6.6-stable review patch. If anyone has any objections, please let me know.
------------------
From: Aleksandr Loktionov <aleksandr.loktionov@xxxxxxxxx>
[ Upstream commit f30490e9695ef7da3d0899c6a0293cc7cd373567 ]
Fix a race condition in the i40e driver that leads to MAC/VLAN filters
becoming corrupted and leaking. Address the issue that occurs under
heavy load when multiple threads are concurrently modifying MAC/VLAN
filters by setting mac and port VLAN.
1. Thread T0 allocates a filter in i40e_add_filter() within
i40e_ndo_set_vf_port_vlan().
2. Thread T1 concurrently frees the filter in __i40e_del_filter() within
i40e_ndo_set_vf_mac().
3. Subsequently, i40e_service_task() calls i40e_sync_vsi_filters(), which
refers to the already freed filter memory, causing corruption.
Reproduction steps:
1. Spawn multiple VFs.
2. Apply a concurrent heavy load by running parallel operations to change
MAC addresses on the VFs and change port VLANs on the host.
3. Observe errors in dmesg:
"Error I40E_AQ_RC_ENOSPC adding RX filters on VF XX,
please set promiscuous on manually for VF XX".
Exact code for stable reproduction Intel can't open-source now.
The fix involves implementing a new intermediate filter state,
I40E_FILTER_NEW_SYNC, for the time when a filter is on a tmp_add_list.
These filters cannot be deleted from the hash list directly but
must be removed using the full process.
Fixes: 278e7d0b9d68 ("i40e: store MAC/VLAN filters in a hash with the MAC Address as key")
Signed-off-by: Aleksandr Loktionov <aleksandr.loktionov@xxxxxxxxx>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@xxxxxxxxx> (A Contingent worker at Intel)
Reviewed-by: Michal Schmidt <mschmidt@xxxxxxxxxx>
Tested-by: Michal Schmidt <mschmidt@xxxxxxxxxx>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@xxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/net/ethernet/intel/i40e/i40e.h | 1 +
drivers/net/ethernet/intel/i40e/i40e_debugfs.c | 1 +
drivers/net/ethernet/intel/i40e/i40e_main.c | 12 ++++++++++--
3 files changed, 12 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 3e6839ac1f0f1..68f403dd2f526 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -763,6 +763,7 @@ enum i40e_filter_state {
I40E_FILTER_ACTIVE, /* Added to switch by FW */
I40E_FILTER_FAILED, /* Rejected by FW */
I40E_FILTER_REMOVE, /* To be removed */
+ I40E_FILTER_NEW_SYNC, /* New, not sent yet, is in i40e_sync_vsi_filters() */
/* There is no 'removed' state; the filter struct is freed */
};
struct i40e_mac_filter {
diff --git a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
index 999c9708def53..a2fca58a91c33 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_debugfs.c
@@ -106,6 +106,7 @@ static char *i40e_filter_state_string[] = {
"ACTIVE",
"FAILED",
"REMOVE",
+ "NEW_SYNC",
};
/**
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 80472aa1deba4..806cec458a072 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -1265,6 +1265,7 @@ int i40e_count_filters(struct i40e_vsi *vsi)
hash_for_each_safe(vsi->mac_filter_hash, bkt, h, f, hlist) {
if (f->state == I40E_FILTER_NEW ||
+ f->state == I40E_FILTER_NEW_SYNC ||
f->state == I40E_FILTER_ACTIVE)
++cnt;
}
@@ -1451,6 +1452,8 @@ static int i40e_correct_mac_vlan_filters(struct i40e_vsi *vsi,
new->f = add_head;
new->state = add_head->state;
+ if (add_head->state == I40E_FILTER_NEW)
+ add_head->state = I40E_FILTER_NEW_SYNC;
/* Add the new filter to the tmp list */
hlist_add_head(&new->hlist, tmp_add_list);
@@ -1560,6 +1563,8 @@ static int i40e_correct_vf_mac_vlan_filters(struct i40e_vsi *vsi,
return -ENOMEM;
new_mac->f = add_head;
new_mac->state = add_head->state;
+ if (add_head->state == I40E_FILTER_NEW)
+ add_head->state = I40E_FILTER_NEW_SYNC;
/* Add the new filter to the tmp list */
hlist_add_head(&new_mac->hlist, tmp_add_list);
@@ -2445,7 +2450,8 @@ static int
i40e_aqc_broadcast_filter(struct i40e_vsi *vsi, const char *vsi_name,
struct i40e_mac_filter *f)
{
- bool enable = f->state == I40E_FILTER_NEW;
+ bool enable = f->state == I40E_FILTER_NEW ||
+ f->state == I40E_FILTER_NEW_SYNC;
struct i40e_hw *hw = &vsi->back->hw;
int aq_ret;
@@ -2619,6 +2625,7 @@ int i40e_sync_vsi_filters(struct i40e_vsi *vsi)
/* Add it to the hash list */
hlist_add_head(&new->hlist, &tmp_add_list);
+ f->state = I40E_FILTER_NEW_SYNC;
}
/* Count the number of active (current and new) VLAN
@@ -2770,7 +2777,8 @@ int i40e_sync_vsi_filters(struct i40e_vsi *vsi)
spin_lock_bh(&vsi->mac_filter_hash_lock);
hlist_for_each_entry_safe(new, h, &tmp_add_list, hlist) {
/* Only update the state if we're still NEW */
- if (new->f->state == I40E_FILTER_NEW)
+ if (new->f->state == I40E_FILTER_NEW ||
+ new->f->state == I40E_FILTER_NEW_SYNC)
new->f->state = new->state;
hlist_del(&new->hlist);
netdev_hw_addr_refcnt(new->f, vsi->netdev, -1);
--
2.43.0
