[Intel-wired-lan] [PATCH iwl-next v2 5/6] ice: Add MDD logging via devlink health

From: Mateusz Polchlopek
Date: Fri Jul 12 2024 - 05:46:43 EST


From: Ben Shelton <benjamin.h.shelton@xxxxxxxxx>

Add a devlink health reporter for MDD events. The 'dump' handler will
return the information captured in each call to ice_handle_mdd_event().
A device reset (CORER/PFR) will put the reporter back in healthy state.

Signed-off-by: Ben Shelton <benjamin.h.shelton@xxxxxxxxx>
Co-developed-by: Przemek Kitszel <przemyslaw.kitszel@xxxxxxxxx>
Signed-off-by: Przemek Kitszel <przemyslaw.kitszel@xxxxxxxxx>
Reviewed-by: Igor Bagnucki <igor.bagnucki@xxxxxxxxx>
Reviewed-by: Wojciech Drewek <wojciech.drewek@xxxxxxxxx>
Reviewed-by: Simon Horman <horms@xxxxxxxxxx>
Signed-off-by: Mateusz Polchlopek <mateusz.polchlopek@xxxxxxxxx>
---
.../intel/ice/devlink/devlink_health.c | 77 +++++++++++++++++++
.../intel/ice/devlink/devlink_health.h | 11 +++
drivers/net/ethernet/intel/ice/ice_main.c | 6 ++
3 files changed, 94 insertions(+)

diff --git a/drivers/net/ethernet/intel/ice/devlink/devlink_health.c b/drivers/net/ethernet/intel/ice/devlink/devlink_health.c
index 2fbf6f08d0f6..f9edfabc9be8 100644
--- a/drivers/net/ethernet/intel/ice/devlink/devlink_health.c
+++ b/drivers/net/ethernet/intel/ice/devlink/devlink_health.c
@@ -33,6 +33,79 @@ static void ice_devlink_health_report(struct devlink_health_reporter *reporter,
}
}

+struct ice_mdd_event {
+ enum ice_mdd_src src;
+ u16 vf_num;
+ u16 queue;
+ u8 pf_num;
+ u8 event;
+};
+
+static const char *ice_mdd_src_to_str(enum ice_mdd_src src)
+{
+ switch (src) {
+ case ICE_MDD_SRC_TX_PQM:
+ return "tx_pqm";
+ case ICE_MDD_SRC_TX_TCLAN:
+ return "tx_tclan";
+ case ICE_MDD_SRC_TX_TDPU:
+ return "tx_tdpu";
+ case ICE_MDD_SRC_RX:
+ return "rx";
+ default:
+ return "invalid";
+ }
+}
+
+static int
+ice_mdd_reporter_dump(struct devlink_health_reporter *reporter,
+ struct devlink_fmsg *fmsg, void *priv_ctx,
+ struct netlink_ext_ack *extack)
+{
+ struct ice_mdd_event *mdd_event = priv_ctx;
+ const char *src;
+
+ if (!mdd_event)
+ return 0;
+
+ src = ice_mdd_src_to_str(mdd_event->src);
+
+ devlink_fmsg_obj_nest_start(fmsg);
+ devlink_fmsg_put(fmsg, "src", src);
+ ICE_DEVLINK_FMSG_PUT_FIELD(fmsg, mdd_event, pf_num);
+ ICE_DEVLINK_FMSG_PUT_FIELD(fmsg, mdd_event, vf_num);
+ ICE_DEVLINK_FMSG_PUT_FIELD(fmsg, mdd_event, event);
+ ICE_DEVLINK_FMSG_PUT_FIELD(fmsg, mdd_event, queue);
+ devlink_fmsg_obj_nest_end(fmsg);
+
+ return 0;
+}
+
+/**
+ * ice_devlink_report_mdd_event - Report an MDD event through devlink health
+ * @pf: the PF device structure
+ * @src: the HW block that was the source of this MDD event
+ * @pf_num: the pf_num on which the MDD event occurred
+ * @vf_num: the vf_num on which the MDD event occurred
+ * @event: the event type of the MDD event
+ * @queue: the queue on which the MDD event occurred
+ *
+ * Report an MDD event that has occurred on this PF.
+ */
+void ice_devlink_report_mdd_event(struct ice_pf *pf, enum ice_mdd_src src,
+ u8 pf_num, u16 vf_num, u8 event, u16 queue)
+{
+ struct ice_mdd_event ev = {
+ .src = src,
+ .pf_num = pf_num,
+ .vf_num = vf_num,
+ .event = event,
+ .queue = queue,
+ };
+
+ ice_devlink_health_report(pf->health_reporters.mdd, "MDD event", &ev);
+}
+
static void ice_dump_ethtool_stats_to_fmsg(struct devlink_fmsg *fmsg,
struct net_device *netdev)
{
@@ -154,6 +227,7 @@ ice_init_devlink_rep(struct ice_pf *pf,
.dump = ice_ ## _name ## _reporter_dump, \
}

+ICE_DEFINE_HEALTH_REPORTER_OPS(mdd);
ICE_DEFINE_HEALTH_REPORTER_OPS(tx_hang);

/**
@@ -166,6 +240,7 @@ void ice_health_init(struct ice_pf *pf)
{
struct ice_health *reps = &pf->health_reporters;

+ reps->mdd = ice_init_devlink_rep(pf, &ice_mdd_reporter_ops);
reps->tx_hang = ice_init_devlink_rep(pf, &ice_tx_hang_reporter_ops);
}

@@ -187,6 +262,7 @@ static void ice_deinit_devl_reporter(struct devlink_health_reporter *reporter)
*/
void ice_health_deinit(struct ice_pf *pf)
{
+ ice_deinit_devl_reporter(pf->health_reporters.mdd);
ice_deinit_devl_reporter(pf->health_reporters.tx_hang);
}

@@ -206,5 +282,6 @@ void ice_health_assign_healthy_state(struct devlink_health_reporter *reporter)
*/
void ice_health_clear(struct ice_pf *pf)
{
+ ice_health_assign_healthy_state(pf->health_reporters.mdd);
ice_health_assign_healthy_state(pf->health_reporters.tx_hang);
}
diff --git a/drivers/net/ethernet/intel/ice/devlink/devlink_health.h b/drivers/net/ethernet/intel/ice/devlink/devlink_health.h
index 984b8f9f56d4..01abd3d8f31e 100644
--- a/drivers/net/ethernet/intel/ice/devlink/devlink_health.h
+++ b/drivers/net/ethernet/intel/ice/devlink/devlink_health.h
@@ -16,18 +16,29 @@
struct ice_pf;
struct ice_tx_ring;

+enum ice_mdd_src {
+ ICE_MDD_SRC_TX_PQM,
+ ICE_MDD_SRC_TX_TCLAN,
+ ICE_MDD_SRC_TX_TDPU,
+ ICE_MDD_SRC_RX,
+};
+
/**
* struct ice_health - stores ice devlink health reporters and accompanied data
* @tx_hang: devlink health reporter for tx_hang event
+ * @mdd: devlink health reporter for MDD detection event
*/
struct ice_health {
struct devlink_health_reporter *tx_hang;
+ struct devlink_health_reporter *mdd;
};

void ice_health_init(struct ice_pf *pf);
void ice_health_deinit(struct ice_pf *pf);
void ice_health_clear(struct ice_pf *pf);

+void ice_devlink_report_mdd_event(struct ice_pf *pf, enum ice_mdd_src src,
+ u8 pf_num, u16 vf_num, u8 event, u16 queue);
void ice_report_tx_hang(struct ice_pf *pf, struct ice_tx_ring *tx_ring,
u16 vsi_num, u32 head, u32 intr);

diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 1fae7d34f2e4..a7c186fb902b 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -1819,6 +1819,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
if (netif_msg_tx_err(pf))
dev_info(dev, "Malicious Driver Detection event %d on TX queue %d PF# %d VF# %d\n",
event, queue, pf_num, vf_num);
+ ice_devlink_report_mdd_event(pf, ICE_MDD_SRC_TX_PQM, pf_num,
+ vf_num, event, queue);
wr32(hw, GL_MDET_TX_PQM, 0xffffffff);
}

@@ -1832,6 +1834,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
if (netif_msg_tx_err(pf))
dev_info(dev, "Malicious Driver Detection event %d on TX queue %d PF# %d VF# %d\n",
event, queue, pf_num, vf_num);
+ ice_devlink_report_mdd_event(pf, ICE_MDD_SRC_TX_TCLAN, pf_num,
+ vf_num, event, queue);
wr32(hw, GL_MDET_TX_TCLAN_BY_MAC(hw), U32_MAX);
}

@@ -1845,6 +1849,8 @@ static void ice_handle_mdd_event(struct ice_pf *pf)
if (netif_msg_rx_err(pf))
dev_info(dev, "Malicious Driver Detection event %d on RX queue %d PF# %d VF# %d\n",
event, queue, pf_num, vf_num);
+ ice_devlink_report_mdd_event(pf, ICE_MDD_SRC_RX, pf_num,
+ vf_num, event, queue);
wr32(hw, GL_MDET_RX, 0xffffffff);
}

--
2.38.1