[PATCHv2,net] net: thunderx: use spin_lock_bh in nicvf_set_rx_mode_task()
diff mbox series

Message ID 9809b6b99bf5ba2dc8d1440c7d4fc93d04a7504a.1596524550.git.lucien.xin@gmail.com
State Mainlined
Commit bab9693a9a8c6dd19f670408ec1e78e12a320682
Headers show
Series
  • [PATCHv2,net] net: thunderx: use spin_lock_bh in nicvf_set_rx_mode_task()
Related show

Commit Message

Xin Long Aug. 4, 2020, 7:02 a.m. UTC
A dead lock was triggered on thunderx driver:

        CPU0                    CPU1
        ----                    ----
   [01] lock(&(&nic->rx_mode_wq_lock)->rlock);
                           [11] lock(&(&mc->mca_lock)->rlock);
                           [12] lock(&(&nic->rx_mode_wq_lock)->rlock);
   [02] <Interrupt> lock(&(&mc->mca_lock)->rlock);

The path for each is:

  [01] worker_thread() -> process_one_work() -> nicvf_set_rx_mode_task()
  [02] mld_ifc_timer_expire()
  [11] ipv6_add_dev() -> ipv6_dev_mc_inc() -> igmp6_group_added() ->
  [12] dev_mc_add() -> __dev_set_rx_mode() -> nicvf_set_rx_mode()

To fix it, it needs to disable bh on [1], so that the timer on [2]
wouldn't be triggered until rx_mode_wq_lock is released. So change
to use spin_lock_bh() instead of spin_lock().

Thanks to Paolo for helping with this.

v1->v2:
  - post to netdev.

Reported-by: Rafael P. <rparrazo@redhat.com>
Tested-by: Dean Nelson <dnelson@redhat.com>
Fixes: 469998c861fa ("net: thunderx: prevent concurrent data re-writing by nicvf_set_rx_mode")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
---
 drivers/net/ethernet/cavium/thunder/nicvf_main.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Comments

David Miller Aug. 4, 2020, 8:04 p.m. UTC | #1
From: Xin Long <lucien.xin@gmail.com>
Date: Tue,  4 Aug 2020 15:02:30 +0800

> A dead lock was triggered on thunderx driver:
> 
>         CPU0                    CPU1
>         ----                    ----
>    [01] lock(&(&nic->rx_mode_wq_lock)->rlock);
>                            [11] lock(&(&mc->mca_lock)->rlock);
>                            [12] lock(&(&nic->rx_mode_wq_lock)->rlock);
>    [02] <Interrupt> lock(&(&mc->mca_lock)->rlock);
> 
> The path for each is:
> 
>   [01] worker_thread() -> process_one_work() -> nicvf_set_rx_mode_task()
>   [02] mld_ifc_timer_expire()
>   [11] ipv6_add_dev() -> ipv6_dev_mc_inc() -> igmp6_group_added() ->
>   [12] dev_mc_add() -> __dev_set_rx_mode() -> nicvf_set_rx_mode()
> 
> To fix it, it needs to disable bh on [1], so that the timer on [2]
> wouldn't be triggered until rx_mode_wq_lock is released. So change
> to use spin_lock_bh() instead of spin_lock().
> 
> Thanks to Paolo for helping with this.
> 
> v1->v2:
>   - post to netdev.
> 
> Reported-by: Rafael P. <rparrazo@redhat.com>
> Tested-by: Dean Nelson <dnelson@redhat.com>
> Fixes: 469998c861fa ("net: thunderx: prevent concurrent data re-writing by nicvf_set_rx_mode")
> Signed-off-by: Xin Long <lucien.xin@gmail.com>

Applied and queued up for -stable, thank you.

Patch
diff mbox series

diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_main.c b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
index 2ba0ce1..a82c708 100644
--- a/drivers/net/ethernet/cavium/thunder/nicvf_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
@@ -2042,11 +2042,11 @@  static void nicvf_set_rx_mode_task(struct work_struct *work_arg)
 	/* Save message data locally to prevent them from
 	 * being overwritten by next ndo_set_rx_mode call().
 	 */
-	spin_lock(&nic->rx_mode_wq_lock);
+	spin_lock_bh(&nic->rx_mode_wq_lock);
 	mode = vf_work->mode;
 	mc = vf_work->mc;
 	vf_work->mc = NULL;
-	spin_unlock(&nic->rx_mode_wq_lock);
+	spin_unlock_bh(&nic->rx_mode_wq_lock);
 
 	__nicvf_set_rx_mode_task(mode, mc, nic);
 }