netdev
[Top] [All Lists]

RE: conflicting alignment requirements

To: "Alexey Kuznetsov" <kuznet@xxxxxxxxxxxxx>
Subject: RE: conflicting alignment requirements
From: "Jacob Avraham" <jacoba@xxxxxxxxx>
Date: Sun, 29 Jul 2001 20:17:41 +0200
Cc: <netdev@xxxxxxxxxxx>
Importance: Normal
In-reply-to: <200107260647.KAA00350@mops.inr.ac.ru>
Sender: owner-netdev@xxxxxxxxxxx
>
> Hello!
>
> > > > On the  other hand, some upper layer code, like tc
> > > > (for non x86/68k), checks if the IP header is 4 byte align,
>
> Do you know any architecture except for these ones which are able
> to generate unaligned packets? I do not. And if you do, please,
> enlighten me about them.
>
I was referring to packets coming from the Tulip h/w (not generated by the
CPU),
on which the IP header is alway unalinged. If the driver is configured NOT
to
copy the packet to a fresh skb (rx_copybreak = 0), the packet will
traverse the net layer with unalinged IP header.

-Jacob


<Prev in Thread] Current Thread [Next in Thread>