Frank Kjul Larsen

17 ポイント. 0 回答 4 質問 0 フォロワー 0 フォロー

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.

フォロー

統計

0 answers, 0 selected
4 questions, 0 solved
6 年, 9 ヶ月, 29 日のメンバー
13360 プロフィール閲覧
Last seen 10時間 ago

ポイント

Total 17

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0

1月 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

1月 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

1月 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

1月 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

1月 22, 2021 Asked printf() usage in RX toolchain 20.04 change when optimisation change

投票数

0 投票獲得 0/100 0up 0down

1 票獲得 100/100 1up 0down

Top 回答

回答がまだありません

新しい質問

0 4月 30, 2021 10 回答

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

0 10月 2, 2018 0 回答

Will fixed-point math be supported?

0 10月 16, 2017 1 回答

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

0 8月 22, 2018 1 回答

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

Frank Kjul Larsen のほかの質問閲覧→