[Bug testsuite/30650] New: [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[Bug testsuite/30650] New: [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

            Bug ID: 30650
           Summary: [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print
                    all_bfds
           Product: gdb
           Version: HEAD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: testsuite
          Assignee: unassigned at sourceware dot org
          Reporter: vries at gcc dot gnu.org
  Target Milestone: ---

With a gdb build with -O2 -flto=auto using gcc 9.3.1, I get:
...
(outer-gdb) PASS: gdb.gdb/python-helper.exp: print *type->main_type
print all_bfds^M
$7 = <optimized out>^M
(outer-gdb) FAIL: gdb.gdb/python-helper.exp: print all_bfds
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

Tom de Vries <vries at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|testsuite                   |symtab

--- Comment #1 from Tom de Vries <vries at gcc dot gnu.org> ---
Hmm, doesn't happen with -readnow, this seems to be a symtab issue.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #2 from Tom de Vries <vries at gcc dot gnu.org> ---
It doesn't reproduce 100%:
...
$ gdb -q -batch ./gdb-x -ex "print all_bfds"
$1 = (htab_t) 0x0
$ gdb -q -batch ./gdb-x -ex "print all_bfds"
$1 = <optimized out>
$ gdb -q -batch ./gdb-x -ex "print all_bfds"
$1 = <optimized out>
$ gdb -q -batch ./gdb-x -ex "print all_bfds"
$1 = <optimized out>
$ gdb -q -batch ./gdb-x -ex "print all_bfds"
$1 = <optimized out>
$ gdb -q -batch ./gdb-x -ex "print all_bfds"
$1 = (htab_t) 0x0
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

Tom de Vries <vries at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tromey at sourceware dot org

--- Comment #3 from Tom de Vries <vries at gcc dot gnu.org> ---
Reproduces with gdb-13-branch, doesn't reproduce with gdb-12-branch.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #4 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #2)
> It doesn't reproduce 100%:

Reproduces 100% with "maint set worker-threads 0":
...
$ for n in $(seq 1 10); do \
    gdb -q -batch -eiex "maint set worker-threads 0" ./gdb-x \
      -ex "print all_bfds"; \
  done
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #5 from Tom de Vries <vries at gcc dot gnu.org> ---
In terms of debug info, we're dealing with a forward inter-CU reference to an
abstract origin:
...
 <0><14b64e8>: Abbrev Number: 1 (DW_TAG_compile_unit)
  ...
 <1><14fc06f>: Abbrev Number: 45 (DW_TAG_variable)
    <14fc070>   DW_AT_abstract_origin: <0x5355817>
    <14fc074>   DW_AT_location    : (DW_OP_addr: 1fd8be8)
  ...
 <0><530411b>: Abbrev Number: 235 (DW_TAG_compile_unit)
  ...
 <1><5355817>: Abbrev Number: 95 (DW_TAG_variable)
    <5355818>   DW_AT_name        : all_bfds
    <535581c>   DW_AT_decl_file   : 186
    <535581d>   DW_AT_decl_line   : 58
    <535581e>   DW_AT_decl_column : 15
    <535581f>   DW_AT_type        : <0x5344aaa>
...

[ FWIW, the address used in the DW_AT_location attribute looks correct:
...
$ nm ./gdb-x | grep all_bfds
0000000001fd8be8 b _ZL8all_bfds
... ]

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #6 from Tom de Vries <vries at gcc dot gnu.org> ---
Resulting cooked index entry in failing case:
...
    [618404] ((cooked_index_entry *) 0x15aeac00)
    name:       all_bfds
    canonical:  all_bfds
    DWARF tag:  DW_TAG_variable
    flags:      0x2 [IS_STATIC]
    DIE offset: 0x14fc06f
    parent:     ((cooked_index_entry *) 0)
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #7 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #6)
> Resulting cooked index entry in failing case:

Same for passing case.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #8 from Tom de Vries <vries at gcc dot gnu.org> ---
I observe the following:
- in set_initial_language we do a lookup of main, which expands the CU
  containing the all_bfds decl.
- when setting a breakpoint on new_symbol conditional on sect_off ==
  all_bfds def, it doesn't trigger

So, there seem to be two independent problems:
- new_symbol should have been called for the all_bfds def
- the all_bfds decl was not combined with the minimal symbol to get the
  address.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #9 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #8)
> - the all_bfds decl was not combined with the minimal symbol to get the
>   address.

I forgot, this is done for global symbols, and all_bfds is static.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #10 from Tom de Vries <vries at gcc dot gnu.org> ---
Looks like a dup of PR25857.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #11 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #10)
> Looks like a dup of PR25857.

Well, at least the "maint set worker-threads 0" case.  With the tentative patch
from that PR we have:
...
$ for n in $(seq 1 10); do gdb -q -batch ./gdb-x -eiex "maint set
worker-threads 0" -ex "print all_bfds"; done
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$1 = (htab_t) 0x0
$ for n in $(seq 1 10); do gdb -q -batch ./gdb-x -ex "print all_bfds"; done
$1 = <optimized out>
$1 = <optimized out>
$1 = (htab_t) 0x0
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
$1 = <optimized out>
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

--- Comment #12 from Tom de Vries <vries at gcc dot gnu.org> ---
Created attachment 15043
  --> https://sourceware.org/bugzilla/attachment.cgi?id=15043&action=edit
Testsuite reproducer

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/30650] [gdb, lto] FAIL: gdb.gdb/python-helper.exp: print all_bfds

[jamborm at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=30650

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

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

-- 
You are receiving this mail because:
You are on the CC list for the bug.