+++ This bug was initially created as a clone of Bug #424581 +++

The network utilization graphics in ganglia report shows peaks of petabytes,
when the network interface is just Gigabit eth

Broadcom developer (Michael Chan) confirmed that NIC BCM5706 and BCM5708
are very similar in design so BCM5706 could suffer the same problem
described in http://lists.openwall.net/netdev/2007/06/04/81 he actually
provide a patch but it's not possible to test locally.

The host is rhel4.5 but it is running latest security updates, including
the kernel, from rhel4.7. here is the kernel info:

<root@446 (~)>$ uname -a
Linux 446 2.6.9-78.0.22.ELsmp #1 SMP Fri Apr 24 12:48:19 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux
<root@446 (~)>$ modinfo bnx2
filename:       /lib/modules/2.6.9-78.0.22.ELsmp/kernel/drivers/net/bnx2.ko
parm:           disable_msi:Disable Message Signaled Interrupt (MSI)
version:        1.6.9 478DCC31494367381918DC4
license:        GPL
description:    Broadcom NetXtreme II BCM5706/5708 Driver
author:         Michael Chan <mchan@broadcom.com>
alias:          pci:v000014E4d0000163Asv*sd*bc*sc*i*
alias:          pci:v000014E4d00001639sv*sd*bc*sc*i*
alias:          pci:v000014E4d000016ACsv*sd*bc*sc*i*
alias:          pci:v000014E4d000016AAsv*sd*bc*sc*i*
alias:          pci:v000014E4d000016AAsv0000103Csd00003102bc*sc*i*
alias:          pci:v000014E4d0000164Csv*sd*bc*sc*i*
alias:          pci:v000014E4d0000164Asv*sd*bc*sc*i*
alias:          pci:v000014E4d0000164Asv0000103Csd00003106bc*sc*i*
alias:          pci:v000014E4d0000164Asv0000103Csd00003101bc*sc*i*
depends:        
vermagic:       2.6.9-78.0.22.ELsmp SMP gcc-3.4
<root@446 (~)>$

as you can see, the bnx2 driver version is > 1.5.

I just want to point out that the fix that was included in version 1.5 is 
for NIC BCM5708 (you can see it from the patch that was submitted to the 
kernel at that time), nic BCM5706 was not detected having the same issue,
and according to Broadcom maintainer, both NIC has similar design so it 
is possible they both suffer same issue, that is why his patch is minor
correction to the original patch (to include nic 5706 id in the check). 
Patch diff shows that wherever 5708 is checked 5706 must be checked too, 
like this:

- if (CHIP_NUM(bp) == CHIP_NUM_5708)
+ if (CHIP_NUM(bp) == CHIP_NUM_5706 || CHIP_NUM(bp) == CHIP_NUM_5708) 

Full patch:
diff --git a/drivers/net/bnx2.c b/drivers/net/bnx2.c
index 2486a65..5079806 100644
--- a/drivers/net/bnx2.c
+++ b/drivers/net/bnx2.c
@@ -4589,7 +4589,7 @@ bnx2_init_chip(struct bnx2 *bp)
        REG_WR(bp, BNX2_HC_CMD_TICKS,
               (bp->cmd_ticks_int << 16) | bp->cmd_ticks);

-       if (CHIP_NUM(bp) == CHIP_NUM_5708)
+       if (CHIP_NUM(bp) == CHIP_NUM_5706 || CHIP_NUM(bp) == CHIP_NUM_5708)
                REG_WR(bp, BNX2_HC_STATS_TICKS, 0);
        else
                REG_WR(bp, BNX2_HC_STATS_TICKS, bp->stats_ticks);
@@ -5700,7 +5700,8 @@ bnx2_timer(unsigned long data)
                bnx2_reg_rd_ind(bp, BNX2_FW_RX_DROP_COUNT);

        /* workaround occasional corrupted counters */
-       if (CHIP_NUM(bp) == CHIP_NUM_5708 && bp->stats_ticks)
+       if ((CHIP_NUM(bp) == CHIP_NUM_5706 || CHIP_NUM(bp) == CHIP_NUM_5708) &&
+           bp->stats_ticks)
                REG_WR(bp, BNX2_HC_COMMAND, bp->hc_cmd |
                                            BNX2_HC_COMMAND_STATS_NOW);
@@ -6609,7 +6610,7 @@ bnx2_set_coalesce(struct net_device *dev, struct ethtool_coalesce *coal)
                0xff;

        bp->stats_ticks = coal->stats_block_coalesce_usecs;
-       if (CHIP_NUM(bp) == CHIP_NUM_5708) {
+       if (CHIP_NUM(bp) == CHIP_NUM_5706 || CHIP_NUM(bp) == CHIP_NUM_5708) {
                if (bp->stats_ticks != 0 && bp->stats_ticks != USEC_PER_SEC)
                        bp->stats_ticks = USEC_PER_SEC;
        }