[Bug target/55303] New: [SH] Add support for clips / clipu instructions

[Bug target/55303] New: [SH] Add support for clips / clipu instructions

[olegendo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55303

             Bug #: 55303
           Summary: [SH] Add support for clips / clipu instructions
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P3
         Component: target
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: olegendo@gcc.gnu.org
            Target: sh2a*-*-*


Support for the following SH2A specific instructions should be added:

clips.b  Rn
clips.w  Rn
clipu.b  Rn
clipu.w  Rn

These can be used to implement saturating arithmetic, for example.

[Bug target/55303] [SH] Add support for clips / clipu instructions

[olegendo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55303

Oleg Endo <olegendo at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2013-03-02
         AssignedTo|unassigned at gcc dot       |olegendo at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #1 from Oleg Endo <olegendo at gcc dot gnu.org> 2013-03-02 16:16:41 UTC ---
Created attachment 29567
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=29567
Working patch with a thinko.

This patch, albeit working, has a thinko.
The idea was to reduce the constraints of the clips/clipu insn comparison
constants by adding/subtracting a constant offset value before/after the actual
clipping insn.  For example:

int
test_02 (int a)
{
  return max (0, min (255, a));
}

becomes:

_test_02:
        movi20  #128,r1
        sub     r1,r4
        mov     r4,r0
        clips.b r0
        rts
        add     r1,r0

The problem with this is that it won't work for values that will wrap-around
before/after the offset subtraction/addition.

E.g. plugging the value 0x80000000 (−2147483648) into the above case:

        movi20  #128,r1
        sub     r1,r4     // r4 =  0x80000000 - 128 = 0x7FFFFF80
        mov     r4,r0
        clips.b r0        // !(r0 < -128) && (r0 > 127) -> r0 = 127
        rts
        add     r1,r0     // r0 = 127 + 128 = 255
                          // expected result: 0

Maybe this case could be handled by using subv/addv insns to catch
over/underflows somehow, but probably the resulting code would be more complex
(and thus slower) than two straight forward compare-and-branch sequences.

On the other hand, if it is known that the input value is in a certain range
(e.g. a sign/zero extended HImode or QImode), the offset approach should work
fine.

I will modify the attached patch so that it will allow only the HW clip
constants for now.

[Bug target/55303] [SH] Add support for clips / clipu instructions

[olegendo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55303

--- Comment #2 from Oleg Endo <olegendo at gcc dot gnu.org> 2013-03-02 16:23:57 UTC ---
For non-SH2A targets there is an opportunity to generate better insn sequences
for the special case

unsigned int test (unsigned int a)
{
  return a > 1 ? 1 : a;
}

on SH2A:
   tst     r4,r4
   movrt   r0

if zero-displacement branches are not good:
   tst     r4,r4
   mov     #-1,r0
   negc    r0,r0

if zero-displacement branches are good:
   tst     r4,r4
   bt      0f
   mov     #1,r1
0f:

This can be done by implementing a pattern
  (umin:SI (match_operand:SI 1 "arith_reg_operand")
           (const_int 1))

as it is already done for SH2A in attachment 29567.

[Bug target/55303] [SH] Add support for clips / clipu instructions

[olegendo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55303

--- Comment #3 from Oleg Endo <olegendo at gcc dot gnu.org> 2013-05-06 05:48:18 UTC ---
(In reply to comment #1)
> I will modify the attached patch so that it will allow only the HW clip
> constants for now.

This has been committed as rev 198617:
http://gcc.gnu.org/viewcvs/gcc?view=revision&revision=198617

    PR target/55303
    * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
    * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
    related expanders.
    * config/sh/iterators.md (SMIN_SMAX): New code iterator.
    * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
    clips_min_const_int, clips_max_const_int, clipu_max_const_int):
    New predicates.

    PR target/55303
    * gcc.target/sh/pr55303-1.c: New.
    * gcc.target/sh/pr55303-2.c: New.
    * gcc.target/sh/pr55303-3.c: New.