[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#380497: marked as done (problems with network on smp kernel when having more than 1 network card)



Your message dated Sun, 30 Jul 2006 18:52:04 +0200
with message-id <20060730165204.GM10814@baikonur.stro.at>
and subject line Bug#380497: problems with network on smp kernel when having more than 1 network card
has caused the attached Bug report to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what I am
talking about this indicates a serious mail system misconfiguration
somewhere.  Please contact me immediately.)

Debian bug tracking system administrator
(administrator, Debian Bugs database)

--- Begin Message ---
Package: linux-image-2.6.17-1-em64t-p4-smp
Version: 2.6.17-4
Architecture: x86_64
Severity: serious

I have 2 network cards. In 50% boot is successfull. In this case it initializes network cards same way if i use kernel with no smp: in my example RTL8169 is eth0, Tigon3 is eth1. But in other 50% boot hangs about 3 times for about 15 seconds, and each time prints error message: "rename_net_if: error changing interface name eth1_temp to eth0: timeout". I find out that in this case Tigon3 initializes first and takes eth0. Finally after boot finishes i have eth0 is what should be eth1, but eth1 doesn't exist at all. I never get this when using no-smp kernel.

 kern.log when boot successfull:
Jul 30 17:02:09 viiv kernel: r8169 Gigabit Ethernet driver 2.2LK-NAPI loaded
Jul 30 17:02:09 viiv kernel: eth0: Identified chip type is 'RTL8169s/8110s'.
Jul 30 17:02:09 viiv kernel: eth0: RTL8169 at 0xffffc20000002000, 00:80:48:3d:54:91, IRQ 217
...
Jul 30 17:02:09 viiv kernel: eth1: Tigon3 [partno(BCM95789) rev 4101 PHY(5750)] (PCI Express) 10/100/1000BaseT Ethernet 00:14:85:c2:86:32
Jul 30 17:02:09 viiv kernel: eth1: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] Split[0] WireSpeed[1] TSOcap[1]
Jul 30 17:02:09 viiv kernel: eth1: dma_rwctrl[76180000] dma_mask[64-bit]


 not successfull:
Jul 30 03:30:27 viiv kernel: eth0: Tigon3 [partno(BCM95789) rev 4101 PHY(5750)] (PCI Express) 10/100/1000BaseT Ethernet 00:14:85:c2:86:32
Jul 30 03:30:27 viiv kernel: eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] Split[0] WireSpeed[1] TSOcap[1]
Jul 30 03:30:27 viiv kernel: eth0: dma_rwctrl[76180000] dma_mask[64-bit]
Jul 30 03:30:27 viiv kernel: r8169 Gigabit Ethernet driver 2.2LK-NAPI loaded
Jul 30 03:30:27 viiv kernel: eth1: Identified chip type is 'RTL8169s/8110s'.
Jul 30 03:30:27 viiv kernel: eth1: RTL8169 at 0xffffc20000004000, 00:80:48:3d:54:91, IRQ 217

I tested on kernel 2.6.16-17 smp and no-smp kernels - same story.

I marked severity as serious cos our web-servers auto-reboot and will become not usable at all.

CPU: Pentium D 940

Sergey Nivarov


--- End Message ---
--- Begin Message ---
On Sun, Jul 30, 2006 at 08:00:27PM +0400, Sergey Nivarov wrote:
> Package: linux-image-2.6.17-1-em64t-p4-smp
> Version: 2.6.17-4
> Architecture: x86_64
> Severity: serious

nack the kernel _never_ guaranteed device order.
 
> I have 2 network cards. In 50% boot is successfull. In this case it
> initializes network cards same way if i use kernel with no smp: in my
> example RTL8169 is eth0, Tigon3 is eth1. But in other 50% boot hangs
> about 3 times for about 15 seconds, and each time prints error
> message: "rename_net_if: error changing interface name eth1_temp to
> eth0: timeout". I find out that in this case Tigon3 initializes first
> and takes eth0. Finally after boot finishes i have eth0 is what should
> be eth1, but eth1 doesn't exist at all. I never get this when using
> no-smp kernel.

policy belongs to userspace:
use persistent naming of the ethernet devices thanks to udev.
take a look at /lib/udev/write_net_rules

-- 
maks

--- End Message ---

Reply to: