[PATCH net v4 0/4] forcedeth: minor fixes for stats, rmmod, sparse

[PATCH net v4 0/4] forcedeth: minor fixes for stats, rmmod, sparse

[David Decotigny]

David,

Thanks for your feedback. Here is v4, with only the (minor) bug fixes
for net. I am setting the "new features" patches aside for later
(net-next).

Since I am still very much of a newbie with the way net & net-next
work, could you please redirect me to some pointer on how/when the
schedule works?

Regards,

Tested: 16-way x86_64 + forcedeth

Changes since v3:
 - removed feature additions, this leaves minor sparse and stats
   fixes. Feature additions shipped previously will go to net-next

Changes since v2:
 - removed "Fix a race during rmmod of forcedeth" from the series
   (will look at it separately with original author)
 - added "remove unneeded stats updates" and "64-bit stats"
 - reordered patches

Changes since v1:
 - rebased on top of netdev tip
 - do not repeat name of device in netdev_dbg
 - do not completely mute TX timeout messages when debug_tx_timeout is
   not set
 - make debug_tx_timeout writable in /sys/module
 Note: I am re-submitting "expose module parameters in /sys/module" as
       it can be useful in production and I was assured it doesn't add
       much memory overhead by the sysfs maintainers.

Tested:
  16-way x86_64 SMP, dual forcedeth ->
  RX bytes:7244556582 (7.2 GB)  TX bytes:181904254 (181.9 MB)


############################################
# Patch Set Summary:

David Decotigny (2):
  forcedeth: remove unneeded stats updates
  forcedeth: fix a few sparse warnings (variable shadowing)

Mandeep Baines (1):
  forcedeth: Improve stats counters

Mike Ditto (1):
  forcedeth: Acknowledge only interrupts that are being processed

 drivers/net/ethernet/nvidia/forcedeth.c |   86 +++++++++++--------------------
 1 files changed, 30 insertions(+), 56 deletions(-)

-- 
1.7.3.1

[PATCH net v4 1/4] forcedeth: Acknowledge only interrupts that are being processed

[David Decotigny]

From: Mike Ditto <mditto@google.com>

This is to avoid a race, accidentally acknowledging an interrupt that
we didn't notice and won't immediately process.  This is based solely
on code inspection; it is not known if there was an actual bug here.



Signed-off-by: David Decotigny <david.decotigny@google.com>
---
 drivers/net/ethernet/nvidia/forcedeth.c |   13 ++++++++-----
 1 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/nvidia/forcedeth.c b/drivers/net/ethernet/nvidia/forcedeth.c
index 1e37eb9..056986b 100644
--- a/drivers/net/ethernet/nvidia/forcedeth.c
+++ b/drivers/net/ethernet/nvidia/forcedeth.c
@@ -3398,7 +3398,8 @@ static irqreturn_t nv_nic_irq_tx(int foo, void *data)
 
 	for (i = 0;; i++) {
 		events = readl(base + NvRegMSIXIrqStatus) & NVREG_IRQ_TX_ALL;
-		writel(NVREG_IRQ_TX_ALL, base + NvRegMSIXIrqStatus);
+		writel(events, base + NvRegMSIXIrqStatus);
+		netdev_dbg(dev, "tx irq events: %08x\n", events);
 		if (!(events & np->irqmask))
 			break;
 
@@ -3509,7 +3510,8 @@ static irqreturn_t nv_nic_irq_rx(int foo, void *data)
 
 	for (i = 0;; i++) {
 		events = readl(base + NvRegMSIXIrqStatus) & NVREG_IRQ_RX_ALL;
-		writel(NVREG_IRQ_RX_ALL, base + NvRegMSIXIrqStatus);
+		writel(events, base + NvRegMSIXIrqStatus);
+		netdev_dbg(dev, "rx irq events: %08x\n", events);
 		if (!(events & np->irqmask))
 			break;
 
@@ -3553,7 +3555,8 @@ static irqreturn_t nv_nic_irq_other(int foo, void *data)
 
 	for (i = 0;; i++) {
 		events = readl(base + NvRegMSIXIrqStatus) & NVREG_IRQ_OTHER;
-		writel(NVREG_IRQ_OTHER, base + NvRegMSIXIrqStatus);
+		writel(events, base + NvRegMSIXIrqStatus);
+		netdev_dbg(dev, "irq events: %08x\n", events);
 		if (!(events & np->irqmask))
 			break;
 
@@ -3617,10 +3620,10 @@ static irqreturn_t nv_nic_irq_test(int foo, void *data)
 
 	if (!(np->msi_flags & NV_MSI_X_ENABLED)) {
 		events = readl(base + NvRegIrqStatus) & NVREG_IRQSTAT_MASK;
-		writel(NVREG_IRQ_TIMER, base + NvRegIrqStatus);
+		writel(events & NVREG_IRQ_TIMER, base + NvRegIrqStatus);
 	} else {
 		events = readl(base + NvRegMSIXIrqStatus) & NVREG_IRQSTAT_MASK;
-		writel(NVREG_IRQ_TIMER, base + NvRegMSIXIrqStatus);
+		writel(events & NVREG_IRQ_TIMER, base + NvRegMSIXIrqStatus);
 	}
 	pci_push(base);
 	if (!(events & NVREG_IRQ_TIMER))
-- 
1.7.3.1

[PATCH net v4 2/4] forcedeth: remove unneeded stats updates

[David Decotigny]

Function ndo_get_stats() updates most of the stats from hardware
registers, making the manual updates un-needed. This change removes
these manual updates. Main exception is rx_missed_errors which needs
manual update. Another exception is rx_packets, still updated manually
in this commit to make sure this patch doesn't change behavior of
driver (this is addressed by a later commit in this series).



Signed-off-by: David Decotigny <david.decotigny@google.com>
---
 drivers/net/ethernet/nvidia/forcedeth.c |   35 +------------------------------
 1 files changed, 1 insertions(+), 34 deletions(-)

diff --git a/drivers/net/ethernet/nvidia/forcedeth.c b/drivers/net/ethernet/nvidia/forcedeth.c
index 056986b..0bcb18e 100644
--- a/drivers/net/ethernet/nvidia/forcedeth.c
+++ b/drivers/net/ethernet/nvidia/forcedeth.c
@@ -2374,16 +2374,8 @@ static int nv_tx_done(struct net_device *dev, int limit)
 		if (np->desc_ver == DESC_VER_1) {
 			if (flags & NV_TX_LASTPACKET) {
 				if (flags & NV_TX_ERROR) {
-					if (flags & NV_TX_UNDERFLOW)
-						dev->stats.tx_fifo_errors++;
-					if (flags & NV_TX_CARRIERLOST)
-						dev->stats.tx_carrier_errors++;
 					if ((flags & NV_TX_RETRYERROR) && !(flags & NV_TX_RETRYCOUNT_MASK))
 						nv_legacybackoff_reseed(dev);
-					dev->stats.tx_errors++;
-				} else {
-					dev->stats.tx_packets++;
-					dev->stats.tx_bytes += np->get_tx_ctx->skb->len;
 				}
 				dev_kfree_skb_any(np->get_tx_ctx->skb);
 				np->get_tx_ctx->skb = NULL;
@@ -2392,16 +2384,8 @@ static int nv_tx_done(struct net_device *dev, int limit)
 		} else {
 			if (flags & NV_TX2_LASTPACKET) {
 				if (flags & NV_TX2_ERROR) {
-					if (flags & NV_TX2_UNDERFLOW)
-						dev->stats.tx_fifo_errors++;
-					if (flags & NV_TX2_CARRIERLOST)
-						dev->stats.tx_carrier_errors++;
 					if ((flags & NV_TX2_RETRYERROR) && !(flags & NV_TX2_RETRYCOUNT_MASK))
 						nv_legacybackoff_reseed(dev);
-					dev->stats.tx_errors++;
-				} else {
-					dev->stats.tx_packets++;
-					dev->stats.tx_bytes += np->get_tx_ctx->skb->len;
 				}
 				dev_kfree_skb_any(np->get_tx_ctx->skb);
 				np->get_tx_ctx->skb = NULL;
@@ -2434,9 +2418,7 @@ static int nv_tx_done_optimized(struct net_device *dev, int limit)
 		nv_unmap_txskb(np, np->get_tx_ctx);
 
 		if (flags & NV_TX2_LASTPACKET) {
-			if (!(flags & NV_TX2_ERROR))
-				dev->stats.tx_packets++;
-			else {
+			if (flags & NV_TX2_ERROR) {
 				if ((flags & NV_TX2_RETRYERROR) && !(flags & NV_TX2_RETRYCOUNT_MASK)) {
 					if (np->driver_data & DEV_HAS_GEAR_MODE)
 						nv_gear_backoff_reseed(dev);
@@ -2636,7 +2618,6 @@ static int nv_rx_process(struct net_device *dev, int limit)
 					if ((flags & NV_RX_ERROR_MASK) == NV_RX_ERROR4) {
 						len = nv_getlen(dev, skb->data, len);
 						if (len < 0) {
-							dev->stats.rx_errors++;
 							dev_kfree_skb(skb);
 							goto next_pkt;
 						}
@@ -2650,11 +2631,6 @@ static int nv_rx_process(struct net_device *dev, int limit)
 					else {
 						if (flags & NV_RX_MISSEDFRAME)
 							dev->stats.rx_missed_errors++;
-						if (flags & NV_RX_CRCERR)
-							dev->stats.rx_crc_errors++;
-						if (flags & NV_RX_OVERFLOW)
-							dev->stats.rx_over_errors++;
-						dev->stats.rx_errors++;
 						dev_kfree_skb(skb);
 						goto next_pkt;
 					}
@@ -2670,7 +2646,6 @@ static int nv_rx_process(struct net_device *dev, int limit)
 					if ((flags & NV_RX2_ERROR_MASK) == NV_RX2_ERROR4) {
 						len = nv_getlen(dev, skb->data, len);
 						if (len < 0) {
-							dev->stats.rx_errors++;
 							dev_kfree_skb(skb);
 							goto next_pkt;
 						}
@@ -2682,11 +2657,6 @@ static int nv_rx_process(struct net_device *dev, int limit)
 					}
 					/* the rest are hard errors */
 					else {
-						if (flags & NV_RX2_CRCERR)
-							dev->stats.rx_crc_errors++;
-						if (flags & NV_RX2_OVERFLOW)
-							dev->stats.rx_over_errors++;
-						dev->stats.rx_errors++;
 						dev_kfree_skb(skb);
 						goto next_pkt;
 					}
@@ -2704,7 +2674,6 @@ static int nv_rx_process(struct net_device *dev, int limit)
 		skb->protocol = eth_type_trans(skb, dev);
 		napi_gro_receive(&np->napi, skb);
 		dev->stats.rx_packets++;
-		dev->stats.rx_bytes += len;
 next_pkt:
 		if (unlikely(np->get_rx.orig++ == np->last_rx.orig))
 			np->get_rx.orig = np->first_rx.orig;
@@ -2787,9 +2756,7 @@ static int nv_rx_process_optimized(struct net_device *dev, int limit)
 				__vlan_hwaccel_put_tag(skb, vid);
 			}
 			napi_gro_receive(&np->napi, skb);
-
 			dev->stats.rx_packets++;
-			dev->stats.rx_bytes += len;
 		} else {
 			dev_kfree_skb(skb);
 		}
-- 
1.7.3.1

[PATCH net v4 3/4] forcedeth: Improve stats counters

[David Decotigny]

From: Mandeep Baines <msb@google.com>

Rx byte count was off; instead use the hardware's count.  Tx packet
count was counting pre-TSO packets; instead count on-the-wire packets.
Report hardware dropped frame count as rx_fifo_errors.

- The count of transmitted packets reported by the forcedeth driver
  reports pre-TSO (TCP Segmentation Offload) packet counts and not the
  count of the number of packets sent on the wire. This change fixes
  the forcedeth driver to report the correct count. Fixed the code by
  copying the count stored in the NIC H/W to the value reported by the
  driver.

- Count rx_drop_frame errors as rx_fifo_errors:
  We see a lot of rx_drop_frame errors if we disable the rx bottom-halves
  for too long.  Normally, rx_fifo_errors would be counted in this case.
  The rx_drop_frame error count is private to forcedeth and is not
  reported by ifconfig or sysfs.  The rx_fifo_errors count is currently
  unused in the forcedeth driver.  It is reported by ifconfig as overruns.
  This change reports rx_drop_frame errors as rx_fifo_errors.



Signed-off-by: David Decotigny <david.decotigny@google.com>
---
 drivers/net/ethernet/nvidia/forcedeth.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/nvidia/forcedeth.c b/drivers/net/ethernet/nvidia/forcedeth.c
index 0bcb18e..fe2ba70 100644
--- a/drivers/net/ethernet/nvidia/forcedeth.c
+++ b/drivers/net/ethernet/nvidia/forcedeth.c
@@ -1682,6 +1682,7 @@ static void nv_get_hw_stats(struct net_device *dev)
 		np->estats.tx_pause += readl(base + NvRegTxPause);
 		np->estats.rx_pause += readl(base + NvRegRxPause);
 		np->estats.rx_drop_frame += readl(base + NvRegRxDropFrame);
+		np->estats.rx_errors_total += np->estats.rx_drop_frame;
 	}
 
 	if (np->driver_data & DEV_HAS_STATISTICS_V3) {
@@ -1706,11 +1707,14 @@ static struct net_device_stats *nv_get_stats(struct net_device *dev)
 		nv_get_hw_stats(dev);
 
 		/* copy to net_device stats */
+		dev->stats.tx_packets = np->estats.tx_packets;
+		dev->stats.rx_bytes = np->estats.rx_bytes;
 		dev->stats.tx_bytes = np->estats.tx_bytes;
 		dev->stats.tx_fifo_errors = np->estats.tx_fifo_errors;
 		dev->stats.tx_carrier_errors = np->estats.tx_carrier_errors;
 		dev->stats.rx_crc_errors = np->estats.rx_crc_errors;
 		dev->stats.rx_over_errors = np->estats.rx_over_errors;
+		dev->stats.rx_fifo_errors = np->estats.rx_drop_frame;
 		dev->stats.rx_errors = np->estats.rx_errors_total;
 		dev->stats.tx_errors = np->estats.tx_errors_total;
 	}
-- 
1.7.3.1

[PATCH net v4 4/4] forcedeth: fix a few sparse warnings (variable shadowing)

[David Decotigny]

This fixes the following sparse warnings:
drivers/net/ethernet/nvidia/forcedeth.c:2113:7: warning: symbol 'size' shadows an earlier one
drivers/net/ethernet/nvidia/forcedeth.c:2102:6: originally declared here
drivers/net/ethernet/nvidia/forcedeth.c:2155:7: warning: symbol 'size' shadows an earlier one
drivers/net/ethernet/nvidia/forcedeth.c:2102:6: originally declared here
drivers/net/ethernet/nvidia/forcedeth.c:2227:7: warning: symbol 'size' shadows an earlier one
drivers/net/ethernet/nvidia/forcedeth.c:2215:6: originally declared here
drivers/net/ethernet/nvidia/forcedeth.c:2271:7: warning: symbol 'size' shadows an earlier one
drivers/net/ethernet/nvidia/forcedeth.c:2215:6: originally declared here
drivers/net/ethernet/nvidia/forcedeth.c:2986:20: warning: symbol 'addr' shadows an earlier one
drivers/net/ethernet/nvidia/forcedeth.c:2963:6: originally declared here



Signed-off-by: David Decotigny <david.decotigny@google.com>
---
 drivers/net/ethernet/nvidia/forcedeth.c |   34 +++++++++++++++---------------
 1 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/drivers/net/ethernet/nvidia/forcedeth.c b/drivers/net/ethernet/nvidia/forcedeth.c
index fe2ba70..d32cd4b 100644
--- a/drivers/net/ethernet/nvidia/forcedeth.c
+++ b/drivers/net/ethernet/nvidia/forcedeth.c
@@ -2103,10 +2103,10 @@ static netdev_tx_t nv_start_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	/* add fragments to entries count */
 	for (i = 0; i < fragments; i++) {
-		u32 size = skb_frag_size(&skb_shinfo(skb)->frags[i]);
+		u32 frag_size = skb_frag_size(&skb_shinfo(skb)->frags[i]);
 
-		entries += (size >> NV_TX2_TSO_MAX_SHIFT) +
-			   ((size & (NV_TX2_TSO_MAX_SIZE-1)) ? 1 : 0);
+		entries += (frag_size >> NV_TX2_TSO_MAX_SHIFT) +
+			   ((frag_size & (NV_TX2_TSO_MAX_SIZE-1)) ? 1 : 0);
 	}
 
 	spin_lock_irqsave(&np->lock, flags);
@@ -2145,13 +2145,13 @@ static netdev_tx_t nv_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	/* setup the fragments */
 	for (i = 0; i < fragments; i++) {
 		const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
-		u32 size = skb_frag_size(frag);
+		u32 frag_size = skb_frag_size(frag);
 		offset = 0;
 
 		do {
 			prev_tx = put_tx;
 			prev_tx_ctx = np->put_tx_ctx;
-			bcnt = (size > NV_TX2_TSO_MAX_SIZE) ? NV_TX2_TSO_MAX_SIZE : size;
+			bcnt = (frag_size > NV_TX2_TSO_MAX_SIZE) ? NV_TX2_TSO_MAX_SIZE : frag_size;
 			np->put_tx_ctx->dma = skb_frag_dma_map(
 							&np->pci_dev->dev,
 							frag, offset,
@@ -2163,12 +2163,12 @@ static netdev_tx_t nv_start_xmit(struct sk_buff *skb, struct net_device *dev)
 			put_tx->flaglen = cpu_to_le32((bcnt-1) | tx_flags);
 
 			offset += bcnt;
-			size -= bcnt;
+			frag_size -= bcnt;
 			if (unlikely(put_tx++ == np->last_tx.orig))
 				put_tx = np->first_tx.orig;
 			if (unlikely(np->put_tx_ctx++ == np->last_tx_ctx))
 				np->put_tx_ctx = np->first_tx_ctx;
-		} while (size);
+		} while (frag_size);
 	}
 
 	/* set last fragment flag  */
@@ -2217,10 +2217,10 @@ static netdev_tx_t nv_start_xmit_optimized(struct sk_buff *skb,
 
 	/* add fragments to entries count */
 	for (i = 0; i < fragments; i++) {
-		u32 size = skb_frag_size(&skb_shinfo(skb)->frags[i]);
+		u32 frag_size = skb_frag_size(&skb_shinfo(skb)->frags[i]);
 
-		entries += (size >> NV_TX2_TSO_MAX_SHIFT) +
-			   ((size & (NV_TX2_TSO_MAX_SIZE-1)) ? 1 : 0);
+		entries += (frag_size >> NV_TX2_TSO_MAX_SHIFT) +
+			   ((frag_size & (NV_TX2_TSO_MAX_SIZE-1)) ? 1 : 0);
 	}
 
 	spin_lock_irqsave(&np->lock, flags);
@@ -2261,13 +2261,13 @@ static netdev_tx_t nv_start_xmit_optimized(struct sk_buff *skb,
 	/* setup the fragments */
 	for (i = 0; i < fragments; i++) {
 		skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
-		u32 size = skb_frag_size(frag);
+		u32 frag_size = skb_frag_size(frag);
 		offset = 0;
 
 		do {
 			prev_tx = put_tx;
 			prev_tx_ctx = np->put_tx_ctx;
-			bcnt = (size > NV_TX2_TSO_MAX_SIZE) ? NV_TX2_TSO_MAX_SIZE : size;
+			bcnt = (frag_size > NV_TX2_TSO_MAX_SIZE) ? NV_TX2_TSO_MAX_SIZE : frag_size;
 			np->put_tx_ctx->dma = skb_frag_dma_map(
 							&np->pci_dev->dev,
 							frag, offset,
@@ -2280,12 +2280,12 @@ static netdev_tx_t nv_start_xmit_optimized(struct sk_buff *skb,
 			put_tx->flaglen = cpu_to_le32((bcnt-1) | tx_flags);
 
 			offset += bcnt;
-			size -= bcnt;
+			frag_size -= bcnt;
 			if (unlikely(put_tx++ == np->last_tx.ex))
 				put_tx = np->first_tx.ex;
 			if (unlikely(np->put_tx_ctx++ == np->last_tx_ctx))
 				np->put_tx_ctx = np->first_tx_ctx;
-		} while (size);
+		} while (frag_size);
 	}
 
 	/* set last fragment flag  */
@@ -2933,11 +2933,11 @@ static void nv_set_multicast(struct net_device *dev)
 				struct netdev_hw_addr *ha;
 
 				netdev_for_each_mc_addr(ha, dev) {
-					unsigned char *addr = ha->addr;
+					unsigned char *hw_addr = ha->addr;
 					u32 a, b;
 
-					a = le32_to_cpu(*(__le32 *) addr);
-					b = le16_to_cpu(*(__le16 *) (&addr[4]));
+					a = le32_to_cpu(*(__le32 *) hw_addr);
+					b = le16_to_cpu(*(__le16 *) (&hw_addr[4]));
 					alwaysOn[0] &= a;
 					alwaysOff[0] &= ~a;
 					alwaysOn[1] &= b;
-- 
1.7.3.1