Skip to contentSkip to conversation

ixgbe: Fix compiler warning about variable being used uninitialized

PATCH In review Jeff Kirsher 4 participants 10 messages Dec 12, 2009 – Dec 14, 2009

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/3 ixgbe: Fix compiler warning about variable being used uninitialized

From: PJ Waskiewicz <[email protected]>

tc is still throwing a warning that is could be used uninitialized. This fixes it.

---

 drivers/net/ixgbe/ixgbe_main.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/net/ixgbe/ixgbe_main.c +1 -1
index 35ea8c9..d561fdf 100644
@@ -258,7 +258,7 @@ static inline bool ixgbe_tx_is_paused(struct ixgbe_adapter *adapter,
 
 #ifdef CONFIG_IXGBE_DCB
 	if (adapter->flags & IXGBE_FLAG_DCB_ENABLED) {
-		int tc;
+		int tc = 0;
 		int reg_idx = tx_ring->reg_idx;
 		int dcb_i = adapter->ring_feature[RING_F_DCB].indices;
 
#1 lore raw

PATCH 2/3 ixgbe: Fix tx_restart_queue/non_eop_desc statistics counters

From: Mallikarjuna R Chilakala <[email protected]>

Fix tx_restart_queue & non_eop_desc stats counters, counters were not getting incremented as expected with the traffic

---

 drivers/net/ixgbe/ixgbe_main.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)
drivers/net/ixgbe/ixgbe_main.c +5 -2
index d561fdf..65eec2d 100644
@@ -4511,6 +4511,7 @@ void ixgbe_update_stats(struct ixgbe_adapter *adapter)
 	struct ixgbe_hw *hw = &adapter->hw;
 	u64 total_mpc = 0;
 	u32 i, missed_rx = 0, mpc, bprc, lxon, lxoff, xon_off_tot;
+	u64 non_eop_descs = 0, restart_queue = 0;
 
 	if (adapter->flags2 & IXGBE_FLAG2_RSC_ENABLED) {
 		u64 rsc_count = 0;
@@ -4528,10 +4529,12 @@ void ixgbe_update_stats(struct ixgbe_adapter *adapter)
 
 	/* gather some stats to the adapter struct that are per queue */
 	for (i = 0; i < adapter->num_tx_queues; i++)
-		adapter->restart_queue += adapter->tx_ring[i].restart_queue;
+		restart_queue += adapter->tx_ring[i]->restart_queue;
+	adapter->restart_queue = restart_queue;
 
 	for (i = 0; i < adapter->num_rx_queues; i++)
-		adapter->non_eop_descs += adapter->tx_ring[i].non_eop_descs;
+		non_eop_descs += adapter->rx_ring[i]->non_eop_descs;
+	adapter->non_eop_descs = non_eop_descs;
 
 	adapter->stats.crcerrs += IXGBE_READ_REG(hw, IXGBE_CRCERRS);
 	for (i = 0; i < 8; i++) {
maintainer
#2 lore raw

Re: [net-2.6 PATCH 2/3] ixgbe: Fix tx_restart_queue/non_eop_desc statistics counters

From: Jeff Kirsher <[email protected]>
Date: Fri, 11 Dec 2009 23:51:41 -0800

drivers/net/ixgbe/ixgbe_main.c void ixgbe_update_stats(struct ixgbe_adapter *adapter) · line 4533

Quoted 7 lines
drivers/net/ixgbe/ixgbe_main.c +2 -1
@@ -4528,10 +4529,12 @@ void ixgbe_update_stats(struct ixgbe_adapter *adapter)
 
 	/* gather some stats to the adapter struct that are per queue */
 	for (i = 0; i < adapter->num_tx_queues; i++)
-		adapter->restart_queue += adapter->tx_ring[i].restart_queue;
+		restart_queue += adapter->tx_ring[i]->restart_queue;
+	adapter->restart_queue = restart_queue;

I don't see how these two versions of the code can behave differently unless there is very broken locking on the adapter->* statistic counters.

I'm not applying this without at least a better commit log message.

#3 lore raw

Re: [net-2.6 PATCH 2/3] ixgbe: Fix tx_restart_queue/non_eop_desc statistics counters

Quoted 5 lines · David Miller

From: Jeff Kirsher <[email protected]>
Date: Fri, 11 Dec 2009 23:51:41 -0800

Quoted 2 lines
(quoted diff) +0 -0
@@ -4528,10 +4529,12 @@ void ixgbe_update_stats(struct ixgbe_adapter *adapter)
 
Quoted 9 lines

drivers/net/ixgbe/ixgbe_main.c void ixgbe_update_stats(struct ixgbe_adapter *adapter) · line 4533

Quoted 5 lines
drivers/net/ixgbe/ixgbe_main.c +2 -1
 	/* gather some stats to the adapter struct that are per queue */
 	for (i = 0; i < adapter->num_tx_queues; i++)
-		adapter->restart_queue += adapter->tx_ring[i].restart_queue;
+		restart_queue += adapter->tx_ring[i]->restart_queue;
+	adapter->restart_queue = restart_queue;

I don't see how these two versions of the code can behave differently unless there is very broken locking on the adapter->* statistic counters.

The problem is the adapter->restart_queue is being double-counted. If we accumulate from each ring's restart_queue, which are cumulative, then we don't want to add it to the previous update_stats() values in adapter->restart_queue.

Quoted 2 lines

I'm not applying this without at least a better commit log message.

We can fix the commit log message to better explain what the problem is.

Thanks,
-PJ

maintainer
#4 lore raw

Re: [net-2.6 PATCH 2/3] ixgbe: Fix tx_restart_queue/non_eop_desc statistics counters

From: Peter P Waskiewicz Jr <[email protected]>
Date: Sun, 13 Dec 2009 20:57:05 -0800

Quoted 9 lines · David Miller
Quoted 3 lines

I don't see how these two versions of the code can behave differently unless there is very broken locking on the adapter->* statistic counters.

The problem is the adapter->restart_queue is being double-counted. If we accumulate from each ring's restart_queue, which are cumulative, then we don't want to add it to the previous update_stats() values in adapter->restart_queue.

Now I understand, thanks.

Quoted 4 lines
Quoted 2 lines

I'm not applying this without at least a better commit log message.

We can fix the commit log message to better explain what the problem is.

Thanks PJ.

#5 lore raw

PATCH 3/3 Fix 82598 premature copper PHY link indicatation

In reply to Jeff Kirsher

From: Mallikarjuna R Chilakala <[email protected]>

Fix 82598 copper link issue, where the phy prematurely indicates link before it is ready to process packets. The new function looks for phy link and indicates that, when it is available. If phy is not ready within few seconds of MAC indicating link, the function will return failure which translates to link down indication

---

 drivers/net/ixgbe/ixgbe_82598.c |   38 ++++++++++++++++++++++++++++++++++++++
 drivers/net/ixgbe/ixgbe_type.h  |    2 ++
 2 files changed, 40 insertions(+), 0 deletions(-)
drivers/net/ixgbe/ixgbe_82598.c +38 -0
index e2d5343..1cc91f8 100644
@@ -510,6 +510,40 @@ static s32 ixgbe_start_mac_link_82598(struct ixgbe_hw *hw,
 }
 
 /**
+ *  ixgbe_validate_link_ready - Function looks for phy link
+ *  @hw: pointer to hardware structure
+ *
+ *  Function indicates success when phy link is available. If phy is not ready
+ *  within 5 seconds of MAC indicating link, the function returns error.
+ **/
+static s32 ixgbe_validate_link_ready(struct ixgbe_hw *hw)
+{
+	u32 timeout;
+	u16 an_reg;
+
+	if (hw->device_id != IXGBE_DEV_ID_82598AT2)
+		return 0;
+
+	for (timeout = 0;
+	     timeout < IXGBE_VALIDATE_LINK_READY_TIMEOUT; timeout++) {
+		hw->phy.ops.read_reg(hw, MDIO_STAT1, MDIO_MMD_AN, &an_reg);
+
+		if ((an_reg & MDIO_AN_STAT1_COMPLETE) &&
+		    (an_reg & MDIO_STAT1_LSTATUS))
+			break;
+
+		mdelay(100);
+	}
+
+	if (timeout == IXGBE_VALIDATE_LINK_READY_TIMEOUT) {
+		hw_dbg(hw, "Link was indicated but link is down\n");
+		return IXGBE_ERR_LINK_SETUP;
+	}
+
+	return 0;
+}
+
+/**
  *  ixgbe_check_mac_link_82598 - Get link/speed status
  *  @hw: pointer to hardware structure
  *  @speed: pointer to link speed
@@ -589,6 +623,10 @@ static s32 ixgbe_check_mac_link_82598(struct ixgbe_hw *hw,
 	else
 		*speed = IXGBE_LINK_SPEED_1GB_FULL;
 
+	if ((hw->device_id == IXGBE_DEV_ID_82598AT2) && (*link_up == true) &&
+	    (ixgbe_validate_link_ready(hw) != 0))
+		*link_up = false;
+
 	/* if link is down, zero out the current_mode */
 	if (*link_up == false) {
 		hw->fc.current_mode = ixgbe_fc_none;
drivers/net/ixgbe/ixgbe_type.h +2 -0
index f3e8d52..84650c6 100644
@@ -841,6 +841,8 @@
 #define IXGBE_MPVC      0x04318
 #define IXGBE_SGMIIC    0x04314
 
+#define IXGBE_VALIDATE_LINK_READY_TIMEOUT 50
+
 /* Omer CORECTL */
 #define IXGBE_CORECTL           0x014F00
 /* BARCTRL */
maintainer
#6 lore raw

Re: [net-2.6 PATCH 3/3] Fix 82598 premature copper PHY link indicatation

From: Jeff Kirsher <[email protected]>
Date: Fri, 11 Dec 2009 23:52:01 -0800

drivers/net/ixgbe/ixgbe_82598.c static s32 ixgbe_start_mac_link_82598(struct ixgbe_hw *hw, · line 536

Quoted 10 lines
drivers/net/ixgbe/ixgbe_82598.c +10 -0
+	for (timeout = 0;
+	     timeout < IXGBE_VALIDATE_LINK_READY_TIMEOUT; timeout++) {
+		hw->phy.ops.read_reg(hw, MDIO_STAT1, MDIO_MMD_AN, &an_reg);
+
+		if ((an_reg & MDIO_AN_STAT1_COMPLETE) &&
+		    (an_reg & MDIO_STAT1_LSTATUS))
+			break;
+
+		mdelay(100);
+	}

An up to 5000 msec non-preemptable spin loop.

If this used "msleep()" that would make it OK.

But as it is, sorry, no way.

#7 lore raw

Re: [net-2.6 PATCH 3/3] Fix 82598 premature copper PHY link indicatation

Thanks Dave. I'll respin the patch, didn't mean to wait in non-preemptive loop.

Quoted 24 lines · David Miller

From: Jeff Kirsher <[email protected]>
Date: Fri, 11 Dec 2009 23:52:01 -0800

drivers/net/ixgbe/ixgbe_82598.c static s32 ixgbe_start_mac_link_82598(struct ixgbe_hw *hw, · line 536

Quoted 10 lines

+     for (timeout = 0;
+          timeout < IXGBE_VALIDATE_LINK_READY_TIMEOUT; timeout++) {
+             hw->phy.ops.read_reg(hw, MDIO_STAT1, MDIO_MMD_AN, &an_reg);

drivers/net/ixgbe/ixgbe_82598.c +7 -0
+
+             if ((an_reg & MDIO_AN_STAT1_COMPLETE) &&
+                 (an_reg & MDIO_STAT1_LSTATUS))
+                     break;
+
+             mdelay(100);
+     }

An up to 5000 msec non-preemptable spin loop.

If this used "msleep()" that would make it OK.

But as it is, sorry, no way.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

maintainer
#8 lore raw
In reply to Jeff Kirsher

From: Jeff Kirsher <[email protected]>
Date: Fri, 11 Dec 2009 23:51:23 -0800

Quoted 7 lines

From: PJ Waskiewicz <[email protected]>

tc is still throwing a warning that is could be used uninitialized. This fixes it.

It's warning because it absolutely can be used uninitialized.

The two code blocks only handle two chip variants. Why not make what's happening here explicit by using a switch statement and having a 'default' case?

Thanks.

#9 lore raw
Quoted 8 lines · David Miller

From: Jeff Kirsher <[email protected]>
Date: Fri, 11 Dec 2009 23:51:23 -0800

Quoted 5 lines

From: PJ Waskiewicz <[email protected]>

tc is still throwing a warning that is could be used uninitialized. This fixes it.

Quoted 8 lines
Quoted 2 lines

It's warning because it absolutely can be used uninitialized.

The two code blocks only handle two chip variants. Why not make what's happening here explicit by using a switch statement and having a 'default' case?

That I can do, and makes more sense. I'll respin the patch.

Thanks Dave,
-PJ