From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dirk Behme Date: Wed, 08 Aug 2007 06:23:15 +0200 Subject: [U-Boot-Users] [ARM] TI DaVinci support, hopefully final [0/5] In-Reply-To: References: <5CBE65F7D9232C47861CB09B0954861C66A45F@MAIL.infinitychannel.local> Message-ID: <46B94533.5080304@googlemail.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de ksi at koi8.net wrote: > On Tue, 7 Aug 2007, Zach Sadecki wrote: > > That's MontaVista toolchain. No, that's not a MontaVista specific issue. I have an other toolchain and it has similiar issues. Nor is it an ARM specific issue. nand_utils.c uses 64bit division, and if you use stuff from nand_utils.c you may get these issues. So solution is to make U-Boot independent from toolchain for 64bit division used in nand_utils.c as already done with other math helpers in lib_arm. Please apply http://article.gmane.org/gmane.comp.boot-loaders.u-boot/30484 http://article.gmane.org/gmane.comp.boot-loaders.u-boot/30476 as well. Best regards Dirk P.S.: Wolfgang: Sorry if I missed anything. Do I have to do anything more to get above patches applied? Thanks. > I can post a proper RPM .spec file and all required patches for GCC 4.2.0 > based toolchain (arm-elf, for standalone applications e.g. U-Boot with > almost complete multilib less EP9312 FPU) if somebody wants them... .... >>But something breaks if I change the dvevm.h file to #define >>CFG_USE_NAND. The compile breaks at the end with warnings (warnings >>that cause make to error out). They have to do with "function >>'__udivsi3'" being redifined (and the warning that stops the compile is >>this: "arm_v5t_le-ld: Warning: size of symbol `__udivsi3' changed from >>152 in lib_arm/libarm.a(_udivsi3.o) to 496 in >>/opt/montavista-arm-v4.00-alt/pro/devkit/arm/v5t_le/bin/../lib/gcc/armv5 >>tl-montavista-linuxeabi/3.4.3/libgcc.a(_udivsi3.oS)" >> >>I got this same error last time I made a quick and dirty attempt to get >>1.2.0 to compile for Davinci.