[Bug testsuite/111850] New: [14 regression]

[Bug testsuite/111850] New: [14 regression]

[seurer at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111850

            Bug ID: 111850
           Summary: [14 regression]
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: testsuite
          Assignee: unassigned at gcc dot gnu.org
          Reporter: seurer at gcc dot gnu.org
  Target Milestone: ---

g:04c9cf5c786b94fbe3f6f21f06cae73a7575ff7a, r14-4664-g04c9cf5c786b94

make  -k check-gcc
RUNTESTFLAGS="powerpc.exp=gcc.target/powerpc/fold-vec-extract-char.p7.c"
FAIL: gcc.target/powerpc/fold-vec-extract-char.p7.c scan-assembler-times
\\maddi\\M 9
# of expected passes            7
# of unexpected failures        1


spawn -ignore SIGHUP /home/seurer/gcc/git/build/gcc-test/gcc/xgcc
-B/home/seurer/gcc/git/build/gcc-test/gcc/
/home/seurer/gcc/git/gcc-test/gcc/testsuite/gcc.target/powerpc/fold-vec-extract-char.p7.c
-fdiagnostics-plain-output -mdejagnu-cpu=power7 -O2 -ffat-lto-objects
-fno-ident -S -o fold-vec-extract-char.p7.s
PASS: gcc.target/powerpc/fold-vec-extract-char.p7.c (test for excess errors)
gcc.target/powerpc/fold-vec-extract-char.p7.c: \\maddi\\M found 6 times
FAIL: gcc.target/powerpc/fold-vec-extract-char.p7.c scan-assembler-times
\\maddi\\M 9
PASS: gcc.target/powerpc/fold-vec-extract-char.p7.c scan-assembler-times
\\mli\\M 6
PASS: gcc.target/powerpc/fold-vec-extract-char.p7.c scan-assembler-times
\\mstxvw4x\\M|\\mstvx\\M|\\mstxv\\M 6
PASS: gcc.target/powerpc/fold-vec-extract-char.p7.c scan-assembler-times
\\mrldicl\\M|\\mrlwinm\\M 3
PASS: gcc.target/powerpc/fold-vec-extract-char.p7.c scan-assembler-times
\\madd\\M 3
PASS: gcc.target/powerpc/fold-vec-extract-char.p7.c scan-assembler-times
\\mlbz\\M 6
PASS: gcc.target/powerpc/fold-vec-extract-char.p7.c scan-assembler-times
\\mextsb\\M 2



This is one of those tests that just counts assembler operations and so may
need adjusting.

seurer@ltcden2-lp1:~/gcc/git/build/gcc-test$ diff
fold-vec-extract-char.p7.s.r14-4663 fold-vec-extract-char.p7.s.r14-4664
12d11
<       rldicl 5,5,0,60
14,15d12
<       addi 10,5,-48
<       add 5,10,1
18c15,17
<       lbz 3,32(5)
---
> 	rldicl 5,5,0,60
> 	add 5,5,1
> 	lbz 3,-16(5)
32d30
<       rldicl 5,5,0,60
34,35d31
<       addi 10,5,-48
<       add 5,10,1
38c34,36
<       lbz 3,32(5)
---
> 	rldicl 5,5,0,60
> 	add 5,5,1
> 	lbz 3,-16(5)
53d50
<       rldicl 5,5,0,60
55,56d51
<       addi 10,5,-48
<       add 5,10,1
59c54,56
<       lbz 3,32(5)
---
> 	rldicl 5,5,0,60
> 	add 5,5,1
> 	lbz 3,-16(5)



commit 04c9cf5c786b94fbe3f6f21f06cae73a7575ff7a (HEAD)
Author: Manolis Tsamis <manolis.tsamis@vrull.eu>
Date:   Mon Oct 16 13:08:12 2023 -0600

    Implement new RTL optimizations pass: fold-mem-offsets

[Bug testsuite/111850] [14 regression] gcc.target/powerpc/fold-vec-extract-char.p7.c fails after r14-4664-g04c9cf5c786b94

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111850

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
           Keywords|                            |testsuite-fail
   Last reconfirmed|                            |2023-10-17
   Target Milestone|---                         |14.0

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---

It looks like the new pass is creating better code.
Before:
<       addi 10,5,-48
<       add 5,10,1
18c15,17
<       lbz 3,32(5)

After:

> 	add 5,5,1
> 	lbz 3,-16(5)


Removing the addi by doing a quick reassociation .

[Bug testsuite/111850] [14 regression] gcc.target/powerpc/fold-vec-extract-char.p7.c fails after r14-4664-g04c9cf5c786b94

[seurer at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111850

--- Comment #2 from seurer at gcc dot gnu.org ---
Yes, it does look like an improvement ad I doubt the number of addis really
matters.  That's why I said the test probably just needs to be updated.

[Bug testsuite/111850] [14 regression] gcc.target/powerpc/fold-vec-extract-char.p7.c fails after r14-4664-g04c9cf5c786b94

[linkw at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111850

Kewen Lin <linkw at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |linkw at gcc dot gnu.org

--- Comment #3 from Kewen Lin <linkw at gcc dot gnu.org> ---
Nice, the mentioned commit make the sub-optimal code gen issue gone (pr101169
#c4). Previously there was one thread on updating these fold-vec-extract*
cases: https://gcc.gnu.org/pipermail/gcc-patches/2023-May/619388.html
, and there were some discussion on the current insn count checking way.

And as my testings before, insn count changes need to be verified on aix as
well.

[Bug testsuite/111850] [14 regression] gcc.target/powerpc/fold-vec-extract-char.p7.c fails after r14-4664-g04c9cf5c786b94

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111850

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug testsuite/111850] [14 regression] gcc.target/powerpc/fold-vec-extract-char.p7.c fails after r14-4664-g04c9cf5c786b94

[linkw at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111850

Kewen Lin <linkw at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |linkw at gcc dot gnu.org

--- Comment #4 from Kewen Lin <linkw at gcc dot gnu.org> ---
Just realized that we also escalated test issue to P1, I'm going to make a
patch for the test case update.

[Bug testsuite/111850] [14 regression] gcc.target/powerpc/fold-vec-extract-char.p7.c fails after r14-4664-g04c9cf5c786b94

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111850

--- Comment #5 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Kewen Lin <linkw@gcc.gnu.org>:

https://gcc.gnu.org/g:f4156fbf7d6d641f8aade8028e87cf302350c3c0

commit r14-8201-gf4156fbf7d6d641f8aade8028e87cf302350c3c0
Author: Kewen Lin <linkw@linux.ibm.com>
Date:   Thu Jan 18 00:00:52 2024 -0600

    testsuite, rs6000: Adjust fold-vec-extract-char.p7.c [PR111850]

    As PR101169 comment #c4 shows, previsouly the addi count
    update on fold-vec-extract-char.p7.c covered a sub-optimal
    code gen issue.  On trunk, pass fold-mem-offsets helps to
    recover the best code sequence, so this patch is to
    revert the count back to the original which matches the
    optimal addi count.

            PR testsuite/111850

    gcc/testsuite/ChangeLog:

            * gcc.target/powerpc/fold-vec-extract-char.p7.c: Update the
            checking count of addi to 6.

[Bug testsuite/111850] [14 regression] gcc.target/powerpc/fold-vec-extract-char.p7.c fails after r14-4664-g04c9cf5c786b94

[linkw at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111850

Kewen Lin <linkw at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #6 from Kewen Lin <linkw at gcc dot gnu.org> ---
Should be fixed on trunk.