bnxt_en: improve fw diagnose devlink health messages

Add firmware event counters as well as health state severity. In
the unhealthy state, recommend a remedy and inform the user as to
its impact.

Readability of the devlink tool's output is negatively impacted by
adding these fields to the diagnosis. The single line of text, as
rendered by devlink health diagnose, benefits from more terse
descriptions, which can be substituted without loss of clarity, even
in pretty printed JSON mode.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 16 insertions, 3 deletions

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 1251d78ffd46..b4d9374548f8 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -2138,10 +2138,12 @@ static int bnxt_async_event_process(struct bnxt *bp,
 			set_bit(BNXT_STATE_FW_ACTIVATE_RESET, &bp->state);
 		} else if (EVENT_DATA1_RESET_NOTIFY_FATAL(data1)) {
 			type_str = "Fatal";
+			bp->fw_health->fatalities++;
 			set_bit(BNXT_STATE_FW_FATAL_COND, &bp->state);
 		} else if (data2 && BNXT_FW_STATUS_HEALTHY !=
 			   EVENT_DATA2_RESET_NOTIFY_FW_STATUS_CODE(data2)) {
 			type_str = "Non-fatal";
+			bp->fw_health->survivals++;
 			set_bit(BNXT_STATE_FW_NON_FATAL_COND, &bp->state);
 		}
 		netif_warn(bp, hw, bp->dev,
@@ -7604,6 +7606,7 @@ static int __bnxt_alloc_fw_health(struct bnxt *bp)
 	if (!bp->fw_health)
 		return -ENOMEM;
 
+	mutex_init(&bp->fw_health->lock);
 	return 0;
 }
 
@@ -7650,12 +7653,16 @@ static void bnxt_inv_fw_health_reg(struct bnxt *bp)
 	struct bnxt_fw_health *fw_health = bp->fw_health;
 	u32 reg_type;
 
-	if (!fw_health || !fw_health->status_reliable)
+	if (!fw_health)
 		return;
 
 	reg_type = BNXT_FW_HEALTH_REG_TYPE(fw_health->regs[BNXT_FW_HEALTH_REG]);
 	if (reg_type == BNXT_FW_HEALTH_REG_TYPE_GRC)
 		fw_health->status_reliable = false;
+
+	reg_type = BNXT_FW_HEALTH_REG_TYPE(fw_health->regs[BNXT_FW_RESET_CNT_REG]);
+	if (reg_type == BNXT_FW_HEALTH_REG_TYPE_GRC)
+		fw_health->resets_reliable = false;
 }
 
 static void bnxt_try_map_fw_health_reg(struct bnxt *bp)
@@ -7712,6 +7719,7 @@ static int bnxt_map_fw_health_regs(struct bnxt *bp)
 	int i;
 
 	bp->fw_health->status_reliable = false;
+	bp->fw_health->resets_reliable = false;
 	/* Only pre-map the monitoring GRC registers using window 3 */
 	for (i = 0; i < 4; i++) {
 		u32 reg = fw_health->regs[i];
@@ -7725,6 +7733,7 @@ static int bnxt_map_fw_health_regs(struct bnxt *bp)
 		fw_health->mapped_regs[i] = BNXT_FW_HEALTH_WIN_OFF(reg);
 	}
 	bp->fw_health->status_reliable = true;
+	bp->fw_health->resets_reliable = true;
 	if (reg_base == 0xffffffff)
 		return 0;
 
@@ -11264,14 +11273,18 @@ static void bnxt_fw_health_check(struct bnxt *bp)
 	}
 
 	val = bnxt_fw_health_readl(bp, BNXT_FW_HEARTBEAT_REG);
-	if (val == fw_health->last_fw_heartbeat)
+	if (val == fw_health->last_fw_heartbeat) {
+		fw_health->arrests++;
 		goto fw_reset;
+	}
 
 	fw_health->last_fw_heartbeat = val;
 
 	val = bnxt_fw_health_readl(bp, BNXT_FW_RESET_CNT_REG);
-	if (val != fw_health->last_fw_reset_cnt)
+	if (val != fw_health->last_fw_reset_cnt) {
+		fw_health->discoveries++;
 		goto fw_reset;
+	}
 
 	fw_health->tmr_counter = fw_health->tmr_multiplier;
 	return;