[llvm-dev] ARM baremetal linking

[Peter Smith via llvm-dev]

On Fri, 5 Oct 2018 at 19:52, Goran Mekić via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
>
> On Thu, Oct 04, 2018 at 10:25:41AM +0200, Goran Mekić via llvm-dev wrote:
> > arm-none-eabi-ld --entry=__start -T/usr/home/meka/repos/nuttx/nuttx/configs/nucleo-f4x1re/scripts/f401re.ld -L"/usr/home/meka/repos/nuttx/nuttx/staging" -L"/usr/home/meka/repos/nuttx/nuttx/arch/arm/src/board" -o "/usr/home/meka/repos/nuttx/nuttx/nuttx" --start-group -lsched -ldrivers -lconfigs -lc -lmm -larch -lxx -lapps -lfs -lbinfmt -lxx -lboard  "/usr/lib/libgcc.a" --end-group
> I lost few days staring at the debugger only to realize that
> "/usr/lib/libgcc.a" is my system lib. If I'm right, it comes from
> compiler-rt, and what I would need is compiler-rt for baremetal ARM. If
> I'm right, as CFLAGS for crosscompile are
>
> -target arm-none-eabi -march=armv7-m -mcpu=cortex-m4
>
> that's the exact chip I would need to build compiler-rt for. How could
> this be done? Also, if anyone is working on compiling for armv7-m without
> using GNU tools, I'm more than interested in joining forces (hopefully
> someone has success).
>
> The reason I think this is true are these commands:
>
> clang -print-libgcc-file-name # using the system lib
> /usr/lib/libgcc.a
>
> clang -target arm-none-eabi -print-libgcc-file-name # using the baremetal lib
> /usr/lib/clang/6.0.1/lib/libclang_rt.builtins-arm.a
>
> As that's where "/usr/lib/libgcc.a" in original command comes from, I
> really hope I figured this right, but do correct me if I'm wrong.
>
> Regards,
> meka
>
> PS. I tried to do the same last year, but I felt I need to learn a lot
> before dealing with LLVM-only baremetal development, and I did. If you
> feel I need some more info/literature, please do point it out, I'm more
> than willing to learn.

Hello again,

I've written

It is possible to build compiler-rt for v6-m and v7-m although it can
be a bit of a struggle to set up. There are a couple of ways that I
know of. The first is to use the BaremetalARM.cmake cache file in
tools/clang/ and the second is to do a standalone build of
compiler-rt.

There isn't a lot of documentation about using the BaremetalARM.cmake
cache file. The best that I know of is the commit message that I've
quoted below:
Don't defer to the GCC driver for linking arm-baremetal

Also comes with a cmake cache for building the runtime bits:

 $ cmake <normal cmake flags> \
   -DBAREMETAL_ARMV6M_SYSROOT=/path/to/sysroot \
   -DBAREMETAL_ARMV7M_SYSROOT=/path/to/sysroot \
   -DBAREMETAL_ARMV7EM_SYSROOT=/path/to/sysroot \
   -C /path/to/clang/cmake/caches/BaremetalARM.cmake \
   /path/to/llvm

https://reviews.llvm.org/D33259

I got this to work a year or so ago. The main missing bit of
information was that this required compiler-rt to be in the runtimes
directory and not in the projects directory as suggested by
https://llvm.org/docs/GettingStarted.html it also uses hardware
floating point and the hard-floating point float-abi for v7m which may
not be what you want depending on whether your cortex-m4 has floating
point support.

For building compiler-rt standalone for Arm there is the docs page
https://llvm.org/docs/HowToCrossCompileBuiltinsOnArm.html . That will
be a good place to start. The hardest part is usually fighting your
way past the CMAKE try compile step. I strongly recommend using a
version of cmake >= 3.6 so that you can use the
-DCMAKE_TRY_COMPILE_TARGET_TYPE=STATIC_LIBRARY .

Hope that this gets you a little bit further.

Peter


> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev