[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250710162403.402739-1-bigeasy@linutronix.de>
Date: Thu, 10 Jul 2025 18:24:02 +0200
From: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To: netdev@...r.kernel.org,
linux-rt-devel@...ts.linux.dev,
linux-ppp@...r.kernel.org
Cc: "David S. Miller" <davem@...emloft.net>,
Andrew Lunn <andrew+netdev@...n.ch>,
Clark Williams <clrkwllms@...nel.org>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
Simon Horman <horms@...nel.org>,
Steven Rostedt <rostedt@...dmis.org>,
Thomas Gleixner <tglx@...utronix.de>,
Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Subject: [PATCH net-next v2 0/1] ppp: Replace per-CPU recursion counter with lock-owner field
This is another approach to avoid relying on local_bh_disable() for
locking of per-CPU in ppp.
I redid it with the per-CPU lock and local_lock_nested_bh() as discussed
in v1. The xmit_recursion counter has been removed since it served the
same purpose as the owner field. Both were updated and checked.
The xmit_recursion looks like a counter in ppp_channel_push() but at
this point, the counter should always be 0 so it always serves as a
boolean. Therefore I removed it.
I do admit that this looks easier to review. On the other hand v1 had a
negative diffstat :)
v1…v2 https://lore.kernel.org/all/20250627105013.Qtv54bEk@linutronix.de/
- Instead of rewriting the sequence and adding two owner fields to
the two variables that may recurse it now adds a per-CPU variable
for locking and keeping mostly the old code flow.
Sebastian Andrzej Siewior (1):
ppp: Replace per-CPU recursion counter with lock-owner field
drivers/net/ppp/ppp_generic.c | 38 ++++++++++++++++++++++++++---------
1 file changed, 29 insertions(+), 9 deletions(-)
--
2.50.0
Powered by blists - more mailing lists