From: "dodji at redhat dot com" <sourceware-bugzilla@sourceware.org>
To: libabigail@sourceware.org
Subject: [Bug default/29610] New: support reading an ocaml abixml symbol table table where a function symbol aliases a variable symbol
Date: Sun, 25 Sep 2022 03:59:11 +0000 [thread overview]
Message-ID: <bug-29610-9487@http.sourceware.org/bugzilla/> (raw)
https://sourceware.org/bugzilla/show_bug.cgi?id=29610
Bug ID: 29610
Summary: support reading an ocaml abixml symbol table table
where a function symbol aliases a variable symbol
Product: libabigail
Version: unspecified
Status: NEW
Severity: normal
Priority: P2
Component: default
Assignee: dodji at redhat dot com
Reporter: dodji at redhat dot com
CC: libabigail at sourceware dot org
Target Milestone: ---
Created attachment 14357
--> https://sourceware.org/bugzilla/attachment.cgi?id=14357&action=edit
Shell script reproducer
Apparently in ELF files generated by Ocaml, a function symbol can alias a
variable symbol, at least on s390 and ppc64le.
This is not yet supported by libabigail's abixml reader as it assumes that a
function symbol can only alias another function symbol, and similarly for
variable symbols.
In abg-reader.cc, build_elf_symbol_db is called to build either a variable or
function symbol and return it. It expects that the resulting function symbol
map or variable symbol map is self-contained. In other words, if a symbol from
a given map aliases another one, then the aliased symbol is also part of the
same map.
To support having one kind of symbol alias the other kind, build_elf_symbol
should also return a map of symbols aliases that are not yet resolved.
Then, read_symbol_db_from_input (which calls build_elf_symbol_db) would then do
the work of resolving the still-not-resolved aliased symbol, i.e, making a
function symbol potentially alias a variable symbol and vice versa.
A reproducer shell script is attached. It reproduces the issue on the
usr/bin/virt-get-kernel binary of the guestfs-tools-1.48.2-1.fc36.s390x.rpm
package.
So, it downloads the package from the Fedora Koji build system, unpacks it and
runs abidw --abidiff on it.
The execution violates an assert in build_elf_symbol_db
--
You are receiving this mail because:
You are on the CC list for the bug.
next reply other threads:[~2022-09-25 3:59 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-25 3:59 dodji at redhat dot com [this message]
2022-10-11 20:19 ` [Bug default/29610] " woodard at redhat dot com
2023-01-07 0:43 ` woodard at redhat dot com
2023-03-24 17:41 ` dodji at redhat dot com
2023-03-27 19:13 ` woodard at redhat dot com
2023-04-11 14:44 ` dodji at redhat dot com
2023-04-11 14:46 ` dodji at redhat dot com
2023-04-11 14:49 ` dodji at redhat dot com
2023-04-11 15:41 ` dodji at redhat dot com
2024-04-04 9:35 ` dodji at seketeli dot org
2024-04-04 15:14 ` dodji at seketeli dot org
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=bug-29610-9487@http.sourceware.org/bugzilla/ \
--to=sourceware-bugzilla@sourceware.org \
--cc=libabigail@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).