[Bug target/54089] [SH] Refactor shift patterns

["olegendo at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54089

--- Comment #55 from Oleg Endo <olegendo at gcc dot gnu.org> ---
(In reply to Alexander Klepikov from comment #54)
> Regarding testsuite. There's execute fails, but this is due to lack of
> multilib. I'll rebuild and retest.
> 
> There's also fail in pr64345-1.c, in this function:
> [...]
> 
> But it looks more like it's not a fail, but an optimization.

Yeah, that looks like an improvement.  There might be some SH specific tests
that scan for particular assembler outputs like that one.  Those tests would
need to be adjusted of course.

In that test you can see the unnecessary push/pop of PR.  This is because
initially it wanted to expand as a library call, but then your patterns decided
to change the insns. This can or can't be avoided, depending on the case.

> 
> But also there's tests that pass on patched but fail on clean. I'll take a
> closer look on them later after GCC and multilibs rebuild.

Yes, there are some (well ... quite a lot actually) tests that will also fail
on vanilla GCC on SH.  Hence the need to look at the test result delta
before/after patch.