arm: fix fallout from recent thumb2 detection patch

The recent change to correct the detection of thumb2 object files
resulted in a ld test for veneering starting to fail.  The problem was
the test itself, which was incorrectly expecting thumb1 type far-call
veneers instead of the thumb2 flavour.  We already have a dump file of
the expected form, so the fix is to change the expected output
accordingly.

ld/

	* testsuite/ld-arm/arm-elf.exp (farcall test for v8-m.mainline):
	Correct expected output.

0 files changed, 0 insertions, 0 deletions