Skip to contentSkip to conversation

ice: Reset VF on Tx MDD event

PATCH Superseded Marcin Szycik 2 participants 4 messages Apr 2, 2024 – Apr 3, 2024

Subsystems (3)

v4 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH v4 1/1 ice: Reset VF on Tx MDD event

In cases when VF sends malformed packets that are classified as malicious, sometimes it causes Tx queue to freeze. This frozen queue can be stuck for several minutes being unusable. This behavior can be reproduced with a faulty userspace app running on VF.

When any Malicious Driver Detection event occurs and the mdd-auto-reset-vf private flag is set, perform a graceful VF reset to quickly bring VF back to operational state. Add a log message to notify about the cause of the reset. Add a helper for this to be reused for both TX and RX events.

---
v4: Only perform auto-reset once per VF
v3 [1]: Only auto reset VF if the mdd-auto-reset-vf flag is set v2 [2]: Revert an unneeded formatting change, fix commit message, fix a log
message with a correct event name

[1] https://lore.kernel.org/intel-wired-lan/[email protected]
[2] https://lore.kernel.org/netdev/[email protected]

---
 drivers/net/ethernet/intel/ice/ice_main.c  | 57 +++++++++++++++++-----
 drivers/net/ethernet/intel/ice/ice_sriov.c | 25 +++++++---
 drivers/net/ethernet/intel/ice/ice_sriov.h |  2 +
 3 files changed, 67 insertions(+), 17 deletions(-)
drivers/net/ethernet/intel/ice/ice_main.c +46 -11
index 185c9b13efcf..80bc83f6e1ab 100644
@@ -1745,6 +1745,39 @@ static void ice_service_timer(struct timer_list *t)
 	ice_service_task_schedule(pf);
 }
 
+/**
+ * ice_mdd_maybe_reset_vf - reset VF after MDD event
+ * @pf: pointer to the PF structure
+ * @vf: pointer to the VF structure
+ * @reset_vf_tx: whether Tx MDD has occurred
+ * @reset_vf_rx: whether Rx MDD has occurred
+ *
+ * Since the queue can get stuck on VF MDD events, the PF can be configured to
+ * automatically reset the VF by enabling the private ethtool flag
+ * mdd-auto-reset-vf.
+ */
+static void ice_mdd_maybe_reset_vf(struct ice_pf *pf, struct ice_vf *vf,
+				   bool reset_vf_tx, bool reset_vf_rx)
+{
+	struct device *dev = ice_pf_to_dev(pf);
+
+	if (!test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags))
+		return;
+
+	/* VF MDD event counters will be cleared by reset, so print the event
+	 * prior to reset.
+	 */
+	if (reset_vf_tx)
+		ice_print_vf_tx_mdd_event(vf);
+
+	if (reset_vf_rx)
+		ice_print_vf_rx_mdd_event(vf);
+
+	dev_info(dev, "PF-to-VF reset on PF %d VF %d due to MDD event\n",
+		 pf->hw.pf_id, vf->vf_id);
+	ice_reset_vf(vf, ICE_VF_RESET_NOTIFY | ICE_VF_RESET_LOCK);
+}
+
 /**
  * ice_handle_mdd_event - handle malicious driver detect event
  * @pf: pointer to the PF structure
@@ -1838,6 +1871,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 	 */
 	mutex_lock(&pf->vfs.table_lock);
 	ice_for_each_vf(pf, bkt, vf) {
+		bool reset_vf_tx = false, reset_vf_rx = false;
+
 		reg = rd32(hw, VP_MDET_TX_PQM(vf->vf_id));
 		if (reg & VP_MDET_TX_PQM_VALID_M) {
 			wr32(hw, VP_MDET_TX_PQM(vf->vf_id), 0xFFFF);
@@ -1846,6 +1881,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection event TX_PQM detected on VF %d\n",
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_TX_TCLAN(vf->vf_id));
@@ -1856,6 +1893,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection event TX_TCLAN detected on VF %d\n",
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_TX_TDPU(vf->vf_id));
@@ -1866,6 +1905,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection event TX_TDPU detected on VF %d\n",
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_RX(vf->vf_id));
@@ -1877,18 +1918,12 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 				dev_info(dev, "Malicious Driver Detection event RX detected on VF %d\n",
 					 vf->vf_id);
 
-			/* Since the queue is disabled on VF Rx MDD events, the
-			 * PF can be configured to reset the VF through ethtool
-			 * private flag mdd-auto-reset-vf.
-			 */
-			if (test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags)) {
-				/* VF MDD event counters will be cleared by
-				 * reset, so print the event prior to reset.
-				 */
-				ice_print_vf_rx_mdd_event(vf);
-				ice_reset_vf(vf, ICE_VF_RESET_LOCK);
-			}
+			reset_vf_rx = true;
 		}
+
+		if (reset_vf_tx || reset_vf_rx)
+			ice_mdd_maybe_reset_vf(pf, vf, reset_vf_tx,
+					       reset_vf_rx);
 	}
 	mutex_unlock(&pf->vfs.table_lock);
 
drivers/net/ethernet/intel/ice/ice_sriov.c +19 -6
index fb2e96db647e..a60dacf8942a 100644
@@ -1861,6 +1861,24 @@ void ice_print_vf_rx_mdd_event(struct ice_vf *vf)
 			  ? "on" : "off");
 }
 
+/**
+ * ice_print_vf_tx_mdd_event - print VF Tx malicious driver detect event
+ * @vf: pointer to the VF structure
+ */
+void ice_print_vf_tx_mdd_event(struct ice_vf *vf)
+{
+	struct ice_pf *pf = vf->pf;
+	struct device *dev;
+
+	dev = ice_pf_to_dev(pf);
+
+	dev_info(dev, "%d Tx Malicious Driver Detection events detected on PF %d VF %d MAC %pM. mdd-auto-reset-vfs=%s\n",
+		 vf->mdd_tx_events.count, pf->hw.pf_id, vf->vf_id,
+		 vf->dev_lan_addr,
+		 test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags)
+			  ? "on" : "off");
+}
+
 /**
  * ice_print_vfs_mdd_events - print VFs malicious driver detect event
  * @pf: pointer to the PF structure
@@ -1869,8 +1887,6 @@ void ice_print_vf_rx_mdd_event(struct ice_vf *vf)
  */
 void ice_print_vfs_mdd_events(struct ice_pf *pf)
 {
-	struct device *dev = ice_pf_to_dev(pf);
-	struct ice_hw *hw = &pf->hw;
 	struct ice_vf *vf;
 	unsigned int bkt;
 
@@ -1897,10 +1913,7 @@ void ice_print_vfs_mdd_events(struct ice_pf *pf)
 		if (vf->mdd_tx_events.count != vf->mdd_tx_events.last_printed) {
 			vf->mdd_tx_events.last_printed =
 							vf->mdd_tx_events.count;
-
-			dev_info(dev, "%d Tx Malicious Driver Detection events detected on PF %d VF %d MAC %pM.\n",
-				 vf->mdd_tx_events.count, hw->pf_id, vf->vf_id,
-				 vf->dev_lan_addr);
+			ice_print_vf_tx_mdd_event(vf);
 		}
 	}
 	mutex_unlock(&pf->vfs.table_lock);
drivers/net/ethernet/intel/ice/ice_sriov.h +2 -0
index 4ba8fb53aea1..8f22313474d6 100644
@@ -58,6 +58,7 @@ void
 ice_vf_lan_overflow_event(struct ice_pf *pf, struct ice_rq_event_info *event);
 void ice_print_vfs_mdd_events(struct ice_pf *pf);
 void ice_print_vf_rx_mdd_event(struct ice_vf *vf);
+void ice_print_vf_tx_mdd_event(struct ice_vf *vf);
 bool
 ice_vc_validate_pattern(struct ice_vf *vf, struct virtchnl_proto_hdrs *proto);
 u32 ice_sriov_get_vf_total_msix(struct pci_dev *pdev);
@@ -69,6 +70,7 @@ static inline
 void ice_vf_lan_overflow_event(struct ice_pf *pf, struct ice_rq_event_info *event) { }
 static inline void ice_print_vfs_mdd_events(struct ice_pf *pf) { }
 static inline void ice_print_vf_rx_mdd_event(struct ice_vf *vf) { }
+static inline void ice_print_vf_tx_mdd_event(struct ice_vf *vf) { }
 static inline void ice_restore_all_vfs_msi_state(struct ice_pf *pf) { }
 
 static inline int
Signature · 1 line
2.41.0
#1 lore raw
Quoted 16 lines

-----Original Message-----
From: Intel-wired-lan <[email protected]> On Behalf Of Marcin Szycik
Sent: Tuesday, April 2, 2024 6:52 PM
To: [email protected]
Cc: Drewek, Wojciech <[email protected]>; Wang, Liang-min <[email protected]>; [email protected]; Chmielewski, Pawel <[email protected]>; Marcin Szycik <[email protected]>; Nguyen, Anthony L <[email protected]>; [email protected]; Kitszel, Przemyslaw <[email protected]>
Subject: [Intel-wired-lan] [PATCH iwl-next v4] ice: Reset VF on Tx MDD event

Please state in the title explicitly the purpose of the patch:
Do you fix a bug? say fix
Do you add functionality? say add
Do you refactor? say refactor

Quoted 11 lines

In cases when VF sends malformed packets that are classified as malicious, sometimes it causes Tx queue to freeze. This frozen queue can be stuck for several minutes being unusable. This behavior can be reproduced with a faulty userspace app running on VF.

When any Malicious Driver Detection event occurs and the mdd-auto- reset-vf private flag is set, perform a graceful VF reset to quickly bring VF back to operational state. Add a log message to notify about the cause of the reset. Add a helper for this to be reused for both TX and RX events.

Please describe your changes explicitly:
Do you just add a helper function without adding a new functionality?
Do you add functionality?
Do you fix?
Please explain what are the changes in driver behavior and the changes in the driver sources you make. Thank you

Quoted 224 lines

---
v4: Only perform auto-reset once per VF
v3 [1]: Only auto reset VF if the mdd-auto-reset-vf flag is set v2 [2]: Revert an unneeded formatting change, fix commit message, fix a log
message with a correct event name

[1] https://lore.kernel.org/intel-wired-lan/20240326164455.735739-
[email protected]
[2] https://lore.kernel.org/netdev/20231102155149.2574209-1-
[email protected]

---
 drivers/net/ethernet/intel/ice/ice_main.c  | 57 +++++++++++++++++-
----  drivers/net/ethernet/intel/ice/ice_sriov.c | 25 +++++++---
drivers/net/ethernet/intel/ice/ice_sriov.h |  2 +
 3 files changed, 67 insertions(+), 17 deletions(-)
drivers/net/ethernet/intel/ice/ice_main.c +20 -0
index 185c9b13efcf..80bc83f6e1ab 100644
@@ -1745,6 +1745,39 @@ static void ice_service_timer(struct
 timer_list *t)
 	ice_service_task_schedule(pf);
 }
 
+/**
+ * ice_mdd_maybe_reset_vf - reset VF after MDD event
+ * @pf: pointer to the PF structure
+ * @vf: pointer to the VF structure
+ * @reset_vf_tx: whether Tx MDD has occurred
+ * @reset_vf_rx: whether Rx MDD has occurred
+ *
+ * Since the queue can get stuck on VF MDD events, the PF can be
+configured to
+ * automatically reset the VF by enabling the private ethtool flag
+ * mdd-auto-reset-vf.
+ */
+static void ice_mdd_maybe_reset_vf(struct ice_pf *pf, struct
 ice_vf *vf,
+				   bool reset_vf_tx, bool reset_vf_rx) {
+	struct device *dev = ice_pf_to_dev(pf);
+
+	if (!test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags))
+		return;
+
+	/* VF MDD event counters will be cleared by reset, so print

the event

drivers/net/ethernet/intel/ice/ice_main.c +9 -0
+	 * prior to reset.
+	 */
+	if (reset_vf_tx)
+		ice_print_vf_tx_mdd_event(vf);
+
+	if (reset_vf_rx)
+		ice_print_vf_rx_mdd_event(vf);
+
+	dev_info(dev, "PF-to-VF reset on PF %d VF %d due to MDD

event\n",

drivers/net/ethernet/intel/ice/ice_main.c +5 -0
+		 pf->hw.pf_id, vf->vf_id);
+	ice_reset_vf(vf, ICE_VF_RESET_NOTIFY | ICE_VF_RESET_LOCK); }
+
 /**
  * ice_handle_mdd_event - handle malicious driver detect event
  * @pf: pointer to the PF structure
@@ -1838,6 +1871,8 @@ static void ice_handle_mdd_event(struct
 ice_pf *pf)
 	 */
 	mutex_lock(&pf->vfs.table_lock);
 	ice_for_each_vf(pf, bkt, vf) {
+		bool reset_vf_tx = false, reset_vf_rx = false;
+
 		reg = rd32(hw, VP_MDET_TX_PQM(vf->vf_id));
 		if (reg & VP_MDET_TX_PQM_VALID_M) {

wr32(hw, VP_MDET_TX_PQM(vf->vf_id), 0xFFFF); @@ -
1846,6 +1881,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)

drivers/net/ethernet/intel/ice/ice_main.c +0 -0
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_PQM detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_TX_TCLAN(vf->vf_id)); @@ -1856,6
+1893,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_TCLAN detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_TX_TDPU(vf->vf_id)); @@ -1866,6
+1905,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_TDPU detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_RX(vf->vf_id)); @@ -1877,18
+1918,12 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 				dev_info(dev, "Malicious Driver Detection

event RX detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +0 -1
 					 vf->vf_id);
 
-			/* Since the queue is disabled on VF Rx MDD

events, the

drivers/net/ethernet/intel/ice/ice_main.c +0 -1
-			 * PF can be configured to reset the VF through

ethtool

drivers/net/ethernet/intel/ice/ice_main.c +0 -3
-			 * private flag mdd-auto-reset-vf.
-			 */
-			if (test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf-
Quoted 1 line

flags)) {

drivers/net/ethernet/intel/ice/ice_main.c +0 -2
-				/* VF MDD event counters will be cleared by
-				 * reset, so print the event prior to

reset.

drivers/net/ethernet/intel/ice/ice_main.c +5 -4
-				 */
-				ice_print_vf_rx_mdd_event(vf);
-				ice_reset_vf(vf, ICE_VF_RESET_LOCK);
-			}
+			reset_vf_rx = true;
 		}
+
+		if (reset_vf_tx || reset_vf_rx)
+			ice_mdd_maybe_reset_vf(pf, vf, reset_vf_tx,
+					       reset_vf_rx);
 	}
 	mutex_unlock(&pf->vfs.table_lock);
drivers/net/ethernet/intel/ice/ice_sriov.c +18 -0
index fb2e96db647e..a60dacf8942a 100644
@@ -1861,6 +1861,24 @@ void ice_print_vf_rx_mdd_event(struct ice_vf
 *vf)
 			  ? "on" : "off");
 }
 
+/**
+ * ice_print_vf_tx_mdd_event - print VF Tx malicious driver detect
+event
+ * @vf: pointer to the VF structure
+ */
+void ice_print_vf_tx_mdd_event(struct ice_vf *vf) {
+	struct ice_pf *pf = vf->pf;
+	struct device *dev;
+
+	dev = ice_pf_to_dev(pf);
+
+	dev_info(dev, "%d Tx Malicious Driver Detection events
 detected on PF %d VF %d MAC %pM. mdd-auto-reset-vfs=%s\n",
+		 vf->mdd_tx_events.count, pf->hw.pf_id, vf->vf_id,
+		 vf->dev_lan_addr,
+		 test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags)
+			  ? "on" : "off");
+}
+
 /**

* ice_print_vfs_mdd_events - print VFs malicious driver detect
event
* @pf: pointer to the PF structure

drivers/net/ethernet/intel/ice/ice_sriov.c +1 -6
@@ -1869,8 +1887,6 @@ void ice_print_vf_rx_mdd_event(struct ice_vf
 *vf)
  */
 void ice_print_vfs_mdd_events(struct ice_pf *pf)  {
-	struct device *dev = ice_pf_to_dev(pf);
-	struct ice_hw *hw = &pf->hw;
 	struct ice_vf *vf;
 	unsigned int bkt;
 
@@ -1897,10 +1913,7 @@ void ice_print_vfs_mdd_events(struct ice_pf
 *pf)
 		if (vf->mdd_tx_events.count != vf-
 >mdd_tx_events.last_printed) {
 			vf->mdd_tx_events.last_printed =
 							vf->mdd_tx_events.count;
-
-			dev_info(dev, "%d Tx Malicious Driver Detection
 events detected on PF %d VF %d MAC %pM.\n",
-				 vf->mdd_tx_events.count, hw->pf_id, vf-
 >vf_id,
-				 vf->dev_lan_addr);
+			ice_print_vf_tx_mdd_event(vf);
 		}
 	}
 	mutex_unlock(&pf->vfs.table_lock);
drivers/net/ethernet/intel/ice/ice_sriov.h +0 -0
index 4ba8fb53aea1..8f22313474d6 100644
@@ -58,6 +58,7 @@ void
 ice_vf_lan_overflow_event(struct ice_pf *pf, struct

ice_rq_event_info *event); void ice_print_vfs_mdd_events(struct ice_pf *pf); void ice_print_vf_rx_mdd_event(struct ice_vf *vf); +void ice_print_vf_tx_mdd_event(struct ice_vf *vf);
bool
ice_vc_validate_pattern(struct ice_vf *vf, struct
virtchnl_proto_hdrs *proto);
u32 ice_sriov_get_vf_total_msix(struct pci_dev *pdev); @@ -69,6
+70,7 @@ static inline void ice_vf_lan_overflow_event(struct ice_pf *pf, struct ice_rq_event_info *event) { } static inline void ice_print_vfs_mdd_events(struct ice_pf *pf) { } static inline void ice_print_vf_rx_mdd_event(struct ice_vf *vf) { } +static inline void ice_print_vf_tx_mdd_event(struct ice_vf *vf) { }
static inline void ice_restore_all_vfs_msi_state(struct ice_pf
*pf) { }

static inline int
--
2.41.0

#2 lore raw
Quoted 16 lines · Aleksandr
Quoted 16 lines

-----Original Message-----
From: Intel-wired-lan <[email protected]> On Behalf Of Marcin Szycik
Sent: Tuesday, April 2, 2024 6:52 PM
To: [email protected]
Cc: Drewek, Wojciech <[email protected]>; Wang, Liang-min <[email protected]>; [email protected]; Chmielewski, Pawel <[email protected]>; Marcin Szycik <[email protected]>; Nguyen, Anthony L <[email protected]>; [email protected]; Kitszel, Przemyslaw <[email protected]>
Subject: [Intel-wired-lan] [PATCH iwl-next v4] ice: Reset VF on Tx MDD event

Quoted 4 lines

Please state in the title explicitly the purpose of the patch:
Do you fix a bug? say fix
Do you add functionality? say add
Do you refactor? say refactor

IMO it's neither of those.

It's not a bugfix because the driver itself doesn't do anything wrong; the patch provides the user with a kind of optional workaround for an userspace app that's behaving incorrectly.

It's not really new functionality, since the mdd-auto-reset-vf flag was already implemented and auto reset was already happening for Rx events. I'm just extending this to Tx as well.

And it's not strictly a refactor because I'm changing behaviour.

Please also note that I've picked up a patch not originally made by me [2], and kept the original title and most of the description. I'm open to suggestions to title though.

Quoted 11 lines
Quoted 11 lines

In cases when VF sends malformed packets that are classified as malicious, sometimes it causes Tx queue to freeze. This frozen queue can be stuck for several minutes being unusable. This behavior can be reproduced with a faulty userspace app running on VF.

When any Malicious Driver Detection event occurs and the mdd-auto- reset-vf private flag is set, perform a graceful VF reset to quickly bring VF back to operational state. Add a log message to notify about the cause of the reset. Add a helper for this to be reused for both TX and RX events.

Quoted 6 lines

Please describe your changes explicitly:
Do you just add a helper function without adding a new functionality?
Do you add functionality?
Do you fix?
Please explain what are the changes in driver behavior and the changes in the driver sources you make. Thank you

I'll try to make the description more explicit.

Thanks,
Marcin

Quoted 225 lines
Quoted 224 lines

---
v4: Only perform auto-reset once per VF
v3 [1]: Only auto reset VF if the mdd-auto-reset-vf flag is set v2 [2]: Revert an unneeded formatting change, fix commit message, fix a log
message with a correct event name

[1] https://lore.kernel.org/intel-wired-lan/20240326164455.735739-
[email protected]
[2] https://lore.kernel.org/netdev/20231102155149.2574209-1-
[email protected]

---
 drivers/net/ethernet/intel/ice/ice_main.c  | 57 +++++++++++++++++-
----  drivers/net/ethernet/intel/ice/ice_sriov.c | 25 +++++++---
drivers/net/ethernet/intel/ice/ice_sriov.h |  2 +
 3 files changed, 67 insertions(+), 17 deletions(-)
drivers/net/ethernet/intel/ice/ice_main.c +20 -0
index 185c9b13efcf..80bc83f6e1ab 100644
@@ -1745,6 +1745,39 @@ static void ice_service_timer(struct
 timer_list *t)
 	ice_service_task_schedule(pf);
 }
 
+/**
+ * ice_mdd_maybe_reset_vf - reset VF after MDD event
+ * @pf: pointer to the PF structure
+ * @vf: pointer to the VF structure
+ * @reset_vf_tx: whether Tx MDD has occurred
+ * @reset_vf_rx: whether Rx MDD has occurred
+ *
+ * Since the queue can get stuck on VF MDD events, the PF can be
+configured to
+ * automatically reset the VF by enabling the private ethtool flag
+ * mdd-auto-reset-vf.
+ */
+static void ice_mdd_maybe_reset_vf(struct ice_pf *pf, struct
 ice_vf *vf,
+				   bool reset_vf_tx, bool reset_vf_rx) {
+	struct device *dev = ice_pf_to_dev(pf);
+
+	if (!test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags))
+		return;
+
+	/* VF MDD event counters will be cleared by reset, so print

the event

drivers/net/ethernet/intel/ice/ice_main.c +9 -0
+	 * prior to reset.
+	 */
+	if (reset_vf_tx)
+		ice_print_vf_tx_mdd_event(vf);
+
+	if (reset_vf_rx)
+		ice_print_vf_rx_mdd_event(vf);
+
+	dev_info(dev, "PF-to-VF reset on PF %d VF %d due to MDD

event\n",

drivers/net/ethernet/intel/ice/ice_main.c +5 -0
+		 pf->hw.pf_id, vf->vf_id);
+	ice_reset_vf(vf, ICE_VF_RESET_NOTIFY | ICE_VF_RESET_LOCK); }
+
 /**
  * ice_handle_mdd_event - handle malicious driver detect event
  * @pf: pointer to the PF structure
@@ -1838,6 +1871,8 @@ static void ice_handle_mdd_event(struct
 ice_pf *pf)
 	 */
 	mutex_lock(&pf->vfs.table_lock);
 	ice_for_each_vf(pf, bkt, vf) {
+		bool reset_vf_tx = false, reset_vf_rx = false;
+
 		reg = rd32(hw, VP_MDET_TX_PQM(vf->vf_id));
 		if (reg & VP_MDET_TX_PQM_VALID_M) {

wr32(hw, VP_MDET_TX_PQM(vf->vf_id), 0xFFFF); @@ -
1846,6 +1881,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)

drivers/net/ethernet/intel/ice/ice_main.c +0 -0
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_PQM detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_TX_TCLAN(vf->vf_id)); @@ -1856,6
+1893,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_TCLAN detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_TX_TDPU(vf->vf_id)); @@ -1866,6
+1905,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_TDPU detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_RX(vf->vf_id)); @@ -1877,18
+1918,12 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 				dev_info(dev, "Malicious Driver Detection

event RX detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +0 -1
 					 vf->vf_id);
 
-			/* Since the queue is disabled on VF Rx MDD

events, the

drivers/net/ethernet/intel/ice/ice_main.c +0 -1
-			 * PF can be configured to reset the VF through

ethtool

drivers/net/ethernet/intel/ice/ice_main.c +0 -3
-			 * private flag mdd-auto-reset-vf.
-			 */
-			if (test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf-
Quoted 1 line

flags)) {

drivers/net/ethernet/intel/ice/ice_main.c +0 -2
-				/* VF MDD event counters will be cleared by
-				 * reset, so print the event prior to

reset.

drivers/net/ethernet/intel/ice/ice_main.c +5 -4
-				 */
-				ice_print_vf_rx_mdd_event(vf);
-				ice_reset_vf(vf, ICE_VF_RESET_LOCK);
-			}
+			reset_vf_rx = true;
 		}
+
+		if (reset_vf_tx || reset_vf_rx)
+			ice_mdd_maybe_reset_vf(pf, vf, reset_vf_tx,
+					       reset_vf_rx);
 	}
 	mutex_unlock(&pf->vfs.table_lock);
drivers/net/ethernet/intel/ice/ice_sriov.c +18 -0
index fb2e96db647e..a60dacf8942a 100644
@@ -1861,6 +1861,24 @@ void ice_print_vf_rx_mdd_event(struct ice_vf
 *vf)
 			  ? "on" : "off");
 }
 
+/**
+ * ice_print_vf_tx_mdd_event - print VF Tx malicious driver detect
+event
+ * @vf: pointer to the VF structure
+ */
+void ice_print_vf_tx_mdd_event(struct ice_vf *vf) {
+	struct ice_pf *pf = vf->pf;
+	struct device *dev;
+
+	dev = ice_pf_to_dev(pf);
+
+	dev_info(dev, "%d Tx Malicious Driver Detection events
 detected on PF %d VF %d MAC %pM. mdd-auto-reset-vfs=%s\n",
+		 vf->mdd_tx_events.count, pf->hw.pf_id, vf->vf_id,
+		 vf->dev_lan_addr,
+		 test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags)
+			  ? "on" : "off");
+}
+
 /**

* ice_print_vfs_mdd_events - print VFs malicious driver detect
event
* @pf: pointer to the PF structure

drivers/net/ethernet/intel/ice/ice_sriov.c +1 -6
@@ -1869,8 +1887,6 @@ void ice_print_vf_rx_mdd_event(struct ice_vf
 *vf)
  */
 void ice_print_vfs_mdd_events(struct ice_pf *pf)  {
-	struct device *dev = ice_pf_to_dev(pf);
-	struct ice_hw *hw = &pf->hw;
 	struct ice_vf *vf;
 	unsigned int bkt;
 
@@ -1897,10 +1913,7 @@ void ice_print_vfs_mdd_events(struct ice_pf
 *pf)
 		if (vf->mdd_tx_events.count != vf-
 > mdd_tx_events.last_printed) {
 			vf->mdd_tx_events.last_printed =
 							vf->mdd_tx_events.count;
-
-			dev_info(dev, "%d Tx Malicious Driver Detection
 events detected on PF %d VF %d MAC %pM.\n",
-				 vf->mdd_tx_events.count, hw->pf_id, vf-
 > vf_id,
-				 vf->dev_lan_addr);
+			ice_print_vf_tx_mdd_event(vf);
 		}
 	}
 	mutex_unlock(&pf->vfs.table_lock);
drivers/net/ethernet/intel/ice/ice_sriov.h +0 -0
index 4ba8fb53aea1..8f22313474d6 100644
@@ -58,6 +58,7 @@ void
 ice_vf_lan_overflow_event(struct ice_pf *pf, struct

ice_rq_event_info *event); void ice_print_vfs_mdd_events(struct ice_pf *pf); void ice_print_vf_rx_mdd_event(struct ice_vf *vf); +void ice_print_vf_tx_mdd_event(struct ice_vf *vf);
bool
ice_vc_validate_pattern(struct ice_vf *vf, struct
virtchnl_proto_hdrs *proto);
u32 ice_sriov_get_vf_total_msix(struct pci_dev *pdev); @@ -69,6
+70,7 @@ static inline void ice_vf_lan_overflow_event(struct ice_pf *pf, struct ice_rq_event_info *event) { } static inline void ice_print_vfs_mdd_events(struct ice_pf *pf) { } static inline void ice_print_vf_rx_mdd_event(struct ice_vf *vf) { } +static inline void ice_print_vf_tx_mdd_event(struct ice_vf *vf) { }
static inline void ice_restore_all_vfs_msi_state(struct ice_pf
*pf) { }

static inline int
--
2.41.0

#3 lore raw
Quoted 45 lines

-----Original Message-----
From: Marcin Szycik <[email protected]>
Sent: Wednesday, April 3, 2024 11:21 AM
To: Loktionov, Aleksandr <[email protected]>; intel- [email protected]
Cc: Drewek, Wojciech <[email protected]>; Wang, Liang-min <[email protected]>; [email protected]; Chmielewski, Pawel <[email protected]>; Nguyen, Anthony L <[email protected]>; [email protected]; Kitszel, Przemyslaw <[email protected]>
Subject: Re: [Intel-wired-lan] [PATCH iwl-next v4] ice: Reset VF on Tx MDD event

Quoted 4 lines · Aleksandr
Quoted 4 lines

-----Original Message-----
From: Intel-wired-lan <[email protected]> On

Behalf

Quoted 4 lines
Quoted 4 lines

Of Marcin Szycik
Sent: Tuesday, April 2, 2024 6:52 PM
To: [email protected]
Cc: Drewek, Wojciech <[email protected]>; Wang, Liang-

min

Quoted 4 lines
Quoted 4 lines

<[email protected]>; [email protected]; Chmielewski, Pawel <[email protected]>; Marcin Szycik <[email protected]>; Nguyen, Anthony L <[email protected]>; [email protected]; Kitszel,

Przemyslaw

Quoted 2 lines
Quoted 2 lines

<[email protected]>
Subject: [Intel-wired-lan] [PATCH iwl-next v4] ice: Reset VF on

Tx

Quoted 6 lines
Quoted 2 lines

MDD event

Please state in the title explicitly the purpose of the patch:
Do you fix a bug? say fix
Do you add functionality? say add
Do you refactor? say refactor

IMO it's neither of those.

Quoted 7 lines

It's not a bugfix because the driver itself doesn't do anything wrong; the patch provides the user with a kind of optional workaround for an userspace app that's behaving incorrectly.

It's not really new functionality, since the mdd-auto-reset-vf flag was already implemented and auto reset was already happening for Rx events. I'm just extending this to Tx as well.

For me it looks like adding new feature, but I'm not sure yet.

Quoted 2 lines

And it's not strictly a refactor because I'm changing behaviour.

Please describe explicitly what behavior do you change, and the motivation, especially if it's not a bugfix.

Quoted 22 lines

Please also note that I've picked up a patch not originally made by me [2], and kept the original title and most of the description. I'm open to suggestions to title though.

Quoted 2 lines
Quoted 2 lines

In cases when VF sends malformed packets that are classified as malicious, sometimes it causes Tx queue to freeze. This frozen

queue

Quoted 1 line
Quoted 1 line

can be stuck for several minutes being unusable. This behavior

can be

Quoted 3 lines
Quoted 3 lines

reproduced with a faulty userspace app running on VF.

When any Malicious Driver Detection event occurs and the mdd-

auto-

Quoted 1 line
Quoted 1 line

reset-vf private flag is set, perform a graceful VF reset to

quickly

Quoted 1 line
Quoted 1 line

bring VF back to operational state. Add a log message to notify

about

Quoted 1 line
Quoted 1 line

the cause of the reset. Add a helper for this to be reused for

both

Quoted 3 lines
Quoted 1 line

TX and RX events.

Please describe your changes explicitly:
Do you just add a helper function without adding a new

Quoted 8 lines

functionality?

Quoted 3 lines

Do you add functionality?
Do you fix?
Please explain what are the changes in driver behavior and the

changes in the driver sources you make.

Quoted 1 line

Thank you

I'll try to make the description more explicit.

Thank you

Quoted 246 lines

Thanks,
Marcin

Quoted 9 lines
Quoted 9 lines

---
v4: Only perform auto-reset once per VF
v3 [1]: Only auto reset VF if the mdd-auto-reset-vf flag is set v2 [2]: Revert an unneeded formatting change, fix commit

message, fix

Quoted 4 lines
Quoted 4 lines

a log
message with a correct event name

[1] https://lore.kernel.org/intel-wired-

lan/20240326164455.735739-

Quoted 5 lines
Quoted 5 lines

[email protected]
[2] https://lore.kernel.org/netdev/20231102155149.2574209-1-
[email protected]

---
 drivers/net/ethernet/intel/ice/ice_main.c  | 57

+++++++++++++++++-

Quoted 14 lines
Quoted 14 lines

---- drivers/net/ethernet/intel/ice/ice_sriov.c | 25 +++++++--- drivers/net/ethernet/intel/ice/ice_sriov.h | 2 +

 3 files changed, 67 insertions(+), 17 deletions(-)
drivers/net/ethernet/intel/ice/ice_main.c +0 -0
index 185c9b13efcf..80bc83f6e1ab 100644
@@ -1745,6 +1745,39 @@ static void ice_service_timer(struct
 timer_list *t)
 	ice_service_task_schedule(pf);
 }
 
Quoted 8 lines

drivers/net/ethernet/intel/ice/ice_main.c static void ice_service_timer(struct timer_list *t)

Quoted 8 lines
drivers/net/ethernet/intel/ice/ice_main.c +8 -0
+/**
+ * ice_mdd_maybe_reset_vf - reset VF after MDD event
+ * @pf: pointer to the PF structure
+ * @vf: pointer to the VF structure
+ * @reset_vf_tx: whether Tx MDD has occurred
+ * @reset_vf_rx: whether Rx MDD has occurred
+ *
+ * Since the queue can get stuck on VF MDD events, the PF can

be

Quoted 2 lines

drivers/net/ethernet/intel/ice/ice_main.c static void ice_service_timer(struct · line 1758

Quoted 2 lines
drivers/net/ethernet/intel/ice/ice_main.c +2 -0
+configured to
+ * automatically reset the VF by enabling the private ethtool

flag

Quoted 21 lines
Quoted 21 lines
drivers/net/ethernet/intel/ice/ice_main.c +10 -0
+ * mdd-auto-reset-vf.
+ */
+static void ice_mdd_maybe_reset_vf(struct ice_pf *pf, struct
 ice_vf *vf,
+				   bool reset_vf_tx, bool reset_vf_rx) {
+	struct device *dev = ice_pf_to_dev(pf);
+
+	if (!test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags))
+		return;
+
+	/* VF MDD event counters will be cleared by reset, so print

the event

drivers/net/ethernet/intel/ice/ice_main.c +9 -0
+	 * prior to reset.
+	 */
+	if (reset_vf_tx)
+		ice_print_vf_tx_mdd_event(vf);
+
+	if (reset_vf_rx)
+		ice_print_vf_rx_mdd_event(vf);
+
+	dev_info(dev, "PF-to-VF reset on PF %d VF %d due to MDD
Quoted 8 lines

drivers/net/ethernet/intel/ice/ice_main.c static void ice_handle_mdd_event(struct

Quoted 8 lines

event\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
+		 pf->hw.pf_id, vf->vf_id);
+	ice_reset_vf(vf, ICE_VF_RESET_NOTIFY | ICE_VF_RESET_LOCK); }
+
 /**
  * ice_handle_mdd_event - handle malicious driver detect event
  * @pf: pointer to the PF structure
@@ -1838,6 +1871,8 @@ static void ice_handle_mdd_event(struct

ice_pf

Quoted 5 lines
Quoted 5 lines
drivers/net/ethernet/intel/ice/ice_main.c +1 -0
 *pf)
 	 */
 	mutex_lock(&pf->vfs.table_lock);
 	ice_for_each_vf(pf, bkt, vf) {
+		bool reset_vf_tx = false, reset_vf_rx = false;
Quoted 8 lines

drivers/net/ethernet/intel/ice/ice_main.c

Quoted 8 lines
drivers/net/ethernet/intel/ice/ice_main.c +1 -0
+
 		reg = rd32(hw, VP_MDET_TX_PQM(vf->vf_id));
 		if (reg & VP_MDET_TX_PQM_VALID_M) {

wr32(hw, VP_MDET_TX_PQM(vf->vf_id), 0xFFFF); @@ -
1846,6 +1881,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)

drivers/net/ethernet/intel/ice/ice_main.c +0 -0
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_PQM detected

Quoted 10 lines

drivers/net/ethernet/intel/ice/ice_main.c static void ice_handle_mdd_event(struct ice_pf *pf)

Quoted 10 lines

on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_TX_TCLAN(vf->vf_id)); @@ -1856,6
+1893,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_TCLAN

Quoted 10 lines

drivers/net/ethernet/intel/ice/ice_main.c static void ice_handle_mdd_event(struct ice_pf *pf)

Quoted 10 lines

detected on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_TX_TDPU(vf->vf_id)); @@ -1866,6
+1905,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 			if (netif_msg_tx_err(pf))
 				dev_info(dev, "Malicious Driver Detection

event TX_TDPU detected

Quoted 9 lines

drivers/net/ethernet/intel/ice/ice_main.c static void ice_handle_mdd_event(struct ice_pf *pf)

Quoted 9 lines

on VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +3 -0
 					 vf->vf_id);
+
+			reset_vf_tx = true;
 		}
 
 		reg = rd32(hw, VP_MDET_RX(vf->vf_id)); @@ -1877,18
+1918,12 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
 				dev_info(dev, "Malicious Driver Detection

event RX detected on

Quoted 27 lines
Quoted 27 lines

VF %d\n",

drivers/net/ethernet/intel/ice/ice_main.c +0 -1
 					 vf->vf_id);
 
-			/* Since the queue is disabled on VF Rx MDD

events, the

drivers/net/ethernet/intel/ice/ice_main.c +0 -1
-			 * PF can be configured to reset the VF through

ethtool

drivers/net/ethernet/intel/ice/ice_main.c +0 -3
-			 * private flag mdd-auto-reset-vf.
-			 */
-			if (test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf-
Quoted 1 line

flags)) {

drivers/net/ethernet/intel/ice/ice_main.c +0 -2
-				/* VF MDD event counters will be cleared by
-				 * reset, so print the event prior to

reset.

drivers/net/ethernet/intel/ice/ice_main.c +5 -4
-				 */
-				ice_print_vf_rx_mdd_event(vf);
-				ice_reset_vf(vf, ICE_VF_RESET_LOCK);
-			}
+			reset_vf_rx = true;
 		}
+
+		if (reset_vf_tx || reset_vf_rx)
+			ice_mdd_maybe_reset_vf(pf, vf, reset_vf_tx,
+					       reset_vf_rx);
 	}
 	mutex_unlock(&pf->vfs.table_lock);
Quoted 6 lines

drivers/net/ethernet/intel/ice/ice_sriov.c void ice_print_vf_rx_mdd_event(struct

Quoted 6 lines
drivers/net/ethernet/intel/ice/ice_sriov.c +0 -0
index fb2e96db647e..a60dacf8942a 100644
@@ -1861,6 +1861,24 @@ void ice_print_vf_rx_mdd_event(struct

ice_vf

Quoted 6 lines

drivers/net/ethernet/intel/ice/ice_sriov.c void ice_print_vf_rx_mdd_event(struct ice_vf *vf)

Quoted 6 lines
drivers/net/ethernet/intel/ice/ice_sriov.c +2 -0
 *vf)
 			  ? "on" : "off");
 }
 
+/**
+ * ice_print_vf_tx_mdd_event - print VF Tx malicious driver

detect

Quoted 14 lines
Quoted 14 lines
drivers/net/ethernet/intel/ice/ice_sriov.c +13 -0
+event
+ * @vf: pointer to the VF structure
+ */
+void ice_print_vf_tx_mdd_event(struct ice_vf *vf) {
+	struct ice_pf *pf = vf->pf;
+	struct device *dev;
+
+	dev = ice_pf_to_dev(pf);
+
+	dev_info(dev, "%d Tx Malicious Driver Detection events
 detected on PF %d VF %d MAC %pM. mdd-auto-reset-vfs=%s\n",
+		 vf->mdd_tx_events.count, pf->hw.pf_id, vf->vf_id,
+		 vf->dev_lan_addr,
+		 test_bit(ICE_FLAG_MDD_AUTO_RESET_VF, pf->flags)
Quoted 8 lines

drivers/net/ethernet/intel/ice/ice_sriov.c void ice_print_vf_rx_mdd_event(struct

Quoted 8 lines
drivers/net/ethernet/intel/ice/ice_sriov.c +3 -0
+			  ? "on" : "off");
+}
+
 /**

* ice_print_vfs_mdd_events - print VFs malicious driver detect
event
* @pf: pointer to the PF structure

drivers/net/ethernet/intel/ice/ice_sriov.c +0 -0
@@ -1869,8 +1887,6 @@ void ice_print_vf_rx_mdd_event(struct

ice_vf

Quoted 9 lines

drivers/net/ethernet/intel/ice/ice_sriov.c void ice_print_vfs_mdd_events(struct

Quoted 9 lines
drivers/net/ethernet/intel/ice/ice_sriov.c +0 -2
 *vf)
  */
 void ice_print_vfs_mdd_events(struct ice_pf *pf)  {
-	struct device *dev = ice_pf_to_dev(pf);
-	struct ice_hw *hw = &pf->hw;
 	struct ice_vf *vf;
 	unsigned int bkt;
 
@@ -1897,10 +1913,7 @@ void ice_print_vfs_mdd_events(struct

ice_pf

Quoted 23 lines
Quoted 23 lines
drivers/net/ethernet/intel/ice/ice_sriov.c +1 -4
 *pf)
 		if (vf->mdd_tx_events.count != vf-
 > mdd_tx_events.last_printed) {
 			vf->mdd_tx_events.last_printed =
 							vf->mdd_tx_events.count;
-
-			dev_info(dev, "%d Tx Malicious Driver Detection
 events detected on PF %d VF %d MAC %pM.\n",
-				 vf->mdd_tx_events.count, hw->pf_id, vf-
 > vf_id,
-				 vf->dev_lan_addr);
+			ice_print_vf_tx_mdd_event(vf);
 		}
 	}
 	mutex_unlock(&pf->vfs.table_lock);
drivers/net/ethernet/intel/ice/ice_sriov.h +0 -0
index 4ba8fb53aea1..8f22313474d6 100644
@@ -58,6 +58,7 @@ void
 ice_vf_lan_overflow_event(struct ice_pf *pf, struct

ice_rq_event_info *event); void ice_print_vfs_mdd_events(struct

Quoted 8 lines
Quoted 8 lines

ice_pf *pf); void ice_print_vf_rx_mdd_event(struct ice_vf *vf); +void ice_print_vf_tx_mdd_event(struct ice_vf *vf);
bool
ice_vc_validate_pattern(struct ice_vf *vf, struct
virtchnl_proto_hdrs *proto);
u32 ice_sriov_get_vf_total_msix(struct pci_dev *pdev); @@ -69,6
+70,7 @@ static inline void ice_vf_lan_overflow_event(struct ice_pf *pf, struct ice_rq_event_info *event) { } static inline

void

Quoted 1 line
Quoted 1 line

ice_print_vfs_mdd_events(struct ice_pf *pf) { } static inline

void

Quoted 2 lines
Quoted 2 lines

ice_print_vf_rx_mdd_event(struct ice_vf *vf) { } +static inline void ice_print_vf_tx_mdd_event(struct ice_vf *vf)

{

Quoted 8 lines
Quoted 7 lines

}
static inline void ice_restore_all_vfs_msi_state(struct ice_pf
*pf) { }

static inline int
--
2.41.0