[Bug debug/113382] New: FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[Bug debug/113382] New: FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113382

            Bug ID: 113382
           Summary: FAIL: gcc.dg/debug/btf/btf-bitfields-3.c
                    scan-assembler-times [\t ]0x6000004[\t
                    ]+[^\n]*btt_info 1
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: danglin at gcc dot gnu.org
  Target Milestone: ---
              Host: hppa64-hp-hpux11.11
            Target: hppa64-hp-hpux11.11
             Build: hppa64-hp-hpux11.11

Created attachment 57074
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57074&action=edit
Assembler output

spawn -ignore SIGHUP /home/dave/gnu/gcc/objdir64/gcc/xgcc
-B/home/dave/gnu/gcc/o
bjdir64/gcc/
/home/dave/gnu/gcc/gcc/gcc/testsuite/gcc.dg/debug/btf/btf-bitfields
-3.c -fdiagnostics-plain-output -O0 -gbtf -dA -fno-ident -S -o
btf-bitfields-3.s
PASS: gcc.dg/debug/btf/btf-bitfields-3.c (test for excess errors)
gcc.dg/debug/btf/btf-bitfields-3.c: [\t ]0x6000004[\t ]+[^\n]*btt_info found 0
t
imes
FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t
]+[^\n]*btt_info 1
PASS: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t
]0x84000001[\t ]+[^\n]*btt_info 1
PASS: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times  btm_type:
\\(BTF_KIND_ENUM 'foo' 1

[Bug debug/113382] FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113382

--- Comment #1 from John David Anglin <danglin at gcc dot gnu.org> ---
Similar fails:
FAIL: gcc.dg/debug/btf/btf-enum-1.c scan-assembler-times [\\t ]0x6000004[\\t
]+[^\\n]*btt_info 1
FAIL: gcc.dg/debug/btf/btf-enum-small.c scan-assembler-times [\\t
]0x6000002[\\t ]+[^\\n]*btt_info 1
FAIL: gcc.dg/debug/btf/btf-enum64-1.c scan-assembler-times [\\t ]0x13000003[\\t
]+[^\\n]*btt_info 2
FAIL: gcc.dg/debug/btf/btf-enum64-1.c scan-assembler-times [\\t ]0x93000003[\\t
]+[^\\n]*btt_info 1

[Bug debug/113382] FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113382

--- Comment #2 from John David Anglin <danglin at gcc dot gnu.org> ---
We have:
        .word   0x86000004      ; btt_info: kind=6, kflag=1, vlen=4

This seems correct encoding:

/* Encoding for struct btf_type.info.  */
#define BTF_TYPE_INFO(kind, kflag, vlen) \
  ((((kflag) ? 1 : 0 ) << 31) | ((kind) << 24) | ((vlen) & 0xffff))

Problem seems kflag.

[Bug debug/113382] FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113382

--- Comment #3 from John David Anglin <danglin at gcc dot gnu.org> ---
btf_kflag is set here:

  else if (btf_kind == BTF_KIND_ENUM)
    {
      btf_kflag = dtd->dtd_enum_unsigned
                    ? BTF_KF_ENUM_UNSIGNED
                    : BTF_KF_ENUM_SIGNED;

(gdb) p dtd->dtd_enum_unsigned
$10 = false

/* BTF_KF_ENUM_ holds the flags for kflags in BTF_KIND_ENUM{,64}.  */
#define BTF_KF_ENUM_UNSIGNED    (0)
#define BTF_KF_ENUM_SIGNED      (1 << 0)

Test has:
enum foo
{
  BAR = 0,
  BAZ = 1,
  QUZ = 2,
  QUX = 3
};

I think the choice of signed int is correct:
https://en.cppreference.com/w/c/language/enum

But then there are articles indicating that the type is implementation
dependent:
https://stackoverflow.com/questions/159034/are-c-enums-signed-or-unsigned

Maybe test needs to handle either case.

[Bug debug/113382] FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113382

--- Comment #4 from John David Anglin <danglin at gcc dot gnu.org> ---
dtd->dtd_enum_unsigned is set in ctf_add_enum:
dtd->dtd_enum_unsigned = eunsigned;

  /* Generate a CTF type for the enumeration.  */
  enumeration_type_id = ctf_add_enum (ctfc, CTF_ADD_ROOT,
                                      enum_name, bit_size / 8,
                                      (signedness == DW_ATE_unsigned),
                                      enumeration);

signedness = 0

(gdb) bt
#0  ctf_add_enum (ctfc=0x800003ffbfea7c00, flag=1,
    name=0x800003ffbfe6b188 "foo", size=4, eunsigned=false,
    die=0x800003ffbfea2320) at ../../gcc/gcc/ctfc.cc:591
#1  0x400000000082df34 in gen_ctf_enumeration_type (
    enumeration=0x800003ffbfea2320, ctfc=0x800003ffbfea7c00)
    at ../../gcc/gcc/dwarf2ctf.cc:762
#2  gen_ctf_type (ctfc=0x800003ffbfea7c00, die=0x800003ffbfea2320)
    at ../../gcc/gcc/dwarf2ctf.cc:899
#3  0x400000000082e8b8 in ctf_do_die (die=0x800003ffbfea2320)
    at ../../gcc/gcc/dwarf2ctf.cc:978
#4  0x400000000088f9b0 in ctf_debug_do_cu (die=<optimized out>)
    at ../../gcc/gcc/dwarf2out.cc:33017
#5  ctf_debug_do_cu (die=<optimized out>) at ../../gcc/gcc/dwarf2out.cc:33010
#6  dwarf2out_early_finish (filename=0x800003ffbfea7c00 "▒\362\004\002")
    at ../../gcc/gcc/dwarf2out.cc:33146
#7  0x40000000007de578 in symbol_table::finalize_compilation_unit (
    this=0x800003ffbfe6b188) at ../../gcc/gcc/cgraphunit.cc:2579
#8  0x4000000000d55338 in compile_file () at ../../gcc/gcc/toplev.cc:474

[Bug debug/113382] FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113382

--- Comment #5 from John David Anglin <danglin at gcc dot gnu.org> ---
The problem seems to be DW_AT_encoding is not found in this call:
static ctf_id_t
gen_ctf_enumeration_type (ctf_container_ref ctfc, dw_die_ref enumeration)
{
  const char *enum_name = get_AT_string (enumeration, DW_AT_name);
  unsigned int bit_size = ctf_die_bitsize (enumeration);
  unsigned int signedness = get_AT_unsigned (enumeration, DW_AT_encoding);

get_AT() returns NULL.

This is because dwarf_strict is 1:
      if (!dwarf_strict)
        add_AT_unsigned (type_die, DW_AT_encoding,
                         TYPE_UNSIGNED (type)
                         ? DW_ATE_unsigned
                         : DW_ATE_signed);

I believe we need to add -gno-strict-dwarf option on hppa*64*-*-hpux*.

[Bug debug/113382] FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113382

--- Comment #6 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by John David Anglin <danglin@gcc.gnu.org>:

https://gcc.gnu.org/g:bc77c035c45bb224790b1c03d06a64c8a1cc51c5

commit r14-8338-gbc77c035c45bb224790b1c03d06a64c8a1cc51c5
Author: John David Anglin <danglin@gcc.gnu.org>
Date:   Mon Jan 22 19:07:32 2024 +0000

    Add -gno-strict-dwarf to dg-options in various btf enum tests

    The -gno-strict-dwarf option is needed to ensure enum signedness
    is added to type_die.

    2024-01-22  John David Anglin  <danglin@gcc.gnu.org>

    gcc/testsuite/ChangeLog:

            PR debug/113382
            * gcc.dg/debug/btf/btf-bitfields-3.c: Add -gno-strict-dwarf
            option to dg-options.
            * gcc.dg/debug/btf/btf-enum-1.c: Likewise.
            * gcc.dg/debug/btf/btf-enum-small.c: Likewise.
            * gcc.dg/debug/btf/btf-enum64-1.c: Likewise.

[Bug debug/113382] FAIL: gcc.dg/debug/btf/btf-bitfields-3.c scan-assembler-times [\t ]0x6000004[\t ]+[^\n]*btt_info 1

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113382

John David Anglin <danglin at gcc dot gnu.org> changed:

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

--- Comment #7 from John David Anglin <danglin at gcc dot gnu.org> ---
Fixed on trunk.