Re: gcc 3.3.5 arm-elf .. which multilib for soft-float?

[Richard Earnshaw <rearnsha at gcc dot gnu dot org>]

On Tue, 2004-11-02 at 00:21, Joel Sherrill  wrote:
> Hi,
> 
> It looks like btween gcc 3.3.3 and 3.3.5, the file config/arm/t-arm-elf
> was reworked.  I am getting linking errors which indicate that the
> default libc.a is hard-float.
> 
> The gcc documentation states that -mhard-float is the default for the
> compiler and I only see these libc.a's
> 

-mhard-float has *never* been the default for the arm-elf
configuration.  In fact, AFAIK the only supported configuration of gcc
on ARM that still uses the FPA hard-float format is Linux.

> ./lib/thumb/libc.a
> ./lib/libc.a
> 

Hard float doesn't (can't) work with Thumb (there are no instructions to
access the co-processor space).  Building the hard-float libraries is
disabled by default because virtually nobody wants them (why would they?
soft-float code runs faster when you don't have an FPA, and the only
chip to have that was the ARM7500FE).

> I think the soft-float multilib was deleted and it shouldn't have been.
> I think someone wanted to revert back to the 3.2.x behavior.
> In our gcc 3.2.x toolset, we had this:
> 
> ./lib/fpu/libc.a
> ./lib/thumb/fpu/libc.a
> ./lib/thumb/libc.a
> ./lib/libc.a
> 

This must have been a local change -- FSF gcc has never created
thumb+fpu - it's meaningless.

> But in 3.3.3, there were these
> 
> be/fpu/interwork/libc.a
> be/fpu/libc.a
> be/fpu/nofmult/libc.a
> be/interwork/libc.a
> be/libc.a
> be/nofmult/libc.a
> fpu/interwork/libc.a
> fpu/libc.a
> fpu/nofmult/libc.a
> interwork/libc.a
> libc.a
> nofmult/libc.a
> thumb/be/interwork/libc.a
> thumb/be/libc.a
> thumb/interwork/libc.a
> thumb/libc.a
> 
> Can we have a happy medium and go back to the gcc 3.2.x behavior?
> 
Not the way you've described it.  It's wrong.

R.