[Bug testsuite/27920] New: [gdb/testsuite] hardcoding -readnow skips testing of partial symbols

[Bug testsuite/27920] New: [gdb/testsuite] hardcoding -readnow skips testing of partial symbols

[vries at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27920

            Bug ID: 27920
           Summary: [gdb/testsuite] hardcoding -readnow skips testing of
                    partial symbols
           Product: gdb
           Version: 10.1
            Status: NEW
          Severity: normal
          Priority: P2
         Component: testsuite
          Assignee: unassigned at sourceware dot org
          Reporter: vries at gcc dot gnu.org
  Target Milestone: ---

I looked into test-case gdb.dwarf2/dw2-icycle.exp, and found:
...
if { [prepare_for_testing "failed to prepare" ${testfile} \
          [list $srcfile $srcfile2] {nodebug}] } {
    return -1
}

# We are trying to verify that the partial symtab to symtab expansion
# for the debugging info hand-coded in our assembly file does not cause
# the debugger to crash (infinite recursion).  To facilitate the test,
# start the debugger with -readnow.  This force expansion as soon as
# the objfile is loaded.                                                        

set saved_gdbflags $GDBFLAGS
set GDBFLAGS "$GDBFLAGS -readnow"
clean_restart ${testfile}
set GDBFLAGS $saved_gdbflags
...

There seems to be a mismatch between comment and code.

It's true that -readnow forces full symtabs to be generated.

But it specifically skips expansion from partial symtab to full symtab.

I we want to exercise that path, we should use "maint expand-symtabs" instead.

Generally, hardcoding -readnow in test-cases is undesirable because it skips
testing the default way of handling things in gdb using partial symtabs.

Testing readnow can be done using the target board.

And if we really want a test-case to be tested with readnow with any target
board we can f.i. add -readnow.exp variant (see
gdb.base/list-ambiguous-readnow.exp).  There are also test-cases that
internally iterate over different loading modes like
partial-syms/readnow/index.

I did a first scan in gdb.dwarf2/*.exp, removing -readnow from test-cases, and
ran into PR27919.

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

[Bug testsuite/27920] [gdb/testsuite] hardcoding -readnow skips testing of partial symbols

[cvs-commit at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=27920

--- Comment #1 from cvs-commit at gcc dot gnu.org <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Tom de Vries <vries@sourceware.org>:

https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=248f7165005096fc618c5fd9bf3588719fc589ef

commit 248f7165005096fc618c5fd9bf3588719fc589ef
Author: Tom de Vries <tdevries@suse.de>
Date:   Thu May 27 15:22:38 2021 +0200

    [gdb/symtab] Fix segfault in process_psymtab_comp_unit

    When running test-case gdb.dwarf2/dw2-dummy-cu.exp without -readnow, we run
    into:
    ...
    (gdb) file outputs/gdb.dwarf2/dw2-dummy-cu/dw2-dummy-cu^M
    Reading symbols from outputs/gdb.dwarf2/dw2-dummy-cu/dw2-dummy-cu...^M
    ERROR: Couldn't load dw2-dummy-cu into GDB (eof).
    ...

    The problem is that we're running into a segfault:
    ...
    Thread 1 "gdb" received signal SIGSEGV, Segmentation fault.
    process_psymtab_comp_unit (this_cu=0x2141090, per_objfile=0x1aa4140,
        want_partial_unit=false, pretend_language=language_minimal)
        at /home/vries/gdb_versions/devel/src/gdb/dwarf2/read.c:7023
    7023      switch (reader.comp_unit_die->tag)
    ...
    due to reader.comp_unit_die == nullptr:
    ...
    (gdb) p reader.comp_unit_die
    $1 = (die_info *) 0x0
    ...

    Indeed, there's no CU DIE in the test-case:
    ...
    $ readelf -wi outputs/gdb.dwarf2/dw2-dummy-cu/dw2-dummy-cu
    Contents of the .debug_info section:

      Compilation Unit @ offset 0x0:
       Length:        0x7 (32-bit)
       Version:       2
       Abbrev Offset: 0x0
       Pointer Size:  4
    $
    ...

    Fix this by handling reader.comp_unit_die == nullptr in
    process_psymtab_comp_unit.

    Update the test-case to trigger this PR, as per PR27920 - "[gdb/testsuite]
    hardcoding -readnow skips testing of partial symbols".

    Tested on x86_64-linux.

    gdb/ChangeLog:

    2021-05-27  Tom de Vries  <tdevries@suse.de>

            PR symtab/27919
            * dwarf2/read.c (process_psymtab_comp_unit):

    gdb/testsuite/ChangeLog:

    2021-05-27  Tom de Vries  <tdevries@suse.de>

            PR symtab/27919
            PR testsuite/27920
            * gdb.dwarf2/dw2-dummy-cu.exp: Use maint expand-symtabs instead of
            -readnow.

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