Bug 210274

Summary: forcedeth driver faulty when on 1Gbps
Product: Red Hat Enterprise Linux 3 Reporter: Colin Coe <colin.coe>
Component: kernelAssignee: Red Hat Kernel Manager <kernel-mgr>
Status: CLOSED WONTFIX QA Contact: Brian Brock <bbrock>
Severity: high Docs Contact:
Priority: medium    
Version: 3.8CC: dhoward
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2007-10-19 18:40:49 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
sysreport of an effected workstation
none
Sysreport of the Stuttgart Workstation which does not have the problems none

Description Colin Coe 2006-10-11 07:24:37 UTC 
Description of problem:
Linux workstations running on a fast network (1Gb) whilst using Landmark 
SeisWorks or Shell 123DI have sluggish performance when interpreting a horizon 
(interp take time to register and does not keep up with cursor clicks)
SeisWorks takes long time to scroll to a new line


Version-Release number of selected component (if applicable):
RHEL3u8 (2.4.21-47)

How reproducible:


Steps to Reproduce:
1. Open Seisworks project
2. Load horizons
3. Start interpreting
  
Actual results:
Interpreting is slow (up to 5-10 seconds to finish interpreting)

Expected results:
Intpreting should take well under a second to complete. 

Additional info:
I disabled the on board NVidia NIC and installed an Intel EPro 1000 and 
repeated the above tests and found no performance problems.  When using a stock 
xw9300 (i.e. using the forcedeth driver) on a 100Mbps port does not show any of 
the above symptoms.

The hardware is HP xw9300 Workstation and a sysreport from a couple of days ago 
is attached.
Comment 1 Colin Coe 2006-10-11 07:24:37 UTC 
Created attachment 138223 [details]
sysreport of an effected workstation
Comment 2 Colin Coe 2006-10-11 22:46:35 UTC 
Created attachment 138290 [details]
Sysreport of the Stuttgart Workstation which does not have the problems
Comment 3 Colin Coe 2006-10-11 22:48:09 UTC 
Attached is a sysreport from an engineering sample Sun Stuttgart workstation 
which is similar in spec to the xw9300 including having onboard NVidia NICs.
Comment 4 Prarit Bhargava 2007-09-14 15:48:21 UTC 
Colin,

Is this still a problem?

P.
Comment 5 Colin Coe 2007-09-16 23:04:13 UTC 
Yes this is still a problem on stock RHEL3 kernels.  The problems only seem 
apparent when doing *lots* of very small updates to a file on an NFS server 
(Solaris in this case) when the card is operating in 1G mode.  100M is fine.  
Applying the NVidia patches to take the driver to v0.60+ fixes the problem.

We are now rolling our own kernels to get around these problems.  

I know EL5 uses 0.56 of the driver but I can't test this as the apps we run 
don't work properly outside EL3.

CC
Comment 6 RHEL Program Management 2007-10-19 18:40:49 UTC 
This bug is filed against RHEL 3, which is in maintenance phase.
During the maintenance phase, only security errata and select mission
critical bug fixes will be released for enterprise products. Since
this bug does not meet that criteria, it is now being closed.
 
For more information of the RHEL errata support policy, please visit:
http://www.redhat.com/security/updates/errata/
 
If you feel this bug is indeed mission critical, please contact your
support representative. You may be asked to provide detailed
information on how this bug is affecting you.
Comment 7 Colin Coe 2007-10-21 03:22:32 UTC 
Nice, especially considering it was logged before EL3 went into maintenance 
phase.