Floating Point errors using powerpc-eabivle-4_9 compilers on MPC5777 e200z7 core

sreekant866 — Tue, 05 Jul 2022 07:01:10 GMT

Hello Team,

I am using use the compiler -powerpc-eabivle-gcc-4.9.4 located in "\nxp\S32DS_Power_v2.1\S32DS\build_tools\powerpc-eabivle-4_9\bin". We are getting following errors while we are compiling our code:
1. undefined reference to `__extendsfdf2'
2. undefined reference to `__extendsfdf2'
3. undefined reference to `__floatunsidf'
4. undefined reference to `__umoddi3'
5. undefined reference to `__gtdf2'
6. undefined reference to `__udivdi3'

I am using MPC5777 e200z7 core. Am I missing any compiler flags to address this issue or are we missing anything in the compiler to add the above mentioned floating point functions?. Please help

Below are the compilation flags used:

Arch Flags:

TGT_FLAGS = -mno-eabi -mspe -mabi=spe -memb -meabi -mlra -mcpu=e200z7 -mbig -mno-vle -Wa,-mbooke32 -mregnames -Wa,-mregnames -mhard-float

COMMON_FLAGS = $(TGT_FLAGS) -fno-common
C Flags:

CFLAGS = $(COMMON_FLAGS) -DUINT_SIZE=32 -c $(C_OPTIMIZATION) -fweb -ffunction-sections -frename-registers -funswitch-loops -fgcse-after-reload \
-g3 $(ASSERTFLAGS) $(CEXTFLAGS) -nostdinc -fno-builtin -std=gnu99 -Wimplicit -Wall -Wstrict-prototypes \
-Wparentheses -Wmissing-braces -Wsequence-point -Wfloat-equal -Wmissing-prototypes -Wmissing-declarations \
-Wold-style-definition -Wimplicit-function-declaration -Wno-long-long -Wunreachable-code -fsingle-precision-constant \

CPP Flags:

CPPFLAGS = $(COMMON_FLAGS) -DUINT_SIZE=32 -c $(CPP_OPTIMIZATION) -fweb -ffunction-sections -frename-registers -funswitch-loops -fgcse-after-reload \
-g3 $(ASSERTFLAGS) $(CPPEXTFLAGS) -fno-default-inline --param max-inline-insns-auto=0 -fno-inline-functions \
-fno-rtti -fno-exceptions -fno-enforce-eh-specs -fno-builtin -fno-threadsafe-statics -fno-gnu-keywords \
-fno-nonansi-builtins -fno-operator-names -nostdinc++ -Wabi -Wreorder -Woverloaded-virtual -Wsign-promo -Winit-self \
-Wswitch-enum -Wextra -Wfloat-equal -Wunreachable-code -fsingle-precision-constant

Regards,

Sreekant Kollur

Re: Floating Point errors using powerpc-eabivle-4_9 compilers on MPC5777 e200z7 core

jiri_kral — Wed, 13 Jul 2022 13:31:33 GMT

Hi,

you may find helpful this StackOverflow article:
https://stackoverflow.com/questions/53402307/powerpc-gcc-floating-point-instructions

Also make sure, that HW float point is enabled in your project properties:

Jiri

S32 Design StudioのトピックFloating Point errors using powerpc-eabivle-4_9 compilers on MPC5777 e200z7 core

Floating Point errors using powerpc-eabivle-4_9 compilers on MPC5777 e200z7 core

Re: Floating Point errors using powerpc-eabivle-4_9 compilers on MPC5777 e200z7 core