: RT8139 Too much work at interrupt, IntrStatus=....
Hi all,
yesterday I upgraded a firewall/router of a *busy* customer site from version 2.4.7 -> 2.4.17. Only 1-2 seconds after the reboot the network card to the DMZ stopped working. I noticed this messages in /var/log/messages:
Jan 8 17:07:45 liquice kernel: 8139too: rx stop wait too long Jan 8 17:07:46 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0050. Jan 8 17:07:46 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0010. Jan 8 17:07:49 liquice kernel: 8139too: rx stop wait too long Jan 8 17:07:49 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0001. Jan 8 17:07:51 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0001. Jan 8 17:07:57 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0050. Jan 8 17:07:57 liquice kernel: eth1: Setting 100mbps half-duplex based on auto-negotiated partner ability 40a1. Jan 8 17:08:09 liquice kernel: NETDEV WATCHDOG: eth1: transmit timed out Jan 8 17:08:09 liquice kernel: eth1: Setting 100mbps half-duplex based on auto-negotiated partner ability ffff.
after a reboot the same problem occured.
Switching back to v2.4.7 solved the problem so I don't think that it's a hardware issue.
So has anybody an idea whats happening here? Hardware info below. Let me know if you need any additional info.
Thanks
Dennis
liquice:~# lspci 00:00.0 Host bridge: VIA Technologies, Inc.: Unknown device 0305 (rev 03) 00:01.0 PCI bridge: VIA Technologies, Inc.: Unknown device 8305 00:04.0 ISA bridge: VIA Technologies, Inc. VT82C686 [Apollo Super] (rev 40) 00:04.1 IDE interface: VIA Technologies, Inc. VT82C586 IDE [Apollo] (rev 06) 00:04.2 USB Controller: VIA Technologies, Inc. VT82C586B USB (rev 16) 00:04.3 USB Controller: VIA Technologies, Inc. VT82C586B USB (rev 16) 00:04.4 Host bridge: VIA Technologies, Inc. VT82C686 [Apollo Super ACPI] (rev 40) 00:0a.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RT8139 (rev 10) 00:0c.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RT8139 (rev 10) 00:0d.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RT8139 (rev 10) 00:11.0 Unknown mass storage controller: Promise Technology, Inc.: Unknown device 0d30 (rev 02) 01:00.0 VGA compatible controller: ATI Technologies Inc Rage XL AGP (rev 65)
liquice:~# cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 4 model name : AMD Athlon(tm) Processor stepping : 2 cpu MHz : 1210.814 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 mmx fxsr syscall mmxext 3dnowext 3dnow bogomips : 2411.72
any reason why you posted that to me and not to the list?
Quoted 5 lines · Mark Hahn
Quoted 1 line
yesterday I upgraded a firewall/router of a *busy* customer site
...
Quoted 1 line
Jan 8 17:07:45 liquice kernel: 8139too: rx stop wait too longsurely a "*busy*" can pay an extra $20 and get a real network card...
sure, but as I said: with kernel v2.4.7 it works! So I guess it must be a bug in something later than version 2.4.7.
Quoted 3 lines
Quoted 1 line
Jan 8 17:07:46 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0050.did you happen to enable the use of local apic (in the new kernel's config)?
nope. I'll try that.
Dennis
Quoted 16 lines · Dennis Schoen
yesterday I upgraded a firewall/router of a *busy* customer site from version 2.4.7 -> 2.4.17. Only 1-2 seconds after the reboot the network card to the DMZ stopped working. I noticed this messages in
/var/log/messages:Jan 8 17:07:45 liquice kernel: 8139too: rx stop wait too long Jan 8 17:07:46 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0050. Jan 8 17:07:46 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0010. Jan 8 17:07:49 liquice kernel: 8139too: rx stop wait too long Jan 8 17:07:49 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0001. Jan 8 17:07:51 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0001. Jan 8 17:07:57 liquice kernel: eth1: Too much work at interrupt, IntrStatus=0x0050. Jan 8 17:07:57 liquice kernel: eth1: Setting 100mbps half-duplex based on auto-negotiated partner ability 40a1. Jan 8 17:08:09 liquice kernel: NETDEV WATCHDOG: eth1: transmit timed out Jan 8 17:08:09 liquice kernel: eth1: Setting 100mbps half-duplex based on auto-negotiated partner ability ffff.
Quoted 4 lines
after a reboot the same problem occured.
Switching back to v2.4.7 solved the problem so I don't think that it's a hardware issue.
Hi dennis, I'm getting these messages too with an identical card... 8139 (rev 10)... seems to have started happening after I went to 2.4.17... strangely enough I'm getting very similar m�ssages on the machine right beside it.. running 2.4.13, SMP, and a 3com 3c905b (Boomerang).
Anyone have any insight on this one?
Signature · 4 lines
Patrick Cole - Debian Developer <[email protected]> - John Curtin, ANU <[email protected]> - Linear-G Network Solutions <[email protected]> - PGP 1024R/60D74C7D C8E0BC7969BE7899AA0FEB16F84BFE5A
Quoted 5 lines · Patrick Cole
Hi dennis, I'm getting these messages too with an identical card... 8139 (rev 10)... seems to have started happening after I went to 2.4.17... strangely enough I'm getting very similar mæssages on the machine right beside it.. running 2.4.13, SMP, and a 3com 3c905b (Boomerang).
The problem with the 8139 is the fact that the chip design is really dumb, when I say dumb here, I mean *really* dumb ;-(
The result is that when the card is bombarded with a lot of small packets it can livelock because the queues fill up faster than the kernel code can empty and process them.
I made some changes to improve the situation in the driver, for 2.2 kernels, which I have submitted to Jeff and he plans to integrated them into the next release.
My patched drivers are available from
http://www.wildopensource.com/proj/download/SG_drivers.html however I haven't tested them on 2.4.x, so I am not going to bet on them working.
Cheers,
Jes