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

e1000 tx hangs after kernel 2.6.11 to 2.6.16 update



Hello all.
I have Redhat Fedora Core 4 (x86_64, smp kernel) installed on pure
intel computer - Intel D945PSN board, Intel Processor, onboard
Intel 1000Mbs NIC (547V chip), 1GB RAM. Cables and switches are
100Mbs.
This PC serves as a Linux-Samba file server for XP Pro stations.
All updates are from Redhat only (YUM UPDATE).
No extra 3rd party updates except installed webmin.
After updating kernel from 2.6.11-1.1369_FC4smp to
2.6.16-1.2108_FC4smp, there is a problem with onboard NIC. When the
network traffic grows, there are very often "Detected Tx Unit Hang"
messages like this one:

May 19 10:43:37 bigorange kernel: e1000: eth0: e1000_clean_tx_irq:
Detected Tx Unit Hang May 19 10:43:37 bigorange kernel:   Tx Queue 
           <0>
May 19 10:43:37 bigorange kernel:   TDH                  <4c>
May 19 10:43:37 bigorange kernel:   TDT                  <9e>
May 19 10:43:37 bigorange kernel:   next_to_use          <9e>
May 19 10:43:37 bigorange kernel:   next_to_clean        <4c>
May 19 10:43:37 bigorange kernel: buffer_info[next_to_clean]
May 19 10:43:37 bigorange kernel:   time_stamp          
<10103d852> May 19 10:43:37 bigorange kernel:   next_to_watch      
 <50>
May 19 10:43:37 bigorange kernel:   jiffies             
<10103d9fa> May 19 10:43:37 bigorange kernel:  
next_to_watch.status <0>
May 19 10:43:39 bigorange kernel: e1000: eth0: e1000_clean_tx_irq:
Detected Tx Unit Hang May 19 10:43:39 bigorange kernel:   Tx Queue 
           <0>
May 19 10:43:39 bigorange kernel:   TDH                  <4c>
May 19 10:43:39 bigorange kernel:   TDT                  <9e>
May 19 10:43:39 bigorange kernel:   next_to_use          <9e>
May 19 10:43:39 bigorange kernel:   next_to_clean        <4c>
May 19 10:43:39 bigorange kernel: buffer_info[next_to_clean]
May 19 10:43:39 bigorange kernel:   time_stamp          
<10103d852> May 19 10:43:39 bigorange kernel:   next_to_watch      
 <50>
May 19 10:43:39 bigorange kernel:   jiffies             
<10103dbee> May 19 10:43:39 bigorange kernel:  
next_to_watch.status <0>
May 19 10:43:41 bigorange kernel: e1000: eth0: e1000_clean_tx_irq:
Detected Tx Unit Hang May 19 10:43:41 bigorange kernel:   Tx Queue 
           <0>
May 19 10:43:41 bigorange kernel:   TDH                  <4c>
May 19 10:43:41 bigorange kernel:   TDT                  <9e>
May 19 10:43:41 bigorange kernel:   next_to_use          <9e>
May 19 10:43:41 bigorange kernel:   next_to_clean        <4c>
May 19 10:43:41 bigorange kernel: buffer_info[next_to_clean]
May 19 10:43:41 bigorange kernel:   time_stamp          
<10103d852> May 19 10:43:41 bigorange kernel:   next_to_watch      
 <50>
May 19 10:43:41 bigorange kernel:   jiffies             
<10103dde2> May 19 10:43:41 bigorange kernel:  
next_to_watch.status <0>
May 19 10:43:42 bigorange kernel: NETDEV WATCHDOG: eth0: transmit
timed out May 19 10:43:44 bigorange kernel: e1000: eth0:
e1000_watchdog_task: NIC Link is Up 100 Mbps Full Duplex

Few seconds after the network connections is lost, the NIC restarts
itself, and all is working again, till next hang. Depends on the
traffic - the higher traffic, the more hangs, the more network
freezes. With kernel 2.6.11-1.1369_FC4smp it worked fine.

Is there any help please?
Thanks a lot, Jan Kusy


-- 
NOVINKA: SUPER SPALOVAC - Zbav se kil EXTRA!
Carnitin+Chrom_500ml a Synephrine Thermo Tabs
- vysoce účinné SPALOVANI TUKU, nyni SLEVA 200 Kc
http://www.sportobchod.cz/survival/spalovani-tuku&redir=32




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