[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20260120004720.1886632-1-p@1g4.org>
Date: Tue, 20 Jan 2026 00:48:21 +0000
From: Paul Moses <p@....org>
To: netdev@...r.kernel.org
Cc: Jamal Hadi Salim <jhs@...atatu.com>, Cong Wang <xiyou.wangcong@...il.com>, Jiri Pirko <jiri@...nulli.us>, "David S . Miller" <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>, Simon Horman <horms@...nel.org>, linux-kernel@...r.kernel.org, Paul Moses <p@....org>
Subject: [PATCH net v2 0/2] net: sched: act_gate: fix update races and infoleak
This series fixes act_gate schedule update races by switching to a
prepare-then-swap model with an RCU-protected params pointer, so the
hrtimer/datapath never observe partially updated or freed schedules.
Old params are freed via call_rcu() after the swap.
It also zero-initializes the netlink dump struct to prevent padding
information leaks, and tightens schedule/timing validation to avoid
misprogramming the hrtimer on invalid inputs.
Changes since v1:
- Drop tc-testing changes; no test updates required
- Validation fixes: base/cycle range checks + derived cycle overflow guard
- Fix create/update corner cases: avoid oldp deref on create, publish params
only after full init, fix partial schedule copy cleanup
- Timer handling: cancel/reprogram only when required
- Keep dump struct zero-init without unrelated code motion
Patches:
1/2 net/sched: act_gate: fix schedule updates with RCU swap
2/2 net/sched: act_gate: zero-initialize netlink dump struct
--
2.52.GIT
Powered by blists - more mailing lists