Hi Michael, 2015-07-16 15:37 GMT+02:00 Michael Olbrich <[email protected]>: > On Thu, Jul 16, 2015 at 03:27:57PM +0200, Guillermo Rodriguez Garcia wrote: >> 2015-07-16 15:14 GMT+02:00 Jean-Claude Monnin <[email protected]>: >> Uhm. I am using a SAMA5D3 Xplained board (so no NEON). However >> Linaro's README file for the cortex-strings project states that these >> should also be available for CPUs without NEON support: >> https://git.linaro.org/toolchain/cortex-strings.git/blob/HEAD:/README >> >> As stated, I used the arm-v7a-linux-gnueabihf from >> OSELAS.Toolchain-2013.12.2 without any modifications or changes and >> the system boots and seems to work just fine. I'll keep an eye on >> this. > > If the cortex-strings are actually used, then any memcpy should crash. > However, it is quite possible that these where ignored by the glibc > build system for ARM v7a. I don't remember the details here. Something like > that happend with some version but I don't remember which. > >> It would be good to konw why cortex-strings was disabled in more >> recent versions of the OSELAS.Toolchain. Can someone shed some light >> here? > > As a said in my other mail glibc no includes a newer version of this code. > I left it there for cortex-a8 because my benchmarks showed that the new > versions did not perform very well there. > I'll revisit the topic for the next major toolchain version.
Thank you very much; this is clear now. Guillermo -- ptxdist mailing list [email protected]
