[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 Richard Biener changed: What|Removed |Added Status|ASSIGNED|RESOLVED Resolution|--- |FIXED --- Comment #12 from Richard Biener --- Not really.
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 Arseny Solokha changed: What|Removed |Added CC||asolokha at gmx dot com --- Comment #11 from Arseny Solokha --- Is the backport to GCC 8 still pending?
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 Bug 85960 depends on bug 86124, which changed state. Bug 86124 Summary: [9 Regression] ICE in create_variable_info_for, at tree-ssa-structalias.c:6123 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86124 What|Removed |Added Status|ASSIGNED|RESOLVED Resolution|--- |FIXED
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 --- Comment #10 from Gianni Tedesco --- Confirmed. Code from trunk works splendidly on the code-base where this bug was discovered.
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 --- Comment #9 from Gianni Tedesco --- Sure, I can test it on real world cases over the weekend I think. Nice work guys. Thanks.
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 Richard Biener changed: What|Removed |Added Known to work||9.0 Known to fail||8.1.0 --- Comment #8 from Richard Biener --- Fixed on trunk sofar. As it is unlikely a regression I'm only considering a backport to GCC 8 but also only after some soaking on trunk. Gianni - can you put the fix (that was actually committed, it has some fixed over the attached patch) on more testing on your code? It should trivially apply to GCC 8 as well.
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 --- Comment #7 from Richard Biener --- Author: rguenth Date: Fri Jun 1 08:20:08 2018 New Revision: 261056 URL: https://gcc.gnu.org/viewcvs?rev=261056=gcc=rev Log: 2018-06-01 Richard Biener PR ipa/85960 * tree-ssa-structalias.c (get_function_part_constraint): Handle NULL fi->decl. (find_func_aliases_for_call): Properly handle indirect fi from direct call. (find_func_clobbers): Likewise. (ipa_pta_execute): Likewise. (create_variable_info_for): For functions that are ifunc_resolver resolve to a varinfo that contains the result of the resolver call. (associate_varinfo_to_alias): Do not treat ifunc resolvers as aliases. * gcc.dg/ipa/ipa-pta-19.c: New testcase. Added: trunk/gcc/testsuite/gcc.dg/ipa/ipa-pta-19.c Modified: trunk/gcc/ChangeLog trunk/gcc/testsuite/ChangeLog trunk/gcc/tree-ssa-structalias.c
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 Richard Biener changed: What|Removed |Added Assignee|marxin at gcc dot gnu.org |rguenth at gcc dot gnu.org --- Comment #6 from Richard Biener --- Created attachment 44211 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=44211=edit patch I am testing the following complete patch.
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960
Richard Biener changed:
What|Removed |Added
Keywords||wrong-code
CC||hubicka at gcc dot gnu.org
Version|unknown |8.1.1
--- Comment #5 from Richard Biener ---
The issue seems to be that the multiver stuff is
multiver_test/35 (multiver_test) @0x7ff2d74c62e0
Type: function definition analyzed alias
Visibility: externally_visible public
References: resolve_multiver/34 (alias)
Referring:
Availability: overwritable
First run: 0
Function flags:
Called by: main/37 (955630223 (estimated locally),8.09 per call)
Calls:
resolve_multiver/34 (resolve_multiver) @0x7ff2d74c6170
Type: function definition analyzed
Visibility: prevailing_def_ironly
References: multiver_default/33 (addr)
Referring: multiver_test/35 (alias)
Availability: available
First run: 0
Function flags: count: 1073741826 (estimated locally) body nonfreeing_fn
Called by:
Calls:
multiver_default/33 (multiver_default) @0x7ff2d74c6000
Type: function definition analyzed
Visibility: prevailing_def_ironly
Address is taken.
References:
Referring: resolve_multiver/34 (addr)
Availability: available
First run: 0
Function flags: count: 1073741826 (estimated locally) body nonfreeing_fn hot
Called by:
Calls:
so multiver_test is an alias of resolve_multiver which means we get constraints
for resolve_multiver instead of the union of resolve_multiver and
multiver_default.
In IPA PTA we do
/* Build the constraints. */
FOR_EACH_DEFINED_FUNCTION (node)
{
...
node->call_for_symbol_thunks_and_aliases
(associate_varinfo_to_alias, vi, true);
where we "merge" aliases into the DEFINED copy. How we represent ifuncs wrecks
with this it seems. We use the resolve_multiver data for calls to
multiver_test. We probably can add a ifunc attribute lookup but IMHO
the cgraph representation is just bogus. It looks like the multiver_test
cgraph node has ifunc_resolver set to 1(!?) but resolve_multiver doesn't.
/* Set when symbol is an IFUNC resolver. */
unsigned ifunc_resolver : 1;
I expected resolve_multiver to be the IFUNC resolver.
So without actually understanding this mess the following "fixes" the issue:
Index: gcc/tree-ssa-structalias.c
===
--- gcc/tree-ssa-structalias.c (revision 260950)
+++ gcc/tree-ssa-structalias.c (working copy)
@@ -7713,7 +7713,8 @@ associate_varinfo_to_alias (struct cgrap
if ((node->alias
|| (node->thunk.thunk_p
&& ! node->global.inlined_to))
- && node->analyzed)
+ && node->analyzed
+ && !node->ifunc_resolver)
insert_vi_for_tree (node->decl, (varinfo_t)data);
return false;
}
My point is that multiver_test doesn't call resolve_multiver but it calls
one of the functions returned by the resolver.
So - a different fix from the above would handle this at the point
we generate constraints for the call to multiver_test and emit constraints
for fn = resolve_multiver (); (*fn) (args...). That looks needed because
if you make multiver_test static the constraints still look wrong. We still
need to avoid associating one decl with the other like above.
And we need to get the clobbers on the ifunc call correct...
Half-way patch where the points-to sets look good but the clobbers/uses are
still broken:
Index: gcc/tree-ssa-structalias.c
===
--- gcc/tree-ssa-structalias.c (revision 260950)
+++ gcc/tree-ssa-structalias.c (working copy)
@@ -3903,7 +3903,7 @@ get_function_part_constraint (varinfo_t
c.offset = 0;
c.type = SCALAR;
}
- else if (TREE_CODE (fi->decl) == FUNCTION_DECL)
+ else if (fi->decl && TREE_CODE (fi->decl) == FUNCTION_DECL)
{
varinfo_t ai = first_vi_for_offset (fi, part);
if (ai)
@@ -4732,6 +4732,30 @@ find_func_aliases_for_call (struct funct
return;
fi = get_fi_for_callee (t);
+ /* First process an ifunc resolver if we run into one. */
+ if (fndecl)
+{
+ cgraph_node *node = cgraph_node::get (fndecl);
+ if (node->ifunc_resolver)
+ {
+ fi = get_vi_for_tree (node->get_alias_target ()->decl);
+ if (fi->is_fn_info)
+ {
+ constraint_expr rhs
+ = get_function_part_constraint (fi, fi_result);
+ fi = new_var_info (NULL_TREE, "ifuncres", true);
+ fi->is_reg_var = true;
+ constraint_expr lhs;
+ lhs.type = SCALAR;
+ lhs.var = fi->id;
+ lhs.offset = 0;
+ process_constraint (new_constraint (lhs, rhs));
+ }
+ else
+ fi = get_varinfo (anything_id);
+ fndecl = NULL_TREE;
+ }
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 Martin Liška changed: What|Removed |Added Status|WAITING |ASSIGNED Assignee|unassigned at gcc dot gnu.org |marxin at gcc dot gnu.org --- Comment #4 from Martin Liška --- Confirmed with: $ gcc pr85960.c -O2 -Wall -fipa-pta && ./a.out 2 Let me take a look.
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 --- Comment #3 from Gianni Tedesco --- Both gcc-7.3 and gcc-8 from ubuntu on x86-64: gcc version 8.0.1 20180414 (experimental) [trunk revision 259383] (Ubuntu 8-20180414-1ubuntu2) gcc version 7.3.0 (Ubuntu 7.3.0-16ubuntu3)
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 --- Comment #2 from Gianni Tedesco --- Both gcc-7.3 and gcc-8 from ubuntu on x86-64: gcc version 8.0.1 20180414 (experimental) [trunk revision 259383] (Ubuntu 8-20180414-1ubuntu2) gcc version 7.3.0 (Ubuntu 7.3.0-16ubuntu3)
[Bug ipa/85960] -fipa-pta and ifunc are incompatible
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85960 Richard Biener changed: What|Removed |Added Status|UNCONFIRMED |WAITING Last reconfirmed||2018-05-29 CC||rguenth at gcc dot gnu.org Ever confirmed|0 |1 --- Comment #1 from Richard Biener --- Please specify the GCC version and architecture you are using. I cannot reproduce this on x86_64-linux with 4.8.5, 4.9.4, 5.5, 6.4, 7.3, 8.1 or current trunk.
