Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 14 Dec 2016 19:06:57 +0100
From: "Jason A. Donenfeld" <Jason@...c4.com>
To: David Miller <davem@...emloft.net>
Cc: David Laight <David.Laight@...lab.com>, Netdev <netdev@...r.kernel.org>, 
	kernel-hardening@...ts.openwall.com, Andi Kleen <ak@...ux.intel.com>, 
	LKML <linux-kernel@...r.kernel.org>, 
	Linux Crypto Mailing List <linux-crypto@...r.kernel.org>
Subject: Re: [PATCH v2 3/4] secure_seq: use siphash24 instead of md5_transform

Hi David,

On Wed, Dec 14, 2016 at 6:56 PM, David Miller <davem@...emloft.net> wrote:
> Just marking the structure __packed, whether necessary or not, makes
> the compiler assume that the members are not aligned and causes
> byte-by-byte accesses to be performed for words.
> Never, _ever_, use __packed unless absolutely necessary, it pessimizes
> the code on cpus that require proper alignment of types.

Oh, jimminy cricket, I did not realize that it made assignments
byte-by-byte *always*. So what options am I left with? What
immediately comes to mind are:

1)

struct {
    u64 a;
    u32 b;
    u32 c;
    u16 d;
    u8 end[];
} a = {
    .a = a,
    .b = b,
    .c = c,
    .d = d
};
siphash24(&a, offsetof(typeof(a), end), key);

2)

u8 bytes[sizeof(u64) + sizeof(u32) * 2 + sizeof(u16)];
*(u64 *)&bytes[0] = a;
*(u32 *)&bytes[sizeof(u64)] = b;
*(u32 *)&bytes[sizeof(u64) + sizeof(u32)] = c;
*(u16 *)&bytes[sizeof(u64) + sizeof(u32) * 2] = d;
siphash24(bytes, sizeof(bytes), key);


Personally I find (1) a bit neater than (2). What's your opinion?

Jason

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.