I don't know about that problem, but I don know that the arm-elf-gcc 3.2.1 compiler does generate bad code sometimes. I switched to 3.4.3 and no more bad code generation. It's been too long for me to remember the specifics. I think it had something to do with not properly manipulating the stack pointer allowing for a race condition that could corrupt the stack.
Jay -----Original Message----- From: Grant Edwards [mailto:[email protected]] Sent: Wednesday, February 25, 2009 12:38 PM To: [email protected] Subject: [ECOS] arm-elf-gcc 3.2.1 bug: "unsigned" vs. "unsigned long"? Somebody has reported to me that he thinks the arm-elf-gcc 3.2.1 toolchain we use for eCos development has a bug such that declaring a variable as "unsigned" doesn't always generate correct code, but changing the declaration to "unsigned long" will fix the problem. Moreover he thinks that files containing "unsigned" variables will compile/link correctly sometimes and incorrectly at other times. But, changing "unsigned" declarations to "unsigned long" makes things better. No, there aren't any source/object files that demonstrate the problem. :/ I've been using this toolchain for a long time (at least 6-7 years), and I've never observed this problem or heard anybody else report it. Google hasn't found anything that sounds similar, but I thought I should ask here if anybody has heard of such a problem. -- Grant Edwards grante Yow! at BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI-BI- visi.com -- Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss -- Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss
