Proper cpu-vendor-os triplet?
Joel Sherrill
joel.sherrill@oarcorp.com
Tue Oct 21 14:02:00 GMT 2014
On 10/21/2014 8:03 AM, Freddie Chopin wrote:
> Hello!
>
> I know this list might not be a perfect fit for my question, but I hope
> someone would share some knowledge with me.
>
> I'm developing an open-source RTOS for higher-end microcontrollers (*).
> My plan is to have it as a proper OS for newlib and GCC - like RTEMS
> (for example). I'm having a hard time deciding what should be the proper
> triplet of such toolchain for ARM chips, because either everyone does
> things differently, or I don't understand something... For example for
> "bare metal ARM" the triplet is "arm-none-eabi", so I assume that this
> should be actually read as "arm" cpu, "none" vendor, "none" OS (missing)
> and "eabi" ABI. Or maybe it's the "vendor" that is missing? In newlib's
> configure.host file the OS is taken from the third element in the
> triplet, for example:
>
>> case "${host}" in
>> *-*-cygwin*)
>> posix_dir=posix
>> stdio64_dir=stdio64
>> xdr_dir=xdr
>> ;;
>> *-*-netware*)
>> signal_dir=
>> sys_dir=netware
>> ;;
>> *-*-rtems*) # generic RTEMS support
>> sys_dir=rtems
>> posix_dir=posix
>> unix_dir=unix
>> ;;
>> *-*-tirtos*)
>> sys_dir=tirtos
>> have_crt0="no"
>> ;;
> On the other hand linaro's GCC toolchain for ARM processors is named
> "arm-linux-gnueabihf", so I assume that in this case the "vendor" field
> is missing...
>
> So I wonder if I should name the toolchain "arm-none-distortos-eabi",
> "arm-distortos-eabi", "arm-none-distortos", "arm-distortos" or maybe
> something else?
IMO arm-distortos is what the user should configure.
Internally, the canonical name will be arm-none-distortos.
> Thx in advance!
>
> Regards,
> FCh
>
> (*) - in case anyone is interested - the repo is here -
> https://github.com/DISTORTEC/distortos
>
--
Joel Sherrill, Ph.D. Director of Research & Development
joel.sherrill@OARcorp.com On-Line Applications Research
Ask me about RTEMS: a free RTOS Huntsville AL 35805
Support Available (256) 722-9985
More information about the Newlib
mailing list