[U-Boot-Users] GCC3.4.0 Coldfire 52xx Build problems
Hello All,
I am new to U-Boot. We develop appl. on coldfire M5275EVB with uclinux 2.6, and intent to use U-Boot as boot loader.
- Use the gcc3.4.0 toolchain, binutils binutils-2.15.90.0.1.1 from http://www.uclinux.org/pub/uClinux/m68k-elf-tools/gcc-3/
- I download u-boot from CVS, so latest version.
- Tried to compile for M5272C3_config did make M5272C3_config make
- got following error: ---------------------- make[1]: Leaving directory `/cygdrive/c/project/U-Boot_20050803/u-boot/post/cpu' m68k-elf-gcc -Wa,-gstabs -D__ASSEMBLY__ -g -Os -ffixed-d7 -msep-data -D__ KERNEL__ -DTEXT_BASE=0xffe00000 -I/cygdrive/c/project/U-Boot_20050803/u-boot /include -fno-builtin -ffreestanding -nostdinc -isystem /usr/lib/gcc/i686-pc-cygwin/3.4.4/include -pipe -DCONFIG_M68K -D__M68K__ -m 5307 -c -o cpu/mcf52x2/start.o /cygdrive/c/project/U-Boot_20050803/u-boot/cpu/mcf52x2/start.S /cygdrive/c/project/U-Boot_20050803/u-boot/cpu/mcf52x2/start.S: Assembler messages: /cygdrive/c/project/U-Boot_20050803/u-boot/cpu/mcf52x2/start.S:345: Fatal error: Tried to convert PC relative conditional branch to absolute jump make: *** [cpu/mcf52x2/start.o] Error 1 ----------------------- - found the same error in the mail list, see http://sourceforge.net/mailarchive/message.php?msg_id=10000351 but no real solution mentioned.
- found that the compiler option -msep-data is the reason for this error.
- removed -msep-data from cpu/mcf52x2/config.mk
- now compiling proceeds, but next error occurs:
make -C `dirname board/m5272c3/libm5272c3.a` make[1]: Entering directory `/cygdrive/c/project/U-Boot_20050803/u-boot/board/m5272c3' m68k-elf-gcc -g -Os -ffixed-d7 -D__KERNEL__ -DTEXT_BASE=0xffe00000 -I/cyg drive/c/project/U-Boot_20050803/u-boot/include -fno-builtin -ffreestanding - nostdinc -isystem /usr/local/m68k-uclinux-tools/lib/gcc/m68k-uclinux/3.4.0/include -pipe -DCO NFIG_M68K -D__M68K__ -m5307 -Wall -Wstrict-prototypes -c -o m5272c3.o m5272c3.c m68k-elf-gcc -g -Os -ffixed-d7 -D__KERNEL__ -DTEXT_BASE=0xffe00000 -I/cyg drive/c/project/U-Boot_20050803/u-boot/include -fno-builtin -ffreestanding - nostdinc -isystem /usr/local/m68k-uclinux-tools/lib/gcc/m68k-uclinux/3.4.0/include -pipe -DCO NFIG_M68K -D__M68K__ -m5307 -Wall -Wstrict-prototypes -c -o flash.o flash.c flash.c: In function `flash_print_info': flash.c:68: error: label at end of compound statement make[1]: *** [flash.o] Error 1 make[1]: Leaving directory `/cygdrive/c/project/U-Boot_20050803/u-boot/board/m5272c3' make: *** [board/m5272c3/libm5272c3.a] Error 2
this flash.c file is version 1.1 from 2003/12/08.
- ok, made patch in flash.c compiling proceeds, but next error:
UNDEF_SYM=`m68k-elf-objdump -x lib_generic/libgeneric.a board/m5272c3/libm5272c3.a cpu/mcf52x2/libmcf52x2.a lib_m68k/libm68k.a fs/cramfs/libcramfs.a fs/fat/libfat.a fs/fdos/libfdos.a fs/jffs2/libjffs2.a fs/reiserfs/libreiserfs.a fs/ext2/libext2fs.a net/libnet.a disk/libdisk.a rtc/librtc.a dtt/libdtt.a drivers/libdrivers.a drivers/sk98lin/libsk98lin.a post/libpost.a post/cpu/libcpu.a common/libcommon.a |sed -n -e 's/.*(__u_boot_cmd_.*)/-u\1/p'|sort|uniq`;\ m68k-elf-ld -Bstatic -T /cygdrive/c/project/U-Boot_20050803/u-boot/board/m5272c3/u-boot.lds -Ttext 0xffe00000 -n $UNDEF_SYM cpu/mcf52x2/start.o \ --start-group lib_generic/libgeneric.a board/m5272c3/libm5272c3.a cpu/mcf52x2/libmcf52x2.a lib_m68k/libm68k.a fs/cramfs/libcramfs.a fs/fat/libfat.a fs/fdos/libfdos.a fs/jffs2/libjffs2.a fs/reiserfs/libreiserfs.a fs/ext2/libext2fs.a net/libnet.a disk/libdisk.a rtc/librtc.a dtt/libdtt.a drivers/libdrivers.a drivers/sk98lin/libsk98lin.a post/libpost.a post/cpu/libcpu.a common/libcommon.a --end-group -L /usr/local/m68k-uclinux-tools/lib/gcc/m68k-uclinux/3.4.0/m5307 -lgcc \ -Map u-boot.map -o u-boot /cygdrive/x/uCLinux/install/bin/m68k-elf-ld.real: internal error /cygdrive/x/uCLinux/binutils-2.15.90.0.1.1/ld/ldlang.c 1977 make: *** [u-boot] Error 1
code from binutils:
int lang_symbol_definition_iteration (const char *name) { struct lang_definedness_hash_entry *defentry = (struct lang_definedness_hash_entry *) bfd_hash_lookup (&lang_definedness_table, name, FALSE, FALSE);
/* We've already created this one on the presence of DEFINED in the script, so it can't be NULL unless something is borked elsewhere in the code. */ if (defentry == NULL) line 1977: FAIL (); ^^^^^^^^^^^^^^^^^^^^^^
return defentry->iteration; }
======================================================
so my overall questions, Is then main reason for this problems that I use GCC3.4.0 toolchain ?? Are other people using U-Boot for coldfire M52xx ?? and did run into same problems ?? I like to hear from their solutions/advices.
=======================================================
Thanks in advance.
Teun Grinwis
Adesys BV Molenweer 4, 2291 NR Postbus 60, 2290 AB Wateringen (NL) Tel.: +31 174 296389 Fax : +31 174 293807 www.adesys.nl
In message OKECIIPCKHCPJAOEEJNMMEONCAAA.teun@adesys.nl you wrote:
Is then main reason for this problems that I use GCC3.4.0 toolchain ??
Probably.
Best regards,
Wolfgang Denk
On Thu, 04 Aug 2005 12:31:27 +0200, Wolfgang Denk writes:
In message OKECIIPCKHCPJAOEEJNMMEONCAAA.teun@adesys.nl you wrote:
Is then main reason for this problems that I use GCC3.4.0 toolchain ??
Probably.
I use GCC-3.4.3 to build U-Boot for our MPC8560 based board. I had an internal linker error which I traced to the following in my linker script:
. = + SIZEOF_HEADERS;
I changed this to:
. = + 1024;
and it prevented the core dump in the linker.
I know your target is very different, but I thought this info might help anyway (if not you, then maybe someone else). Cheers! Murray...
Thanks for your time Murray,
but ... it didn't workout successful.
but, the reason for my problems is the gcc version. I now compiled with version 2.95.3 from: http://www.uclinux.org/pub/uClinux/m68k-elf-tools/m68k-elf-tools-20030314.sh
I have some minor compile problem ( var declaration in middle of compound statement)
I am able to build u-boot.srec with use of patch below.
Here is patch: --- common/cmd_flash.c.orig 2005-08-05 09:56:58.659430808 +0200 +++ common/cmd_flash.c 2005-08-05 09:55:34.344248664 +0200 @@ -117,11 +117,11 @@ addr_spec(char *arg1, char *arg2, ulong *addr_first, ulong *addr_last) { char *ep; + char len_used = 0; /* indicates if the "start +length" form used */ *addr_first = simple_strtoul(arg1, &ep, 16); if (ep == arg1 || *ep != '\0') return -1;
- char len_used = 0; /* indicates if the "start +length" form used */ if (arg2 && *arg2 == '+'){ len_used = 1; ++arg2; @@ -132,6 +132,8 @@ return -1;
if (len_used){ + char found = 0; + ulong bank; /* * *addr_last has the length, compute correct *addr_last * XXX watch out for the integer overflow! Right now it is @@ -146,8 +148,6 @@ */
/* find the end addr of the sector where the *addr_last is */ - char found = 0; - ulong bank; for (bank = 0; bank < CFG_MAX_FLASH_BANKS && !found; ++bank){ int i; flash_info_t *info = &flash_info[bank];
On Thu, 04 Aug 2005 12:31:27 +0200, Wolfgang Denk writes:
In message OKECIIPCKHCPJAOEEJNMMEONCAAA.teun@adesys.nl you wrote:
Is then main reason for this problems that I use GCC3.4.0 toolchain ??
Probably.
I use GCC-3.4.3 to build U-Boot for our MPC8560 based board. I had an internal linker error which I traced to the following in my linker script:
. = + SIZEOF_HEADERS;
I changed this to:
. = + 1024;
and it prevented the core dump in the linker.
Best Regards,
Teun Grinwis
In message OKECIIPCKHCPJAOEEJNMIEPACAAA.teun@adesys.nl you wrote:
I have some minor compile problem ( var declaration in middle of compound statement)
Thanks for pointing out. Fixed and checked in on CVS and git trees.
Best regards,
Wolfgang Denk
In message 14067.1123225365@gerd you wrote:
I use GCC-3.4.3 to build U-Boot for our MPC8560 based board. I had an internal linker error which I traced to the following in my linker script:
. = + SIZEOF_HEADERS;
I changed this to:
. = + 1024;
and it prevented the core dump in the linker.
Of course this can be no more than part of a bug report to the binutils folk. I would't even consider this as a workaround.
Best regards,
Wolfgang Denk
On Fri, 05 Aug 2005 11:07:09 +0200, Wolfgang Denk writes:
In message 14067.1123225365@gerd you wrote:
I use GCC-3.4.3 to build U-Boot for our MPC8560 based board. I had an internal linker error which I traced to the following in my linker script:
. = + SIZEOF_HEADERS;
I changed this to:
. = + 1024;
and it prevented the core dump in the linker.
Of course this can be no more than part of a bug report to the binutils folk. I would't even consider this as a workaround.
Umm.. I think I worked out what was a reasonable value for SIZEOF_HEADERS and substituted it. The build succeeds and runs fine on our MPC8560 board. So I would consider this a reasonable work-around, until they fix binutils (maybe they already have - I haven't tried later versions). Cheers! Murray...
In message 20838.1123235458@gerd you wrote:
Umm.. I think I worked out what was a reasonable value for SIZEOF_HEADERS and substituted it. The build succeeds and runs fine on our MPC8560 board.
Replacing a Builtin Function which is supposed to return the correct size of the output file's headers by a hardwired value which just happens to be correct in one case is definitely not acceptable, not even as a temporary workaround.
Simply don't use tools that are broken on such a basic level. And if you want to use them, so either bug the binutils folks iuuintil they fix it, or fix it yourself. But don;t break perfectly legal and working code.
Best regards,
Wolfgang Denk
participants (3)
-
Murray.Jensen＠csiro.au -
Teun Grinwis -
Wolfgang Denk