[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20211117025956.79616-1-dave@stgolabs.net>
Date: Tue, 16 Nov 2021 18:59:53 -0800
From: Davidlohr Bueso <dave@...olabs.net>
To: martin.petersen@...cle.com, jejb@...ux.ibm.com
Cc: hare@...e.de, bigeasy@...utronix.de, tglx@...utronix.de,
linux-scsi@...r.kernel.org, linux-rt-users@...r.kernel.org,
linux-kernel@...r.kernel.org, dave@...olabs.net
Subject: [PATCH -next 0/3] scsi/fcoe: Play nicer with PREEMPT_RT
Hi,
The following are the result of trying to get rid of the out-of-tree
equivalent[1].
Patches 1 and 2 remove the actual scheduling while atomic scenarios.
Patch 3 could be considered optional because afaict the stats do not
actually have blocking calls while having preemption disabled. But
from an RT perspective it's still beneficial as the region remains
preemptible.
Compile tested only.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git/tree/patches/scsi_fcoe__Make_RT_aware..patch?h=linux-5.15.y-rt-patches
Thanks!
Davidlohr Bueso (3):
scsi/libfc: Remove get_cpu() semantics in fc_exch_em_alloc()
scsi/fcoe: Add a local_lock to fcoe_percpu
scsi/fcoe: Add a local_lock to percpu localport statistics
drivers/scsi/bnx2fc/bnx2fc_fcoe.c | 16 ++++++++------
drivers/scsi/bnx2fc/bnx2fc_io.c | 5 +++--
drivers/scsi/fcoe/fcoe.c | 35 +++++++++++++++++++------------
drivers/scsi/fcoe/fcoe_ctlr.c | 24 +++++++++++++--------
drivers/scsi/libfc/fc_exch.c | 3 +--
drivers/scsi/libfc/fc_fcp.c | 31 +++++++++++++++++++--------
drivers/scsi/qedf/qedf_main.c | 7 ++++---
include/scsi/libfc.h | 7 +++++++
include/scsi/libfcoe.h | 2 ++
9 files changed, 86 insertions(+), 44 deletions(-)
--
2.26.2
Powered by blists - more mailing lists