[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1317115409.2541.6.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC>
Date: Tue, 27 Sep 2011 11:23:29 +0200
From: Eric Dumazet <eric.dumazet@...il.com>
To: christoph.paasch@...ouvain.be
Cc: David Miller <davem@...emloft.net>, netdev@...r.kernel.org,
hkchu@...gle.com, ilpo.jarvinen@...sinki.fi, jhs@...atatu.com
Subject: Re: [PATCH net-next] tcp: unalias tcp_skb_cb flags and ip_dsfield
Le mardi 27 septembre 2011 à 11:38 +0300, Christoph Paasch a écrit :
> On Tuesday 27 September 2011 wrote David Miller:
> > From: Christoph Paasch <christoph.paasch@...ouvain.be>
> > Date: Tue, 27 Sep 2011 10:37:38 +0300
> >
> > > However, (as far as I can see) before Eric's change it was 42 bytes long
> > > and now it became 43 bytes.
> >
> > Then, I suggest that you go buy a pair of glasses. :-)
>
> Yeah, maybe it's time to... ;-)
>
> This two-byte hole, is it because "flags" and "sacked" are both __u8 and
> fields are aligned to 32-bit ?
Not exactly.
Reason for the hole is the following :
struct whatever {
u8 first;
u8 second;
u32 third;
};
"third" field has an alignement requirement of 4 bytes on most arches.
Compiler inserts a 2 bytes hole before "third" to respect this
requirement.
There is no hole between "first" and "second", since second has no
alignment requirement.
sizeof(struct whatever) = 8
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists