RE: NIC lockup in 2.4.17 (SMP/APIC/Intel 82557)

[Robbert Kouprie]

Not much new, but still:

Today I got the same problem again with 2.4.18-pre3-ac2. Network
connections stuck, NFS mounts stuck. Bringing down/up the interface
doesn't help. Seems like the NIC is really in trouble here. Only a
reboot would bring the nick back in use.

Still no testcase though, and I have no idea on how to investigate this
:(
Can anyone give a hint as where to seek?

Regards,
- Robbert Kouprie

> -----Original Message-----
> From: James Bourne [mailto:[e-mail]] 
> Sent: dinsdag 22 januari 2002 15:54
> To: Robbert Kouprie
> Cc: [e-mail]; [e-mail]
> Subject: Re: NIC lockup in 2.4.17 (SMP/APIC/Intel 82557)
> 
> 
> On Tue, 22 Jan 2002, Robbert Kouprie wrote:
> 
> >
> > Jussi Laako wrote:
> >
> > > Robbert Kouprie wrote:
> > > >
> > > > Thanks for the quick reply :) Just checked it, and it's 
> in slot 2, so
> > > > that's not the problem. It doesn't share the HPT366 
> IRQ. This is my
> > > > /proc/interrupts:
> > >
> > > Driver is eepro100? I suspect there is something in 
> eepro100 driver that
> > > should be protected by a spinlock but is not. I haven't 
> got time to
> > > analyze it further, yet...
> > >
> > >  - Jussi Laako
> >
> > Yes, eepro100.c. Let me know if I can test something, 
> although I would
> > need a reproducible testcase also. Still doing some tests with high
> > network load, as this caused the similar lockup in the other thread.
> >
> > - Robbert
> >
> 
> Perhaps this will help.  Yesterday we had a strange error on 
> an eepro100
> NIC.  System is 4-way Xeon, 4G RAM, 4 eepro100 nics (2 in 
> use), Dell PE6400.
> Kernel is 2.4.17, no additional patches.  The system has not locked up
> though.
> 
> The error was
> eth0: can't fill rx buffer (force 0)!
> eth0: Tx ring dump,  Tx queue 3013060 / 3013060:
> eth0:     0 200ca000.
> eth0:     1 000ca000.
> eth0:     2 000ca000.
> eth0:     3 400ca000.
> eth0:  *= 4 000ca000.
> eth0:     5 000ca000.
> [... all the same ...]
> eth0:    30 000ca000.
> eth0:    31 000ca000.
> eth0: Printing Rx ring (next to receive into 2522947, dirty 
> index 2522946).
> eth0:     0 00000001.
> eth0: l   1 c0000001.
> eth0:  *  2 00000000.
> eth0:   = 3 00000001.
> eth0:     4 00000001.
> eth0:     5 00000001.
> [... all the same ...]
> eth0:    30 00000001.
> eth0:    31 00000001.
> 
> System has only 4 days uptime, eth0 output is:
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:3049032 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:3566542 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:100
>           RX bytes:580129470 (553.2 Mb)  TX bytes:2775810991 
> (2647.2 Mb)
>           Interrupt:26
> 
> /proc/interrupts
> loki:bash# cat /proc/interrupts
>            CPU0       CPU1       CPU2       CPU3
>   0:   10332192   10317829   10310969   10381268    
> IO-APIC-edge  timer
>   1:        292        307        250        273    
> IO-APIC-edge  keyboard
>   2:          0          0          0          0          
> XT-PIC  cascade
>   8:         26         32         30         27    IO-APIC-edge  rtc
>  17:          6          4          3          3   
> IO-APIC-level  aic7xxx
>  18:          2          5          2          7   
> IO-APIC-level  aic7xxx
>  22:     766030     765530     764892     765127   IO-APIC-level  eth1
>  23:     388900     388509     388022     388376   
> IO-APIC-level  megaraid
>  26:    1395108    1395240    1394961    1396555   IO-APIC-level  eth0
> NMI:          0          0          0          0
> LOC:   41347538   41347536   41347536   41347494
> ERR:          0
> MIS:          0
> 
> lspci
> loki:bash# lspci
> 00:00.0 Host bridge: ServerWorks CNB20HE Host Bridge (rev 21)
> 00:00.1 Host bridge: ServerWorks CNB20HE Host Bridge (rev 01)
> 00:00.2 Host bridge: ServerWorks: Unknown device 0006
> 00:00.3 Host bridge: ServerWorks: Unknown device 0006
> 00:04.0 VGA compatible controller: ATI Technologies Inc 3D 
> Rage IIC (rev 7a)
> 00:05.0 SCSI storage controller: Adaptec 7899P (rev 01)
> 00:05.1 SCSI storage controller: Adaptec 7899P (rev 01)
> 00:08.0 Ethernet controller: Intel Corporation 82557 
> [Ethernet Pro 100] (rev 08)
> 00:0f.0 ISA bridge: ServerWorks OSB4 South Bridge (rev 50)
> 00:0f.1 IDE interface: ServerWorks OSB4 IDE Controller
> 00:0f.2 USB Controller: ServerWorks OSB4/CSB5 OHCI USB 
> Controller (rev 04)
> 03:08.0 Ethernet controller: Intel Corporation 82557 
> [Ethernet Pro 100] (rev 08)
> 03:09.0 PCI bridge: Digital Equipment Corporation DECchip 
> 21154 (rev 05)
> 03:0a.0 Ethernet controller: Intel Corporation 82557 
> [Ethernet Pro 100] (rev 08)
> 03:0b.0 Ethernet controller: Intel Corporation 82557 
> [Ethernet Pro 100] (rev 08)
> 04:00.0 PCI bridge: Digital Equipment Corporation DECchip 
> 21154 (rev 05)
> 04:01.0 SCSI storage controller: Q Logic QLA12160 (rev 06)
> 05:00.0 RAID bus controller: American Megatrends Inc. 
> MegaRAID (rev 20)
> 
> 
> Although I haven't had much time to track this yet (was planning later
> today) I thought it might be related to the above...  If any other
> information would help, please let me know.
> 
> Regards
> James Bourne
> 
> -- 
> James Bourne, Supervisor Data Centre Operations
> Mount Royal College, Calgary, AB, CA
> www.mtroyal.ab.ca
> 
> **************************************************************
> ****************
> This communication is intended for the use of the recipient 
> to which it is
> addressed, and may contain confidential, personal, and or privileged
> information. Please contact the sender immediately if you are not the
> intended recipient of this communication, and do not copy, 
> distribute, or
> take action relying on it. Any communication received in error, or
> subsequent reply, should be deleted or destroyed.
> **************************************************************
> ****************
>