Re: [PATCH 2.6.14-rc2] aoe [1/3]: explicitly set minimum packetlength to ETH_ZLEN

[Ed L Cashin]

Alan Cox <alan@xxxxxxxxxxxxxxxxxxx> writes:

> On Iau, 2005-09-29 at 12:45 -0400, Ed L. Cashin wrote:
>> Signed-off-by: "Ed L. Cashin" <ecashin@xxxxxxxxxx>
>> 
>> Explicitly set the minimum packet length to ETH_ZLEN and zero the
>> packet data.
>
> You still haven't explained why this is neccessary. The drivers should
> be doing it for you.

I did respond to your earlier request post, but I forgot to make sure
you were in the CC list, sorry.  I've quoted my response below.

Ed L Cashin <ecashin@xxxxxxxxxx> writes:

> Alan Cox <alan@xxxxxxxxxxxxxxxxxxx> writes:
...
>> The network driver is supposed to pad frames if the hardware cannot and
>> to blank the spare bits. 
>
> Ah ha.
>
>> If it isn't occurring please try and trace down
>> the offender.
>
> My colleague Sam observed problems with the e1000 driver in the
> 2.6.11.4-21.9-smp kernel from Suse 9.3 and also the e1000 driver in
> 2.6.12-1.1398_FC4smp from Fedora Core 4.  
>
> The problems aren't fully characterized, but AoE ATA read packets
> appeared to be getting dropped and/or corrupted.
>
> When using the tg3 driver instead of e1000 the problems went away, and
> making the aoe driver alloc_skb with a minimum length of ETH_ZLEN also
> made the problems go away.

We suspect that the e1000 driver is misbehaving when given short
packets, but we have not had time to pinpoint what part of the e1000
driver is involved or what the specific problem is.

-- 
  Ed L Cashin <ecashin@xxxxxxxxxx>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/