chelsio: cxgb: Disable the card on error in threaded interrupt

t1_fatal_err() is invoked from the interrupt handler. The bad part is
that it invokes (via t1_sge_stop()) del_timer_sync() and tasklet_kill().
Both functions must not be called from an interrupt because it is
possible that it will wait for the completion of the timer/tasklet it
just interrupted.

In case of a fatal error, use t1_interrupts_disable() to disable all
interrupt sources and then wake the interrupt thread with
F_PL_INTR_SGE_ERR as pending flag. The threaded-interrupt will stop the
card via t1_sge_stop() and not re-enable the interrupts again.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

5 files changed, 44 insertions, 27 deletions

diff --git a/drivers/net/ethernet/chelsio/cxgb/common.h b/drivers/net/ethernet/chelsio/cxgb/common.h
index e999a9b9fe6c..0321be77366c 100644
--- a/drivers/net/ethernet/chelsio/cxgb/common.h
+++ b/drivers/net/ethernet/chelsio/cxgb/common.h
@@ -346,7 +346,6 @@ int t1_get_board_rev(adapter_t *adapter, const struct board_info *bi,
 int t1_init_hw_modules(adapter_t *adapter);
 int t1_init_sw_modules(adapter_t *adapter, const struct board_info *bi);
 void t1_free_sw_modules(adapter_t *adapter);
-void t1_fatal_err(adapter_t *adapter);
 void t1_link_changed(adapter_t *adapter, int port_id);
 void t1_link_negotiated(adapter_t *adapter, int port_id, int link_stat,
 			    int speed, int duplex, int pause);
diff --git a/drivers/net/ethernet/chelsio/cxgb/cxgb2.c b/drivers/net/ethernet/chelsio/cxgb/cxgb2.c
index bd6f3c532d72..512da98019c6 100644
--- a/drivers/net/ethernet/chelsio/cxgb/cxgb2.c
+++ b/drivers/net/ethernet/chelsio/cxgb/cxgb2.c
@@ -917,16 +917,6 @@ static void mac_stats_task(struct work_struct *work)
 	spin_unlock(&adapter->work_lock);
 }
 
-void t1_fatal_err(struct adapter *adapter)
-{
-	if (adapter->flags & FULL_INIT_DONE) {
-		t1_sge_stop(adapter->sge);
-		t1_interrupts_disable(adapter);
-	}
-	pr_alert("%s: encountered fatal error, operation suspended\n",
-		 adapter->name);
-}
-
 static const struct net_device_ops cxgb_netdev_ops = {
 	.ndo_open		= cxgb_open,
 	.ndo_stop		= cxgb_close,
diff --git a/drivers/net/ethernet/chelsio/cxgb/sge.c b/drivers/net/ethernet/chelsio/cxgb/sge.c
index 5aef9ae1ecfe..cda01f22c71c 100644
--- a/drivers/net/ethernet/chelsio/cxgb/sge.c
+++ b/drivers/net/ethernet/chelsio/cxgb/sge.c
@@ -940,10 +940,11 @@ void t1_sge_intr_clear(struct sge *sge)
 /*
  * SGE 'Error' interrupt handler
  */
-int t1_sge_intr_error_handler(struct sge *sge)
+bool t1_sge_intr_error_handler(struct sge *sge)
 {
 	struct adapter *adapter = sge->adapter;
 	u32 cause = readl(adapter->regs + A_SG_INT_CAUSE);
+	bool wake = false;
 
 	if (adapter->port[0].dev->hw_features & NETIF_F_TSO)
 		cause &= ~F_PACKET_TOO_BIG;
@@ -967,11 +968,14 @@ int t1_sge_intr_error_handler(struct sge *sge)
 		sge->stats.pkt_mismatch++;
 		pr_alert("%s: SGE packet mismatch\n", adapter->name);
 	}
-	if (cause & SGE_INT_FATAL)
-		t1_fatal_err(adapter);
+	if (cause & SGE_INT_FATAL) {
+		t1_interrupts_disable(adapter);
+		adapter->pending_thread_intr |= F_PL_INTR_SGE_ERR;
+		wake = true;
+	}
 
 	writel(cause, adapter->regs + A_SG_INT_CAUSE);
-	return 0;
+	return wake;
 }
 
 const struct sge_intr_counts *t1_sge_get_intr_counts(const struct sge *sge)
@@ -1635,6 +1639,14 @@ irqreturn_t t1_interrupt_thread(int irq, void *data)
 	if (pending_thread_intr & F_PL_INTR_EXT)
 		t1_elmer0_ext_intr_handler(adapter);
 
+	/* This error is fatal, interrupts remain off */
+	if (pending_thread_intr & F_PL_INTR_SGE_ERR) {
+		pr_alert("%s: encountered fatal error, operation suspended\n",
+			 adapter->name);
+		t1_sge_stop(adapter->sge);
+		return IRQ_HANDLED;
+	}
+
 	spin_lock_irq(&adapter->async_lock);
 	adapter->slow_intr_mask |= F_PL_INTR_EXT;
 
diff --git a/drivers/net/ethernet/chelsio/cxgb/sge.h b/drivers/net/ethernet/chelsio/cxgb/sge.h
index 76516d2a8aa9..716705b96f26 100644
--- a/drivers/net/ethernet/chelsio/cxgb/sge.h
+++ b/drivers/net/ethernet/chelsio/cxgb/sge.h
@@ -82,7 +82,7 @@ netdev_tx_t t1_start_xmit(struct sk_buff *skb, struct net_device *dev);
 void t1_vlan_mode(struct adapter *adapter, netdev_features_t features);
 void t1_sge_start(struct sge *);
 void t1_sge_stop(struct sge *);
-int t1_sge_intr_error_handler(struct sge *);
+bool t1_sge_intr_error_handler(struct sge *sge);
 void t1_sge_intr_enable(struct sge *);
 void t1_sge_intr_disable(struct sge *);
 void t1_sge_intr_clear(struct sge *);
diff --git a/drivers/net/ethernet/chelsio/cxgb/subr.c b/drivers/net/ethernet/chelsio/cxgb/subr.c
index d90ad07ff1a4..310add28fcf5 100644
--- a/drivers/net/ethernet/chelsio/cxgb/subr.c
+++ b/drivers/net/ethernet/chelsio/cxgb/subr.c
@@ -170,7 +170,7 @@ void t1_link_changed(adapter_t *adapter, int port_id)
 	t1_link_negotiated(adapter, port_id, link_ok, speed, duplex, fc);
 }
 
-static int t1_pci_intr_handler(adapter_t *adapter)
+static bool t1_pci_intr_handler(adapter_t *adapter)
 {
 	u32 pcix_cause;
 
@@ -179,9 +179,13 @@ static int t1_pci_intr_handler(adapter_t *adapter)
 	if (pcix_cause) {
 		pci_write_config_dword(adapter->pdev, A_PCICFG_INTR_CAUSE,
 				       pcix_cause);
-		t1_fatal_err(adapter);    /* PCI errors are fatal */
+		/* PCI errors are fatal */
+		t1_interrupts_disable(adapter);
+		adapter->pending_thread_intr |= F_PL_INTR_SGE_ERR;
+		pr_alert("%s: PCI error encountered.\n", adapter->name);
+		return true;
 	}
-	return 0;
+	return false;
 }
 
 #ifdef CONFIG_CHELSIO_T1_1G
@@ -213,10 +217,13 @@ static int fpga_phy_intr_handler(adapter_t *adapter)
 static irqreturn_t fpga_slow_intr(adapter_t *adapter)
 {
 	u32 cause = readl(adapter->regs + A_PL_CAUSE);
+	irqreturn_t ret = IRQ_NONE;
 
 	cause &= ~F_PL_INTR_SGE_DATA;
-	if (cause & F_PL_INTR_SGE_ERR)
-		t1_sge_intr_error_handler(adapter->sge);
+	if (cause & F_PL_INTR_SGE_ERR) {
+		if (t1_sge_intr_error_handler(adapter->sge))
+			ret = IRQ_WAKE_THREAD;
+	}
 
 	if (cause & FPGA_PCIX_INTERRUPT_GMAC)
 		fpga_phy_intr_handler(adapter);
@@ -231,13 +238,18 @@ static irqreturn_t fpga_slow_intr(adapter_t *adapter)
 		/* Clear TP interrupt */
 		writel(tp_cause, adapter->regs + FPGA_TP_ADDR_INTERRUPT_CAUSE);
 	}
-	if (cause & FPGA_PCIX_INTERRUPT_PCIX)
-		t1_pci_intr_handler(adapter);
+	if (cause & FPGA_PCIX_INTERRUPT_PCIX) {
+		if (t1_pci_intr_handler(adapter))
+			ret = IRQ_WAKE_THREAD;
+	}
 
 	/* Clear the interrupts just processed. */
 	if (cause)
 		writel(cause, adapter->regs + A_PL_CAUSE);
 
+	if (ret != IRQ_NONE)
+		return ret;
+
 	return cause == 0 ? IRQ_NONE : IRQ_HANDLED;
 }
 #endif
@@ -850,14 +862,18 @@ static irqreturn_t asic_slow_intr(adapter_t *adapter)
 	cause &= adapter->slow_intr_mask;
 	if (!cause)
 		return IRQ_NONE;
-	if (cause & F_PL_INTR_SGE_ERR)
-		t1_sge_intr_error_handler(adapter->sge);
+	if (cause & F_PL_INTR_SGE_ERR) {
+		if (t1_sge_intr_error_handler(adapter->sge))
+			ret = IRQ_WAKE_THREAD;
+	}
 	if (cause & F_PL_INTR_TP)
 		t1_tp_intr_handler(adapter->tp);
 	if (cause & F_PL_INTR_ESPI)
 		t1_espi_intr_handler(adapter->espi);
-	if (cause & F_PL_INTR_PCIX)
-		t1_pci_intr_handler(adapter);
+	if (cause & F_PL_INTR_PCIX) {
+		if (t1_pci_intr_handler(adapter))
+			ret = IRQ_WAKE_THREAD;
+	}
 	if (cause & F_PL_INTR_EXT) {
 		/* Wake the threaded interrupt to handle external interrupts as
 		 * we require a process context. We disable EXT interrupts in