Bug 6084 - sky2 hangs on large data transfers
Summary: sky2 hangs on large data transfers
Status: RESOLVED PATCH_ALREADY_AVAILABLE
Alias: None
Product: Drivers
Classification: Unclassified
Component: Network (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: Stephen Hemminger
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-02-15 22:25 UTC by Wolfgang Hoffmann
Modified: 2008-05-13 04:34 UTC (History)
3 users (show)

See Also:
Kernel Version: 2.6.16-rc3
Subsystem:
Regression: ---
Bisected commit-id:


Attachments
config (36.13 KB, text/plain)
2006-02-15 22:26 UTC, Wolfgang Hoffmann
Details
lspci -vx (11.78 KB, text/plain)
2006-02-15 22:27 UTC, Wolfgang Hoffmann
Details
dmesg (24.82 KB, text/plain)
2006-02-15 22:29 UTC, Wolfgang Hoffmann
Details

Description Wolfgang Hoffmann 2006-02-15 22:25:21 UTC
Most recent kernel where this bug did not occur: none yet  
Distribution: Ubuntu  
Hardware Environment: Marvel 88E8053 onboard NIC 
Software Environment:  
Problem Description: 
 
When transfering large amounts of data (e.g. rsyncing between two local 
machines), network hangs after some time (few minutes). dmesg doesn't give any 
messages. Removing the sky2 module and modprobing it back in cures the hang, 
i.e. transfer continues until the next hang. 
  
Steps to reproduce:  
 
As above
Comment 1 Wolfgang Hoffmann 2006-02-15 22:26:29 UTC
Created attachment 7364 [details]
config
Comment 2 Wolfgang Hoffmann 2006-02-15 22:27:13 UTC
Created attachment 7365 [details]
lspci -vx
Comment 3 Wolfgang Hoffmann 2006-02-15 22:29:31 UTC
Created attachment 7366 [details]
dmesg

The last messages in dmesg come from rmmod and modprobing the sky2 module.

Please tell me what more info I should provide.
Comment 4 Wolfgang Hoffmann 2006-02-27 08:36:44 UTC
2.6.16-rc5 still does hang OOTB, but reverting commit 
 
4d52b48b43d0d1d5959fa722ee0046e3542e5e1b is first bad commit 
    [PATCH] sky2: support msi interrupt (revised) 
 
_or_ turning off msi via module parameter (modprobe sky2.ko disable_msi=1) 
fixes the problem. 
 
Unless a fix for msi is at hand, may I suggest for 2.6.16 to revert this commit 
or switch the default to disable_msi=1? 
 
Comment 5 Stephen Hemminger 2006-03-06 22:14:27 UTC
MSI reverted in current drive 2.6.16.

Note You need to log in before you can comment on or make changes to this bug.