[Bug target/80505] FAIL: gcc.dg/ipa/iinline-attr.c scan-ipa-dump inline "hooray[^\\n]*inline copy in test"
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80505 Segher Boessenkool changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #6 from Segher Boessenkool --- This all works now. Please reopen if you still see FAILs.
[Bug target/80505] FAIL: gcc.dg/ipa/iinline-attr.c scan-ipa-dump inline "hooray[^\\n]*inline copy in test"
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80505 Segher Boessenkool changed: What|Removed |Added Target|ppc64le-linux-gnu |powerpc*-*-* Host|ppc64le-linux-gnu |powerpc*-*-* Build|ppc64le-linux-gnu |powerpc*-*-* --- Comment #5 from Segher Boessenkool --- It is broken on any powerpc target (also BE, 32-bit, AIX, and Darwin).
[Bug target/80505] FAIL: gcc.dg/ipa/iinline-attr.c scan-ipa-dump inline "hooray[^\\n]*inline copy in test"
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80505 --- Comment #4 from Martin Jambor --- (In reply to Martin Liška from comment #3) > > And by the way, I tend to think that the testcase should use > > -fno-indirect-inlining option, the indirect call is direct when the > > IPA layer sees it. > > Why should we add -fno-indirect-inlining to the test? I meant -fno-early-inlining, sorry. The reason is that early inlining inlines hiphip() to test() and then constant propagation turns the indirect call into a direct one before IPA inliner (well, IPA-CP, but anyway) sees it. So, at least as far as internal representation is concerned, we never actually do inlining of an indirect call. However, it really depends on why such a test using indirect calls was added in the first place and that (i.e. why direct calls are not enough) is not clear to me. Maybe we are fine.
[Bug target/80505] FAIL: gcc.dg/ipa/iinline-attr.c scan-ipa-dump inline "hooray[^\\n]*inline copy in test"
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80505 Martin Liška changed: What|Removed |Added CC||dje at gcc dot gnu.org, ||segher at gcc dot gnu.org --- Comment #3 from Martin Liška --- (In reply to Martin Jambor from comment #2) > The testcase was added with a i386 specific patch to test that > inlining of indirect calls in presence of __optimize__ attribute. > > I am not sure whether inlining, even direct, in presence of the > attribute is supposed to work on ppc64le, but it shows that it does > not. Even early inliner dump shows that an attempt to inline hiphip/1 > into test/2 fails because of "optimization level attribute mismatch." > I suppose we need something like r222559 for ppc. Yes, I really believe so. I'm CCing port maintainers. > > If not, the testcase should be moved to a i386 specific directory. > > And by the way, I tend to think that the testcase should use > -fno-indirect-inlining option, the indirect call is direct when the > IPA layer sees it. Why should we add -fno-indirect-inlining to the test?
[Bug target/80505] FAIL: gcc.dg/ipa/iinline-attr.c scan-ipa-dump inline "hooray[^\\n]*inline copy in test"
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80505 Martin Jambor changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2017-04-27 Ever confirmed|0 |1
[Bug target/80505] FAIL: gcc.dg/ipa/iinline-attr.c scan-ipa-dump inline "hooray[^\\n]*inline copy in test"
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80505 Martin Jambor changed: What|Removed |Added CC||jamborm at gcc dot gnu.org Component|testsuite |target --- Comment #2 from Martin Jambor --- The testcase was added with a i386 specific patch to test that inlining of indirect calls in presence of __optimize__ attribute. I am not sure whether inlining, even direct, in presence of the attribute is supposed to work on ppc64le, but it shows that it does not. Even early inliner dump shows that an attempt to inline hiphip/1 into test/2 fails because of "optimization level attribute mismatch." I suppose we need something like r222559 for ppc. If not, the testcase should be moved to a i386 specific directory. And by the way, I tend to think that the testcase should use -fno-indirect-inlining option, the indirect call is direct when the IPA layer sees it.