[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20140728.171213.1219772609929469928.davem@davemloft.net>
Date: Mon, 28 Jul 2014 17:12:13 -0700 (PDT)
From: David Miller <davem@...emloft.net>
To: pablo@...filter.org
Cc: ast@...mgrid.com, dborkman@...hat.com, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, willemb@...gle.com,
netfilter-devel@...r.kernel.org
Subject: Re: [PATCH net-next] net: filter: rename 'struct sk_filter' to
'struct bpf_prog'
From: Pablo Neira Ayuso <pablo@...filter.org>
Date: Mon, 28 Jul 2014 23:45:52 +0200
> By renaming this, you're not fixing up things the semantics. It seems
> to me you just want to find a quick path to solve inconsistencies in
> your code.
Agreed, this looks just like messing around with naming to me.
But to the original issue, that of xt_bpf, I wonder about a few things:
1) If we have a kernel pointer embedded in a user provided datastructure,
what takes care of 32-bit compat applications uploading xt_bpf rules
on a 64-bit kernel? Won't the size be wrong or does it not matter
and is in some way helped by that 8-byte alignment thing there?
2) The user can't care about the type of "filter" in xt_bpf_info, so
we can use whatever name we want for the type.
Therefore you can just do something like:
struct bpf_prog;
struct xt_bpf_info {
__u16 bpf_program_num_elem;
struct sock_filter bpf_program[XT_BPF_MAX_NUM_INSTR];
/* only used in the kernel */
struct bpf_prog *filter __attribute__((aligned(8)));
};
and then you won't need any casting.
--
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