linux-kernel - Re: [PATCH] binder: rust: pin zero-to-one ref increments to initiating thread

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: <aUMV1Kv8dtHLVWIc@google.com>
Date: Wed, 17 Dec 2025 20:43:00 +0000
From: Alice Ryhl <aliceryhl@...gle.com>
To: Pieter-Louis Schoeman <pl.schoeman@...ketlaw.net>
Cc: arve@...roid.com, brauner@...nel.org, cmllamas@...gle.com, 
	gregkh@...uxfoundation.org, tkjos@...roid.com, linux-kernel@...r.kernel.org, 
	ojeda@...nel.org, paul@...l-moore.com, tamird@...il.com, 
	vitaly.wool@...sulko.se, wedsonaf@...il.com, yury.norov@...il.com
Subject: Re: [PATCH] binder: rust: pin zero-to-one ref increments to
 initiating thread

On Mon, Dec 15, 2025 at 09:45:41PM +0200, Pieter-Louis Schoeman wrote:
> The refcount update path must use the zero-to-one aware API on the first
> increment to preserve correct delivery semantics.
> 
> Pass the initiating thread through Process::update_ref() and NodeRef so
> zero->one increments are pinned when available, while keeping decrement
> behavior unchanged.
> 
> Also harden update_refcount_locked() to reject increments.
> 
> Signed-off-by: Pieter-Louis Schoeman <pl.schoeman@...ketlaw.net>

Currently the zero-to-one aware API is used only when a process sends an
outgoing transaction containing a node owned by the sender. Can you
explain what other scenario requires the zero-to-one aware API?

Alice