Frank Kjul Larsen

19 Rep. 0 Answers 5 Questions 0 Followers 0 Following

Male GNU tools user for H8SX1648 and RX111, RX71M under Windows and Linux 64 bit for several small targets with no OS (bare metal, super loop systems). Using makefile for build.

Stats

0 answers, 0 selected
5 questions, 0 solved
Member for 7 years, 8 months, 16 days
15525 profile views
Last seen April 12, 2024

Reputation

Total 19

0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0,0,0,0,0,0,0,0,0,0,0,0

April 12, 2024 Asked Which CPU to select for RX26T compilation.

January 26, 2021 Commented Hello Robert OK. I changed to use newlib-nano by exchanging use of optimised lib C and lib Math. OPTIONS += -loptc OPTIONS += -loptm with OPTIONS += -lc OPTIONS += -lm OPTIONS += --specs=nano.specs OPTIONS are CFLAGS and I activate the linker through GCC. After changing to newlib nano, I had to implement two more functions in order to link: void *sbrk(intptr_t increment) { (void)increment; return 0; } _Noreturn void _exit(int status) { exit(status); } The good thing is that now the number warnings does not depend on the optimisation level:-) The code size is less than before:-) The code still executes:-) BR Frank

January 26, 2021 Commented Hello NoMaY-jp I just compile ordinary C-code, not C++. I compile from command line using makefile not using E2 studio. I call the compiler like this (I've remove some include path to make it less messy): rx-elf-gcc -c -Isource -Drx71m -mcpu=rx71m -mlittle-endian-data -D__RX_LITTLE_ENDIAN__=1 -std=c11 -nostartfiles -Wall -Wextra -Wl,--warn-common -Wl,--gc-sections -ffunction-sections -fdata-sections -DROMSTART -loptc -loptm -DPRINTF -O3 -g2 -o obj/source/main.o source/main.c My project consists of several files - makes a binary of >100 kB at -O3, thus it makes no sense to post them. I just wanted to tell GNU tools that I experience this behaviour. It seems that in their investigation they see same behaviour regarding link errors when using printf using -O0 and -O3. I therefore suspect another setting in my setup to cause my trouble. But I know how to handle it, so I will leave it for now. This discussion will show other users what they can do. Thanks for your effort. BR Frank

January 26, 2021 Commented Hello all I cannot send all my source code, but I can be more specific on how I compile and link. I compile for a RXv2 MCU - RX71m. I use my own start file. I execute from flash. OPTIONS += -mcpu=rx71m OPTIONS += -mlittle-endian-data # Select little endian operation OPTIONS += -D__RX_LITTLE_ENDIAN__=1 OPTIONS += -std=c11 OPTIONS += -nostartfiles OPTIONS += -Wall OPTIONS += -Wextra OPTIONS += -Wl,--warn-common OPTIONS += -Wl,--gc-sections OPTIONS += -ffunction-sections OPTIONS += -fdata-sections OPTIONS += -DROMSTART OPTIONS += -loptc OPTIONS += -loptm # Optimisation OPTIONS += -O3 or # Optimisation OPTIONS += -O0 BR Frank

January 25, 2021 Commented To be honest, I have not used printf() before with the RX tool chain. Last time was with GNU H8 toolchain. The core of my question is why the behaviour is different at different optimisation levels on the newest RX toolchain. BR Frank

Votes

0 votes received 0/100 0up 0down

1 votes casted 100/100 1up 0down

New questions

0 April 15, 2024 1 answers

Which CPU to select for RX26T compilation.

0 April 30, 2021 10 answers

printf() usage in RX toolchain 20.04 change when optimisation change

0 October 2, 2018 0 answers

Will fixed-point math be supported?

0 October 16, 2017 1 answers

Difference between Windows and Linux RX 1703 toolchain when using optimised libraries

0 August 22, 2018 1 answers

Where are the 12.02 Linux and Windows versions of the H8 GNU tool chain?

More questions by Frank Kjul Larsen view→

Frank Kjul Larsen

Stats

Reputation

Votes

Top Answers

New questions