On Tue, 24 Feb 2004, Wolfgang Denk wrote:
In message Pine.LNX.4.56.0402241626510.21705@mag.sysgo.com Marius Groeger wrote:
This is a problem with the gcc/binutils toochain, which doesn't handle -msoft-float the way we want it. The error you're seeing isn't really one, since neither the U-Boot code nor the quoted functions in libgcc.a actually use FP instructions. It's only the ELF binary tags that don't match.
I'm not sure. I don't know for certain for ARM, but on other archi- tectures GCC will happily generate FP instructions for example to perform fast memory copy operations (like when assigning structures of the "right" size).
I'm having similar doubts, as I wrote in my previous post. Please read it again (to the end.)
In practice, however, I have not seen this kind of optimising as of yet. Maybe on ARM there is no gain in performance. Unless there is a proper (read: for all ARM architectures, and working like we know it from powerpc-linux) -msoft-float patch for GCC 3, I'm inclined to just (carefully) drop -msoft-float from *boot.
I'm curious about Robert Schwebel's assessment of this issue, since I know he's been workin on this.
Regards, Marius