public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/98282] [8/9/10/11 Regression] Segmentation fault when compiling with optimization >= 2
Date: Tue, 15 Dec 2020 10:05:51 +0000 [thread overview]
Message-ID: <bug-98282-4-W7Ec5VD28x@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-98282-4@http.gcc.gnu.org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98282
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rguenth at gcc dot gnu.org
--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
So, the ICE is because compute_avail gets called twice on VIEW_CONVERT_EXPR
from NULL to decltype(nullptr):
stmt:
_7 = VIEW_CONVERT_EXPR<<<< Unknown tree: nullptr_type >>>>(0B);
rhs1:
<view_convert_expr 0x7fffe9ddd0e0
type <nullptr_type 0x7fffea1a3c78 decltype(nullptr) unsigned type_6 DI
size <integer_cst 0x7fffea041eb8 constant 64>
unit-size <integer_cst 0x7fffea041ed0 constant 8>
align:64 warn_if_not_align:0 symtab:0 alias-set 14 canonical-type
0x7fffea1a3c78
pointer_to_this <pointer_type 0x7fffea21db28> reference_to_this
<reference_type 0x7fffea1eb738>>
constant
arg:0 <integer_cst 0x7fffea1b0b28 type <pointer_type 0x7fffea1c3a80>
constant 0>
pr98282.ii:19:13 start: pr98282.ii:19:13 finish: pr98282.ii:19:16>
TREE_TYPE (TREE_OPERAND (rhs1, 0)):
<pointer_type 0x7fffea1c3a80
type <integer_type 0x7fffea1c35e8 l sizes-gimplified type_6 SI
size <integer_cst 0x7fffea062108 constant 32>
unit-size <integer_cst 0x7fffea062120 constant 4>
align:32 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0x7fffea05f5e8 precision:32 min <integer_cst 0x7fffea0620c0 -2147483648> max
<integer_cst 0x7fffea0620d8 2147483647>
pointer_to_this <pointer_type 0x7fffea1c3a80>>
public unsigned DI
size <integer_cst 0x7fffea041eb8 type <integer_type 0x7fffea05f0a8
bitsizetype> constant 64>
unit-size <integer_cst 0x7fffea041ed0 type <integer_type 0x7fffea05f000
sizetype> constant 8>
align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0x7fffea0679d8>
TYPE_CANONICAL (TREE_TYPE (TREE_OPERAND (rhs1, 0))):
<pointer_type 0x7fffea0679d8
type <integer_type 0x7fffea05f5e8 int sizes-gimplified public type_6 SI
size <integer_cst 0x7fffea062108 constant 32>
unit-size <integer_cst 0x7fffea062120 constant 4>
align:32 warn_if_not_align:0 symtab:0 alias-set 10 canonical-type
0x7fffea05f5e8 precision:32 min <integer_cst 0x7fffea0620c0 -2147483648> max
<integer_cst 0x7fffea0620d8 2147483647>
pointer_to_this <pointer_type 0x7fffea0679d8>>
unsigned DI
size <integer_cst 0x7fffea041eb8 type <integer_type 0x7fffea05f0a8
bitsizetype> constant 64>
unit-size <integer_cst 0x7fffea041ed0 type <integer_type 0x7fffea05f000
sizetype> constant 8>
align:64 warn_if_not_align:0 symtab:0 alias-set 4 canonical-type
0x7fffea0679d8>
The alias-set 4 above means get_alias_set on the VCE returns 4.
Second case:
stmt:
_8 = VIEW_CONVERT_EXPR<<<< Unknown tree: nullptr_type >>>>(0B);
rhs1:
<view_convert_expr 0x7fffe9ddd120
type <nullptr_type 0x7fffea1a3c78 decltype(nullptr) unsigned type_6 DI
size <integer_cst 0x7fffea041eb8 constant 64>
unit-size <integer_cst 0x7fffea041ed0 constant 8>
align:64 warn_if_not_align:0 symtab:0 alias-set 14 canonical-type
0x7fffea1a3c78
pointer_to_this <pointer_type 0x7fffea21db28> reference_to_this
<reference_type 0x7fffea1eb738>>
constant
arg:0 <integer_cst 0x7fffea1b09f0 type <pointer_type 0x7fffea1bf0a8 i>
constant 0>
pr98282.ii:15:15 start: pr98282.ii:15:15 finish: pr98282.ii:15:15>
TREE_TYPE (TREE_OPERAND (rhs1, 0)):
<pointer_type 0x7fffea1bf0a8 i
type <record_type 0x7fffea1b5dc8 h sizes-gimplified addressable
needs-constructing cxx-odr-p type_5 type_6 BLK
size <integer_cst 0x7fffea1d35b8 constant 96>
unit-size <integer_cst 0x7fffea1d35d0 constant 12>
align:32 warn_if_not_align:0 symtab:0 alias-set 17 canonical-type
0x7fffea1b5dc8
fields <function_decl 0x7fffea213800 operator= type <method_type
0x7fffea212d20>
addressable used nothrow public static weak autoinline decl_3
decl_5 QI defer-output pr98282.ii:60:8 align:16 warn_if_not_align:0 context
<record_type 0x7fffea1b5dc8 h> initial <block 0x7fffea1e3600> result
<result_decl 0x7fffea1fa438 D.3000>
full-name "h& h::operator=(h&&)" chain <function_decl
0x7fffea213700 operator=>> context <translation_unit_decl 0x7fffea04e168
pr98282.ii>
full-name "struct h"
needs-constructor X() X(constX&) this=(X&) n_parents=0 use_template=0
interface-unknown
pointer_to_this <pointer_type 0x7fffea1bf000> reference_to_this
<reference_type 0x7fffea2120a8> chain <type_decl 0x7fffea1be000 h>>
public unsigned DI
size <integer_cst 0x7fffea041eb8 type <integer_type 0x7fffea05f0a8
bitsizetype> constant 64>
unit-size <integer_cst 0x7fffea041ed0 type <integer_type 0x7fffea05f000
sizetype> constant 8>
align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0x7fffea1bf000>
TYPE_CANONICAL (TREE_TYPE (TREE_OPERAND (rhs1, 0))):
<pointer_type 0x7fffea1bf000
type <record_type 0x7fffea1b5dc8 h sizes-gimplified addressable
needs-constructing cxx-odr-p type_5 type_6 BLK
size <integer_cst 0x7fffea1d35b8 constant 96>
unit-size <integer_cst 0x7fffea1d35d0 constant 12>
align:32 warn_if_not_align:0 symtab:0 alias-set 17 canonical-type
0x7fffea1b5dc8
fields <function_decl 0x7fffea213800 operator= type <method_type
0x7fffea212d20>
addressable used nothrow public static weak autoinline decl_3
decl_5 QI defer-output pr98282.ii:60:8 align:16 warn_if_not_align:0 context
<record_type 0x7fffea1b5dc8 h> initial <block 0x7fffea1e3600> result
<result_decl 0x7fffea1fa438 D.3000>
full-name "h& h::operator=(h&&)" chain <function_decl
0x7fffea213700 operator=>> context <translation_unit_decl 0x7fffea04e168
pr98282.ii>
full-name "struct h"
needs-constructor X() X(constX&) this=(X&) n_parents=0 use_template=0
interface-unknown
pointer_to_this <pointer_type 0x7fffea1bf000> reference_to_this
<reference_type 0x7fffea2120a8> chain <type_decl 0x7fffea1be000 h>>
public unsigned type_6 DI
size <integer_cst 0x7fffea041eb8 type <integer_type 0x7fffea05f0a8
bitsizetype> constant 64>
unit-size <integer_cst 0x7fffea041ed0 type <integer_type 0x7fffea05f000
sizetype> constant 8>
align:64 warn_if_not_align:0 symtab:0 alias-set 1 canonical-type
0x7fffea1bf000>
And the alias-set 1 means get_alias_set on the second VCE returns 1.
I think at least for the VCE of constant case (but VCE of SSA_NAME etc. would
as well) we shouldn't consider alias set at all.
So, one possible fix is:
--- gcc/tree-ssa-pre.c.jj 2020-11-30 10:48:31.000000000 +0100
+++ gcc/tree-ssa-pre.c 2020-12-15 10:51:51.061206572 +0100
@@ -4176,13 +4176,17 @@ compute_avail (void)
&& ref2->opcode != MEM_REF
&& ref2 != &operands[0])
--ref2;
- if ((ref1->opcode == TARGET_MEM_REF
- || ref1->opcode == MEM_REF)
- && (TYPE_ALIGN (ref1->type)
- > TYPE_ALIGN (ref2->type)))
- ref1->type
- = build_aligned_type (ref1->type,
- TYPE_ALIGN (ref2->type));
+ if (ref1->opcode == TARGET_MEM_REF
+ || ref1->opcode == MEM_REF)
+ {
+ if (TYPE_ALIGN (ref1->type)
+ > TYPE_ALIGN (ref2->type))
+ ref1->type
+ = build_aligned_type (ref1->type,
+ TYPE_ALIGN (ref2->type));
+ }
+ else
+ set = ref->set;
/* TBAA behavior is an obvious part so make sure
that the hashtable one covers this as well
by adjusting the ref alias set and its base. */
i.e. ignore the alias set if there is no MEM_REF or TARGET_MEM_REF, after all,
the code below it relies on the alias sets to be equal except for MEM_REF or
TARGET_MEM_REF.
Another possibility would be for ao_ref_alias_set and ao_ref_base_alias_set to
return 0 if they aren't really memory reads, such as VCE of constant or VCE of
a SSA_NAME.
And yet another would be not to call ao_ref_alias_set at all in the
compute_avail code for such references, but do something different.
next prev parent reply other threads:[~2020-12-15 10:05 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-14 19:25 [Bug c++/98282] New: " bengilgit at gmail dot com
2020-12-14 19:33 ` [Bug c++/98282] " bengilgit at gmail dot com
2020-12-14 22:06 ` jakub at gcc dot gnu.org
2020-12-14 23:30 ` jakub at gcc dot gnu.org
2020-12-14 23:34 ` [Bug c++/98282] [8/9/10/11 Regression] " jakub at gcc dot gnu.org
2020-12-15 10:05 ` jakub at gcc dot gnu.org [this message]
2020-12-16 9:03 ` [Bug tree-optimization/98282] " rguenth at gcc dot gnu.org
2020-12-16 9:14 ` jakub at gcc dot gnu.org
2021-01-04 10:40 ` rguenth at gcc dot gnu.org
2021-01-04 11:59 ` cvs-commit at gcc dot gnu.org
2021-01-04 12:05 ` [Bug tree-optimization/98282] [8/9/10 " rguenth at gcc dot gnu.org
2021-01-12 14:51 ` cvs-commit at gcc dot gnu.org
2021-03-17 9:35 ` [Bug tree-optimization/98282] [8/9 " cvs-commit at gcc dot gnu.org
2021-03-17 10:20 ` [Bug tree-optimization/98282] [8 " cvs-commit at gcc dot gnu.org
2021-03-17 10:21 ` rguenth at gcc dot gnu.org
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=bug-98282-4-W7Ec5VD28x@http.gcc.gnu.org/bugzilla/ \
--to=gcc-bugzilla@gcc.gnu.org \
--cc=gcc-bugs@gcc.gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).