[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <6d4a0450-9be1-4d91-ba18-5e9bd750fa40@gmail.com>
Date: Tue, 14 May 2024 08:49:20 +0200
From: Heiner Kallweit <hkallweit1@...il.com>
To: Paolo Abeni <pabeni@...hat.com>, Jakub Kicinski <kuba@...nel.org>,
Eric Dumazet <edumazet@...gle.com>, David Miller <davem@...emloft.net>,
Realtek linux nic maintainers <nic_swsd@...ltek.com>
Cc: "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
Ken Milmore <ken.milmore@...il.com>
Subject: [PATCH net 0/2] r8169: Fix GRO-related issue with not disabled device
interrupts
Ken reported that RTL8125b can lock up if gro_flush_timeout has the
default value of 20000 and napi_defer_hard_irqs is set to 0.
In this scenario device interrupts aren't disabled, what seems to
trigger some silicon bug under heavy load. I was able to reproduce this
behavior on RTL8168h.
Disabling device interrupts if NAPI is scheduled from a place other than
the driver's interrupt handler is a necessity in r8169, for other
drivers it may still be a performance optimization.
Therefore add a variant of napi_schedule_prep() with a more granular
return value.
Patch was verified to fix the issue for RTL8168h.
Heiner Kallweit (2):
net: add napi_schedule_prep variant with more granular return value
r8169: disable interrupts also for GRO-scheduled NAPI
drivers/net/ethernet/realtek/r8169_main.c | 6 ++++--
include/linux/netdevice.h | 7 ++++++-
net/core/dev.c | 12 ++++++------
3 files changed, 16 insertions(+), 9 deletions(-)
--
2.45.0
Powered by blists - more mailing lists