Hi, I am using the standard OpenEmbedded ARM EABI toolchain and I am not > seeing this issue on the NEO1973 phone (Samsung ARM SOC), or a Xscale > platform. Both kernel configs include: > > CONFIG_AEABI=y > CONFIG_OABI_COMPAT=y > > As I understand it, floating point is emulated at the library level, > and not in the kernel with this setup.
That sounds fine to me, there was some talk on the openezx ML about including CONFIG_OABI_COMPAT=y as the default for the ezx kernel builds; it was suggested that for apps that are eabi compliant the OABI compat should not be needed, but if the neo kernels are being built w/ this config then i'm happy to build my ezx kernels the same way. Are you building the mono-1.2.5pre5 package? I never did find a > solution to the ARM floating point issues with 1.2.4. yes, i'm using the 1.2.5pre5 version, although i've supplemented the bb recipe for the EXTRA_OECONF_arm to include --with-x (for libgdi testing?) and --with-static-mono (you can probably tell i'm new to this and kind of fishing in the dark...) i was right in the middle of testing 1.2.4 and getting an illegal instruction error when i noticed the 1.2.5pre5 version, however i think i was testing with a non-oabi compat kernel at the time so i'm not sure if 1.2.4 would work once the oabi compat was set. with the oabi compat configured a basic helloworld.exe app will run for me, but as soon as i start trying something very simple that needs Windows.Forms (and in turn libgdiplus and X) it bombs with an error from mono/mini/inssel.c (line 3645) which is related to SOFT_FLOAT_IMPL, and the stack trace mentioned nothing about libgdiplus (so i assumed it didn't get that far) so i thought it might have been related to the eabi issue. Is anything known about the status of a libgdiplus_1.2.5 that would correspond to the 1.2.5pre5 version of mono? thanks again for the help, and its great to know that mono is running on the NEO, now i will have to get one ;-) cheers, Mike
_______________________________________________ Mono-devel-list mailing list Mono-devel-list@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-devel-list