On Thu, Feb 26, 2015 at 5:32 PM, Shmulik Ladkani
[off-list ref] wrote:

On Thu, 26 Feb 2015 12:48:14 +0200 Eyal Birger [off-list ref] wrote:

quoted

After giving it some additional thought and research, I don't think
this is necessary.

The sizeof operator on struct sock_skb_cb would give a padded result
which would take care of the structure alignment.

This is under the assumption that skb->cb[] size is of a proper multiple (which
I think is rather safe).

Yes, you are correct.

With the assumption that skb->cb and 'cb' size are properly aligned and
sized (compile time asserted, btw?), you can simply:

struct your_thing {     // name it
        u32 dropcount;
};

struct sock_skb_cb {
        /* protocol families specifc CBs */
        u8 pf_reserved[PAD];
        struct your_thing sock_cb;
};

where:

#define PAD (FIELD_SIZEOF(struct sk_buff, cb) - sizeof(struct your_thing))

struct sock_skb_cb as I see it represents the sock generic metadata
required in the skb->cb[] and its placement at the end of skb->cb[]
is an implementation detail - so it, in my view, is what you referred to as
"struct your_thing".

Coding-wise I would rather avoid the additional enclosing struct and stick
with the proposed implementation.

Eyal.