| Message ID | 20200205081616.18378-1-kai.heng.feng@canonical.com |
|---|---|
| State | Superseded |
| Delegated to: | Jeff Kirsher |
| Headers | show |
| Series | [v2,1/2] igb: Use device_lock() insead of rtnl_lock() | expand |
Hi Kai-Heng,
url: https://github.com/0day-ci/linux/commits/Kai-Heng-Feng/igb-Use-device_lock-insead-of-rtnl_lock/20200205-174208
base: https://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue.git dev-queue
If you fix the issue, kindly add following tag
Reported-by: kbuild test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
New smatch warnings:
drivers/net/ethernet/intel/igb/igb_main.c:4036 igb_close() warn: inconsistent returns 'dev->mutex'.
# https://github.com/0day-ci/linux/commit/905877ae7d44efc1d9c5c1ae9f4489f56bcb13a6
git remote add linux-review https://github.com/0day-ci/linux
git remote update linux-review
git checkout 905877ae7d44efc1d9c5c1ae9f4489f56bcb13a6
vim +4036 drivers/net/ethernet/intel/igb/igb_main.c
9d5c824399dea8 drivers/net/igb/igb_main.c Auke Kok 2008-01-24 4026
46eafa59e18d03 drivers/net/ethernet/intel/igb/igb_main.c Stefan Assmann 2016-02-03 4027 int igb_close(struct net_device *netdev)
749ab2cd127046 drivers/net/ethernet/intel/igb/igb_main.c Yan, Zheng 2012-01-04 4028 {
905877ae7d44ef drivers/net/ethernet/intel/igb/igb_main.c Kai-Heng Feng 2020-02-05 4029 struct igb_adapter *adapter = netdev_priv(netdev);
905877ae7d44ef drivers/net/ethernet/intel/igb/igb_main.c Kai-Heng Feng 2020-02-05 4030 struct device *dev = &adapter->pdev->dev;
905877ae7d44ef drivers/net/ethernet/intel/igb/igb_main.c Kai-Heng Feng 2020-02-05 4031
905877ae7d44ef drivers/net/ethernet/intel/igb/igb_main.c Kai-Heng Feng 2020-02-05 4032 device_lock(dev);
888f22931478a0 drivers/net/ethernet/intel/igb/igb_main.c Lyude Paul 2017-12-12 4033 if (netif_device_present(netdev) || netdev->dismantle)
749ab2cd127046 drivers/net/ethernet/intel/igb/igb_main.c Yan, Zheng 2012-01-04 4034 return __igb_close(netdev, false);
^^^^^^
Lock held for this return.
905877ae7d44ef drivers/net/ethernet/intel/igb/igb_main.c Kai-Heng Feng 2020-02-05 4035 device_unlock(dev);
9474933caf21a4 drivers/net/ethernet/intel/igb/igb_main.c Todd Fujinaka 2016-11-15 @4036 return 0;
749ab2cd127046 drivers/net/ethernet/intel/igb/igb_main.c Yan, Zheng 2012-01-04 4037 }
749ab2cd127046 drivers/net/ethernet/intel/igb/igb_main.c Yan, Zheng 2012-01-04 4038
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org Intel Corporation
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c index b46bff8fe056..3750e2b926b1 100644 --- a/drivers/net/ethernet/intel/igb/igb_main.c +++ b/drivers/net/ethernet/intel/igb/igb_main.c @@ -4026,8 +4026,13 @@ static int __igb_close(struct net_device *netdev, bool suspending) int igb_close(struct net_device *netdev) { + struct igb_adapter *adapter = netdev_priv(netdev); + struct device *dev = &adapter->pdev->dev; + + device_lock(dev); if (netif_device_present(netdev) || netdev->dismantle) return __igb_close(netdev, false); + device_unlock(dev); return 0; } @@ -8760,7 +8765,6 @@ static int __igb_shutdown(struct pci_dev *pdev, bool *enable_wake, u32 wufc = runtime ? E1000_WUFC_LNKC : adapter->wol; bool wake; - rtnl_lock(); netif_device_detach(netdev); if (netif_running(netdev)) @@ -8769,7 +8773,6 @@ static int __igb_shutdown(struct pci_dev *pdev, bool *enable_wake, igb_ptp_suspend(adapter); igb_clear_interrupt_scheme(adapter); - rtnl_unlock(); status = rd32(E1000_STATUS); if (status & E1000_STATUS_LU) @@ -8897,13 +8900,11 @@ static int __maybe_unused igb_resume(struct device *dev) wr32(E1000_WUS, ~0); - rtnl_lock(); if (!err && netif_running(netdev)) err = __igb_open(netdev, true); if (!err) netif_device_attach(netdev); - rtnl_unlock(); return err; }
Commit 9474933caf21 ("igb: close/suspend race in netif_device_detach") fixed race condition between close and power management ops by using rtnl_lock(). However we can achieve the same by using device_lock() since all power management ops are protected by device_lock(). This fix is a preparation for next patch, to prevent a dead lock under rtnl_lock() when calling runtime resume routine. Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> --- v2: - No change. drivers/net/ethernet/intel/igb/igb_main.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)