[PATCH v2 0/5] [gdb/symtab] Handle invalid .gdb_index better

[Tom de Vries <tdevries@suse.de>]

I ran into a failure with test-case gdb.tui/tui-missing-src.exp and target
board gold-gdb-index on aarch64-linux.

Investigation showed that this was due to an incorrect address table entry in
the .gdb_index section, with the start of the entry positioned in a hole
between two sections.

The fourth patch detects such entries.

The third patch detects overlapping entries, and the second patch adds some
infrastructure for that patch.

The first patch rejects any .gdb_index with an incorrect adress table.

The last patch changes the related complaints into warnings, making sure that
rejecting the .gdb_index is reported to the user.

The first and fourth patch fix aforementioned test-case.

A v1 was submitted here [1].

New in v2:
- moved the approved patch to be the first patch
- added approved-by tag in first patch
- rewrite commit messages to not refer to other patches, to better allow
  committing patches independently
- add two patches for detecting overlapping entries
- add missing function comment update for create_addrmap_from_gdb_index
- fixed fourth patch to use relocated addresses, after review comment by Simon

[1] https://sourceware.org/pipermail/gdb-patches/2025-August/220087.html

Tom de Vries (5):
  [gdb/symtab] Bail out of create_addrmap_from_gdb_index on error
  [gdb] Make addrmap_mutable::insert_empty return bool
  [gdb/symtab] Detect overlapping ranges in
    create_addrmap_from_gdb_index
  [gdb/symtab] Improve invalid range check in
    create_addrmap_from_gdb_index
  [gdb/symtab] Turn complaints in create_addrmap_from_gdb_index into
    warnings

 gdb/addrmap.c               | 23 +++++++++++----
 gdb/addrmap.h               |  4 +--
 gdb/dwarf2/read-gdb-index.c | 57 +++++++++++++++++++++++++++++--------
 3 files changed, 65 insertions(+), 19 deletions(-)


base-commit: 8186f0d31ddf339ba6803cde97ef6a6310ca2587
-- 
2.43.0