[Bug ipa/103083] New: Wrong code due to ipa-cp's value range propagation

[Bug ipa/103083] New: Wrong code due to ipa-cp's value range propagation

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

            Bug ID: 103083
           Summary: Wrong code due to ipa-cp's value range propagation
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: ipa
          Assignee: unassigned at gcc dot gnu.org
          Reporter: hubicka at gcc dot gnu.org
                CC: marxin at gcc dot gnu.org
  Target Milestone: ---

struct b {int b;};
struct a {int a; struct b b;};

long i;

__attribute__ ((noinline))
static void test2 (struct b *b)
{
        if (((int)b)&4)
                __builtin_abort ();
}

__attribute__ ((noinline))
static void
test (struct a *a)
{
         test2(a? &a->b : 0);
}
void
main()
{
        test(0);
}

aborts when built with -O2.

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's value range propagation since r10-5538-gc7ac9a0c7e3916f1

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Wrong code due to ipa-cp's  |[10/11/12 Regression] Wrong
                   |value range propagation     |code due to ipa-cp's value
                   |                            |range propagation since
                   |                            |r10-5538-gc7ac9a0c7e3916f1
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2021-11-04

--- Comment #1 from Martin Liška <marxin at gcc dot gnu.org> ---
If I see correctly, started with r10-5538-gc7ac9a0c7e3916f1.

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's value range propagation since r10-5538-gc7ac9a0c7e3916f1

[jamborm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

Martin Jambor <jamborm at gcc dot gnu.org> changed:

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

--- Comment #2 from Martin Jambor <jamborm at gcc dot gnu.org> ---
Mine.

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's value range propagation since r10-5538-gc7ac9a0c7e3916f1

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |10.4
           Keywords|                            |wrong-code
           Priority|P3                          |P2

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's value range propagation since r10-5538-gc7ac9a0c7e3916f1

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

--- Comment #3 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Just for the record, the problem is the ancestor jump function being used to
model two things - pointer plus which originates from ADDR_EXPR of component
reference and also C++ casts which checks whether parameter is NULL and does
pointer plus inly for non-NULL.

In this case the VR propgataion expects the first interpretation (without
special casing NULL) while the testcase does the second.

Honza

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's bits value propagation since r10-5538-gc7ac9a0c7e3916f1

[jamborm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

Martin Jambor <jamborm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[10/11/12 Regression] Wrong |[10/11/12 Regression] Wrong
                   |code due to ipa-cp's value  |code due to ipa-cp's bits
                   |range propagation since     |value propagation since
                   |r10-5538-gc7ac9a0c7e3916f1  |r10-5538-gc7ac9a0c7e3916f1

--- Comment #4 from Martin Jambor <jamborm at gcc dot gnu.org> ---
The last proposal to fix this is in
https://gcc.gnu.org/pipermail/gcc-patches/2021-November/585668.html

I will have another look, re-test and ping it soon.

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's bits value propagation since r10-5538-gc7ac9a0c7e3916f1

[jamborm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

--- Comment #5 from Martin Jambor <jamborm at gcc dot gnu.org> ---
I have changed the patch a bit and re-submitted for review:
https://gcc.gnu.org/pipermail/gcc-patches/2022-February/590341.html

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's bits value propagation since r10-5538-gc7ac9a0c7e3916f1

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

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

https://gcc.gnu.org/g:7ea3a73c195a79e6740ae594ee1a14c8bf7a938d

commit r12-7939-g7ea3a73c195a79e6740ae594ee1a14c8bf7a938d
Author: Martin Jambor <mjambor@suse.cz>
Date:   Thu Mar 31 17:22:34 2022 +0200

    ipa: Careful processing ANCESTOR jump functions and NULL pointers (PR
103083)

    IPA_JF_ANCESTOR jump functions are constructed also when the formal
    parameter of the caller is first checked whether it is NULL and left
    as it is if it is NULL, to accommodate C++ casts to an ancestor class.

    The jump function type was invented for devirtualization and IPA-CP
    propagation of tree constants is also careful to apply it only to
    existing DECLs(*) but as PR 103083 shows, the part propagating "known
    bits" was not careful about this, which can lead to miscompilations.

    This patch introduces a flag to the ancestor jump functions which
    tells whether a NULL-check was elided when creating it and makes the
    bits propagation behave accordingly, masking any bits otherwise would
    be known to be one.  This should safely preserve alignment info, which
    is the primary ifnormation that we keep in bits for pointers.

    (*) There still may remain problems when a DECL resides on address
    zero (with -fno-delete-null-pointer-checks ...I hope it cannot happen
    otherwise).  I am looking into that now but I think it will be easier
    for everyone if I do so in a follow-up patch.

    gcc/ChangeLog:

    2022-02-11  Martin Jambor  <mjambor@suse.cz>

            PR ipa/103083
            * ipa-prop.h (ipa_ancestor_jf_data): New flag keep_null;
            (ipa_get_jf_ancestor_keep_null): New function.
            * ipa-prop.cc (ipa_set_ancestor_jf): Initialize keep_null field of
the
            ancestor function.
            (compute_complex_assign_jump_func): Pass false to keep_null
            parameter of ipa_set_ancestor_jf.
            (compute_complex_ancestor_jump_func): Pass true to keep_null
            parameter of ipa_set_ancestor_jf.
            (update_jump_functions_after_inlining): Carry over keep_null from
the
            original ancestor jump-function or merge them.
            (ipa_write_jump_function): Stream keep_null flag.
            (ipa_read_jump_function): Likewise.
            (ipa_print_node_jump_functions_for_edge): Print the new flag.
            * ipa-cp.cc (class ipcp_bits_lattice): Make various getters const. 
New
            member function known_nonzero_p.
            (ipcp_bits_lattice::known_nonzero_p): New.
            (ipcp_bits_lattice::meet_with_1): New parameter drop_all_ones,
            observe it.
            (ipcp_bits_lattice::meet_with): Likewise.
            (propagate_bits_across_jump_function): Simplify.  Pass true in
            drop_all_ones when it is necessary.
            (propagate_aggs_across_jump_function): Take care of keep_null
            flag.
            (ipa_get_jf_ancestor_result): Propagate NULL accross keep_null
            jump functions.

    gcc/testsuite/ChangeLog:

    2021-11-25  Martin Jambor  <mjambor@suse.cz>

            * gcc.dg/ipa/pr103083-1.c: New test.
            * gcc.dg/ipa/pr103083-2.c: Likewise.

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's bits value propagation since r10-5538-gc7ac9a0c7e3916f1

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Martin Jambor
<jamborm@gcc.gnu.org>:

https://gcc.gnu.org/g:4f939ac1e295f38624c82fa75fa798e83d825232

commit r11-9775-g4f939ac1e295f38624c82fa75fa798e83d825232
Author: Martin Jambor <mjambor@suse.cz>
Date:   Mon Apr 4 19:31:42 2022 +0200

    ipa: Careful processing ANCESTOR jump functions and NULL pointers (PR
103083)

    IPA_JF_ANCESTOR jump functions are constructed also when the formal
    parameter of the caller is first checked whether it is NULL and left
    as it is if it is NULL, to accommodate C++ casts to an ancestor class.

    The jump function type was invented for devirtualization and IPA-CP
    propagation of tree constants is also careful to apply it only to
    existing DECLs(*) but as PR 103083 shows, the part propagating "known
    bits" was not careful about this, which can lead to miscompilations.

    This patch introduces a flag to the ancestor jump functions which
    tells whether a NULL-check was elided when creating it and makes the
    bits propagation behave accordingly, masking any bits otherwise would
    be known to be one.  This should safely preserve alignment info, which
    is the primary ifnormation that we keep in bits for pointers.

    (*) There still may remain problems when a DECL resides on address
    zero (with -fno-delete-null-pointer-checks ...I hope it cannot happen
    otherwise).  I am looking into that now but I think it will be easier
    for everyone if I do so in a follow-up patch.

    gcc/ChangeLog:

    2022-02-11  Martin Jambor  <mjambor@suse.cz>

            PR ipa/103083
            * ipa-prop.h (ipa_ancestor_jf_data): New flag keep_null;
            (ipa_get_jf_ancestor_keep_null): New function.
            * ipa-prop.c (ipa_set_ancestor_jf): Initialize keep_null field of
the
            ancestor function.
            (compute_complex_assign_jump_func): Pass false to keep_null
            parameter of ipa_set_ancestor_jf.
            (compute_complex_ancestor_jump_func): Pass true to keep_null
            parameter of ipa_set_ancestor_jf.
            (update_jump_functions_after_inlining): Carry over keep_null from
the
            original ancestor jump-function or merge them.
            (ipa_write_jump_function): Stream keep_null flag.
            (ipa_read_jump_function): Likewise.
            (ipa_print_node_jump_functions_for_edge): Print the new flag.
            * ipa-cp.c (class ipcp_bits_lattice): Make various getters const. 
New
            member function known_nonzero_p.
            (ipcp_bits_lattice::known_nonzero_p): New.
            (ipcp_bits_lattice::meet_with_1): New parameter drop_all_ones,
            observe it.
            (ipcp_bits_lattice::meet_with): Likewise.
            (propagate_bits_across_jump_function): Simplify.  Pass true in
            drop_all_ones when it is necessary.
            (propagate_aggs_across_jump_function): Take care of keep_null
            flag.
            (ipa_get_jf_ancestor_result): Propagate NULL accross keep_null
            jump functions.

    gcc/testsuite/ChangeLog:

    2021-11-25  Martin Jambor  <mjambor@suse.cz>

            * gcc.dg/ipa/pr103083-1.c: New test.
            * gcc.dg/ipa/pr103083-2.c: Likewise.

    (cherry picked from commit 7ea3a73c195a79e6740ae594ee1a14c8bf7a938d)

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's bits value propagation since r10-5538-gc7ac9a0c7e3916f1

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

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

https://gcc.gnu.org/g:8c3e60933fa56dba47e75c88757395a73d38b04d

commit r10-10523-g8c3e60933fa56dba47e75c88757395a73d38b04d
Author: Martin Jambor <mjambor@suse.cz>
Date:   Tue Apr 5 11:19:49 2022 +0200

    ipa: Careful processing ANCESTOR jump functions and NULL pointers (PR
103083)

    IPA_JF_ANCESTOR jump functions are constructed also when the formal
    parameter of the caller is first checked whether it is NULL and left
    as it is if it is NULL, to accommodate C++ casts to an ancestor class.

    The jump function type was invented for devirtualization and IPA-CP
    propagation of tree constants is also careful to apply it only to
    existing DECLs(*) but as PR 103083 shows, the part propagating "known
    bits" was not careful about this, which can lead to miscompilations.

    This patch introduces a flag to the ancestor jump functions which
    tells whether a NULL-check was elided when creating it and makes the
    bits propagation behave accordingly, masking any bits otherwise would
    be known to be one.  This should safely preserve alignment info, which
    is the primary ifnormation that we keep in bits for pointers.

    (*) There still may remain problems when a DECL resides on address
    zero (with -fno-delete-null-pointer-checks ...I hope it cannot happen
    otherwise).  I am looking into that now but I think it will be easier
    for everyone if I do so in a follow-up patch.

    gcc/ChangeLog:

    2022-02-11  Martin Jambor  <mjambor@suse.cz>

            PR ipa/103083
            * ipa-prop.h (ipa_ancestor_jf_data): New flag keep_null;
            (ipa_get_jf_ancestor_keep_null): New function.
            * ipa-prop.c (ipa_set_ancestor_jf): Initialize keep_null field of
the
            ancestor function.
            (compute_complex_assign_jump_func): Pass false to keep_null
            parameter of ipa_set_ancestor_jf.
            (compute_complex_ancestor_jump_func): Pass true to keep_null
            parameter of ipa_set_ancestor_jf.
            (update_jump_functions_after_inlining): Carry over keep_null from
the
            original ancestor jump-function or merge them.
            (ipa_write_jump_function): Stream keep_null flag.
            (ipa_read_jump_function): Likewise.
            (ipa_print_node_jump_functions_for_edge): Print the new flag.
            * ipa-cp.c (class ipcp_bits_lattice): Make various getters const. 
New
            member function known_nonzero_p.
            (ipcp_bits_lattice::known_nonzero_p): New.
            (ipcp_bits_lattice::meet_with_1): New parameter drop_all_ones,
            observe it.
            (ipcp_bits_lattice::meet_with): Likewise.
            (propagate_bits_across_jump_function): Simplify.  Pass true in
            drop_all_ones when it is necessary.
            (propagate_aggs_across_jump_function): Take care of keep_null
            flag.
            (ipa_get_jf_ancestor_result): Propagate NULL accross keep_null
            jump functions.

    gcc/testsuite/ChangeLog:

    2021-11-25  Martin Jambor  <mjambor@suse.cz>

            * gcc.dg/ipa/pr103083-1.c: New test.
            * gcc.dg/ipa/pr103083-2.c: Likewise.

    (cherry picked from commit 7ea3a73c195a79e6740ae594ee1a14c8bf7a938d)

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's bits value propagation since r10-5538-gc7ac9a0c7e3916f1

[mikpelinux at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

Mikael Pettersson <mikpelinux at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mikpelinux at gmail dot com

--- Comment #9 from Mikael Pettersson <mikpelinux at gmail dot com> ---
Can this be closed now? I cannot reproduce the abort at -O2 for the original
test case with current gcc-10/11 branch snapshots (but I can with snapshots
from March).

[Bug ipa/103083] [10/11/12 Regression] Wrong code due to ipa-cp's bits value propagation since r10-5538-gc7ac9a0c7e3916f1

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103083

Jan Hubicka <hubicka at gcc dot gnu.org> changed:

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

--- Comment #10 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Fixed.