Bug 83975 - [8 Regression] ICE in set_parm_default_def_partition, at tree-ssa-coalesce.c:1919
Summary: [8 Regression] ICE in set_parm_default_def_partition, at tree-ssa-coalesce.c:...
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: fortran (show other bugs)
Version: 8.0
: P4 normal
Target Milestone: 8.0
Assignee: Janne Blomqvist
URL: https://gcc.gnu.org/ml/gcc-patches/20...
Keywords: ice-on-valid-code, patch
Depends on:
Blocks:
 
Reported: 2018-01-22 18:10 UTC by G. Steinmetz
Modified: 2018-02-23 12:02 UTC (History)
1 user (show)

See Also:
Host:
Target:
Build:
Known to work: 7.2.0
Known to fail: 8.0
Last reconfirmed: 2018-01-23 00:00:00


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description G. Steinmetz 2018-01-22 18:10:32 UTC
Changed between 20171217 and 20180107 :


$ cat z1.f90
subroutine s(x)
   character(*) :: x
   associate (y => x)
   end associate
end


$ gfortran-8-20171217 -c z1.f90
$
$ gfortran-8-20180121 -c z1.f90
during RTL pass: expand
z1.f90:1:0:

 subroutine s(x)

internal compiler error: in set_parm_default_def_partition, at tree-ssa-coalesce.c:1919
0xc72efc set_parm_default_def_partition
        ../../gcc/tree-ssa-coalesce.c:1919
0xc72e07 for_all_parms
        ../../gcc/tree-ssa-coalesce.c:1023
0xc74798 get_parm_default_def_partitions(_var_map*)
        ../../gcc/tree-ssa-coalesce.c:1936
0xc2099b remove_ssa_form
        ../../gcc/tree-outof-ssa.c:971
0xc2099b rewrite_out_of_ssa(ssaexpand*)
        ../../gcc/tree-outof-ssa.c:1174
0x805810 execute
        ../../gcc/cfgexpand.c:6218
Comment 1 G. Steinmetz 2018-01-22 18:10:50 UTC
Configured with --enable-checking=yes :

$ gfortran-8-20180121-chk -c z1.f90
z1.f90:1:0:

 subroutine s(x)

internal compiler error: tree check: expected parm_decl, have var_decl in assign_parm_find_data_types, at function.c:2456
0x61e2e1 tree_check_failed(tree_node const*, char const*, int, char const*, ...)
        ../../gcc/tree.c:9337
0x9e9224 tree_check(tree_node*, char const*, int, char const*, tree_code)
        ../../gcc/tree.h:3132
0x9e9224 assign_parm_find_data_types
        ../../gcc/function.c:2456
0x9ef55b gimplify_parameters(gimple**)
        ../../gcc/function.c:4014
0xa5eccc gimplify_body(tree_node*, bool)
        ../../gcc/gimplify.c:12631
0xa5f084 gimplify_function_tree(tree_node*)
        ../../gcc/gimplify.c:12800
0x8aed87 cgraph_node::analyze()
        ../../gcc/cgraphunit.c:670
0x8b2483 analyze_functions
        ../../gcc/cgraphunit.c:1131
0x8b2f92 symbol_table::finalize_compilation_unit()
        ../../gcc/cgraphunit.c:2691
Comment 2 David Malcolm 2018-01-23 01:41:51 UTC
Confirmed.  Seems to have started with r256021.
Comment 3 Dominique d'Humieres 2018-01-23 03:39:56 UTC
> Seems to have started with r256021.

I reached the same conclusion.
Comment 4 Richard Biener 2018-01-23 08:29:40 UTC
Somehow the VAR_DECL for 'y' is in the PARM_DECL chain for the FUNCTION_DECL s:

 <function_decl 0x7ffff6a52f00 s
    type <function_type 0x7ffff6a53f18
        type <void_type 0x7ffff68a8f18 void VOID
            align:8 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type 0x7ffff68a8f18
            pointer_to_this <pointer_type 0x7ffff68af0a8>>
        QI
        size <integer_cst 0x7ffff6894cd8 constant 8>
        unit-size <integer_cst 0x7ffff6894cf0 constant 1>
        align:8 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type 0x7ffff6a53e70
        attributes <tree_list 0x7ffff6a55578
            purpose <identifier_node 0x7ffff6900d48 fn spec>
            value <tree_list 0x7ffff6a55550
                value <string_cst 0x7ffff6a48900 constant ".w">>>
        arg-types <tree_list 0x7ffff6a55528 value <reference_type 0x7ffff6a53dc8>
            chain <tree_list 0x7ffff6a55500 value <integer_type 0x7ffff68a8738 integer(kind=8)>
                chain <tree_list 0x7ffff68a7aa0 value <void_type 0x7ffff68a8f18 void>>>>>
    public static QI t.f90:1:0 align:8 warn_if_not_align:0 context <translation_unit_decl 0x7ffff689e168 t.f90> initial <block 0x7ffff6a58060>
    result <result_decl 0x7ffff689e1e0 D.3758 type <void_type 0x7ffff68a8f18 void>
        ignored VOID t.f90:1:0
        align:8 warn_if_not_align:0 context <function_decl 0x7ffff6a52f00 s>>
    (mem:QI (symbol_ref:DI ("s_") [flags 0x3] <function_decl 0x7ffff6a52f00 s>) [0  S1 A8])
    arguments <parm_decl 0x7ffff6a41280 x
        type <reference_type 0x7ffff6a531f8 type <array_type 0x7ffff6a530a8>
            unsigned restrict DI
            size <integer_cst 0x7ffff6894be8 constant 64>
            unit-size <integer_cst 0x7ffff6894c00 constant 8>
            align:64 warn_if_not_align:0 symtab:0 alias-set -1 structural-equality>
        readonly used unsigned DI passed-by-reference t.f90:1:0 size <integer_cst 0x7ffff6894be8 64> unit-size <integer_cst 0x7ffff6894c00 8>
        align:64 warn_if_not_align:0 context <function_decl 0x7ffff6a52f00 s>
        arg-type <reference_type 0x7ffff6a53dc8 type <array_type 0x7ffff6a53c78>
            unsigned restrict DI size <integer_cst 0x7ffff6894be8 64> unit-size <integer_cst 0x7ffff6894c00 8>
            align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type 0x7ffff6a53dc8>
        chain <parm_decl 0x7ffff6a41200 _x type <integer_type 0x7ffff68a8738 integer(kind=8)>
            readonly used DI t.f90:1:0 size <integer_cst 0x7ffff6894be8 64> unit-size <integer_cst 0x7ffff6894c00 8>
            align:64 warn_if_not_align:0 context <function_decl 0x7ffff6a52f00 s> arg-type <integer_type 0x7ffff68a8738 integer(kind=8)> chain <var_decl 0x7ffff7fefab0 y>>>
Comment 5 Janne Blomqvist 2018-01-23 09:16:44 UTC
I have a "fix", I'll post it shortly.
Comment 6 Janne Blomqvist 2018-01-23 09:34:14 UTC
Patch: https://gcc.gnu.org/ml/gcc-patches/2018-01/msg01917.html
Comment 7 G. Steinmetz 2018-01-29 18:57:41 UTC
Some additional testcases :


$ cat z2.f90
subroutine s(x)
   character(*) :: x
   associate (y => [x])
      print *, size(y), len(y), y
   end associate
end


$ cat z3.f90
subroutine s(x)
   character(*) :: x
   associate (y => [x, x])
      print *, y
   end associate
end


$ cat z4.f90
subroutine s(x)
   character(*) :: x
   associate (y => x)
      y = 'abc'
      print *, y
   end associate
end


$ cat z5.f90
subroutine s(x)
   character(*) :: x
   associate (y => x//x)
      print *, y
   end associate
end
Comment 8 Janne Blomqvist 2018-02-01 19:47:46 UTC
Author: jb
Date: Thu Feb  1 19:47:15 2018
New Revision: 257310

URL: https://gcc.gnu.org/viewcvs?rev=257310&root=gcc&view=rev
Log:
PR 83975 Associate target with non-constant character length

When associating a variable of type character, if the length of the
target isn't known at compile time, generate an error. See PR 83344
for more details.

Regtested on x86_64-pc-linux-gnu.

gcc/fortran/ChangeLog:

2018-02-01  Janne Blomqvist  <jb@gcc.gnu.org>

	PR 83975
	PR 83344
	* resolve.c (resolve_assoc_var): Generate an error if
	target length unknown.

Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/resolve.c
Comment 9 Paul Thomas 2018-02-19 22:09:45 UTC
Author: pault
Date: Mon Feb 19 22:09:13 2018
New Revision: 257827

URL: https://gcc.gnu.org/viewcvs?rev=257827&root=gcc&view=rev
Log:
2018-02-19  Paul Thomas  <pault@gcc.gnu.org>

	PR fortran/83344
	PR fortran/83975
	* resolve.c (resolve_assoc_var): Rearrange the logic for the
	determination of the character length of associate names. If
	the associate name is missing a length expression or the length
	expression is not a constant and the target is not a variable,
	make the associate name allocatable and deferred length.
	* trans-decl.c (gfc_get_symbol_decl): Null the character length
	backend_decl for deferred length associate names that are not
	variables. Set 'length' to gfc_index_zero_node for character
	associate names, whose character length is a PARM_DECL.

2018-02-19  Paul Thomas  <pault@gcc.gnu.org>

	PR fortran/83344
	PR fortran/83975
	* gfortran.dg/associate_22.f90: Enable commented out test.
	* gfortran.dg/associate_36.f90: New test.


Added:
    trunk/gcc/testsuite/gfortran.dg/associate_36.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/resolve.c
    trunk/gcc/fortran/trans-decl.c
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gfortran.dg/associate_22.f90
Comment 10 Janne Blomqvist 2018-02-23 12:02:20 UTC
AFAICT Paul Thomas nailed this in r257827, closing as fixed.