linux-kernel - Re: [PATCH v5 10/19] gendwarfksyms: Limit structure expansion

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20241120215454.GA3512979@google.com>
Date: Wed, 20 Nov 2024 21:54:54 +0000
From: Sami Tolvanen <samitolvanen@...gle.com>
To: Masahiro Yamada <masahiroy@...nel.org>,
	Luis Chamberlain <mcgrof@...nel.org>,
	Miguel Ojeda <ojeda@...nel.org>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc: Matthew Maurer <mmaurer@...gle.com>,
	Alex Gaynor <alex.gaynor@...il.com>, Gary Guo <gary@...yguo.net>,
	Petr Pavlu <petr.pavlu@...e.com>,
	Daniel Gomez <da.gomez@...sung.com>, Neal Gompa <neal@...pa.dev>,
	Hector Martin <marcan@...can.st>, Janne Grunau <j@...nau.net>,
	Miroslav Benes <mbenes@...e.cz>,
	Asahi Linux <asahi@...ts.linux.dev>,
	Sedat Dilek <sedat.dilek@...il.com>, linux-kbuild@...r.kernel.org,
	linux-kernel@...r.kernel.org, linux-modules@...r.kernel.org,
	rust-for-linux@...r.kernel.org
Subject: Re: [PATCH v5 10/19] gendwarfksyms: Limit structure expansion

On Wed, Oct 30, 2024 at 05:01:17PM +0000, Sami Tolvanen wrote:
> 
> For pointers, limit structure expansion after the first pointer
> in the symbol type. This should be plenty for detecting ABI
> differences, but it stops us from pulling in half the kernel for
> types that contain pointers to large kernel data structures, like
> task_struct, for example.

We spent some time backtesting gendwarfksyms with android15-6.6
releases to validate the kABI stability options that are currently
implemented [1]. While we discovered a few missing knobs that are
needed, we also realized that pointer expansion limits don't work
when symbols are removed (which is possible without breaking the
existing ABI when TRIM_UNUSED_KSYMS is used), as the processing
cut-off can change. Petr also had some concerns about this earlier,
so instead of coming up with an alternative solution, I'm going to
drop the pointer expansion limit in v6.

I still think that it would be great to come up with a better way to
limit how deep into internal data structures we go when calculating
versions, but I suspect that's going to be challenging unless C
grows an annotation for public / private struct members.

While looking into another version stability issue compared to
genksyms, we also came to the conclusion that private structure
definitions in .c files should be excluded from processing as they
are opaque outside the CU and cannot change the ABI. I'll include
this in v6 as well.

[1] https://github.com/samitolvanen/linux/commits/android15-6.6-2024-11-gendwarfksyms/

Sami