Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   menglong8.dong@gmail.com
To:     helgaas@kernel.org
Cc:     bhelgaas@google.com, linux-pci@vger.kernel.org,
        linux-kernel@vger.kernel.org, Menglong Dong <imagedong@tencent.com>
Subject: [PATCH v2] pci: call cond_resched() after pci bus config writing
Date:   Fri, 15 Oct 2021 10:40:25 +0800
Message-Id: <20211015024025.2916363-1-imagedong@tencent.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Menglong Dong <imagedong@tencent.com>

While the system is running in KVM, pci config writing for virtio devices
may cost long time (about 1-2ms), as it causes VM-exit. During
__pci_bus_assign_resources(), pci_setup_bridge(), which can write pci
config up to 10 times, can be called many times without any
cond_resched(). So __pci_bus_assign_resources() can cause 25+ms
scheduling latency with !CONFIG_PREEMPT.

To solve this problem, call cond_resched() after pci config writing.

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
v2:
- use cond_resched() instead of _cond_resched()
---
 drivers/pci/access.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/pci/access.c b/drivers/pci/access.c
index 46935695cfb9..4c52a50f2c46 100644
--- a/drivers/pci/access.c
+++ b/drivers/pci/access.c
@@ -57,6 +57,7 @@ int noinline pci_bus_write_config_##size \
 	pci_lock_config(flags);						\
 	res = bus->ops->write(bus, devfn, pos, len, value);		\
 	pci_unlock_config(flags);					\
+	cond_resched();						\
 	return res;							\
 }
 
-- 
2.27.0