0001114: Cannot build ARM eabi toolchain - BusyBox Bug and Patch Tracking

Notes
(0001860) mikewhit 12-13-06 02:43	This is actually a Blocking issue, since cannot use Buildroot without EABI support.

(0001900) bernhardf 12-21-06 03:50	post your egrep -v "^(#\|$)" .config egrep -v "^(#\|$)" uClibc.config egrep -v "^(#\|$)" busybox.config egrep -v "^(#\|$)" kernel.config So i can try to reproduce it locally.

(0001974) stephaneC 01-09-07 05:06	I have the same problem with : buildroot rev 17204 ARM920T / EABI / gcc-3.4.6 see below for my .config and uClibc.config 1st error is bug 1131 : file toolchain_build_arm_nofpu/uClibc/libc/misc/glob/glob.c line 364 #if !defined COMPILE_GLOB64 to #if defined __UCLIBC_HAS_LFS__ && !defined COMPILE_GLOB64 2nd error : --- [snip]buildroot/build_arm_nofpu/staging_dir/arm-linux-uclibcgnueabi/bin/ld: unrecognised emulation mode: armelf_linux Supported emulations: armelf_linux_eabi armelfb_linux_eabi collect2: ld returned 1 exit status make[3]: *** [libgcc/./_udivsi3.oS] Erreur 1 make[3]: quittant le r

(0002106) bernhardf 02-04-07 09:09	I think this is fixed in current trunk. Please retest.

(0002159) bernhardf 02-14-07 03:46	Works with $ egrep "^BR2_(BINUT\|GCC_VER\|SOFT)" .config BR2_BINUTILS_VERSION_2_17_50_0_10=y BR2_BINUTILS_VERSION="2.17.50.0.10" BR2_GCC_VERSION_4_1_2=y BR2_GCC_VERSION="4.1.2" BR2_SOFT_FLOAT=y

(0002162) mikewhit 02-14-07 09:30	Still does not work using original settings (buildroot-20070214), looks like it just won't work for 3.4.3. Still, now we have the external toolchain option ! egrep on .config: >> BR2_HAVE_DOT_CONFIG=y BR2_arm=y BR2_arm926t=y BR2_ARM_TYPE="ARM926T" BR2_ARM_EABI=y BR2_ARCH="arm" BR2_ENDIAN="LITTLE" BR2_WGET="wget --passive-ftp" BR2_SVN="svn co" BR2_ZCAT="zcat" BR2_BZCAT="bzcat" BR2_TAR_OPTIONS="" BR2_DL_DIR="/opt/dl" BR2_SOURCEFORGE_MIRROR="easynews" BR2_ATMEL_MIRROR="distrib@81.80.104.162/AT91_Third_Party_Design_Flow/Linux_Host/"">ftp://at91dist:distrib@81.80.104.162/AT91_Third_Party_Design_Flow/Linux_Host/" [distrib@81.80.104.162/AT91_Third_Party_Design_Flow/Linux_Host/"" target="_blank">^] BR2_AT91_PATCH_MIRROR="http://maxim.org.za/AT91RM9200/2.6/" [^] BR2_STAGING_DIR="$(BUILD_DIR)/staging_dir" BR2_TOPDIR_PREFIX="" BR2_TOPDIR_SUFFIX="" BR2_GNU_BUILD_SUFFIX="pc-linux-gnu" BR2_GNU_TARGET_SUFFIX="linux-uclibcgnueabi" BR2_JLEVEL=1 BR2_TOOLCHAIN_BUILDROOT=y BR2_DEFAULT_KERNEL_HEADERS="2.6.12" BR2_UCLIBC_VERSION_0_9_28=y BR2_PTHREADS_OLD=y BR2_BINUTILS_VERSION_2_16_91_0_7=y BR2_BINUTILS_VERSION="2.16.91.0.7" BR2_EXTRA_BINUTILS_CONFIG_OPTIONS="" BR2_GCC_VERSION_3_4_3=y BR2_GCC_VERSION="3.4.6" BR2_EXTRA_GCC_CONFIG_OPTIONS="" BR2_INSTALL_LIBSTDCPP=y BR2_GCC_SHARED_LIBGCC=y BR2_LARGEFILE=y BR2_SOFT_FLOAT=y BR2_TARGET_OPTIMIZATION="-Os -pipe" BR2_PACKAGE_BUSYBOX=y BR2_BUSYBOX_VERSION_1_2_2_1=y BR2_PACKAGE_BUSYBOX_INSTALL_SYMLINKS=y BR2_PACKAGE_BUSYBOX_CONFIG="package/busybox/busybox-1.2.2.1.config" BR2_EXTRA_TARGET_GCC_CONFIG_OPTIONS="" BR2_NETWORK_SUPPORT=y BR2_BLOCKDEV_SUPPORT=y BR2_PACKAGE_MTD=y BR2_PACKAGE_MTD_ORIG=y BR2_PACKAGE_MTD_ORIG_STRING="mtd_20050122.orig.tar.gz" BR2_PACKAGE_MTD_FLASH_ERASE=y BR2_PACKAGE_MTD_FLASH_ERASEALL=y BR2_PACKAGE_MTD_FLASH_INFO=y BR2_PACKAGE_MTD_FLASH_LOCK=y BR2_PACKAGE_MTD_FLASH_UNLOCK=y BR2_PACKAGE_MTD_FLASHCP=y BR2_PACKAGE_MTD_ERASE=y BR2_PACKAGE_MTD_JFFS2DUMP=y BR2_PACKAGE_MTD_JFFS3DUMP=y BR2_PACKAGE_MTD_SUMTOOL=y BR2_PACKAGE_MTD_FTL_CHECK=y BR2_PACKAGE_MTD_FTL_FORMAT=y BR2_PACKAGE_MTD_NFTL_FORMAT=y BR2_PACKAGE_MTD_NFTLDUMP=y BR2_PACKAGE_MTD_MKFSJFFS2=y BR2_PACKAGE_MTD_MKFSJFFS=y BR2_PACKAGE_MTD_NANDDUMP=y BR2_PACKAGE_MTD_NANDWRITE=y BR2_PACKAGE_MTD_MTD_DEBUG=y BR2_PACKAGE_MTD_DOCFDISK=y BR2_PACKAGE_MTD_DOC_LOADBIOS=y BR2_AUDIO_SUPPORT=y BR2_GRAPHIC_SUPPORT=y BR2_COMPRESSOR_SUPPORT=y BR2_PACKAGE_ZLIB=y BR2_SCRIPTING_SUPPORT=y BR2_TARGET_ROOTFS_JFFS2=y BR2_TARGET_ROOTFS_JFFS2_DEFAULT_PAGESIZE=y BR2_TARGET_ROOTFS_JFFS2_EBSIZE=0x20000 BR2_TARGET_ROOTFS_JFFS2_LE=y BR2_TARGET_ROOTFS_JFFS2_OUTPUT="$(IMAGE).jffs2" BR2_TARGET_ROOTFS_JFFS2_COPYTO="" <<

(0002164) bernhardf 02-14-07 15:22	Let me spell this out: - 3.4 is DEAD. - 4.1, which is the current stable release deals with this fine It will not be fixed for the long outdated and not maintained, deprecated version 3.4.x nor 4.0.x unless _you_ provide a working patch that _you_ support for this alleged glitch. Put short, Not an issue since it works for current, _stable_ stuff. Not an issue since it works for current expreimental (new) toolchains. Will not fix for outdated, unmaintained, deprecated setups. Thanks for your understanding.

(0002166) mikewhit 02-15-07 00:57	Only reopening to add comment - then close again ! Re the previous note (berhhardf) - I was asked to "retest" (2007-02-04), which I did using the same parameters pertaining to my original report. A test does not have meaning if several of the parameters change each time ! If you have seen any of my postings on this subject on the buildroot newsgroup, you will know that I was wanting to use this combination for compatibility with 3rd-party supplied object modules, built with a 3.4.3 EABI Arm toolchain, which were not available in source form, and which the linker refused to link with buildroot toolchain generated objects. However, now that the External Toolchain option is available for Buildroot, it should be possible to rebuild applications using this toolchain, and link against the prebuilt objects. Now that I have "spelt that out" I hope it's clear why I added that note. There is certainly no need for heavy-handed sarcasm: but if I misread the tone of that note, apologies in turn !

(0002180) mikewhit 02-19-07 04:24	Note on compiler version 'improvements'. Looking around comparisons of GCC, I note the following: (http://lists.freebsd.org/pipermail/freebsd-current/2006-December/068108.html) [^] >> >Actually, 4.1.x will produce much worse code than 3.4.6. >You can search the gcc mail listings for extensive comparison >by Clinton Whaley (the author of math/atlas) for details I would call 4.0 the "make it work" stage. gcc 4.1 concentrated on bug fixes, and gcc 4.2 has started to remove the warts (e.g., slow compilation times, bloated generated code, etc.) << so it doesn't follow that 'the bigger the version number, the better the compiler'. From those comments it looks as if 3.4.x is a better bet in the short-term ... FWIW. [OK the thread above does imply that 4.2 has better ARM support as well ;-) ]

Issue History
Date Modified	Username	Field	Change
12-13-06 01:58	mikewhit	New Issue
12-13-06 01:58	mikewhit	Status	new => assigned
12-13-06 01:58	mikewhit	Assigned To	=> uClibc
12-13-06 02:43	mikewhit	Note Added: 0001860
12-21-06 03:50	bernhardf	Note Added: 0001900
01-09-07 05:06	stephaneC	Note Added: 0001974
01-09-07 05:07	stephaneC	Issue Monitored: stephaneC
02-04-07 09:09	bernhardf	Note Added: 0002106
02-04-07 09:09	bernhardf	Status	assigned => feedback
02-12-07 05:33	vapier	Status	feedback => assigned
02-12-07 05:33	vapier	Assigned To	uClibc => buildroot
02-14-07 03:46	bernhardf	Status	assigned => closed
02-14-07 03:46	bernhardf	Note Added: 0002159
02-14-07 03:47	bernhardf	Resolution	open => fixed
02-14-07 09:30	mikewhit	Status	closed => feedback
02-14-07 09:30	mikewhit	Resolution	fixed => reopened
02-14-07 09:30	mikewhit	Note Added: 0002162
02-14-07 15:22	bernhardf	Status	feedback => closed
02-14-07 15:22	bernhardf	Note Added: 0002164
02-14-07 15:22	bernhardf	Resolution	reopened => won't fix
02-15-07 00:57	mikewhit	Status	closed => feedback
02-15-07 00:57	mikewhit	Resolution	won't fix => reopened
02-15-07 00:57	mikewhit	Note Added: 0002166
02-19-07 04:24	mikewhit	Note Added: 0002180
06-25-07 11:04	bernhardf	Status	feedback => closed
06-25-07 11:04	bernhardf	Resolution	reopened => no change required

Relationships