[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250625034021.3650359-1-dqfext@gmail.com>
Date: Wed, 25 Jun 2025 11:40:17 +0800
From: Qingfang Deng <dqfext@...il.com>
To: Andrew Lunn <andrew+netdev@...n.ch>,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
linux-ppp@...r.kernel.org,
netdev@...r.kernel.org,
linux-kernel@...r.kernel.org
Cc: Guillaume Nault <gnault@...hat.com>
Subject: [PATCH net-next 0/3] ppp: improve receive path performance
This patch series improves the performance of the PPPoE receive paths.
Patch 1 converts the ppp->rlock from a spinlock to a read-write lock,
allowing concurrent receive-side processing when no state is being
modified.
Patch 2 optimizes PPPoE receive performance by bypassing sk_receive_skb()
when the socket is in the PPPOX_BOUND state, avoiding unnecessary socket
locking and overhead.
Patch 3 synchronizes all updates to net_device->stats using
DEV_STATS_INC() to prevent data races now that the receive path may run on
multiple CPUs.
Qingfang Deng (3):
ppp: convert rlock to rwlock to improve RX concurrency
pppoe: call ppp_input directly when PPPOX_BOUND
ppp: synchronize netstats updates
drivers/net/ppp/ppp_generic.c | 32 +++++++++++++++++---------------
drivers/net/ppp/pppoe.c | 10 +++++++++-
2 files changed, 26 insertions(+), 16 deletions(-)
--
2.43.0
Powered by blists - more mailing lists