[Bug ipa/83346] inliner crash with attribute always_inline/flatten on a destructor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83346 Jakub Jelinek changed: What|Removed |Added Status|NEW |RESOLVED CC||jakub at gcc dot gnu.org Resolution|--- |FIXED --- Comment #7 from Jakub Jelinek --- Fixed.
[Bug ipa/83346] inliner crash with attribute always_inline/flatten on a destructor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83346 --- Comment #6 from Jakub Jelinek --- Author: jakub Date: Mon Jun 25 17:02:47 2018 New Revision: 262046 URL: https://gcc.gnu.org/viewcvs?rev=262046&root=gcc&view=rev Log: Backported from mainline 2017-12-19 Jakub Jelinek PR ipa/82801 PR ipa/83346 * ipa-inline.c (flatten_remove_node_hook): New function. (ipa_inline): Keep only nodes with flatten attribute at the end of the array in the order from ipa_reverse_postorder, only walk that portion of array for flattening, if there is more than one such node, temporarily register a removal hook and ignore removed nodes. * g++.dg/ipa/pr82801.C: New test. Added: branches/gcc-6-branch/gcc/testsuite/g++.dg/ipa/pr82801.C Modified: branches/gcc-6-branch/gcc/ChangeLog branches/gcc-6-branch/gcc/ipa-inline.c branches/gcc-6-branch/gcc/testsuite/ChangeLog
[Bug ipa/83346] inliner crash with attribute always_inline/flatten on a destructor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83346 --- Comment #5 from Jakub Jelinek --- Author: jakub Date: Fri Dec 22 08:50:30 2017 New Revision: 255968 URL: https://gcc.gnu.org/viewcvs?rev=255968&root=gcc&view=rev Log: Backported from mainline 2017-12-19 Jakub Jelinek PR ipa/82801 PR ipa/83346 * ipa-inline.c (flatten_remove_node_hook): New function. (ipa_inline): Keep only nodes with flatten attribute at the end of the array in the order from ipa_reverse_postorder, only walk that portion of array for flattening, if there is more than one such node, temporarily register a removal hook and ignore removed nodes. * g++.dg/ipa/pr82801.C: New test. Added: branches/gcc-7-branch/gcc/testsuite/g++.dg/ipa/pr82801.C Modified: branches/gcc-7-branch/gcc/ChangeLog branches/gcc-7-branch/gcc/ipa-inline.c branches/gcc-7-branch/gcc/testsuite/ChangeLog
[Bug ipa/83346] inliner crash with attribute always_inline/flatten on a destructor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83346 --- Comment #4 from Jakub Jelinek --- Author: jakub Date: Tue Dec 19 07:41:30 2017 New Revision: 255805 URL: https://gcc.gnu.org/viewcvs?rev=255805&root=gcc&view=rev Log: PR ipa/82801 PR ipa/83346 * ipa-inline.c (flatten_remove_node_hook): New function. (ipa_inline): Keep only nodes with flatten attribute at the end of the array in the order from ipa_reverse_postorder, only walk that portion of array for flattening, if there is more than one such node, temporarily register a removal hook and ignore removed nodes. * g++.dg/ipa/pr82801.C: New test. Added: trunk/gcc/testsuite/g++.dg/ipa/pr82801.C Modified: trunk/gcc/ChangeLog trunk/gcc/ipa-inline.c trunk/gcc/testsuite/ChangeLog
[Bug ipa/83346] inliner crash with attribute always_inline/flatten on a destructor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83346 --- Comment #3 from Andi Kleen --- Fixed by https://gcc.gnu.org/ml/gcc-patches/2017-12/msg00764.html
[Bug ipa/83346] inliner crash with attribute always_inline/flatten on a destructor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83346
Richard Biener changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last reconfirmed||2017-12-11
CC||hubicka at gcc dot gnu.org,
||rguenth at gcc dot gnu.org
Ever confirmed|0 |1
--- Comment #2 from Richard Biener ---
(gdb) p node->debug ()
_ZN12_GLOBAL__N_11bIiNS_12_GLOBAL__N_12acIcccEEED2Ev/1 ({anonymous}::b >::~b()) @0x76db8170
Type: function definition analyzed
Visibility: prevailing_def_ironly artificial
References:
Referring:
Function {anonymous}::b >::~b()/1 is
inline copy in {anonymous}::ae::~ae()/6
Availability: local
First run: 0
Function flags: count: 574129753 (estimated locally) body local
Called by: {anonymous}::ae::~ae()/6 (inlined) (574129753 (estimated
locally),0.53 per call)
Calls: {anonymous}::{anonymous}::ac,
>::~ac() [with = char;
= char; = char]/15 (574129752
(estimated locally),0.53 per call)
so it's an inline clone. BUT ... at the start of the loop it is
_ZN12_GLOBAL__N_11bIiNS_12_GLOBAL__N_12acIcccEEED1Ev/2 ({anonymous}::b >::~b()) @0x76db82e0
Type: function definition analyzed alias cpp_implicit_alias
Visibility: prevailing_def_ironly artificial
References: _ZN12_GLOBAL__N_11bIiNS_12_GLOBAL__N_12acIcccEEED2Ev/1 (alias)
Referring:
Availability: local
First run: 0
Function flags: local
Called by: {anonymous}::ae::~ae()/6 (574129753 (estimated locally),0.53 per
call)
Calls:
so I'm not sure if 'order' stays valid during flatten processing or if we're
just lucky in re-using GC memory with the inline clone.
Honza?
