+       In Windows memcpy is putting too much register pressure on GCC when used by Lowcode instructions"
+       <inline: #never>
+       <option: #LowcodeVM>
+       <var: #destAddress type: #'void*'>
+       <var: #sourceAddress type: #'void*'>
+       <var: #bytes type: #'sqInt'>
+
+       "Using memmove instead of memcpy to avoid crashing GCC in Windows."
+       self mem: destAddress mo: sourceAddress ve: bytes!

Item was changed:
  ----- Method: StackToRegisterMappingCogit>>genLowcodePerformCallStructure (in category 'inline primitive generators generated code') -----
  genLowcodePerformCallStructure
        <option: #LowcodeVM>    "Lowcode instruction generator"

        "Push the result space"
        self ssNativeTop nativeStackPopToReg: TempReg.
        self ssNativePop: 1.
        self PushR: TempReg.
        "Call the function"
        self callSwitchToCStack.
        self MoveCw: extA R: TempReg.
        self CallRT: ceFFICalloutTrampoline.
        "Fetch the result"
        self MoveR: backEnd cResultRegister R: ReceiverResultReg.
        self ssPushNativeRegister: ReceiverResultReg.
        extA := 0.
        extB := 0.
        numExtB := 0.

        ^ 0

  !

Item was changed:
  ----- Method: StackToRegisterMappingCogit>>genLowcodePointerAddConstantOffset (in category 'inline primitive generators generated code') -----
  genLowcodePointerAddConstantOffset
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | base offset |
        offset := extB.

        (base := backEnd availableRegisterOrNoneFor: self liveRegisters) = NoReg ifTrue:
                [self ssAllocateRequiredReg:
                        (base := optStatus isReceiverResultRegLive
                                ifTrue: [Arg0Reg]
                                ifFalse: [ReceiverResultReg])].
        base = ReceiverResultReg ifTrue:
                [ optStatus isReceiverResultRegLive: false ].
        self ssNativeTop nativePopToReg: base.
        self ssNativePop: 1.

        self AddCq: offset R: base.
        self ssPushNativeRegister: base.

        extB := 0.
        numExtB := 0.
        ^ 0

  !




--
_,,,^..^,,,_
best, Eliot
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: VM Maker: VMMaker.oscog-rsf.2083.mcz

Ronie Salgado
 
Hi Eliot,

Isn't it great when one has to work around compiler bugs??  ;-)
This is very annoying. And for me this is not the first time. It seems that with the heavy inlining in GCC we are putting a bit too much stress on its register allocator.

Isn't it great when one has to work around compiler bugs??  ;-)

However, let me suggest that this this is perhaps a case where a macro would be better.  If you added the definition of the macro to StackInterpreter class>>#preambleCCode you'd be able to avoid the overhead with compilers that can correctly inline memcpy.  If required, a sqPlatform.h could define a value, say DontInlineMemcpyForLowcode and then in the preamble you could have

#if DontInlineMemcpyForLowcode
# define memcpy(a,b,c) noinline_memcpy(a,b,c)
#endif

?  And then noinline_memcpy could be defined in some platform support file, sqWin32Main.c perhaps?  The simulator's noinline_memcpy would be defined as <doNotGenerate>.

Anyway, some way of making this platform-dependent is nice as you'll get better performance on the other platforms, and on x64.

And yes, feel free to ignore me as this perhaps does count as a premature optimization.
 
I was thinking on doing something like this, but I did not knew how to do it because of Slang. Later in some time I will fix it.

Best regards,
Ronie

2017-01-11 16:38 GMT-03:00 Eliot Miranda <[hidden email]>:
Hi Ronie,

    I see this :-)

lowcode_mem: destAddress cp: sourceAddress y: bytes
"This method is a workaround a GCC bug.
In Windows memcpy is putting too much register pressure on GCC when used by Lowcode instructions"
<inline: #never>
<option: #LowcodeVM>
<var: #destAddress type: #'void*'>
<var: #sourceAddress type: #'void*'>
<var: #bytes type: #'sqInt'>
"Using memmove instead of memcpy to avoid crashing GCC in Windows."
self mem: destAddress mo: sourceAddress ve: bytes

Isn't it great when one has to work around compiler bugs??  ;-)

However, let me suggest that this this is perhaps a case where a macro would be better.  If you added the definition of the macro to StackInterpreter class>>#preambleCCode you'd be able to avoid the overhead with compilers that can correctly inline memcpy.  If required, a sqPlatform.h could define a value, say DontInlineMemcpyForLowcode and then in the preamble you could have

#if DontInlineMemcpyForLowcode
# define memcpy(a,b,c) noinline_memcpy(a,b,c)
#endif

?  And then noinline_memcpy could be defined in some platform support file, sqWin32Main.c perhaps?  The simulator's noinline_memcpy would be defined as <doNotGenerate>.

Anyway, some way of making this platform-dependent is nice as you'll get better performance on the other platforms, and on x64.

And yes, feel free to ignore me as this perhaps does count as a premature optimization.

On Tue, Jan 10, 2017 at 11:42 PM, <[hidden email]> wrote:

Ronie Salgado Faila uploaded a new version of VMMaker to project VM Maker:
http://source.squeak.org/VMMaker/VMMaker.oscog-rsf.2083.mcz

==================== Summary ====================

Name: VMMaker.oscog-rsf.2083
Author: rsf
Time: 11 January 2017, 4:42:00.330997 am
UUID: 2debebfc-5008-4ab3-b16d-37ab942d9bc0
Ancestors: VMMaker.oscog-eem.2082

Workaround a GCC crash in Windows when building a Lowcode VM. Too much register allocation pressure for calling a builtin memcpy.

=============== Diff against VMMaker.oscog-eem.2082 ===============

Item was changed:
  ----- Method: StackInterpreter>>internalPushShadowCallStackStructure:size: (in category 'internal interpreter access') -----
  internalPushShadowCallStackStructure: structurePointer size: size
        <option: #LowcodeVM>
        shadowCallStackPointer := shadowCallStackPointer - size.
+       self lowcode_mem: shadowCallStackPointer cp: structurePointer y: size!
-       self mem: shadowCallStackPointer cp: structurePointer y: size!

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveInt32ToPointer (in category 'inline primitive generated code') -----
  lowcodePrimitiveInt32ToPointer
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | value result |
        <var: #value type: #'sqInt' >
        <var: #result type: #'char*' >
        value := self internalPopStackInt32.

+       result := self cCoerce: (self cCoerce: value to: 'uintptr_t') to: 'char*'.
-       result := self cCoerce: value to: 'uintptr_t'.

        self internalPushPointer: result.

  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpy32 (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpy32
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | source dest size |
        <var: #source type: #'char*' >
        <var: #dest type: #'char*' >
        <var: #size type: #'sqInt' >
        size := self internalPopStackInt32.
        source := self internalPopStackPointer.
        dest := self internalPopStackPointer.

+       self lowcode_mem: dest cp: source y: size.
-       self mem: dest cp: source y: size.


  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpy64 (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpy64
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | source dest size |
        <var: #source type: #'char*' >
        <var: #dest type: #'char*' >
        <var: #size type: #'sqLong' >
        size := self internalPopStackInt64.
        source := self internalPopStackPointer.
        dest := self internalPopStackPointer.

+       self lowcode_mem: dest cp: source y: size.
-       self mem: dest cp: source y: size.


  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitiveMemcpyFixed (in category 'inline primitive generated code') -----
  lowcodePrimitiveMemcpyFixed
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | source size dest |
        <var: #source type: #'char*' >
        <var: #dest type: #'char*' >
        size := extA.
        source := self internalPopStackPointer.
        dest := self internalPopStackPointer.

+       self lowcode_mem: dest cp: source y: size.
-       self mem: dest cp: source y: size.

        extA := 0.

  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitivePerformCallStructure (in category 'inline primitive generated code') -----
  lowcodePrimitivePerformCallStructure
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | resultPointer result function structureSize |
        <var: #resultPointer type: #'char*' >
        <var: #result type: #'char*' >
        function := extA.
        structureSize := extB.
        result := self internalPopStackPointer.

        self internalPushShadowCallStackPointer: result.
        resultPointer := self lowcodeCalloutPointerResult: (self cCoerce: function to: #'char*').

        self internalPushPointer: resultPointer.
        extA := 0.
        extB := 0.
        numExtB := 0.
+
  !

Item was changed:
  ----- Method: StackInterpreter>>lowcodePrimitivePointerAddConstantOffset (in category 'inline primitive generated code') -----
  lowcodePrimitivePointerAddConstantOffset
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | base offset result |
        <var: #base type: #'char*' >
        <var: #result type: #'char*' >
        offset := extB.
        base := self internalPopStackPointer.

        result := base + offset.

        self internalPushPointer: result.
        extB := 0.
        numExtB := 0.

  !

Item was added:
+ ----- Method: StackInterpreter>>lowcode_mem:cp:y: (in category 'inline primitive support') -----
+ lowcode_mem: destAddress cp: sourceAddress y: bytes
+       "This method is a workaround a GCC bug.
+       In Windows memcpy is putting too much register pressure on GCC when used by Lowcode instructions"
+       <inline: #never>
+       <option: #LowcodeVM>
+       <var: #destAddress type: #'void*'>
+       <var: #sourceAddress type: #'void*'>
+       <var: #bytes type: #'sqInt'>
+
+       "Using memmove instead of memcpy to avoid crashing GCC in Windows."
+       self mem: destAddress mo: sourceAddress ve: bytes!

Item was changed:
  ----- Method: StackToRegisterMappingCogit>>genLowcodePerformCallStructure (in category 'inline primitive generators generated code') -----
  genLowcodePerformCallStructure
        <option: #LowcodeVM>    "Lowcode instruction generator"

        "Push the result space"
        self ssNativeTop nativeStackPopToReg: TempReg.
        self ssNativePop: 1.
        self PushR: TempReg.
        "Call the function"
        self callSwitchToCStack.
        self MoveCw: extA R: TempReg.
        self CallRT: ceFFICalloutTrampoline.
        "Fetch the result"
        self MoveR: backEnd cResultRegister R: ReceiverResultReg.
        self ssPushNativeRegister: ReceiverResultReg.
        extA := 0.
        extB := 0.
        numExtB := 0.

        ^ 0

  !

Item was changed:
  ----- Method: StackToRegisterMappingCogit>>genLowcodePointerAddConstantOffset (in category 'inline primitive generators generated code') -----
  genLowcodePointerAddConstantOffset
        <option: #LowcodeVM>    "Lowcode instruction generator"
        | base offset |
        offset := extB.

        (base := backEnd availableRegisterOrNoneFor: self liveRegisters) = NoReg ifTrue:
                [self ssAllocateRequiredReg:
                        (base := optStatus isReceiverResultRegLive
                                ifTrue: [Arg0Reg]
                                ifFalse: [ReceiverResultReg])].
        base = ReceiverResultReg ifTrue:
                [ optStatus isReceiverResultRegLive: false ].
        self ssNativeTop nativePopToReg: base.
        self ssNativePop: 1.

        self AddCq: offset R: base.
        self ssPushNativeRegister: base.

        extB := 0.
        numExtB := 0.
        ^ 0

  !




--
_,,,^..^,,,_
best, Eliot

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: VM Maker: VMMaker.oscog-rsf.2083.mcz

Holger Freyther
In reply to this post by Eliot Miranda-2
 

> On 11 Jan 2017, at 20:38, Eliot Miranda <[hidden email]> wrote:
>
> Hi Ronie,
>

Hi Ronie,



> "Using memmove instead of memcpy to avoid crashing GCC in Windows."
> self mem: destAddress mo: sourceAddress ve: bytes


do you have an example C-code to illustrate the issue? I am happy to file a bug report for GCC and do the follow-up.

kind regards
        holger
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: VM Maker: VMMaker.oscog-rsf.2083.mcz

Ronie Salgado
 
Hi Holger,

Try with the attached file. It constains the problematic sources passed through the preprocessor to remove the dependencies on the headers.

../../spurlowcodesrc/vm/gcc3x-cointerp.c:32885:1: error: unable to find a register to spill
 }
 ^
../../spurlowcodesrc/vm/gcc3x-cointerp.c:32885:1: error: este es la insn (this is the instruction):
(insn 13277 83596 83595 1281 (parallel [
            (set (reg:SI 27472 [25785])
                (const_int 0 [0]))
            (set (reg/f:SI 27470 [orig:21149 D.104541 ] [21149])
                (plus:SI (ashift:SI (reg:SI 27472 [25785])
                        (const_int 2 [0x2]))
                    (reg/f:SI 27470 [orig:21149 D.104541 ] [21149])))
            (set (reg/f:SI 27471 [orig:21150 structurePointer ] [21150])
                (plus:SI (ashift:SI (reg:SI 27472 [25785])
                        (const_int 2 [0x2]))
                    (reg/f:SI 27471 [orig:21150 structurePointer ] [21150])))
            (set (mem:BLK (reg/f:SI 27470 [orig:21149 D.104541 ] [21149]) [0  A32])
                (mem:BLK (reg/f:SI 27471 [orig:21150 structurePointer ] [21150]) [0  A8]))
            (use (reg:SI 27472 [25785]))
        ]) ../../spurlowcodesrc/vm/gcc3x-cointerp.c:85260 773 {*rep_movsi}
     (expr_list:REG_UNUSED (reg:SI 27472 [25785])
        (nil)))
../../spurlowcodesrc/vm/gcc3x-cointerp.c:32885: confusión por errores previos, saliendo (confussion due to previous errors, exiting)

Build command line:
i686-w64-mingw32-gcc -x c -MT build/vm/gcc3x-cointerp.o -MMD -MP -MF deps/gcc3x-cointerp.Td -msse2 -ggdb2 -m32 -mno-rtd -mms-bitfields -O2 -march=pentium4 -momit-leaf-frame-pointer -funroll-loops -D_MT -fno-builtin-printf -fno-builtin-putchar -fno-builtin-fprintf -Wall -Wno-unused-variable -Wno-unknown-pragmas -Wno-unused-value -Wno-unused-function -Wno-unused-but-set-variable -I. -I../../spurlowcodesrc/vm -I../../platforms/win32/vm -I../../platforms/Cross/vm  -DSTACK_ALIGN_BYTES=16 -DALLOCA_LIES_SO_USE_GETSP=0 -DCOGMTVM=0 -DDEBUGVM=0 -D_WIN32_WINNT=0x0501 -DWINVER=0x0501 -DWIN32 -DWIN32_FILE_SUPPORT -DNO_ISNAN -DNO_SERVICE -DNO_STD_FILE_SUPPORT -D'TZ="CLST"' -DNDEBUG -D'VM_LABEL(foo)=0'  -DLSB_FIRST -D'VM_NAME="Squeak"' -DSQUEAK_BUILTIN_PLUGIN -DCROQUET -c gcc3x-cointerp.prep.c -o build/vm/gcc3x-cointerp.o

GCC version:
i686-w64-mingw32-gcc --version
i686-w64-mingw32-gcc (GCC) 5.4.0

On Linux and Mac I am not having this problem.

Best regards,
Ronie

2017-01-12 19:12 GMT-03:00 Holger Freyther <[hidden email]>:


> On 11 Jan 2017, at 20:38, Eliot Miranda <[hidden email]> wrote:
>
> Hi Ronie,
>

Hi Ronie,



>       "Using memmove instead of memcpy to avoid crashing GCC in Windows."
>       self mem: destAddress mo: sourceAddress ve: bytes


do you have an example C-code to illustrate the issue? I am happy to file a bug report for GCC and do the follow-up.

kind regards
        holger


gcc3x-cointerp.prep.c.zip (417K) Download Attachment
Loading...