topic Re: redhat4 +team +psp 7.4 error in Operating System - Linux

redhat4 +team +psp 7.4 error

expect — Sun, 27 Aug 2006 22:34:33 GMT

Hi,
OS:redhat as 4

Use team(bond0),If install smartstart 7.4 again .
A few days later
Appear and the network will be obstructed,
But ping ok
/var/log/message
Aug 18 09:48:03 rr ifup: Enslaving eth0 to bond0
Aug 18 09:48:03 rr kernel: e100: eth0: e100_watchdog: link up, 100Mbps, half-duplex
Aug 18 09:48:03 rr kernel: Debug: sleeping function called from invalid context at include/asm/uaccess.h:556
Aug 18 09:48:03 rr kernel: in_atomic():1[expected: 0], irqs_disabled():0
Aug 18 09:48:03 rr kernel: [] __might_sleep+0x7d/0x87
Aug 18 09:48:03 rr kernel: [] ethtool_ioctl+0x51/0x270 [e100]
Aug 18 09:48:03 rr kernel: [] netlink_broadcast+0x246/0x2be
Aug 18 09:48:03 rr kernel: [] bond_update_speed_duplex+0x96/0xda [bonding]
Aug 18 09:48:03 rr kernel: [] bond_enslave+0x3b4/0x5c1 [bonding]
Aug 18 09:48:03 rr kernel: [] fib6_clean_node+0x13/0x62 [ipv6]
Aug 18 09:48:03 rr kernel: [] bond_do_ioctl+0x2d9/0x323 [bonding]
Aug 18 09:48:03 rr kernel: [] dev_ioctl+0x105/0x4a2
Aug 18 09:48:03 rr kernel: [] fib6_clean_node+0x0/0x62 [ipv6]
Aug 18 09:48:03 rr kernel: [] vlan_device_event+0x11/0x110 [8021q]
Aug 18 09:48:03 rr kernel: [] dev_ifsioc+0x43c/0x449
Aug 18 09:48:03 rr kernel: [] bond_do_ioctl+0x0/0x323 [bonding]
Aug 18 09:48:03 rr kernel: [] dev_ioctl+0x392/0x4a2
Aug 18 09:48:03 rr kernel: [] udp_ioctl+0x0/0xc5
Aug 18 09:48:03 rr kernel: [] inet_ioctl+0xa0/0xa5
Aug 18 09:48:03 rr kernel: [] sock_ioctl+0x28c/0x2b4

Re: redhat4 +team +psp 7.4 error

Al Licause — Mon, 28 Aug 2006 08:45:49 GMT

What is your question ?

How did you enable debugging ?

Did you specifically enable half duplex as it reports in the log ?

100Mbps or 1000Mbps card runs at half duplex it usually means it failed to autonegotiate and it more than likely operating in a different mode than the network switch which will impede throughput.

Re: redhat4 +team +psp 7.4 error

Rick Beldin — Tue, 29 Aug 2006 13:04:30 GMT

Are you using the HP-supplied e100 driver? If so, try reverting back to the one supplied by RH. There appears to be an issue with the HP-supplied e100 driver and bonding that causes these messages.

The messages are coming from the scheduler (sched.c) when it detects that some ioctl()'s are not responding appropriately.

We detected this on a dl380 G2 with e100 cards running RHEL4 U4 and PSP 7.52. e100 package e100-3.4.14a-1-i386

Note that unless you took steps to save your ORIGINAL RH driver, you might have restore it from a backup.