In article [off-list ref] (at Fri, 20 Feb 2004 18:03:11 +0000), Andrew Walrond [off-list ref] says:
I'm getting loads of these messages in dmesg.
This is a vanilla 2.6.3 kernel with ACPI disabled with acpi=off boot
parameter. No preempt, APIC and IOAPIC enabled
Its a single 2Ghz celeron with Via chipset and Rhine II ethernet. It is under
no load at all. ssh only.
dmesg is short and follows; Any knowledge apprieciated!
:
eth0: Too much work at interrupt, status=0x00000001.
I've got this several times, randomly, and
I had to go to the console to reboot the machine.
Well, sorry, I haven't tracked it down yet...
Others?
--
Hideaki YOSHIFUJI @ USAGI Project [off-list ref]
GPG FP: 9022 65EB 1ECF 3AD1 0BDF 80D8 4807 F894 E062 0EEA
From: Andrew Walrond <hidden> Date: 2004-02-20 19:15:09
On Friday 20 Feb 2004 6:20 pm, YOSHIFUJI Hideaki / 吉藤英明 wrote:
In article [off-list ref] (at Fri, 20 Feb 2004
18:03:11 +0000), Andrew Walrond [off-list ref] says:
I've got this several times, randomly, and
I had to go to the console to reboot the machine.
Well the machine and network (at least with just ssh access) is working fine
I left the machine (which is a remote rack-mounted server) for an hour and now
there is perhaps 120 of these warnings. Net is still working though.
Unless anyone has any ideas, I'll move back to 2.4 on this machine
Andrew
From: Stephen Hemminger <hidden> Date: 2004-02-20 19:43:20
On Fri, 20 Feb 2004 19:15:09 +0000
Andrew Walrond [off-list ref] wrote:
On Friday 20 Feb 2004 6:20 pm, YOSHIFUJI Hideaki / ____________ wrote:
quoted
In article [off-list ref] (at Fri, 20 Feb 2004
18:03:11 +0000), Andrew Walrond [off-list ref] says:
quoted
I've got this several times, randomly, and
I had to go to the console to reboot the machine.
Well the machine and network (at least with just ssh access) is working fine
I left the machine (which is a remote rack-mounted server) for an hour and now
there is perhaps 120 of these warnings. Net is still working though.
Unless anyone has any ideas, I'll move back to 2.4 on this machine
Andrew
The driver needs to be converted to NAPI to sustain high network loads.
The existing driver allocates and copies each receive packet in the interrupt routine.
See the recent revision of 8139too to see what that entails.
From: Andrew Walrond <hidden> Date: 2004-02-20 20:08:44
On Friday 20 Feb 2004 7:34 pm, Stephen Hemminger wrote:
The driver needs to be converted to NAPI to sustain high network loads.
The existing driver allocates and copies each receive packet in the
interrupt routine.
See the recent revision of 8139too to see what that entails.
Yes but...there were no high network loads. Just me typing into ssh.
STOP PRESS....FIX...
That may well be so, but I rebuilt the kernel with the optional MMIO on the
rhine driver, and the messages have disappeared.
I have also used wget to fetch a 100Mb file at an average 1.02Mb/s and it
worked flawlessly and still no warnings in dmesg.
Andrew Walrond