[next,S78-V8,04/12] i40e: don't hold spinlock while resetting VF

Message ID 20170822105754.29486-5-alice.michael@intel.com
State Under Review
Delegated to: Jeff Kirsher
Headers show

Commit Message

Alice Michael Aug. 22, 2017, 10:57 a.m.
From: Jacob Keller <jacob.e.keller@intel.com>

When we refactored handling of the PVID in commit 9af52f60b2d9
("i40e: use (add|rm)_vlan_all_mac helper functions when changing PVID")
we introduced a scheduling while atomic regression.

This occurred because we now held the spinlock across a call to
i40e_reset_vf(), which results in a usleep_range() call that triggers
a scheduling while atomic bug. This was rare as it only occurred if the
user configured a VLAN on a VF and also attempted to reconfigure the VF
from the host system with a port VLAN.

We do need to hold the lock while calling i40e_is_vsi_in_vlan(), but we
should not be holding it while we reset the VF.

We'll fix this by introducing a separate helper function
i40e_vsi_has_vlans which checks whether we have a PVID and whether the
VSI has configured VLANs. This helper function will manage its own need
for the mac_filter_hash_lock.

Then, we can move the acquiring of the spinlock until after we reset the
VF, which ensures that we do not sleep while holding the lock.

Using a separate function like this makes the code more clear and is
easier to read than attempting to release and re-acquire the spinlock
when we reset the VF.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 36 +++++++++++++++++++---
 1 file changed, 32 insertions(+), 4 deletions(-)

Comments

Bowers, AndrewX Aug. 23, 2017, 4:42 p.m. | #1
> -----Original Message-----
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces@osuosl.org] On
> Behalf Of Alice Michael
> Sent: Tuesday, August 22, 2017 3:58 AM
> To: Michael, Alice <alice.michael@intel.com>; intel-wired-
> lan@lists.osuosl.org
> Subject: [Intel-wired-lan] [next PATCH S78-V8 04/12] i40e: don't hold spinlock
> while resetting VF
> 
> From: Jacob Keller <jacob.e.keller@intel.com>
> 
> When we refactored handling of the PVID in commit 9af52f60b2d9
> ("i40e: use (add|rm)_vlan_all_mac helper functions when changing PVID")
> we introduced a scheduling while atomic regression.
> 
> This occurred because we now held the spinlock across a call to
> i40e_reset_vf(), which results in a usleep_range() call that triggers a
> scheduling while atomic bug. This was rare as it only occurred if the user
> configured a VLAN on a VF and also attempted to reconfigure the VF from
> the host system with a port VLAN.
> 
> We do need to hold the lock while calling i40e_is_vsi_in_vlan(), but we
> should not be holding it while we reset the VF.
> 
> We'll fix this by introducing a separate helper function i40e_vsi_has_vlans
> which checks whether we have a PVID and whether the VSI has configured
> VLANs. This helper function will manage its own need for the
> mac_filter_hash_lock.
> 
> Then, we can move the acquiring of the spinlock until after we reset the VF,
> which ensures that we do not sleep while holding the lock.
> 
> Using a separate function like this makes the code more clear and is easier to
> read than attempting to release and re-acquire the spinlock when we reset
> the VF.
> 
> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
> ---
>  drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 36
> +++++++++++++++++++---
>  1 file changed, 32 insertions(+), 4 deletions(-)

Tested-by: Andrew Bowers <andrewx.bowers@intel.com>

Patch

diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index e156096..25628f5 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -2923,6 +2923,34 @@  int i40e_ndo_set_vf_mac(struct net_device *netdev, int vf_id, u8 *mac)
 }
 
 /**
+ * i40e_vsi_has_vlans - True if VSI has configured VLANs
+ * @vsi: pointer to the vsi
+ *
+ * Check if a VSI has configured any VLANs. False if we have a port VLAN or if
+ * we have no configured VLANs. Do not call while holding the
+ * mac_filter_hash_lock.
+ */
+static bool i40e_vsi_has_vlans(struct i40e_vsi *vsi)
+{
+	bool have_vlans;
+
+	/* If we have a port VLAN, then the VSI cannot have any VLANs
+	 * configured, as all MAC/VLAN filters will be assigned to the PVID.
+	 */
+	if (vsi->info.pvid)
+		return false;
+
+	/* Since we don't have a PVID, we know that if the device is in VLAN
+	 * mode it must be because of a VLAN filter configured on this VSI.
+	 */
+	spin_lock_bh(&vsi->mac_filter_hash_lock);
+	have_vlans = i40e_is_vsi_in_vlan(vsi);
+	spin_unlock_bh(&vsi->mac_filter_hash_lock);
+
+	return have_vlans;
+}
+
+/**
  * i40e_ndo_set_vf_port_vlan
  * @netdev: network interface device structure
  * @vf_id: VF identifier
@@ -2974,10 +3002,7 @@  int i40e_ndo_set_vf_port_vlan(struct net_device *netdev, int vf_id,
 		/* duplicate request, so just return success */
 		goto error_pvid;
 
-	/* Locked once because multiple functions below iterate list */
-	spin_lock_bh(&vsi->mac_filter_hash_lock);
-
-	if (le16_to_cpu(vsi->info.pvid) == 0 && i40e_is_vsi_in_vlan(vsi)) {
+	if (i40e_vsi_has_vlans(vsi)) {
 		dev_err(&pf->pdev->dev,
 			"VF %d has already configured VLAN filters and the administrator is requesting a port VLAN override.\nPlease unload and reload the VF driver for this change to take effect.\n",
 			vf_id);
@@ -2990,6 +3015,9 @@  int i40e_ndo_set_vf_port_vlan(struct net_device *netdev, int vf_id,
 		vsi = pf->vsi[vf->lan_vsi_idx];
 	}
 
+	/* Locked once because multiple functions below iterate list */
+	spin_lock_bh(&vsi->mac_filter_hash_lock);
+
 	/* Check for condition where there was already a port VLAN ID
 	 * filter set and now it is being deleted by setting it to zero.
 	 * Additionally check for the condition where there was a port