mklibgcc fallout

[Richard Sandiford]

Alan Modra <amodra@bigpond.net.au> writes:
> 	* mklibgcc.in: Build shared libgcc and shared libunwind in gcc/.
> 	Don't subst shlib_dir for SHLIB_LINK, SHLIBUNWIND_LINK,
> 	SHLIB_INSTALL, and SHLIBUNWIND_INSTALL.
> 	* config/i386/t-nwld (SHLIB_NAME): Use shlib_base_name in place of
> 	shlib_dir and shlib_so_name.
> 	* config/mips/t-slibgcc-irix (SHLIB_NAME): Likewise.
> 	* config/t-libunwind-elf (SHLIB_NAME): Likewise.
> 	* config/t-slibgcc-darwin (SHLIB_NAME): Likewise.
> 	* config/t-slibgcc-sld (SHLIB_NAME): Likewise.
> 	(SHLIB_LINK): Don't use shlib_dir when creating symlink.

This breaks build-dir testing for me on IRIX because (a) we set the
DT_SONAME of every libgcc_s*.so.1 multilib to libgcc.so.1 and (b) there's
now only one libgcc_so.1 in the build directory (the one associated with
the default multilib).  Tests for the non-default multilibs fail to run.

Is this a known problem?  Although IRIX uses its own t-slibgcc-* file,
I couldn't see anything off-hand that would make the IRIX case different
from powerpc64-linux-gnu.

Richard