Re: Question regarding e1000 driver and dropped packets (2.6.5 / 2.6.10)?

From: Bukie Mabayoje
Date: Tue Feb 08 2005 - 19:28:39 EST

Can you do a simple test?
Connect the two box to the same switch. ( No other box should be on the physical bus)
1. Send packets from BoxA -------> BoxB ( Record the stats)

2. Send packets from BoxB -------> BoxA (Record the stats)

3. Send packets simultaneously from BoxB----->BoxA and BoxA -----> BoxB (Record the stats)

if you can find a third box

4. Send packets [BoxA and BoxC] ---------> BoxB and BoxB -----> BoxA (Record the stats)

5. Send packets [BoxB and BoxC] ---------> BoxA and BoxA ------> BoxB (Record the stats)

I don't understand why you received more packet on BoxB. A controlled test will help clarify any ambiguity.
[BoxA] RX packets:196787934 errors:4 dropped:0 overruns:0 frame:2
TX packets:101356779 errors:0 dropped:0 overruns:0 carrier:0

[BoxB] RX packets:446380046 errors:1276833 dropped:1276833 overruns:1276833 frame:0
TX packets:572550636 errors:0 dropped:0 overruns:0 carrier:0


Justin Piszcz wrote:

> I have two identical machines [mobo/hardware wise]:
>
> Each machine is a Dell GX1p (500MHZ).
>
> I have two Intel Gigabit NICs, one in each box, hooked up to a GigE
> switch.
>
> Ethernet controller: Intel Corp. 82541GI/PI Gigabit Ethernet Controller
> Ethernet controller: Intel Corp. 82541GI/PI Gigabit Ethernet Controller
>
> I doubt its the kernel version; does anyone have any suggestions/ideas why
> one machine has virtually NO overruns/errors/drops and the other has tons?
>
> Also, (I doubt this to be the case but I'll ask anyway) - Is the way the
> NIC's are setup in the box next to other cards / alter their PCI/IRQ
> routing which would effect error/drop rates?
>
> IE:
>
> PCI1 - promise card / pata
> PCI2 - promise card / pata
> PCI3 - promise card / sata
> PCI4 - e1000 nic
> PCI5 - 4 port nic

What matters is which INT# [A,B,C,D] line and/or combination the PCI slot 1, 2, 3, 4 is using.
You can find out by running lspci -vv
If they are routed to the same system interrupt and lastly, the interrupt priority issues.

>
>
> Would it make sense to order them in a different direction?

May not help in identifying the problem.

>
>
> Also, is there a correlation between errors on the NIC and ERR
> in /proc/interrupts?

Maybe......

>
>
> Secondly, could loading lm-sensors/temperature modules be causing these
> problems?

You don't have any overrun on this box.

>
>
> dmesg from box2 below:
>
> e1000: eth0: e1000_watchdog: NIC Link is Up 1000 Mbps Full Duplex
> eth1: Setting full-duplex based on MII#1 link partner capability of 45e1.
> eth2: Setting full-duplex based on MII#1 link partner capability of 45e1.
> nfs warning: mount version older than kernel
> nfs warning: mount version older than kernel
> nfs warning: mount version older than kernel
> nfs warning: mount version older than kernel
> i2c /dev entries driver
> piix4_smbus 0000:00:07.3: Found 0000:00:07.3 device
> piix4_smbus 0000:00:07.3: WARNING: SMBus interface has been FORCEFULLY
> ENABLED!
> mtrr: no MTRR for fd000000,800000 found
> spurious 8259A interrupt: IRQ7.
> spurious 8259A interrupt: IRQ15.
>
> I am currently out of ideas, if anyone can suggest anything, I'd be most
> greatful, thanks!
>
> On the first box, there are hardly any problems receiving packets:
>
> Note the errors & dropped on the receiving end:
>
> BOX1: (2.6.5)
>
> eth0 Link encap:Ethernet HWaddr 00:0E:0C:00:CD:B1
> inet addr:10.0.2.254 Bcast:10.0.2.255 Mask:255.255.255.0
> inet6 addr: fe80::20e:cff:fe00:cdb1/64 Scope:Link
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> RX packets:196787934 errors:4 dropped:0 overruns:0 frame:2
> TX packets:101356779 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0 txqueuelen:1000
> RX bytes:2602045376 (2481.5 Mb) TX bytes:4051930608 (3864.2 Mb)
> Base address:0xcc80 Memory:ff020000-ff040000
>
> BOX1 MODULES:
>
> $ lsmod
> Module Size Used by
> ip_nat_ftp 4016 0
> ip_conntrack_ftp 71088 1 ip_nat_ftp
>
> BOX2: (2.6.10)
>
> On another box (same physical HW) I get this:
>
> eth0 Link encap:Ethernet HWaddr 00:0E:0C:00:D2:06
> inet addr:10.0.2.253 Bcast:10.0.2.255 Mask:255.255.255.0
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> ****--> RX packets:446380046 errors:1276833 dropped:1276833 overruns:1276833 frame:0
> TX packets:572550636 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0 txqueuelen:1000
> RX bytes:2351750726 (2.1 GiB) TX bytes:3659840330 (3.4 GiB)
> Base address:0xd8c0 Memory:f8fa0000-f8fc0000
>
> BOX2 MODULES:
>
> $ lsmod
> Module Size Used by
> ip_nat_irc 3408 0
> ip_conntrack_irc 70480 1 ip_nat_irc
> ip_nat_ftp 4112 0
> ip_conntrack_ftp 71344 1 ip_nat_ftp
> adm1021 11060 0
> i2c_piix4 6000 0
> i2c_sensor 2784 1 adm1021
> i2c_dev 7680 0
> i2c_core 18224 4 adm1021,i2c_piix4,i2c_sensor,i2c_dev
>
> I have tried using different cable and ports on the switch, the result is
> the same.
>
> $ tar cvf /box2/4gb_of_stuff.tar 4gb_of_stuff # then the numbers rise rapidly
>
> After copying only 1-2GB on BOX2, this is what I get:
>
> eth0 Link encap:Ethernet HWaddr 00:0E:0C:00:D2:06
> inet addr:10.0.2.253 Bcast:10.0.2.255 Mask:255.255.255.0
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> RX packets:1038733 errors:1459 dropped:1459 overruns:1459 frame:0
> TX packets:560952 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0 txqueuelen:1000
> RX bytes:1491121900 (1.3 GiB) TX bytes:763420385 (728.0 MiB)
> Base address:0xd8c0 Memory:f8fa0000-f8fc0000
>
> The only thing that is different is one has more HDD's and an extra PCI
> controller or so:
>
> BOX1 LSPCI:
>
> 00:00.0 Host bridge: Intel Corp. 440BX/ZX/DX - 82443BX/ZX/DX Host bridge
> (rev 03)
> 00:01.0 PCI bridge: Intel Corp. 440BX/ZX/DX - 82443BX/ZX/DX AGP bridge
> (rev 03)
> 00:07.0 ISA bridge: Intel Corp. 82371AB/EB/MB PIIX4 ISA (rev 02)
> 00:07.1 IDE interface: Intel Corp. 82371AB/EB/MB PIIX4 IDE (rev 01)
> 00:07.2 USB Controller: Intel Corp. 82371AB/EB/MB PIIX4 USB (rev 01)
> 00:07.3 Bridge: Intel Corp. 82371AB/EB/MB PIIX4 ACPI (rev 02)
> 00:0d.0 Ethernet controller: Intel Corp. 82541GI/PI Gigabit Ethernet
> Controller
> 00:0e.0 Unknown mass storage controller: Promise Technology, Inc. 20268
> (rev 02)
> 00:0f.0 PCI bridge: Digital Equipment Corporation DECchip 21152 (rev 03)
> 00:11.0 Ethernet controller: 3Com Corporation 3c905B 100BaseTX [Cyclone]
> (rev 24)
> 01:00.0 VGA compatible controller: ATI Technologies Inc 3D Rage Pro AGP
> 1X/2X (rev 5c)
> 02:09.0 Communication controller: Individual Computers - Jens Schoenfeld
> Intel 537
> 02:0a.0 Ethernet controller: 3Com Corporation 3c590 10BaseT [Vortex]
>
> BOX2 LSPCI:
>
> 0000:00:00.0 Host bridge: Intel Corp. 440BX/ZX/DX - 82443BX/ZX/DX Host
> bridge (rev 03)
> 0000:00:01.0 PCI bridge: Intel Corp. 440BX/ZX/DX - 82443BX/ZX/DX AGP
> bridge (rev 03)
> 0000:00:07.0 ISA bridge: Intel Corp. 82371AB/EB/MB PIIX4 ISA (rev 02)
> 0000:00:07.1 IDE interface: Intel Corp. 82371AB/EB/MB PIIX4 IDE (rev 01)
> 0000:00:07.2 USB Controller: Intel Corp. 82371AB/EB/MB PIIX4 USB (rev 01)
> 0000:00:07.3 Bridge: Intel Corp. 82371AB/EB/MB PIIX4 ACPI (rev 02)
> 0000:00:0d.0 Unknown mass storage controller: Promise Technology, Inc.
> 20269 (rev 02)
> 0000:00:0e.0 Unknown mass storage controller: Promise Technology, Inc.
> 20269 (rev 02)
> 0000:00:0f.0 PCI bridge: Digital Equipment Corporation DECchip 21152 (rev
> 03)
> 0000:01:00.0 VGA compatible controller: ATI Technologies Inc 3D Rage Pro
> AGP 1X/2X (rev 5c)
> 0000:02:09.0 Unknown mass storage controller: Promise Technology, Inc.
> PDC20375 (SATA150 TX2plus) (rev 02)
> 0000:02:0a.0 Ethernet controller: Intel Corp. 82541GI/PI Gigabit Ethernet
> Controller
> 0000:02:0b.0 PCI bridge: Digital Equipment Corporation DECchip 21152 (rev
> 03)
> 0000:03:04.0 Ethernet controller: Digital Equipment Corporation DECchip
> 21140 [FasterNet] (rev 22)
> 0000:03:05.0 Ethernet controller: Digital Equipment Corporation DECchip
> 21140 [FasterNet] (rev 22)
> 0000:03:06.0 Ethernet controller: Digital Equipment Corporation DECchip
> 21140 [FasterNet] (rev 22)
> 0000:03:07.0 Ethernet controller: Digital Equipment Corporation DECchip
> 21140 [FasterNet] (rev 22)
>
> Please CC me as I am not on the list, thanks!
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html