[Bug c/96834] New: ICE： Segmentation fault signal terminated program cc1

[Bug c/96834] New: ICE： Segmentation fault signal terminated program cc1

[duanbo3 at huawei dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

            Bug ID: 96834
           Summary: ICE： Segmentation fault signal terminated program cc1
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: duanbo3 at huawei dot com
  Target Milestone: ---

Hi，gcc-trunk ICEs when compiling the following testcase with -O3 on aarch64 or
x86:

Testcase:
extern int a[][1000000];
int b;
void c() {
  for (int d = 2; d <= 9; d++)
    for (int e = 32; e <= 41; e++)
      b += a[d][5];
}

produces the error:
aarch64：
aarch64-linux-gnu-gcc: internal compiler error: Segmentation fault signal
terminated program cc1
Please submit a full bug report,
with preprocessed source if appropriate.

x86：
x86_64-pc-linux-gnu-gcc: internal compiler error: Segmentation fault signal
terminated program cc1
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://gcc.gnu.org/bugs/> for instructions.

any suggestion

[Bug c/96834] ICE： Segmentation fault signal terminated program cc1

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Seems the ICE is due to too deep recursion like
#76 0x00000000013ea04d in interpret_gimple_assign (loop=0x7ffff0dfc000,
stmt=0x7fffe4e807e8) at ../../gcc/tree-scalar-evolution.c:1918
#77 0x00000000013ea1c1 in analyze_scalar_evolution_1 (loop=0x7ffff0dfc000,
var=0x7fffe4e83000) at ../../gcc/tree-scalar-evolution.c:1969
#78 0x00000000013ea3cb in analyze_scalar_evolution (loop=0x7ffff0dfc000,
var=0x7fffe4e83000) at ../../gcc/tree-scalar-evolution.c:2038
#79 0x00000000013e8f1d in interpret_rhs_expr (loop=0x7ffff0dfc000,
at_stmt=0x7fffe4e80840, type=0x7ffff0dd9498, rhs1=0x7fffe4e83000, 
    code=POINTER_PLUS_EXPR, rhs2=0x7ffff0ca0d38) at
../../gcc/tree-scalar-evolution.c:1704
#80 0x00000000013ea04d in interpret_gimple_assign (loop=0x7ffff0dfc000,
stmt=0x7fffe4e80840) at ../../gcc/tree-scalar-evolution.c:1918
#81 0x00000000013ea1c1 in analyze_scalar_evolution_1 (loop=0x7ffff0dfc000,
var=0x7fffe4e83090) at ../../gcc/tree-scalar-evolution.c:1969
#82 0x00000000013ea3cb in analyze_scalar_evolution (loop=0x7ffff0dfc000,
var=0x7fffe4e83090) at ../../gcc/tree-scalar-evolution.c:2038
#83 0x00000000013e8f1d in interpret_rhs_expr (loop=0x7ffff0dfc000,
at_stmt=0x7fffe4e80898, type=0x7ffff0dd9498, rhs1=0x7fffe4e83090, 
    code=POINTER_PLUS_EXPR, rhs2=0x7ffff0ca0d38) at
../../gcc/tree-scalar-evolution.c:1704
#84 0x00000000013ea04d in interpret_gimple_assign (loop=0x7ffff0dfc000,
stmt=0x7fffe4e80898) at ../../gcc/tree-scalar-evolution.c:1918
#85 0x00000000013ea1c1 in analyze_scalar_evolution_1 (loop=0x7ffff0dfc000,
var=0x7fffe4e83120) at ../../gcc/tree-scalar-evolution.c:1969
#86 0x00000000013ea3cb in analyze_scalar_evolution (loop=0x7ffff0dfc000,
var=0x7fffe4e83120) at ../../gcc/tree-scalar-evolution.c:2038
#87 0x00000000013e8f1d in interpret_rhs_expr (loop=0x7ffff0dfc000,
at_stmt=0x7fffe4e808f0, type=0x7ffff0dd9498, rhs1=0x7fffe4e83120, 
    code=POINTER_PLUS_EXPR, rhs2=0x7ffff0ca0d38) at
../../gcc/tree-scalar-evolution.c:1704
#88 0x00000000013ea04d in interpret_gimple_assign (loop=0x7ffff0dfc000,
stmt=0x7fffe4e808f0) at ../../gcc/tree-scalar-evolution.c:1918
#89 0x00000000013ea1c1 in analyze_scalar_evolution_1 (loop=0x7ffff0dfc000,
var=0x7fffe4e831b0) at ../../gcc/tree-scalar-evolution.c:1969
#90 0x00000000013ea3cb in analyze_scalar_evolution (loop=0x7ffff0dfc000,
var=0x7fffe4e831b0) at ../../gcc/tree-scalar-evolution.c:2038
#91 0x00000000013e8f1d in interpret_rhs_expr (loop=0x7ffff0dfc000,
at_stmt=0x7fffe4e80948, type=0x7ffff0dd9498, rhs1=0x7fffe4e831b0, 
    code=POINTER_PLUS_EXPR, rhs2=0x7ffff0ca0d38) at
../../gcc/tree-scalar-evolution.c:1704
#92 0x00000000013ea04d in interpret_gimple_assign (loop=0x7ffff0dfc000,
stmt=0x7fffe4e80948) at ../../gcc/tree-scalar-evolution.c:1918
#93 0x00000000013ea1c1 in analyze_scalar_evolution_1 (loop=0x7ffff0dfc000,
var=0x7fffe4e83240) at ../../gcc/tree-scalar-evolution.c:1969
#94 0x00000000013ea3cb in analyze_scalar_evolution (loop=0x7ffff0dfc000,
var=0x7fffe4e83240) at ../../gcc/tree-scalar-evolution.c:2038
#95 0x00000000013e8f1d in interpret_rhs_expr (loop=0x7ffff0dfc000,
at_stmt=0x7fffe4e809a0, type=0x7ffff0dd9498, rhs1=0x7fffe4e83240, 
    code=POINTER_PLUS_EXPR, rhs2=0x7ffff0ca0d38) at
../../gcc/tree-scalar-evolution.c:1704
during scev inside predictive commoning.
This is on    vectp_a.8_1000004 = vectp_a.8_1000002 + 16;
    vectp_a.8_1000006 = vectp_a.8_1000004 + 16;
    vectp_a.8_1000008 = vectp_a.8_1000006 + 16;
    vectp_a.8_1000010 = vectp_a.8_1000008 + 16;
    vectp_a.8_1000012 = vectp_a.8_1000010 + 16;
    vectp_a.8_1000014 = vectp_a.8_1000012 + 16;
    vectp_a.8_1000016 = vectp_a.8_1000014 + 16;
    vectp_a.8_1000018 = vectp_a.8_1000016 + 16;
    vectp_a.8_1000020 = vectp_a.8_1000018 + 16;
    vectp_a.8_1000022 = vectp_a.8_1000020 + 16;
    vectp_a.8_1000024 = vectp_a.8_1000022 + 16;
    vectp_a.8_1000026 = vectp_a.8_1000024 + 16;
    vectp_a.8_1000028 = vectp_a.8_1000026 + 16;
    vectp_a.8_1000030 = vectp_a.8_1000028 + 16;
    vectp_a.8_1000032 = vectp_a.8_1000030 + 16;
    vectp_a.8_1000034 = vectp_a.8_1000032 + 16;
    vectp_a.8_1000036 = vectp_a.8_1000034 + 16;
    vectp_a.8_1000038 = vectp_a.8_1000036 + 16;
    vectp_a.8_1000040 = vectp_a.8_1000038 + 16;
like IL, where nothing has really simplified those before predcom.

[Bug c/96834] [9/10/11 Regression] Segmentation fault signal terminated program cc1

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
            Summary|ICE： Segmentation fault     |[9/10/11 Regression]
                   |signal terminated program   |Segmentation fault signal
                   |cc1                         |terminated program cc1
   Target Milestone|---                         |9.4
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2020-08-28

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r9-360-g68435eb293a6b9c161936bccab66e4b1f702372b

[Bug c/96834] [9/10/11 Regression] Segmentation fault signal terminated program cc1

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Blocks|                            |53947
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org
             Status|NEW                         |ASSIGNED
           Keywords|                            |compile-time-hog

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
There's some other bug about this kind of excessive vectorization where
load permutation ends up loading a lot of dead stmts.  It's a bit awkward
to address though.

t.c:4:3: note:   Detected single element interleaving a[d_69][5] step 4000000
...

t.c:4:3: note:   node 0x3e2e450 (max_nunits=2, refcnt=2)
t.c:4:3: note:          stmt 0 _14 = a[d_69][5];
t.c:4:3: note:          stmt 1 _14 = a[d_69][5];
t.c:4:3: note:          stmt 2 _14 = a[d_69][5];
t.c:4:3: note:          stmt 3 _14 = a[d_69][5];
t.c:4:3: note:          stmt 4 _14 = a[d_69][5];
t.c:4:3: note:          stmt 5 _14 = a[d_69][5];
t.c:4:3: note:          stmt 6 _14 = a[d_69][5];
t.c:4:3: note:          stmt 7 _14 = a[d_69][5];
t.c:4:3: note:          stmt 8 _14 = a[d_69][5];
t.c:4:3: note:          stmt 9 _14 = a[d_69][5];
t.c:4:3: note:          load permutation { 0 0 0 0 0 0 0 0 0 0 }


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53947
[Bug 53947] [meta-bug] vectorizer missed-optimizations

[Bug c/96834] [9/10/11 Regression] Segmentation fault signal terminated program cc1

[z.zhanghaijian at huawei dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

z.zhanghaijian at huawei dot com <z.zhanghaijian at huawei dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |z.zhanghaijian at huawei dot com

--- Comment #4 from z.zhanghaijian at huawei dot com <z.zhanghaijian at huawei dot com> ---
The case like:
test.c:
int f300001() { }
 void f300000() {       printf("300000\n");     f300001(); }
 void f299999() {       printf("299999\n");     f300000(); }
 void f299998() {       printf("299998\n");     f299999(); }
 void f299997() {       printf("299997\n");     f299998(); }
...
 void f10() {   printf("10\n");         f11(); }
 void f9() {    printf("9\n");  f10(); }
 void f8() {    printf("8\n");  f9(); }
 void f7() {    printf("7\n");  f8(); }
 void f6() {    printf("6\n");  f7(); }
 void f5() {    printf("5\n");  f6(); }
 void f4() {    printf("4\n");  f5(); }
 void f3() {    printf("3\n");  f4(); }
 void f2() {    printf("2\n");  f3(); }
 void f1() {    printf("1\n");  f2(); }
 int main(int argc, char** argv){       f1(); }

This can also produces the error on aarch64:

gcc test.c -S
gcc: internal compiler error: Segmentation fault signal terminated program cc1
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://gcc.gnu.org/bugs/> for instructions.

[Bug c/96834] [9/10/11 Regression] Segmentation fault signal terminated program cc1

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to z.zhanghaijian@huawei.com from comment #4)
> The case like:
> test.c:
> int f300001() { }
>  void f300000() {       printf("300000\n");     f300001(); }
>  void f299999() {       printf("299999\n");     f300000(); }
>  void f299998() {       printf("299998\n");     f299999(); }
>  void f299997() {       printf("299997\n");     f299998(); }
> ...
>  void f10() {   printf("10\n");         f11(); }
>  void f9() {    printf("9\n");  f10(); }
>  void f8() {    printf("8\n");  f9(); }
>  void f7() {    printf("7\n");  f8(); }
>  void f6() {    printf("6\n");  f7(); }
>  void f5() {    printf("5\n");  f6(); }
>  void f4() {    printf("4\n");  f5(); }
>  void f3() {    printf("3\n");  f4(); }
>  void f2() {    printf("2\n");  f3(); }
>  void f1() {    printf("1\n");  f2(); }
>  int main(int argc, char** argv){       f1(); }
> 
> This can also produces the error on aarch64:
> 
> gcc test.c -S
> gcc: internal compiler error: Segmentation fault signal terminated program
> cc1
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <https://gcc.gnu.org/bugs/> for instructions.

But that's sth entirely different and not vectorization triggered.

[Bug c/96834] [9/10/11 Regression] Segmentation fault signal terminated program cc1

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=91403

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
PR91403 was that other similar bug.

[Bug c/96834] [9/10/11 Regression] Segmentation fault signal terminated program cc1

[z.zhanghaijian at huawei dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

--- Comment #7 from z.zhanghaijian at huawei dot com <z.zhanghaijian at huawei dot com> ---
(In reply to Richard Biener from comment #5)
> (In reply to z.zhanghaijian@huawei.com from comment #4)
> > The case like:
> > test.c:
> > int f300001() { }
> >  void f300000() {       printf("300000\n");     f300001(); }
> >  void f299999() {       printf("299999\n");     f300000(); }
> >  void f299998() {       printf("299998\n");     f299999(); }
> >  void f299997() {       printf("299997\n");     f299998(); }
> > ...
> >  void f10() {   printf("10\n");         f11(); }
> >  void f9() {    printf("9\n");  f10(); }
> >  void f8() {    printf("8\n");  f9(); }
> >  void f7() {    printf("7\n");  f8(); }
> >  void f6() {    printf("6\n");  f7(); }
> >  void f5() {    printf("5\n");  f6(); }
> >  void f4() {    printf("4\n");  f5(); }
> >  void f3() {    printf("3\n");  f4(); }
> >  void f2() {    printf("2\n");  f3(); }
> >  void f1() {    printf("1\n");  f2(); }
> >  int main(int argc, char** argv){       f1(); }
> > 
> > This can also produces the error on aarch64:
> > 
> > gcc test.c -S
> > gcc: internal compiler error: Segmentation fault signal terminated program
> > cc1
> > Please submit a full bug report,
> > with preprocessed source if appropriate.
> > See <https://gcc.gnu.org/bugs/> for instructions.
> 
> But that's sth entirely different and not vectorization triggered.

Yes, I'll re-submit a new PR.

[Bug tree-optimization/96834] [9/10/11 Regression] Segmentation fault signal terminated program cc1

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2
          Component|c                           |tree-optimization

[Bug tree-optimization/96834] [9/10/11 Regression] Segmentation fault signal terminated program cc1

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96834

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |DUPLICATE

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
duplicate

*** This bug has been marked as a duplicate of bug 91403 ***