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

Bug#640091: linux-image-2.6.32-5-amd64: Performance problem with the e1000e driver



On Fri, 2011-09-02 at 10:36 +0200, Thorsten Giese wrote:
> Package: linux-2.6
> Version: 2.6.32-36~~test1
> Severity: important
> 
> I discussed this already with Ben Hutchings, please see this thread: 
> 
> http://lists.debian.org/debian-kernel/2011/08/msg01267.html

To summarise the information you've given about different
configurations:

Affected:
1. e1000e on 2.6.32 connecting to site on different network segment
2. e1000e backport from 2.6.38 connecting to site on different network
   segment

Not affected:
3. e1000e on 2.6.26
4. e1000e on 2.6.32 connecting to site on same network segment
5. e1000e on 2.6.32 during first few minutes after boot
6. e1000e on 2.6.39
7. r8169 on 2.6.32 (not same system)

Case 7 strongly suggests that this issue is related to the driver or the
environment and not the stack (or not only the stack).  But then there
don't seem to be any relevant changes to this driver between 2.6.38 and
2.6.39, so why the difference beween cases 2 and 6?

Case 5 is interesting.  Is there some service started at boot, that may
start to access the network after a few minutes?  In cases 3 and 6, did
you test in the first few minutes after boot, or did you wait a while?

Ben.

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: