Compiling angelscript on ppc64

135

Author

November 05, 2015 12:35 PM

Hi,

when I try to compile angelscript on a ppc64 machine, I get the following error:

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp: In function 'asQWORD X64_CallFunction(const asQWORD*, int, funcptr_t, asQWORD&, bool)':

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%rcx' in 'asm'

"%rdi", "%rsi", "%rax", "%rdx", "%rcx", "%r8", "%r9", "%r10", "%r11", "%r15");

^

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%rdx' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%rax' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%rsi' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%rdi' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%xmm7' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%xmm6' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%xmm5' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%xmm4' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%xmm3' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%xmm2' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%xmm1' in 'asm'

../sdk/angelscript/source/as_callfunc_x64_gcc.cpp:162:82: error: unknown register name '%xmm0' in 'asm'

ninja: build stopped: subcommand failed.

How does one go about compiling it for this platform?

Att.

--

Rafael Fonseca

WitchLord

4,834

November 05, 2015 08:22 PM

Apparently the configuration in as_config.h is not detecting that you're actually compiling for a PPC64 CPU. Instead it is lead to think you're targeting an AMD64 CPU.

The conditions in as_config.h should, based on the pre-defines provided by the compiler, define the AS_PPC_64 macro, instead it is defining the AS_X64_GCC macro (or perhaps it is defining both for some reason).

I assume you're using a gnuc compiler (or one derived from it). If so you should be able to get the list of pre-defines provided by the compiler with the following command:

echo . | g++ -dM -E -

Add whatever extra compiler arguments you use to target your specific platform.

Show me the output from that command and I can help you try to identify what might be wrong in the as_config.h file for your compiler and platform.

AngelCode.com - game development and more - Reference DB - game developer references
AngelScript - free scripting library - BMFont - free bitmap font generator - Tower - free puzzle game

rfonseca

135

Author

November 05, 2015 08:55 PM

I assume you're using a gnuc compiler (or one derived from it). If so you should be able to get the list of pre-defines provided by the compiler with the following command:

echo . | g++ -dM -E -

Add whatever extra compiler arguments you use to target your specific platform.

Show me the output from that command and I can help you try to identify what might be wrong in the as_config.h file for your compiler and platform.

These are all the flags used: CXXFLAGS='-O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong --param=ssp-buffer-size=4 -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -m64 -fno-strict-aliasing'

This is the output of the command:

#define __DBL_MIN_EXP__ (-1021)

#define __builtin_vsx_xvmsubmsp __builtin_vsx_xvmsubsp

#define __UINT_LEAST16_MAX__ 0xffff

#define __ATOMIC_ACQUIRE 2

#define __FLT_MIN__ 1.17549435082228750796873653722225e-38F

#define __GCC_IEC_559_COMPLEX 2

#define __UINT_LEAST8_TYPE__ unsigned char

#define __INTMAX_C(c) c ## L

#define __CHAR_BIT__ 8

#define _ARCH_PPCGR 1

#define __UINT8_MAX__ 0xff

#define __builtin_vsx_xvnmaddadp __builtin_vsx_xvnmadddp

#define __WINT_MAX__ 0xffffffffU

#define __PPC64__ 1

#define __RECIPF__ 1

#define __builtin_vsx_xxlnor __builtin_vec_nor

#define __ORDER_LITTLE_ENDIAN__ 1234

#define __SIZE_MAX__ 0xffffffffffffffffUL

#define __WCHAR_MAX__ 0x7fffffff

#define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1 1

#define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2 1

#define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 1

#define __DBL_DENORM_MIN__ ((double)4.94065645841246544176568792868221e-324L)

#define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 1

#define __GCC_ATOMIC_CHAR_LOCK_FREE 2

#define __GCC_IEC_559 2

#define __FLT_EVAL_METHOD__ 0

#define __unix__ 1

#define __GCC_ATOMIC_CHAR32_T_LOCK_FREE 2

#define pixel pixel

#define _ARCH_PPCSQ 1

#define bool bool

#define __UINT_FAST64_MAX__ 0xffffffffffffffffUL

#define __SIG_ATOMIC_TYPE__ int

#define __DBL_MIN_10_EXP__ (-307)

#define __FINITE_MATH_ONLY__ 0

#define __GNUC_PATCHLEVEL__ 1

#define __builtin_vsx_xvnmaddasp __builtin_vsx_xvnmaddsp

#define __UINT_FAST8_MAX__ 0xff

#define __has_include(STR) __has_include__(STR)

#define __DEC64_MAX_EXP__ 385

#define __INT8_C(c) c

#define __builtin_vsx_xvnmsubmdp __builtin_vsx_xvnmsubdp

#define __UINT_LEAST64_MAX__ 0xffffffffffffffffUL

#define __SHRT_MAX__ 0x7fff

#define __LDBL_MAX__ 1.79769313486231580793728971405301e+308L

#define __UINT_LEAST8_MAX__ 0xff

#define __GCC_ATOMIC_BOOL_LOCK_FREE 2

#define _ARCH_PPC 1

#define __UINTMAX_TYPE__ long unsigned int

#define __linux 1

#define __DEC32_EPSILON__ 1E-6DF

#define __CHAR_UNSIGNED__ 1

#define __UINT32_MAX__ 0xffffffffU

#define __LDBL_MAX_EXP__ 1024

#define __WINT_MIN__ 0U

#define __builtin_vsx_xvmaddmdp __builtin_vsx_xvmadddp

#define __linux__ 1

#define __builtin_vsx_vperm __builtin_vec_perm

#define __SCHAR_MAX__ 0x7f

#define __WCHAR_MIN__ (-__WCHAR_MAX__ - 1)

#define vector vector

#define __builtin_vsx_xvnmsubmsp __builtin_vsx_xvnmsubsp

#define __INT64_C(c) c ## L

#define __DBL_DIG__ 15

#define __GCC_ATOMIC_POINTER_LOCK_FREE 2

#define __SIZEOF_INT__ 4

#define __SIZEOF_POINTER__ 8

#define __USER_LABEL_PREFIX__

#define __STDC_HOSTED__ 1

#define __LDBL_HAS_INFINITY__ 1

#define _ARCH_PWR4 1

#define __builtin_vsx_xvmaddmsp __builtin_vsx_xvmaddsp

#define __CMODEL_MEDIUM__ 1

#define __FLT_EPSILON__ 1.19209289550781250000000000000000e-7F

#define __LDBL_MIN__ 2.00416836000897277799610805135016e-292L

#define __STDC_UTF_16__ 1

#define __DEC32_MAX__ 9.999999E96DF

#define __INT32_MAX__ 0x7fffffff

#define __SIZEOF_LONG__ 8

#define __STDC_IEC_559__ 1

#define __STDC_ISO_10646__ 201304L

#define __UINT16_C(c) c

#define __DECIMAL_DIG__ 33

#define __gnu_linux__ 1

#define __has_include_next(STR) __has_include_next__(STR)

#define __LDBL_HAS_QUIET_NAN__ 1

#define _CALL_LINUX 1

#define __GNUC__ 5

#define __FLT_HAS_DENORM__ 1

#define __SIZEOF_LONG_DOUBLE__ 16

#define __powerpc64__ 1

#define __BIGGEST_ALIGNMENT__ 16

#define __DBL_MAX__ ((double)1.79769313486231570814527423731704e+308L)

#define __INT_FAST32_MAX__ 0x7fffffffffffffffL

#define __DBL_HAS_INFINITY__ 1

#define __builtin_vsx_xvnmaddmdp __builtin_vsx_xvnmadddp

#define __DEC32_MIN_EXP__ (-94)

#define __builtin_vsx_xxlandc __builtin_vec_andc

#define __INT_FAST16_TYPE__ long int

#define __LDBL_HAS_DENORM__ 1

#define __DEC128_MAX__ 9.999999999999999999999999999999999E6144DL

#define __INT_LEAST32_MAX__ 0x7fffffff

#define __DEC32_MIN__ 1E-95DF

#define __DBL_MAX_EXP__ 1024

#define __DEC128_EPSILON__ 1E-33DL

#define __PTRDIFF_MAX__ 0x7fffffffffffffffL

#define __STDC_NO_THREADS__ 1

#define __STRUCT_PARM_ALIGN__ 16

#define __LONG_LONG_MAX__ 0x7fffffffffffffffLL

#define __SIZEOF_SIZE_T__ 8

#define __builtin_vsx_xvnmaddmsp __builtin_vsx_xvnmaddsp

#define __PPC__ 1

#define __SIZEOF_WINT_T__ 4

#define __GXX_ABI_VERSION 1008

#define __FLT_MIN_EXP__ (-125)

#define __INT_FAST64_TYPE__ long int

#define __FP_FAST_FMAF 1

#define __DBL_MIN__ ((double)2.22507385850720138309023271733240e-308L)

#define __LONGDOUBLE128 1

#define __LP64__ 1

#define __DEC128_MIN__ 1E-6143DL

#define __REGISTER_PREFIX__

#define __UINT16_MAX__ 0xffff

#define __DBL_HAS_DENORM__ 1

#define __builtin_vsx_xxland __builtin_vec_and

#define __UINT8_TYPE__ unsigned char

#define __NO_INLINE__ 1

#define __FLT_MANT_DIG__ 24

#define __VERSION__ "5.1.1 20150618 (Red Hat 5.1.1-4)"

#define __UINT64_C(c) c ## UL

#define _STDC_PREDEF_H 1

#define __BIG_ENDIAN__ 1

#define __GCC_ATOMIC_INT_LOCK_FREE 2

#define __pixel __attribute__((altivec(pixel__))) unsigned short

#define __FLOAT_WORD_ORDER__ __ORDER_BIG_ENDIAN__

#define __STDC_IEC_559_COMPLEX__ 1

#define __INT32_C(c) c

#define __DEC64_EPSILON__ 1E-15DD

#define __ORDER_PDP_ENDIAN__ 3412

#define __DEC128_MIN_EXP__ (-6142)

#define __INT_FAST32_TYPE__ long int

#define __powerpc__ 1

#define __UINT_LEAST16_TYPE__ short unsigned int

#define unix 1

#define __INT16_MAX__ 0x7fff

#define __SIZE_TYPE__ long unsigned int

#define __UINT64_MAX__ 0xffffffffffffffffUL

#define __INT8_TYPE__ signed char

#define __ELF__ 1

#define __HAVE_BSWAP__ 1

#define __FLT_RADIX__ 2

#define __INT_LEAST16_TYPE__ short int

#define __LDBL_EPSILON__ 4.94065645841246544176568792868221e-324L

#define __UINTMAX_C(c) c ## UL

#define __GNUC_RH_RELEASE__ 4

#define __SIG_ATOMIC_MAX__ 0x7fffffff

#define __GCC_ATOMIC_WCHAR_T_LOCK_FREE 2

#define __SIZEOF_PTRDIFF_T__ 8

#define __DEC32_SUBNORMAL_MIN__ 0.000001E-95DF

#define __INT_FAST16_MAX__ 0x7fffffffffffffffL

#define __vector __attribute__((altivec(vector__)))

#define __UINT_FAST32_MAX__ 0xffffffffffffffffUL

#define __UINT_LEAST64_TYPE__ long unsigned int

#define __builtin_vsx_xvmsubadp __builtin_vsx_xvmsubdp

#define __builtin_vsx_xxsel __builtin_vec_sel

#define __FLT_HAS_QUIET_NAN__ 1

#define __bool __attribute__((altivec(bool__))) unsigned

#define __FLT_MAX_10_EXP__ 38

#define __LONG_MAX__ 0x7fffffffffffffffL

#define __DEC128_SUBNORMAL_MIN__ 0.000000000000000000000000000000001E-6143DL

#define __FLT_HAS_INFINITY__ 1

#define __unix 1

#define __UINT_FAST16_TYPE__ long unsigned int

#define __DEC64_MAX__ 9.999999999999999E384DD

#define __CHAR16_TYPE__ short unsigned int

#define __PRAGMA_REDEFINE_EXTNAME 1

#define __INT_LEAST16_MAX__ 0x7fff

#define __DEC64_MANT_DIG__ 16

#define __INT64_MAX__ 0x7fffffffffffffffL

#define __UINT_LEAST32_MAX__ 0xffffffffU

#define __GCC_ATOMIC_LONG_LOCK_FREE 2

#define __INT_LEAST64_TYPE__ long int

#define __INT16_TYPE__ short int

#define __INT_LEAST8_TYPE__ signed char

#define __STDC_VERSION__ 201112L

#define __DEC32_MAX_EXP__ 97

#define _BIG_ENDIAN 1

#define __INT_FAST8_MAX__ 0x7f

#define __INTPTR_MAX__ 0x7fffffffffffffffL

#define linux 1

#define __builtin_vsx_xvmsubasp __builtin_vsx_xvmsubsp

#define __LDBL_MANT_DIG__ 106

#define __DBL_HAS_QUIET_NAN__ 1

#define __SIG_ATOMIC_MIN__ (-__SIG_ATOMIC_MAX__ - 1)

#define __INTPTR_TYPE__ long int

#define __UINT16_TYPE__ short unsigned int

#define __WCHAR_TYPE__ int

#define __SIZEOF_FLOAT__ 4

#define __UINTPTR_MAX__ 0xffffffffffffffffUL

#define __DEC64_MIN_EXP__ (-382)

#define __INT_FAST64_MAX__ 0x7fffffffffffffffL

#define __GCC_ATOMIC_TEST_AND_SET_TRUEVAL 1

#define __RSQRTE__ 1

#define __FLT_DIG__ 6

#define __UINT_FAST64_TYPE__ long unsigned int

#define __INT_MAX__ 0x7fffffff

#define __LONG_DOUBLE_128__ 1

#define __INT64_TYPE__ long int

#define __FLT_MAX_EXP__ 128

#define __builtin_vsx_xxlor __builtin_vec_or

#define __ORDER_BIG_ENDIAN__ 4321

#define __DBL_MANT_DIG__ 53

#define __INT_LEAST64_MAX__ 0x7fffffffffffffffL

#define __GCC_ATOMIC_CHAR16_T_LOCK_FREE 2

#define __builtin_vsx_xxlxor __builtin_vec_xor

#define __DEC64_MIN__ 1E-383DD

#define __WINT_TYPE__ unsigned int

#define __UINT_LEAST32_TYPE__ unsigned int

#define __SIZEOF_SHORT__ 2

#define __LDBL_MIN_EXP__ (-968)

#define __INT_LEAST8_MAX__ 0x7f

#define __SIZEOF_INT128__ 16

#define __LDBL_MAX_10_EXP__ 308

#define __ATOMIC_RELAXED 0

#define __DBL_EPSILON__ ((double)2.22044604925031308084726333618164e-16L)

#define _LP64 1

#define __UINT8_C(c) c

#define __INT_LEAST32_TYPE__ int

#define __SIZEOF_WCHAR_T__ 4

#define __UINT64_TYPE__ long unsigned int

#define __builtin_vsx_xvnmsubadp __builtin_vsx_xvnmsubdp

#define __INT_FAST8_TYPE__ signed char

#define __GNUC_STDC_INLINE__ 1

#define __DBL_DECIMAL_DIG__ 17

#define __STDC_UTF_32__ 1

#define __DEC_EVAL_METHOD__ 2

#define __VEC_ELEMENT_REG_ORDER__ __ORDER_BIG_ENDIAN__

#define __builtin_vsx_xvmaddadp __builtin_vsx_xvmadddp

#define __UINT32_C(c) c ## U

#define __INTMAX_MAX__ 0x7fffffffffffffffL

#define __BYTE_ORDER__ __ORDER_BIG_ENDIAN__

#define __FLT_DENORM_MIN__ 1.40129846432481707092372958328992e-45F

#define _CALL_AIX 1

#define __builtin_vsx_xvnmsubasp __builtin_vsx_xvnmsubsp

#define __INT8_MAX__ 0x7f

#define __UINT_FAST32_TYPE__ long unsigned int

#define __CHAR32_TYPE__ unsigned int

#define __FLT_MAX__ 3.40282346638528859811704183484517e+38F

#define __FP_FAST_FMA 1

#define __INT32_TYPE__ int

#define __SIZEOF_DOUBLE__ 8

#define __FLT_MIN_10_EXP__ (-37)

#define _CALL_AIXDESC 1

#define __INTMAX_TYPE__ long int

#define __DEC128_MAX_EXP__ 6145

#define __ATOMIC_CONSUME 1

#define __GNUC_MINOR__ 1

#define __UINTMAX_MAX__ 0xffffffffffffffffUL

#define __DEC32_MANT_DIG__ 7

#define __builtin_vsx_xvmaddasp __builtin_vsx_xvmaddsp

#define __DBL_MAX_10_EXP__ 308

#define __LDBL_DENORM_MIN__ 4.94065645841246544176568792868221e-324L

#define __INT16_C(c) c

#define __STDC__ 1

#define __builtin_vsx_xvmsubmdp __builtin_vsx_xvmsubdp

#define __PTRDIFF_TYPE__ long int

#define __ATOMIC_SEQ_CST 5

#define __UINT32_TYPE__ unsigned int

#define __UINTPTR_TYPE__ long unsigned int

#define __DEC64_SUBNORMAL_MIN__ 0.000000000000001E-383DD

#define _ARCH_PPC64 1

#define __DEC128_MANT_DIG__ 34

#define __LDBL_MIN_10_EXP__ (-291)

#define __SIZEOF_LONG_LONG__ 8

#define _Bool _Bool

#define __GCC_ATOMIC_LLONG_LOCK_FREE 2

#define __LDBL_DIG__ 31

#define __FLT_DECIMAL_DIG__ 9

#define __UINT_FAST16_MAX__ 0xffffffffffffffffUL

#define __GCC_ATOMIC_SHORT_LOCK_FREE 2

#define __UINT_FAST8_TYPE__ unsigned char

#define _CALL_ELF 1

#define __ATOMIC_ACQ_REL 4

#define __ATOMIC_RELEASE 3

WitchLord

4,834

November 06, 2015 12:10 AM

Alright. Based on this I can see that as_config.h is indeed identifying the platform as a Linux with x86/64 CPU.

Please make the following changes to the file and let me know how it works out for you:

On line 847 add a condition not to enter the x86/64 code path like this:

#elif defined(__LP64__) && !defined(__arm64__) && !defined(__PPC64__)

On line 914 add the following before the #else, to create a specific configuration for Linux with PPC64 processors:

#elif defined(__PPC64)

// Support native calling conventions on Linux with PPC64

#define AS_PPC_64

#define SPLIT_OBJS_BY_MEMBER_TYPES

#define THISCALL_RETURN_SIMPLE_IN_MEMORY

#define CDECL_RETURN_SIMPLE_IN_MEMORY

#define STDCALL_RETURN_SIMPLE_IN_MEMORY

#undef STDCALL

#define STDCALL

#else

I'm hoping the Linux calling convention for PPC64 is close enough to what the PlayStation 3 uses that this change will be enough to get native calling conventions working on your platform. If it isn't then we'll need to figure out what the differences are and make the necessary changes in as_callfunc_ppc_64.cpp. Another option would be to try the XBox 360 in as_callfunc_xenon.cpp file, but I sincerely doubt GNUC and Microsoft would use the same ABI.

If you're having trouble getting the native calling conventions to work and you just want to get AngelScript to work, then you can also define the flag AS_MAX_PORTABILITY to turn off all support for native calling conventions and instead use the generic calling convention when registering the application interface.

AngelCode.com - game development and more - Reference DB - game developer references
AngelScript - free scripting library - BMFont - free bitmap font generator - Tower - free puzzle game

rfonseca

135

Author

November 06, 2015 01:14 AM

The compilation went fine, with just the warnings below.

Is there a test you use that I can run to check it actually works on ppc64?

In the meantime, I'll check what modifications are need for ppc64le and send a patch containing both (if possible).

../sdk/angelscript/source/as_scriptengine.cpp: In member function ‘void*

asCScriptEngine::CallGlobalFunctionRetPtr(asSSystemFunctionInterface*,

asCScriptFunction*) const’:

../sdk/angelscript/source/as_scriptengine.cpp:4315:52: warning: ‘stdcall’

attribute directive ignored [-Wattributes]

../sdk/angelscript/source/as_scriptengine.cpp: In member function ‘void*

asCScriptEngine::CallGlobalFunctionRetPtr(asSSystemFunctionInterface*,

asCScriptFunction*, void*) const’:

../sdk/angelscript/source/as_scriptengine.cpp:4337:58: warning: ‘stdcall’

attribute directive ignored [-Wattributes]

../sdk/angelscript/source/as_scriptengine.cpp: In member function ‘void

asCScriptEngine::CallGlobalFunction(void*, void*, asSSystemFunctionInterface*,

asCScriptFunction*) const’:

../sdk/angelscript/source/as_scriptengine.cpp:4438:65: warning: ‘stdcall’

attribute directive ignored [-Wattributes]

../sdk/angelscript/source/as_scriptengine.cpp: In member function ‘bool

asCScriptEngine::CallGlobalFunctionRetBool(void*, void*,

asSSystemFunctionInterface*, asCScriptFunction*) const’:

../sdk/angelscript/source/as_scriptengine.cpp:4464:65: warning: ‘stdcall’

attribute directive ignored [-Wattributes]

WitchLord

4,834

November 06, 2015 09:54 PM

To test if everything is working OK, you can run the regression test suite (available in the SVN http://svn.code.sf.net/p/angelscript/code/trunk/).

Go to the directory <your location>/sdk/tests/test_feature/projects/gnuc/ run make, then execute the testgnuc binary from the directory <your location>/sdk/tests/test_feature/bin/

If everything is fine all the tests will pass successfully. If there is anything wrong one of the tests will fail. Then it will be a matter of debugging the code to figure out what went wrong.

If you look at the main.cpp file in the <your location>/sdk/tests/test_feature/source/ folder, you'll see the list of tests that will be executed. If you're getting lots of errors it is a good idea to comment out all of the tests except for the one that start at line 348. These tests have been designed specifically for figuring out the details of the ABI, and stat with the most simplest case, of calling a function without any arguments or return value, and then increasing the complexity of the function calls to the most complex calls including virtual methods taking and returning objects by value.

AngelCode.com - game development and more - Reference DB - game developer references
AngelScript - free scripting library - BMFont - free bitmap font generator - Tower - free puzzle game

rfonseca

135

Author

November 09, 2015 05:05 PM

Andreas,

the first time I ran the tests I got this output:

AngelScript version: 2.31.0 WIP

AngelScript options: AS_MAX_PORTABILITY AS_64BIT_PTR AS_NO_THISCALL_FUNCTOR_METHOD AS_BIG_ENDIAN AS_LINUX

Compile the tests with AS_MAX_PORTABILITY defined

--------------------------------------------

One of the tests failed, see details above.

then I added -DAS_MAX_PORTABILITY to CXXFLAGS in gnuc/makefile, compiled again, and I'm now getting the following error:

AngelScript version: 2.31.0 WIP

AngelScript options: AS_MAX_PORTABILITY AS_64BIT_PTR AS_NO_THISCALL_FUNCTOR_METHOD AS_BIG_ENDIAN AS_LINUX

Test on line 10 in ../../source/test_addon_scriptfile.cpp skipped

-- Test_Addon_ScriptFile passed

-- Test_Addon_ContextMgr passed

Test on line 12 in ../../source/test_addon_scriptgrid.cpp skipped

-- Test_Addon_ScriptGrid passed

Program received signal SIGSEGV, Segmentation fault.

0x00003fffb7eac884 in ScriptObject_GetWeakRefFlag_Generic (gen=0x3fffffffe300)

at ../sdk/angelscript/source/as_scriptobject.cpp:159

159 *(asILockableSharedBool**)gen->GetAddressOfReturnLocation() = self->GetWeakRefFlag();

Missing separate debuginfos, use: dnf debuginfo-install libgcc-5.2.1-5.fc24.ppc64 libstdc++-5.2.1-5.fc24.ppc64

(gdb) bt

#0 0x00003fffb7eac884 in ScriptObject_GetWeakRefFlag_Generic (

gen=0x3fffffffe300) at ../sdk/angelscript/source/as_scriptobject.cpp:159

#1 0x00003fffb7e93290 in asCScriptEngine::CallObjectMethodRetPtr (

this=0x10472550, obj=0x10478870, func=5)

at ../sdk/angelscript/source/as_scriptengine.cpp:4225

#2 0x00003fffb7e90854 in asCScriptEngine::GetWeakRefFlagOfScriptObject (

this=0x10472550, obj=0x10478870, type=0x10485030)

at ../sdk/angelscript/source/as_scriptengine.cpp:3661

#3 0x00000000101feb88 in CScriptWeakRef::CScriptWeakRef (this=0x1048bad8,

ref=0x10478870, type=0x10485fe0)

at ../../../../add_on/weakref/weakref.cpp:88

#4 0x00000000101fe468 in ScriptWeakRefConstruct2 (type=0x10485fe0,

ref=0x10478870, mem=0x1048bad8) at ../../../../add_on/weakref/weakref.cpp:18

#5 0x0000000010200814 in ScriptWeakRefConstruct2_Generic (gen=0x3fffffffe658)

at ../../../../add_on/weakref/weakref.cpp:259

#6 0x00003fffb7e2d22c in asCContext::CallGeneric (this=0x10478420, descr=

0x104988d0) at ../sdk/angelscript/source/as_context.cpp:5125

#7 0x00003fffb7dcf234 in CallSystemFunction (id=45, context=0x10478420)

at ../sdk/angelscript/source/as_callfunc.cpp:513

#8 0x00003fffb7e250f0 in asCContext::ExecuteNext (this=0x10478420)

at ../sdk/angelscript/source/as_context.cpp:2509

#9 0x00003fffb7e217fc in asCContext::Execute (this=0x10478420)

at ../sdk/angelscript/source/as_context.cpp:1291

#10 0x00000000101ae934 in ExecuteString (engine=0x10472550,

code=0x10231608 "main()", ref=0x0, refTypeId=0, mod=0x104842c0, ctx=0x0)

at ../../../../add_on/scripthelper/scripthelper.cpp:183

#11 0x00000000101ae4f8 in ExecuteString (engine=0x10472550,

code=0x10231608 "main()", mod=0x104842c0, ctx=0x0)

at ../../../../add_on/scripthelper/scripthelper.cpp:137

#12 0x00000000100273a4 in Test_Addon_WeakRef::Test ()

at ../../source/test_addon_weakref.cpp:112

#13 0x0000000010008038 in allTests () at ../../source/main.cpp:227

#14 0x00000000100099ec in main () at ../../source/main.cpp:425

(gdb) p self

$1 = (asCScriptObject *) 0x10478870

(gdb) p self->GetWeakRefFlag()

Cannot access memory at address 0x30

WitchLord

4,834

November 09, 2015 10:28 PM

OK, so you went with the MAX_PORTABILITY mode. That's probably wise and a good start. :)

The first 3 tests passed successfully, so the problem is probably specific for the weak reference scenarios in Test_Addon_WeakRef. Most likely it has something to do with the fact that PPC64 is 64bit and big endian, which is something that would be specific for your platform. No other platform that I know AngelScript is running on has this particularity. Not even PlayStation3 has 64bit pointers, even though the CPU is based on PPC64 architecture. (I'm not sure about Xbox 360, but I'm guessing it too has only 32bit pointers).

Can you debug the code to try to figure out where it is going wrong? Or else, perhaps you can give me remote access to your machine over SSH so I can do a little debugging on my own?

AngelCode.com - game development and more - Reference DB - game developer references
AngelScript - free scripting library - BMFont - free bitmap font generator - Tower - free puzzle game

rfonseca

135

Author

November 10, 2015 02:27 PM

Can you debug the code to try to figure out where it is going wrong? Or else, perhaps you can give me remote access to your machine over SSH so I can do a little debugging on my own?

I'll be investigating the issue. Anyway, if you want access to a ppc64 guest you can fill in the form in https://fit-rhlab.rhcloud.com/powerlinux-openpower-development-hosting/.

Let me know if you need anything.

rfonseca

135

Author

November 11, 2015 01:24 PM

Andreas,

this is the information I was able to gather so far about the error:

the tests segfaults in this function 'void *asCGeneric::GetAddressOfReturnLocation()'

(gdb) p *sysFunction

$11 = {<asIScriptFunction> = {_vptr.asIScriptFunction = 0x3fffb7f5f848 <vtable for asCScriptFunction+16>},

externalRefCount = {value = 0}, internalRefCount = {value = 2}, gcFlag = false, engine = 0x10472550,

module = 0x0, userData = {array = 0x0, length = 0, maxLength = 0, buf = '\000' <repeats 15 times>}, name = {

length = 5, {dynamic = 0x2462656837000000 <error: Cannot access memory at address 0x2462656837000000>,

local = "$beh7\000\000\000\000\000\000"}}, returnType = {tokenType = ttInt, objectType = 0x0,

funcDef = 0x0, isReference = true, isReadOnly = false, isObjectHandle = false, isConstHandle = false,

isAuto = false, isHandleToAsHandleType = false, dummy = 0 '\000'}, parameterTypes = {array = 0x0,

length = 0, maxLength = 0, buf = '\000' <repeats 15 times>}, parameterNames = {array = 0x0, length = 0,

maxLength = 0, buf = '\000' <repeats 15 times>}, inOutFlags = {array = 0x0, length = 0, maxLength = 0,

buf = '\000' <repeats 15 times>}, defaultArgs = {array = 0x0, length = 0, maxLength = 0,

buf = '\000' <repeats 15 times>}, isReadOnly = false, isPrivate = false, isProtected = false,

isFinal = false, isOverride = false, objectType = 0x104725d8, signatureId = 0, id = 5,

funcType = asFUNC_SYSTEM, accessMask = 1, isShared = false, nameSpace = 0x10478100, parentClass = 0x0,

objForDelegate = 0x0, funcForDelegate = 0x0, listPattern = 0x0, scriptData = 0x0,

dontCleanUpOnException = false, vfTableIdx = -1, sysFuncIntf = 0x104791d0}

as we can see, returnType.isReference is true. That means the test 'if( dt.IsObject() && !dt.IsReference() )' is false, taking us to the end of the function

// Primitive types and references are stored in the returnVal property

return &returnVal;

(gdb) p returnVal

$12 = 0

that's why when we try to

*(asILockableSharedBool**)gen->GetAddressOfReturnLocation() = self->GetWeakRefFlag();

we are dereferencing a pointer to null:

(gdb) p *(asILockableSharedBool**)gen->GetAddressOfReturnLocation()

$14 = (asILockableSharedBool *) 0x0

and it segfaults when we try to assign a value to it.

Compiling angelscript on ppc64

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Compiling angelscript on ppc64

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines