[Bug c++/99034] New: [10/11 Regression] ICE in emit_to_new_bb_before, at except.c:932

[Bug c++/99034] New: [10/11 Regression] ICE in emit_to_new_bb_before, at except.c:932

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

            Bug ID: 99034
           Summary: [10/11 Regression] ICE in emit_to_new_bb_before, at
                    except.c:932
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

This changed between 20190929 and 20191006 at -O2+ :


$ g++-10-20190929 -c pr90208-2.c -finstrument-functions -O2
$
$ g++-11-20210207 -c pr90208-2.c -finstrument-functions -O2
during RTL pass: expand
pr90208-2.c: In function 'void baz()':
pr90208-2.c:20:1: internal compiler error: Segmentation fault
   20 | }
      | ^
0xcd922f crash_signal
        ../../gcc/toplev.c:327
0x9bd1e6 emit_to_new_bb_before
        ../../gcc/except.c:932
0x9c185a dw2_build_landing_pads
        ../../gcc/except.c:1013
0x9c185a finish_eh_generation()
        ../../gcc/except.c:1521
0x8e2e7d execute
        ../../gcc/cfgexpand.c:6837

---

pr90208-2.c: In function 'void baz()':
pr90208-2.c:16:1: error: EH landing pad label
   16 | baz (void)
      | ^~~
<L3> is not first in a sequence of labels in bb 5during IPA pass: inline
pr90208-2.c:16:1: internal compiler error: verify_flow_info failed
0xae3ac4 verify_flow_info()
        ../../gcc/cfghooks.c:269
0x1046ba4 checking_verify_flow_info
        ../../gcc/cfghooks.h:212
0x1046ba4 cleanup_tree_cfg_noloop
        ../../gcc/tree-cfgcleanup.c:1136
0x1046ba4 cleanup_tree_cfg(unsigned int)
        ../../gcc/tree-cfgcleanup.c:1187
0xef8a44 execute_function_todo
        ../../gcc/passes.c:2008
0xef97f2 execute_todo
        ../../gcc/passes.c:2096
0xefba43 execute_one_ipa_transform_pass
        ../../gcc/passes.c:2293
0xefba43 execute_all_ipa_transforms(bool)
        ../../gcc/passes.c:2337
0xb22995 cgraph_node::expand()
        ../../gcc/cgraphunit.c:1820
0xb24316 expand_all_functions
        ../../gcc/cgraphunit.c:1995
0xb24316 symbol_table::compile()
        ../../gcc/cgraphunit.c:2359
0xb2784f symbol_table::compile()
        ../../gcc/cgraphunit.c:2543
0xb2784f symbol_table::finalize_compilation_unit()
        ../../gcc/cgraphunit.c:2540

[Bug ipa/99034] [10/11 Regression] ICE in emit_to_new_bb_before, at except.c:932

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
          Component|c++                         |ipa
                 CC|                            |hubicka at gcc dot gnu.org,
                   |                            |marxin at gcc dot gnu.org,
                   |                            |mpolacek at gcc dot gnu.org
             Status|UNCONFIRMED                 |NEW
   Target Milestone|---                         |10.3
   Last reconfirmed|                            |2021-02-09

--- Comment #1 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Started with r276469.

[Bug ipa/99034] [10/11 Regression] ICE in emit_to_new_bb_before, at except.c:932

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

[Bug ipa/99034] [10/11 Regression] ICE in emit_to_new_bb_before, at except.c:932

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |marxin at gcc dot gnu.org
             Status|NEW                         |ASSIGNED

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
I'll take a look.

[Bug ipa/99034] [9/10/11 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org
            Summary|[10/11 Regression] ICE in   |[9/10/11 Regression] error:
                   |emit_to_new_bb_before, at   |EH landing pad label <L3>
                   |except.c:932                |is not first in a sequence
                   |                            |of labels in bb 6during
                   |                            |GIMPLE pass: einline since
                   |                            |r9-6254-gf86624d85f937e03
             Status|ASSIGNED                    |NEW
           Assignee|marxin at gcc dot gnu.org          |unassigned at gcc dot gnu.org

--- Comment #3 from Martin Liška <marxin at gcc dot gnu.org> ---
A slightly modified test-case started with r9-6254-gf86624d85f937e03.

$ cat x.c
void *b[5];
void foo (void);

static void
__attribute__((always_inline))
bar (int d)
{
  while (d)
    foo ();
}

void
baz (void)
{
  bar (2);
  __builtin_setjmp (b);
}

@Jakub: Can you please take a look?

[Bug ipa/99034] [9/10/11 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
No need for -finstrument-functions:

void *b[5];
void foo (void);
struct S { ~S (); };

static inline void
__attribute__((always_inline))
bar (int d)
{
  S s;
  while (d)
    foo ();
}

void
baz (void)
{
  bar (2);
  __builtin_setjmp (b);
}

ICEs just with -O2.

[Bug ipa/99034] [9/10/11 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 50218
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=50218&action=edit
gcc11-pr99034.patch

Untested fix.

[Bug ipa/99034] [9/10/11 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:37bde2f87267908a93c07856317a28827f8284f7

commit r11-7294-g37bde2f87267908a93c07856317a28827f8284f7
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Feb 19 12:14:39 2021 +0100

    tree-cfg: Fix up gimple_merge_blocks FORCED_LABEL handling [PR99034]

    The verifiers require that DECL_NONLOCAL or EH_LANDING_PAD_NR
    labels are always the first label if there is more than one label.

    When merging blocks, we don't honor that though.
    On the following testcase, we try to merge blocks:
    <bb 13> [count: 0]:
    <L2>:
    S::~S (&s);

    and
    <bb 15> [count: 0]:
    <L0>:
    resx 1

    where <L2> is landing pad and <L0> is FORCED_LABEL.  And the code puts
    the FORCED_LABEL before the landing pad label, violating the verification
    requirements.

    The following patch fixes it by moving the FORCED_LABEL after the
    DECL_NONLOCAL or EH_LANDING_PAD_NR label if it is the first label.

    2021-02-19  Jakub Jelinek  <jakub@redhat.com>

            PR ipa/99034
            * tree-cfg.c (gimple_merge_blocks): If bb a starts with eh landing
            pad or non-local label, put FORCED_LABELs from bb b after that
label
            rather than before it.

            * g++.dg/opt/pr99034.C: New test.

[Bug ipa/99034] [9/10 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10/11 Regression] error: |[9/10 Regression] error: EH
                   |EH landing pad label <L3>   |landing pad label <L3> is
                   |is not first in a sequence  |not first in a sequence of
                   |of labels in bb 6during     |labels in bb 6during GIMPLE
                   |GIMPLE pass: einline since  |pass: einline since
                   |r9-6254-gf86624d85f937e03   |r9-6254-gf86624d85f937e03

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed on the trunk so far.

[Bug ipa/99034] [9/10 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:33be24d77d3d8f0c992eb344ce63f78e14cf753d

commit r10-9473-g33be24d77d3d8f0c992eb344ce63f78e14cf753d
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Feb 19 12:14:39 2021 +0100

    tree-cfg: Fix up gimple_merge_blocks FORCED_LABEL handling [PR99034]

    The verifiers require that DECL_NONLOCAL or EH_LANDING_PAD_NR
    labels are always the first label if there is more than one label.

    When merging blocks, we don't honor that though.
    On the following testcase, we try to merge blocks:
    <bb 13> [count: 0]:
    <L2>:
    S::~S (&s);

    and
    <bb 15> [count: 0]:
    <L0>:
    resx 1

    where <L2> is landing pad and <L0> is FORCED_LABEL.  And the code puts
    the FORCED_LABEL before the landing pad label, violating the verification
    requirements.

    The following patch fixes it by moving the FORCED_LABEL after the
    DECL_NONLOCAL or EH_LANDING_PAD_NR label if it is the first label.

    2021-02-19  Jakub Jelinek  <jakub@redhat.com>

            PR ipa/99034
            * tree-cfg.c (gimple_merge_blocks): If bb a starts with eh landing
            pad or non-local label, put FORCED_LABELs from bb b after that
label
            rather than before it.

            * g++.dg/opt/pr99034.C: New test.

    (cherry picked from commit 37bde2f87267908a93c07856317a28827f8284f7)

[Bug ipa/99034] [9 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10 Regression] error: EH |[9 Regression] error: EH
                   |landing pad label <L3> is   |landing pad label <L3> is
                   |not first in a sequence of  |not first in a sequence of
                   |labels in bb 6during GIMPLE |labels in bb 6during GIMPLE
                   |pass: einline since         |pass: einline since
                   |r9-6254-gf86624d85f937e03   |r9-6254-gf86624d85f937e03

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 10.3 too.

[Bug ipa/99034] [9 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.3                        |10.4

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 10.3 is being released, retargeting bugs to GCC 10.4.

[Bug ipa/99034] [9 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

--- Comment #11 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:8a6114146001eafc1921b60e40bf5c0e4f4b8e64

commit r9-9423-g8a6114146001eafc1921b60e40bf5c0e4f4b8e64
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Feb 19 12:14:39 2021 +0100

    tree-cfg: Fix up gimple_merge_blocks FORCED_LABEL handling [PR99034]

    The verifiers require that DECL_NONLOCAL or EH_LANDING_PAD_NR
    labels are always the first label if there is more than one label.

    When merging blocks, we don't honor that though.
    On the following testcase, we try to merge blocks:
    <bb 13> [count: 0]:
    <L2>:
    S::~S (&s);

    and
    <bb 15> [count: 0]:
    <L0>:
    resx 1

    where <L2> is landing pad and <L0> is FORCED_LABEL.  And the code puts
    the FORCED_LABEL before the landing pad label, violating the verification
    requirements.

    The following patch fixes it by moving the FORCED_LABEL after the
    DECL_NONLOCAL or EH_LANDING_PAD_NR label if it is the first label.

    2021-02-19  Jakub Jelinek  <jakub@redhat.com>

            PR ipa/99034
            * tree-cfg.c (gimple_merge_blocks): If bb a starts with eh landing
            pad or non-local label, put FORCED_LABELs from bb b after that
label
            rather than before it.

            * g++.dg/opt/pr99034.C: New test.

    (cherry picked from commit 33be24d77d3d8f0c992eb344ce63f78e14cf753d)

[Bug ipa/99034] [9 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

--- Comment #12 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-8 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:a1d1df435b9b9eb2e78f73ac2ab6558302cb1918

commit r8-10888-ga1d1df435b9b9eb2e78f73ac2ab6558302cb1918
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Feb 19 12:14:39 2021 +0100

    tree-cfg: Fix up gimple_merge_blocks FORCED_LABEL handling [PR99034]

    The verifiers require that DECL_NONLOCAL or EH_LANDING_PAD_NR
    labels are always the first label if there is more than one label.

    When merging blocks, we don't honor that though.
    On the following testcase, we try to merge blocks:
    <bb 13> [count: 0]:
    <L2>:
    S::~S (&s);

    and
    <bb 15> [count: 0]:
    <L0>:
    resx 1

    where <L2> is landing pad and <L0> is FORCED_LABEL.  And the code puts
    the FORCED_LABEL before the landing pad label, violating the verification
    requirements.

    The following patch fixes it by moving the FORCED_LABEL after the
    DECL_NONLOCAL or EH_LANDING_PAD_NR label if it is the first label.

    2021-02-19  Jakub Jelinek  <jakub@redhat.com>

            PR ipa/99034
            * tree-cfg.c (gimple_merge_blocks): If bb a starts with eh landing
            pad or non-local label, put FORCED_LABELs from bb b after that
label
            rather than before it.

            * g++.dg/opt/pr99034.C: New test.

    (cherry picked from commit 33be24d77d3d8f0c992eb344ce63f78e14cf753d)

[Bug ipa/99034] [9 Regression] error: EH landing pad label <L3> is not first in a sequence of labels in bb 6during GIMPLE pass: einline since r9-6254-gf86624d85f937e03

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99034

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

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

--- Comment #13 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.