Various SH fixes; make R_SH_REL32 partial_inplace etc.

[Hans-Peter Nilsson]

On Sun, 23 Sep 2001, kaz Kojima wrote:
> I've tried to bootstrap current GCC (3.1 CVS) using binutils in
> CVS and your patch in
>   http://sources.redhat.com/ml/binutils/2001-09/msg00302.html
> and found a new problem of SH assembler for a merge section.
> A testcase is the following foo.s:

Just to tell you that I've not forgotten.  Thanks for your
report, test-case, patch -- and for your patience.

> The appended patch adjusts inplace relocation against a symbol in
> a merge section, like as for weak symbols. With this patch, gcc-3.1
> can bootstrap for sh-unknown-linux-gnu.

I assume your patch for reg_overlap_mentioned_for_reload_p to
gcc is still needed.

> BTW, make check for g++ (3.1 CVS) reports 15 failures and 2 XPASSs.
> All failures except for g++.jason/thunk3.C and g++.other/enum5.C
> are for g++.dg/opt/vtgc1.C. XPASS cases are g++.pt/decl2.C and
> g++.robertl/eb17.C.

Very cool.

brgds, H-P