Re: [PATCH]: cpmac bugfixes and enhancements

[Matteo Croce <matteo@openwrt.org>]

Il Tuesday 06 May 2008 01:16:34 Andrew Morton ha scritto:
> On Sun, 4 May 2008 19:04:22 +0200
> Matteo Croce <matteo@openwrt.org> wrote:
> 
> > This patch fixes an IRQ storm, a locking issues, moves platform code in the right sections
> > and other small fixes.
> > 
> 
> Please feed this patch (and all future ones) through scripts/checkpatch.pl.
> It picks up rather a lot of simple problems which there is no reason for
> us to retain.
> 
> >
> > ...
> >
> > +	spin_unlock(&priv->rx_lock);
> > +	netif_rx_complete(priv->dev, napi);
> > +	netif_stop_queue(priv->dev);
> > +	napi_disable(&priv->napi);
> > +	
> > +	atomic_inc(&priv->reset_pending);
> > +	cpmac_hw_stop(priv->dev);
> > +	if (!schedule_work(&priv->reset_work))
> > +		atomic_dec(&priv->reset_pending);
> > +	return 0;
> > + 
> >  }
> >  
> >  static int cpmac_start_xmit(struct sk_buff *skb, struct net_device *dev)
> > @@ -456,6 +549,9 @@ static int cpmac_start_xmit(struct sk_buff *skb, struct net_device *dev)
> >  	struct cpmac_desc *desc;
> >  	struct cpmac_priv *priv = netdev_priv(dev);
> >  
> > +	if (unlikely(atomic_read(&priv->reset_pending)))
> > +		return NETDEV_TX_BUSY;
> > +
> 
> This looks a bit strange.  schedule_work() will return zero if the work was
> already scheduled, in which case we arrange for cpmac_start_xmit() to abort
> early.
> 
> But if schedule_work() *doesn't* return zero, there is a time window in
> which the reset is still pending.  Because it takes time for keventd to be
> awoken and to run the work function.
> 
> I would have thought that we would want to prevent cpmac_start_xmit() from
> running within that time window also?
> 
> 
> But that's just a guess - the text which you used to describe your work is
> missing much information, so I don't have a lot to work with here.
> 

This one is cleaner:


Signed-off-by: Matteo Croce <matteo@openwrt.org>
Signed-off-by: Felix Fietkau <nbd@openwrt.org>

diff --git a/drivers/net/cpmac.c b/drivers/net/cpmac.c
index 2b5740b..7f3f62e 100644
--- a/drivers/net/cpmac.c
+++ b/drivers/net/cpmac.c
@@ -38,6 +38,7 @@
 #include <linux/platform_device.h>
 #include <linux/dma-mapping.h>
 #include <asm/gpio.h>
+#include <asm/atomic.h>
 
 MODULE_AUTHOR("Eugene Konev <ejka@imfi.kspu.ru>");
 MODULE_DESCRIPTION("TI AR7 ethernet driver (CPMAC)");
@@ -187,6 +188,7 @@ struct cpmac_desc {
 #define CPMAC_EOQ			0x1000
 	struct sk_buff *skb;
 	struct cpmac_desc *next;
+	struct cpmac_desc *prev;
 	dma_addr_t mapping;
 	dma_addr_t data_mapping;
 };
@@ -208,6 +210,7 @@ struct cpmac_priv {
 	struct work_struct reset_work;
 	struct platform_device *pdev;
 	struct napi_struct napi;
+	atomic_t reset_pending;
 };
 
 static irqreturn_t cpmac_irq(int, void *);
@@ -241,6 +244,16 @@ static void cpmac_dump_desc(struct net_device *dev, struct cpmac_desc *desc)
 	printk("\n");
 }
 
+static void cpmac_dump_all_desc(struct net_device *dev)
+{
+	struct cpmac_priv *priv = netdev_priv(dev);
+	struct cpmac_desc *dump = priv->rx_head;
+	do {
+		cpmac_dump_desc(dev, dump);
+		dump = dump->next;
+	} while (dump != priv->rx_head);
+}
+
 static void cpmac_dump_skb(struct net_device *dev, struct sk_buff *skb)
 {
 	int i;
@@ -412,21 +425,42 @@ static struct sk_buff *cpmac_rx_one(struct cpmac_priv *priv,
 static int cpmac_poll(struct napi_struct *napi, int budget)
 {
 	struct sk_buff *skb;
-	struct cpmac_desc *desc;
-	int received = 0;
+	struct cpmac_desc *desc, *restart;
 	struct cpmac_priv *priv = container_of(napi, struct cpmac_priv, napi);
+	int received = 0, processed = 0;
 
 	spin_lock(&priv->rx_lock);
 	if (unlikely(!priv->rx_head)) {
 		if (netif_msg_rx_err(priv) && net_ratelimit())
 			printk(KERN_WARNING "%s: rx: polling, but no queue\n",
 			       priv->dev->name);
+		spin_unlock(&priv->rx_lock);
 		netif_rx_complete(priv->dev, napi);
 		return 0;
 	}
 
 	desc = priv->rx_head;
+	restart = NULL;
 	while (((desc->dataflags & CPMAC_OWN) == 0) && (received < budget)) {
+		processed++;
+
+		if ((desc->dataflags & CPMAC_EOQ) != 0) {
+			/* The last update to eoq->hw_next didn't happen
+			* soon enough, and the receiver stopped here.
+			*Remember this descriptor so we can restart
+			* the receiver after freeing some space.
+			*/
+			if (unlikely(restart)) {
+				if (netif_msg_rx_err(priv))
+					printk(KERN_ERR "%s: poll found a"
+						" duplicate EOQ: %p and %p\n",
+						priv->dev->name, restart, desc);
+				goto fatal_error;
+			}
+
+			restart = desc->next;
+		}
+
 		skb = cpmac_rx_one(priv, desc);
 		if (likely(skb)) {
 			netif_receive_skb(skb);
@@ -435,19 +469,90 @@ static int cpmac_poll(struct napi_struct *napi, int budget)
 		desc = desc->next;
 	}
 
+	if (desc != priv->rx_head) {
+		/* We freed some buffers, but not the whole ring,
+		 * add what we did free to the rx list */
+		desc->prev->hw_next = (u32)0;
+		priv->rx_head->prev->hw_next = priv->rx_head->mapping;
+	}
+
+	/* Optimization: If we did not actually process an EOQ (perhaps because
+	 * of quota limits), check to see if the tail of the queue has EOQ set.
+	* We should immediately restart in that case so that the receiver can
+	* restart and run in parallel with more packet processing.
+	* This lets us handle slightly larger bursts before running
+	* out of ring space (assuming dev->weight < ring_size) */
+
+	if (!restart &&
+	     (priv->rx_head->prev->dataflags & (CPMAC_OWN|CPMAC_EOQ))
+		    == CPMAC_EOQ &&
+	     (priv->rx_head->dataflags & CPMAC_OWN) != 0) {
+		/* reset EOQ so the poll loop (above) doesn't try to
+		* restart this when it eventually gets to this descriptor.
+		*/
+		priv->rx_head->prev->dataflags &= ~CPMAC_EOQ;
+		restart = priv->rx_head;
+	}
+
+	if (restart) {
+		priv->dev->stats.rx_errors++;
+		priv->dev->stats.rx_fifo_errors++;
+		if (netif_msg_rx_err(priv) && net_ratelimit())
+			printk(KERN_WARNING "%s: rx dma ring overrun\n",
+			       priv->dev->name);
+
+		if (unlikely((restart->dataflags & CPMAC_OWN) == 0)) {
+			if (netif_msg_drv(priv))
+				printk(KERN_ERR "%s: cpmac_poll is trying to "
+					"restart rx from a descriptor that's "
+					"not free: %p\n",
+					priv->dev->name, restart);
+				goto fatal_error;
+		}
+
+		cpmac_write(priv->regs, CPMAC_RX_PTR(0), restart->mapping);
+	}
+
 	priv->rx_head = desc;
 	spin_unlock(&priv->rx_lock);
 	if (unlikely(netif_msg_rx_status(priv)))
 		printk(KERN_DEBUG "%s: poll processed %d packets\n",
 		       priv->dev->name, received);
-	if (desc->dataflags & CPMAC_OWN) {
+	if (processed == 0) {
+		/* we ran out of packets to read,
+		 * revert to interrupt-driven mode */
 		netif_rx_complete(priv->dev, napi);
-		cpmac_write(priv->regs, CPMAC_RX_PTR(0), (u32)desc->mapping);
 		cpmac_write(priv->regs, CPMAC_RX_INT_ENABLE, 1);
 		return 0;
 	}
 
 	return 1;
+
+fatal_error:
+	/* Something went horribly wrong.
+	 * Reset hardware to try to recover rather than wedging. */
+
+	if (netif_msg_drv(priv)) {
+		printk(KERN_ERR "%s: cpmac_poll is confused. "
+				"Resetting hardware\n", priv->dev->name);
+		cpmac_dump_all_desc(priv->dev);
+		printk(KERN_DEBUG "%s: RX_PTR(0)=0x%08x RX_ACK(0)=0x%08x\n",
+			priv->dev->name,
+			cpmac_read(priv->regs, CPMAC_RX_PTR(0)),
+			cpmac_read(priv->regs, CPMAC_RX_ACK(0)));
+	}
+
+	spin_unlock(&priv->rx_lock);
+	netif_rx_complete(priv->dev, napi);
+	netif_stop_queue(priv->dev);
+	napi_disable(&priv->napi);
+
+	atomic_inc(&priv->reset_pending);
+	cpmac_hw_stop(priv->dev);
+	if (!schedule_work(&priv->reset_work))
+		atomic_dec(&priv->reset_pending);
+	return 0;
+
 }
 
 static int cpmac_start_xmit(struct sk_buff *skb, struct net_device *dev)
@@ -456,6 +561,9 @@ static int cpmac_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct cpmac_desc *desc;
 	struct cpmac_priv *priv = netdev_priv(dev);
 
+	if (unlikely(atomic_read(&priv->reset_pending)))
+		return NETDEV_TX_BUSY;
+
 	if (unlikely(skb_padto(skb, ETH_ZLEN)))
 		return NETDEV_TX_OK;
 
@@ -621,8 +729,10 @@ static void cpmac_clear_rx(struct net_device *dev)
 			desc->dataflags = CPMAC_OWN;
 			dev->stats.rx_dropped++;
 		}
+		desc->hw_next = desc->next->mapping;
 		desc = desc->next;
 	}
+	priv->rx_head->prev->hw_next = 0;
 }
 
 static void cpmac_clear_tx(struct net_device *dev)
@@ -635,14 +745,14 @@ static void cpmac_clear_tx(struct net_device *dev)
 		priv->desc_ring[i].dataflags = 0;
 		if (priv->desc_ring[i].skb) {
 			dev_kfree_skb_any(priv->desc_ring[i].skb);
-			if (netif_subqueue_stopped(dev, i))
-			    netif_wake_subqueue(dev, i);
+			priv->desc_ring[i].skb = NULL;
 		}
 	}
 }
 
 static void cpmac_hw_error(struct work_struct *work)
 {
+	int i;
 	struct cpmac_priv *priv =
 		container_of(work, struct cpmac_priv, reset_work);
 
@@ -651,8 +761,48 @@ static void cpmac_hw_error(struct work_struct *work)
 	spin_unlock(&priv->rx_lock);
 	cpmac_clear_tx(priv->dev);
 	cpmac_hw_start(priv->dev);
-	napi_enable(&priv->napi);
-	netif_start_queue(priv->dev);
+	barrier();
+	atomic_dec(&priv->reset_pending);
+
+	for (i = 0; i < CPMAC_QUEUES; i++)
+		netif_wake_subqueue(priv->dev, i);
+	netif_wake_queue(priv->dev);
+	cpmac_write(priv->regs, CPMAC_MAC_INT_ENABLE, 3);
+}
+
+static void cpmac_check_status(struct net_device *dev)
+{
+	struct cpmac_priv *priv = netdev_priv(dev);
+
+	u32 macstatus = cpmac_read(priv->regs, CPMAC_MAC_STATUS);
+	int rx_channel = (macstatus >> 8) & 7;
+	int rx_code = (macstatus >> 12) & 15;
+	int tx_channel = (macstatus >> 16) & 7;
+	int tx_code = (macstatus >> 20) & 15;
+
+	if (rx_code || tx_code) {
+		if (netif_msg_drv(priv) && net_ratelimit()) {
+			/* Can't find any documentation on what these
+			 *error codes actually are. So just log them and hope..
+			 */
+			if (rx_code)
+				printk(KERN_WARNING "%s: host error %d on rx "
+				     "channel %d (macstatus %08x), resetting\n",
+				     dev->name, rx_code, rx_channel, macstatus);
+			if (tx_code)
+				printk(KERN_WARNING "%s: host error %d on tx "
+				     "channel %d (macstatus %08x), resetting\n",
+				     dev->name, tx_code, tx_channel, macstatus);
+		}
+
+		netif_stop_queue(dev);
+		cpmac_hw_stop(dev);
+		if (schedule_work(&priv->reset_work))
+			atomic_inc(&priv->reset_pending);
+		if (unlikely(netif_msg_hw(priv)))
+			cpmac_dump_regs(dev);
+	}
+	cpmac_write(priv->regs, CPMAC_MAC_INT_CLEAR, 0xff);
 }
 
 static irqreturn_t cpmac_irq(int irq, void *dev_id)
@@ -683,49 +833,32 @@ static irqreturn_t cpmac_irq(int irq, void *dev_id)
 
 	cpmac_write(priv->regs, CPMAC_MAC_EOI_VECTOR, 0);
 
-	if (unlikely(status & (MAC_INT_HOST | MAC_INT_STATUS))) {
-		if (netif_msg_drv(priv) && net_ratelimit())
-			printk(KERN_ERR "%s: hw error, resetting...\n",
-			       dev->name);
-		netif_stop_queue(dev);
-		napi_disable(&priv->napi);
-		cpmac_hw_stop(dev);
-		schedule_work(&priv->reset_work);
-		if (unlikely(netif_msg_hw(priv)))
-			cpmac_dump_regs(dev);
-	}
+	if (unlikely(status & (MAC_INT_HOST | MAC_INT_STATUS)))
+		cpmac_check_status(dev);
 
 	return IRQ_HANDLED;
 }
 
 static void cpmac_tx_timeout(struct net_device *dev)
 {
-	struct cpmac_priv *priv = netdev_priv(dev);
 	int i;
+	struct cpmac_priv *priv = netdev_priv(dev);
 
 	spin_lock(&priv->lock);
 	dev->stats.tx_errors++;
 	spin_unlock(&priv->lock);
 	if (netif_msg_tx_err(priv) && net_ratelimit())
 		printk(KERN_WARNING "%s: transmit timeout\n", dev->name);
-	/* 
-	 * FIXME: waking up random queue is not the best thing to
-	 * do... on the other hand why we got here at all?
-	 */
-#ifdef CONFIG_NETDEVICES_MULTIQUEUE
+
+	atomic_inc(&priv->reset_pending);
+	barrier();
+	cpmac_clear_tx(dev);
+	barrier();
+	atomic_dec(&priv->reset_pending);
+
+	netif_wake_queue(priv->dev);
 	for (i = 0; i < CPMAC_QUEUES; i++)
-		if (priv->desc_ring[i].skb) {
-			priv->desc_ring[i].dataflags = 0;
-			dev_kfree_skb_any(priv->desc_ring[i].skb);
-			netif_wake_subqueue(dev, i);
-			break;
-		}
-#else
-	priv->desc_ring[0].dataflags = 0;
-	if (priv->desc_ring[0].skb)
-		dev_kfree_skb_any(priv->desc_ring[0].skb);
-	netif_wake_queue(dev);
-#endif
+		netif_wake_subqueue(dev, i);
 }
 
 static int cpmac_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
@@ -901,9 +1034,12 @@ static int cpmac_open(struct net_device *dev)
 		desc->buflen = CPMAC_SKB_SIZE;
 		desc->dataflags = CPMAC_OWN;
 		desc->next = &priv->rx_head[(i + 1) % priv->ring_size];
+		desc->next->prev = desc;
 		desc->hw_next = (u32)desc->next->mapping;
 	}
 
+	priv->rx_head->prev->hw_next = (u32)0;
+
 	if ((res = request_irq(dev->irq, cpmac_irq, IRQF_SHARED,
 			       dev->name, dev))) {
 		if (netif_msg_drv(priv))
@@ -912,6 +1048,7 @@ static int cpmac_open(struct net_device *dev)
 		goto fail_irq;
 	}
 
+	atomic_set(&priv->reset_pending, 0);
 	INIT_WORK(&priv->reset_work, cpmac_hw_error);
 	cpmac_hw_start(dev);
 
@@ -1007,21 +1144,10 @@ static int __devinit cpmac_probe(struct platform_device *pdev)
 
 	if (phy_id == PHY_MAX_ADDR) {
 		if (external_switch || dumb_switch) {
-			struct fixed_phy_status status = {};
-
-			/*
-			 * FIXME: this should be in the platform code!
-			 * Since there is not platform code at all (that is,
-			 * no mainline users of that driver), place it here
-			 * for now.
-			 */
-			phy_id = 0;
-			status.link = 1;
-			status.duplex = 1;
-			status.speed = 100;
-			fixed_phy_add(PHY_POLL, phy_id, &status);
+			mdio_bus_id = 0; /* fixed phys bus */
+			phy_id = pdev->id;
 		} else {
-			printk(KERN_ERR "cpmac: no PHY present\n");
+			dev_err(&pdev->dev, "no PHY present\n");
 			return -ENODEV;
 		}
 	}
@@ -1064,10 +1190,8 @@ static int __devinit cpmac_probe(struct platform_device *pdev)
 	priv->msg_enable = netif_msg_init(debug_level, 0xff);
 	memcpy(dev->dev_addr, pdata->dev_addr, sizeof(dev->dev_addr));
 
-	snprintf(priv->phy_name, BUS_ID_SIZE, PHY_ID_FMT, mdio_bus_id, phy_id);
-
-	priv->phy = phy_connect(dev, priv->phy_name, &cpmac_adjust_link, 0,
-				PHY_INTERFACE_MODE_MII);
+	priv->phy = phy_connect(dev, cpmac_mii.phy_map[phy_id]->dev.bus_id,
+				&cpmac_adjust_link, 0, PHY_INTERFACE_MODE_MII);
 	if (IS_ERR(priv->phy)) {
 		if (netif_msg_drv(priv))
 			printk(KERN_ERR "%s: Could not attach to PHY\n",