_unofficial_ tg3 rpms available

Ed Martin ed.martin at iop.org
Wed Nov 27 11:55:01 CST 2002

>> Though I should add that the network situation has changed with this server - we've recently moved it from a 
>100MB switch to a GB switch.  And I haven't retried the crashing version since the move.
>> I can confirm that your kernel does show a tg3 version of 1.2.
>Thanks much for testing.

I've good news and bad news.  The good news is that the 2650 stayed up overnight, which the old tg3 never managed before.  I've also applied your kernel to a 2550 which showed the same problem, it's now been up for an hour with no problem.

The bad news is that I tried rebooting the 2650 into the -18 kernel and it also hasn't hung yet.  This is bad because it always did before.  My best guess is that move of network switch has had an effect - was connected to a 100Mbs Extreme Summit 48, now on a GB Summit7i.  I remain happy, but it probably means that my test isn't really valid.  I suggest you ensure that others with the problem have a positive result before being confident that the problem is resolved.  Sorry.

>> Can anyone tell me which broadcom driver you'd recommend in the latest kernel please?
>> (assuming this problem is fixed)
>> Is it bcm (v2.2) or tg3 (v1.2)?
>> Do you think that the answer will be the same in 6 months time?
>Red Hat is strenuously recommending tg3, for both now and the future.
>Personally I feel the same way, but I am also biased, being one of the
>two tg3 authors ;-)
>In addition to concrete features like better performance and better
>hub/switch compatibility, the driver is much better from a source
>code perspective, which makes it easier to maintain and review.
>BroadCom's driver has a history of bugs, both major and minor, that
>is often due to their lack of experience with Linux net drivers.  From
>a source code perspective, their source code has multiple abstraction
>layers which allow them to share their source code across multiple
>OS's -- and also makes it difficult to review and validate.
>Their abstraction layer has also been proven time and again as a
>source of bugs.

Thanks, that is very helpful.  I'm persuaded - we'll stay with the tg3.



Ed Martin
Head of Systems Technology and Support
IOP Publishing Ltd
Dirac House, Temple Back
Bristol  BS1 6BE
ddi: +44 (0)117 930 1102
www:  http://www.iop.org

Institute of Physics
Registered charity No. 293851
76 Portland Place, London, W1B 1NT, England

IOP Publishing Limited
Registered in England under Registration No 467514.
Registered Office: Dirac House, Temple Back, Bristol BS1 6BE England

This e-mail message has been checked by MIMEsweeper using
F-Secure Anti-Virus for the presence of computer viruses.

More information about the Linux-PowerEdge mailing list