wiki:CrossCompilation

Version 11 (modified by nathanhowell, 3 years ago) (diff)

--

Cross Compiling GHC

As of this moment (GHC 6.12) GHC does not support cross-compilation. There are reasons that we would like it to:

  • TakeoffGW is a distribution of Unix tools for Windows, built by cross-compiling on a Linux machine. They would like to be able to build and distribute GHC this way. It might be useful for us to be able to cross-compile a Windows GHC from Linux too.
  • We could build a 64-bit GHC on OS X, by cross-compiling using the 32-bit version.
  • We could port to Win64 (#1884) by cross-compiling using a 32-bit Windows GHC.
  • Other porting tasks might be easier, given a suitable cross-compilation toolchain.

By way of example, let's suppose we have an x86/Linux platform and we want to cross-compile to PPC64/OSX. Then our build is going to look like this:

Compiler Runs on Generates code for
Stage 0 x86/Linux x86/Linux
Stage 1 x86/Linux PPC64/OSX
Stage 2 PPC64/OSX PPC64/OSX

Where stage 0 is the bootstrap compiler (the one you specify using --with-ghc when configuring the build), and stages 1 and 2 are the compilers being built.

Now some general nomenclature:

  • Build platform: the platform on which the software is being built
  • Host platform: the platform on which the software will run
  • Target platform: for a compiler, the platform on which the generated code will run

These correspond to CPP symbols that are defined when compiling both Haskell and C code:

  • xxx_BUILD_ARCH, xxx_BUILD_OS: the build platform
  • xxx_HOST_ARCH, xxx_HOST_OS: the host platform
  • xxx_TARGET_ARCH, xxx_TARGET_OS: the target platform

The important thing to realise about the 2-stage bootstrap is that each stage has a different notion of build/host/target: these CPP symbols will map to different things when compiling stage 1 and stage 2. Furthermore the RTS and libraries also have a notion of build and host (but not target: they don't generate code).

The overall build has a build/host/target, supplied on the configure command line:

$ ./configure --build=build --host=host --target=target

The host and target specified on the configure command line refer to the stage 2 compiler. Specifically, here is how we map the platforms from the configure command line onto the platforms used by the different stages, and the RTS and libraries:

Overall build Stage 0 Stage 1 Stage 2 libs-host libs-target
Build platform build --- build build build build
Host platform host build build host host target
Target platform target build host target --- ---

Where libs-host refers to the libraries and RTS that we are building to link with the stage 2 compiler, and libs-target refers to the libraries and RTS that will be linked with binaries built by the stage 2 compiler to run on the target platform.

In the special case where we are using cross compilation to bootstrap a new platform, as in the above example, we have host == target:

Overall buildStage 1Stage 2libs-host
Build platform build build build build
Host platform target build target target
Target platformtarget target target

Note that with host == target, libs-host == libs-target, so we only need to build the RTS and libraries once (fortunately, because the GHC build system only supports building them once).

Suppose we wanted to build a cross-compiler to run on the current platform. Then we could configure with build == host, but target is different:

Overall buildStage 1Stage 2libs-hostlibs-target
Build platform build build build build build
Host platform build build build build target
Target platformtarget build target

Note in this configuration that we need both libs-host and libs-target, so currently the GHC build system does not support building this kind of cross-compiler. Fortunately, most of the things you would want to do with this kind of cross-compiler are supported by the first kind, the only caveat is that you can't install a cross-compiler that way.

Plan

Here is how it should work:

$ ./configure --build=<here> --host=<there> --target=<there>

note that we're cross-compiling from the build machine to the host machine. The target machine is the same as the host: the GHC that we're trying to create will generate binaries for host.

No doubt we'll also need to specify some additional configuration parameters to tell the build system where to find our cross-compilation tools. Perhaps something like

--with-host-cc=...
--with-host-as=...
--with-host-ld=...
--with-host-ar=...
--with-host-strip=...
  • stage 1: runs on build, compiles for host
  • stage 2: runs on host, compiles for host

Things that probably need fixing

  • The configure script doesn't let you specify different build, host, and target right now
  • The build system has no distinction between the gcc used to compile from build->build and build->host.
  • We can't build anything with stage2 when cross-compiling, e.g. Haddock and DPH must be disabled.