* [Bug c/110556] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
@ 2023-07-05 9:20 ` 19373742 at buaa dot edu.cn
2023-07-05 10:30 ` 19373742 at buaa dot edu.cn
` (13 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: 19373742 at buaa dot edu.cn @ 2023-07-05 9:20 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #1 from CTC <19373742 at buaa dot edu.cn> ---
Created attachment 55474
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=55474&action=edit
The compiler output
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug c/110556] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
2023-07-05 9:20 ` [Bug c/110556] " 19373742 at buaa dot edu.cn
@ 2023-07-05 10:30 ` 19373742 at buaa dot edu.cn
2023-07-05 17:54 ` pinskia at gcc dot gnu.org
` (12 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: 19373742 at buaa dot edu.cn @ 2023-07-05 10:30 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #2 from CTC <19373742 at buaa dot edu.cn> ---
The reduced sequence:
-O3 -fno-inline-small-functions -fno-tree-fre -fno-delete-dead-exceptions
-fexpensive-optimizations -fipa-pure-const -fipa-sra -fthread-jumps -ftree-dce
-ftree-dominator-opts -ftree-forwprop -ftree-pre -ftree-tail-merge
-fnon-call-exceptions
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug c/110556] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
2023-07-05 9:20 ` [Bug c/110556] " 19373742 at buaa dot edu.cn
2023-07-05 10:30 ` 19373742 at buaa dot edu.cn
@ 2023-07-05 17:54 ` pinskia at gcc dot gnu.org
2023-07-05 21:28 ` [Bug middle-end/110556] " pinskia at gcc dot gnu.org
` (11 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-07-05 17:54 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
It is a division by 0 being kept around ...
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug middle-end/110556] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (2 preceding siblings ...)
2023-07-05 17:54 ` pinskia at gcc dot gnu.org
@ 2023-07-05 21:28 ` pinskia at gcc dot gnu.org
2023-07-05 22:32 ` pinskia at gcc dot gnu.org
` (10 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-07-05 21:28 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Reducing this but it is taking a little longer as I get to the case where we
get undefined code on accident so I need to test it using
-fsanitize=undefined,address also.
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug middle-end/110556] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (3 preceding siblings ...)
2023-07-05 21:28 ` [Bug middle-end/110556] " pinskia at gcc dot gnu.org
@ 2023-07-05 22:32 ` pinskia at gcc dot gnu.org
2023-07-06 1:36 ` 19373742 at buaa dot edu.cn
` (9 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-07-05 22:32 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #4)
> Reducing this but it is taking a little longer as I get to the case where we
> get undefined code on accident so I need to test it using
> -fsanitize=undefined,address also.
The reduction I got still got an unintialized variable in it ...
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug middle-end/110556] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (4 preceding siblings ...)
2023-07-05 22:32 ` pinskia at gcc dot gnu.org
@ 2023-07-06 1:36 ` 19373742 at buaa dot edu.cn
2023-07-06 1:38 ` pinskia at gcc dot gnu.org
` (8 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: 19373742 at buaa dot edu.cn @ 2023-07-06 1:36 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #6 from CTC <19373742 at buaa dot edu.cn> ---
(In reply to Andrew Pinski from comment #4)
> Reducing this but it is taking a little longer as I get to the case where we
> get undefined code on accident so I need to test it using
> -fsanitize=undefined,address also.
For the unreduced program and optimizations, I test it with
-fno-strict-aliasing -fwrapv -fno-aggressive-loop-optimizations and
-fsanitize=undefined. Both of them got floating point exception too.
I reduced the program with -fsanitize=undefined. The result doesn't get
floating point exception. But its return value is not 0.
# gcc-13 -I /root/csmith/include/csmith-2.3.0/ -O3 -fno-inline-small-functions
-fno-tree-fre -fno-delete-dead-exceptions -fexpensive-optimizations
-fipa-pure-const -fipa-sra -fthread-jumps -ftree-dce -ftree-dominator-opts
-ftree-forwprop -ftree-pre -ftree-tail-merge -fnon-call-exceptions
-fsanitize=undefined -o fails.o testcase.i -Wfatal-errors
# ./fails.o
# echo $?
64
# gcc-13 -I /root/autodl-fs/csmith/include/csmith-2.3.0/ -o works testcase.i
-Wfatal-errors
# ./works
# echo $?
0
# cat testcase.i
int a;
void main() { int b = a;}
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug middle-end/110556] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (5 preceding siblings ...)
2023-07-06 1:36 ` 19373742 at buaa dot edu.cn
@ 2023-07-06 1:38 ` pinskia at gcc dot gnu.org
2023-07-06 5:30 ` pinskia at gcc dot gnu.org
` (7 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-07-06 1:38 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #7 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to CTC from comment #6)
> (In reply to Andrew Pinski from comment #4)
> > Reducing this but it is taking a little longer as I get to the case where we
> > get undefined code on accident so I need to test it using
> > -fsanitize=undefined,address also.
>
> For the unreduced program and optimizations, I test it with
> -fno-strict-aliasing -fwrapv -fno-aggressive-loop-optimizations and
> -fsanitize=undefined. Both of them got floating point exception too.
>
> I reduced the program with -fsanitize=undefined. The result doesn't get
> floating point exception. But its return value is not 0.
That is still undefined code ... as the return value of the app is not
specified in the source.
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug middle-end/110556] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (6 preceding siblings ...)
2023-07-06 1:38 ` pinskia at gcc dot gnu.org
@ 2023-07-06 5:30 ` pinskia at gcc dot gnu.org
2023-07-06 5:42 ` [Bug middle-end/110556] [12/13/14 Regression] " pinskia at gcc dot gnu.org
` (6 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-07-06 5:30 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Created attachment 55485
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=55485&action=edit
Reduced testcase
Compile with ` -O2 -fno-tree-fre -fno-delete-dead-exceptions
-fnon-call-exceptions` to make sure the division stays around.
Note the interesting thing is if I remove the loop in func_6, the code
generation for safe_mul_func_int64_t_s_s changes (to be correct). (that is just
call func_34/func_30 in func_6).
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug middle-end/110556] [12/13/14 Regression] Floating point exception with specific optimizations
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (7 preceding siblings ...)
2023-07-06 5:30 ` pinskia at gcc dot gnu.org
@ 2023-07-06 5:42 ` pinskia at gcc dot gnu.org
2023-07-06 8:15 ` [Bug tree-optimization/110556] [12/13/14 Regression] division of INT_MIN and -1 happening incorrectly with -fno-delete-dead-exceptions -fnon-call-exceptions rguenth at gcc dot gnu.org
` (5 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-07-06 5:42 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Target Milestone|--- |12.4
Keywords| |wrong-code
Status|UNCONFIRMED |NEW
Summary|Floating point exception |[12/13/14 Regression]
|with specific optimizations |Floating point exception
| |with specific optimizations
Known to work| |11.1.0
Last reconfirmed| |2023-07-06
See Also| |https://gcc.gnu.org/bugzill
| |a/show_bug.cgi?id=100394
Known to fail| |12.1.0
--- Comment #9 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.
Note I think this is a latent bug, which was exposed with
r12-475-g8ebf6b99952ada09bf9ea .
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug tree-optimization/110556] [12/13/14 Regression] division of INT_MIN and -1 happening incorrectly with -fno-delete-dead-exceptions -fnon-call-exceptions
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (8 preceding siblings ...)
2023-07-06 5:42 ` [Bug middle-end/110556] [12/13/14 Regression] " pinskia at gcc dot gnu.org
@ 2023-07-06 8:15 ` rguenth at gcc dot gnu.org
2023-07-06 11:45 ` rguenth at gcc dot gnu.org
` (4 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: rguenth at gcc dot gnu.org @ 2023-07-06 8:15 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rguenth at gcc dot gnu.org
--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
Needs investigation on what actually does an invalid transform.
-fno-thread-jumps avoids the issue.
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug tree-optimization/110556] [12/13/14 Regression] division of INT_MIN and -1 happening incorrectly with -fno-delete-dead-exceptions -fnon-call-exceptions
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (9 preceding siblings ...)
2023-07-06 8:15 ` [Bug tree-optimization/110556] [12/13/14 Regression] division of INT_MIN and -1 happening incorrectly with -fno-delete-dead-exceptions -fnon-call-exceptions rguenth at gcc dot gnu.org
@ 2023-07-06 11:45 ` rguenth at gcc dot gnu.org
2023-07-06 13:24 ` cvs-commit at gcc dot gnu.org
` (3 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: rguenth at gcc dot gnu.org @ 2023-07-06 11:45 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |ASSIGNED
Assignee|unassigned at gcc dot gnu.org |rguenth at gcc dot gnu.org
--- Comment #11 from Richard Biener <rguenth at gcc dot gnu.org> ---
The issue is tail merging and I think we have a duplicate bug for this.
find_duplicates: <bb 5> duplicate of <bb 7>
Removing basic block 7
;; basic block 7, loop depth 0
;; pred:
_27 = -9223372036854775808 / si2_15(D);
goto <bb 10>; [100.00%]
and we have
<bb 4> [local count: 268435457]:
_5 = si2_15(D) > 0;
_6 = _1 | _5;
if (_6 != 0)
goto <bb 6>; [20.00%]
else
goto <bb 5>; [80.00%]
<bb 5> [local count: 214748365]:
_13 = -9223372036854775808 / si1_14(D);
goto <bb 10>; [100.00%]
<bb 6> [local count: 301989889]:
_7 = si1_14(D) > 0;
_9 = _2 | _7;
if (_9 != 0)
goto <bb 8>; [0.00%]
else
goto <bb 7>; [100.00%]
<bb 7> [local count: 301989888]:
_27 = -9223372036854775808 / si2_15(D);
goto <bb 10>; [100.00%]
oddly enough gimple_equal_p looks strange:
case GIMPLE_ASSIGN:
lhs1 = gimple_get_lhs (s1);
lhs2 = gimple_get_lhs (s2);
if (TREE_CODE (lhs1) != SSA_NAME
&& TREE_CODE (lhs2) != SSA_NAME)
return (operand_equal_p (lhs1, lhs2, 0)
&& gimple_operand_equal_value_p (gimple_assign_rhs1 (s1),
gimple_assign_rhs1 (s2)));
else if (TREE_CODE (lhs1) == SSA_NAME
&& TREE_CODE (lhs2) == SSA_NAME)
return operand_equal_p (gimple_assign_rhs1 (s1),
gimple_assign_rhs1 (s2), 0);
that looks pre-tuplish to me.
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug tree-optimization/110556] [12/13/14 Regression] division of INT_MIN and -1 happening incorrectly with -fno-delete-dead-exceptions -fnon-call-exceptions
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (10 preceding siblings ...)
2023-07-06 11:45 ` rguenth at gcc dot gnu.org
@ 2023-07-06 13:24 ` cvs-commit at gcc dot gnu.org
2023-07-07 12:07 ` [Bug tree-optimization/110556] [12 " cvs-commit at gcc dot gnu.org
` (2 subsequent siblings)
14 siblings, 0 replies; 16+ messages in thread
From: cvs-commit at gcc dot gnu.org @ 2023-07-06 13:24 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #12 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Richard Biener <rguenth@gcc.gnu.org>:
https://gcc.gnu.org/g:7b16686ef882ab141276f0e36a9d4ce1d755f64a
commit r14-2363-g7b16686ef882ab141276f0e36a9d4ce1d755f64a
Author: Richard Biener <rguenther@suse.de>
Date: Thu Jul 6 13:51:55 2023 +0200
tree-optimization/110556 - tail merging still pre-tuples
The stmt comparison function for GIMPLE_ASSIGNs for tail merging
still looks like it deals with pre-tuples IL. The following
attempts to fix this, not only comparing the first operand (sic!)
of stmts but all of them plus also compare the operation code.
PR tree-optimization/110556
* tree-ssa-tail-merge.cc (gimple_equal_p): Check
assign code and all operands of non-stores.
* gcc.dg/torture/pr110556.c: New testcase.
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug tree-optimization/110556] [12 Regression] division of INT_MIN and -1 happening incorrectly with -fno-delete-dead-exceptions -fnon-call-exceptions
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (11 preceding siblings ...)
2023-07-06 13:24 ` cvs-commit at gcc dot gnu.org
@ 2023-07-07 12:07 ` cvs-commit at gcc dot gnu.org
2023-11-27 11:35 ` cvs-commit at gcc dot gnu.org
2023-11-27 11:36 ` rguenth at gcc dot gnu.org
14 siblings, 0 replies; 16+ messages in thread
From: cvs-commit at gcc dot gnu.org @ 2023-07-07 12:07 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #13 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-13 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:
https://gcc.gnu.org/g:45e950b542ca598eb8a4a69305fdd96dcb603353
commit r13-7546-g45e950b542ca598eb8a4a69305fdd96dcb603353
Author: Richard Biener <rguenther@suse.de>
Date: Thu Jul 6 13:51:55 2023 +0200
tree-optimization/110556 - tail merging still pre-tuples
The stmt comparison function for GIMPLE_ASSIGNs for tail merging
still looks like it deals with pre-tuples IL. The following
attempts to fix this, not only comparing the first operand (sic!)
of stmts but all of them plus also compare the operation code.
PR tree-optimization/110556
* tree-ssa-tail-merge.cc (gimple_equal_p): Check
assign code and all operands of non-stores.
* gcc.dg/torture/pr110556.c: New testcase.
(cherry picked from commit 7b16686ef882ab141276f0e36a9d4ce1d755f64a)
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug tree-optimization/110556] [12 Regression] division of INT_MIN and -1 happening incorrectly with -fno-delete-dead-exceptions -fnon-call-exceptions
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (12 preceding siblings ...)
2023-07-07 12:07 ` [Bug tree-optimization/110556] [12 " cvs-commit at gcc dot gnu.org
@ 2023-11-27 11:35 ` cvs-commit at gcc dot gnu.org
2023-11-27 11:36 ` rguenth at gcc dot gnu.org
14 siblings, 0 replies; 16+ messages in thread
From: cvs-commit at gcc dot gnu.org @ 2023-11-27 11:35 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
--- Comment #14 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:
https://gcc.gnu.org/g:34bb7282da434e43d29f35733121d8c08a7a1ae1
commit r12-10008-g34bb7282da434e43d29f35733121d8c08a7a1ae1
Author: Richard Biener <rguenther@suse.de>
Date: Thu Jul 6 13:51:55 2023 +0200
tree-optimization/110556 - tail merging still pre-tuples
The stmt comparison function for GIMPLE_ASSIGNs for tail merging
still looks like it deals with pre-tuples IL. The following
attempts to fix this, not only comparing the first operand (sic!)
of stmts but all of them plus also compare the operation code.
PR tree-optimization/110556
* tree-ssa-tail-merge.cc (gimple_equal_p): Check
assign code and all operands of non-stores.
* gcc.dg/torture/pr110556.c: New testcase.
(cherry picked from commit 7b16686ef882ab141276f0e36a9d4ce1d755f64a)
^ permalink raw reply [flat|nested] 16+ messages in thread* [Bug tree-optimization/110556] [12 Regression] division of INT_MIN and -1 happening incorrectly with -fno-delete-dead-exceptions -fnon-call-exceptions
2023-07-05 9:19 [Bug c/110556] New: Floating point exception with specific optimizations 19373742 at buaa dot edu.cn
` (13 preceding siblings ...)
2023-11-27 11:35 ` cvs-commit at gcc dot gnu.org
@ 2023-11-27 11:36 ` rguenth at gcc dot gnu.org
14 siblings, 0 replies; 16+ messages in thread
From: rguenth at gcc dot gnu.org @ 2023-11-27 11:36 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110556
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Known to fail| |12.3.0
Known to work| |12.3.1
Resolution|--- |FIXED
Status|ASSIGNED |RESOLVED
--- Comment #15 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed.
^ permalink raw reply [flat|nested] 16+ messages in thread