http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58730
Bug ID: 58730
Summary: arm-eabi-gcc generate error instructions for armv4t
cpu
Product: gcc
Version: 4.8.1
Status: UNCONFIRMED
Severity: blocker
Priority: P3
Component: c
Assignee: unassigned at gcc dot gnu.org
Reporter: tcliuqiang at msn dot com
I build two versions of gcc: 4.7.3, and 4.8.1. the configure line listed as
followed:
mkdir -p build/gcc_arm; cd build/gcc_arm
../../src/gcc-4.8.1/configure --target=$target --prefix=$prefix
--enable-languages=c,c++ --disable-multilib --with-float=soft
--with-bugurl=tcliuqi...@gmail.com --enable-__cxa_atexit --enable-threads=posix
--disable-nls --enable-symvers=gnu --enable-c99 --enable-long-long CFLAGS="-O2"
CXXFLAGS="-O2" --with-gnu-as --with-gnu-ld --disable-rpath --enable-lto
--enable-shared --enable-libssp --with-arch=armv4t
And i use these toolchains to build my Linux kernel 3.8.13 for Samsung S3C2440
/ armv4t cpu (with the same kernel configs). GCC4.7.3 generate right code, my
embedded system run ok, and GCC4.8.1 may lead to a system-hang: the cpu just
seems to stop suddenly without any useful informations. So I googled for this
condition, and found that GCC4.8.1/ARM toolchain may generate error code for
Cortex M0 CPU, but no reports related to armv4t. I think there may be the same
bugs.
I found the report "49641/48173 [4.6 Regression] Wrong code for ARMv4T and
stmia / bootstrap error on arm-linux-gnueabi (defaulting to armv4t)", and I
think these problems may have been resolved in 4.7 and back again in 4.8?
This is the first time I issue a bug to gcc team; should I provide more
informations? I'am not a arm-architecture professional, but i can do more to
find out what happens to 4.8 version.
