https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117954
Bug ID: 117954
Summary: [15 Regression] gcc.dg/tree-ssa/pr111456-1.c FAILs for
non-LOGICAL_OP_NON_SHORT_CIRCUIT after
r15-5646-gd1cf0d7a0f27fd
Product: gcc
Version: 15.0
Status: UNCONFIRMED
Keywords: testsuite-fail
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: hp at gcc dot gnu.org
CC: rguenth at gcc dot gnu.org
Target Milestone: ---
It appears that r15-5646-gd1cf0d7a0f27fd caused gcc.dg/tree-ssa/pr111456-1.c to
regress for all non-LOGICAL_OP_NON_SHORT_CIRCUIT targets, alternatively
LOGICAL_OP_NON_SHORT_CIRCUIT targets when compiling that test-case with --param
logical-op-non-short-circuit=0 as evidenced by the presence of this line:
FAIL: gcc.dg/tree-ssa/pr111456-1.c scan-tree-dump-not optimized "foo "
in posts to gcc-testresults around that commit for some targets (like
pru-unknown-elf s390x-ibm-linux-gnu, powerpc64le-unknown-linux-gnu
m68k-unknown-linux-gnu) but not for others (armv8l-unknown-linux-gnueabihf,
aarch64-unknown-linux-gnu, x86_64-pc-linux-gnu, i686-pc-linux-gnu). Also seen
for cris-elf: before this commit the whole test-case was optimized to return 0
from main, while after, there's a loop with foo called.
The commit log for r15-5646-gd1cf0d7a0f27fd does mention fallout and has
patches to handle them in various ways, for cases "where we rely on unrolling
even when calls are involved". But, here the call seems "involved" primarily
in that it isn't optimized out. It's not clear whether this particular fallout
is expected.
If it's "expected" I suggest xfailing the test and adding "--param
logical-op-non-short-circuit=0" to make it (x)fail for all targets.
Other "workarounds": set "--param logical-op-non-short-circuit=0" to make it
pass for all targets, or adding a "#pragma GCC unroll 2" before the loop,
making the test pass.
Testresults post examples; affected targets, passing before the triggering
commit
r15-5620-ga095d720acd
pru-unknown-elf
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831214.html
r15-5592-gcdd7171a6b0
s390x-ibm-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831146.html
r15-5638-g97717af3867
powerpc64le-unknown-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831546.html
r15-5637-gcc5779fcaf76ae
m68k-unknown-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831318.html
Failing after that commit:
r15-5663-gf89c7916b97
pru-unknown-elf
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831335.html
r15-5655-ge876acab6cd
s390x-ibm-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831310.html
r15-5649-g0b47d7579e8
powerpc64le-unknown-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831509.html
r15-5662-ge6d26a21e03690
m68k-unknown-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831378.html
Unaffected targets:
15-5663-gf89c7916b97
armv8l-unknown-linux-gnueabihf
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831345.html
15-5650-gd9c908b75039
aarch64-unknown-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831310.html
r15-5649-g0b47d7579e8
x86_64-pc-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831302.html
r15-5648-ge984ba40aa0
i686-pc-linux-gnu
https://gcc.gnu.org/pipermail/gcc-testresults/2024-November/831300.html
