[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <063D6719AE5E284EB5DD2968C1650D6D1C9E258D@AcuExch.aculab.com>
Date: Thu, 30 Oct 2014 17:09:31 +0000
From: David Laight <David.Laight@...LAB.COM>
To: 'Or Gerlitz' <ogerlitz@...lanox.com>
CC: "David S. Miller" <davem@...emloft.net>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
Matan Barak <matanb@...lanox.com>,
Amir Vadai <amirv@...lanox.com>,
Saeed Mahameed <saeedm@...lanox.com>,
Shani Michaeli <shanim@...lanox.com>
Subject: RE: [PATCH net-next 7/8] net: Add calaulation of non folded IPV6
pseudo header checksum
> OK, talking to Matan, hecame up with this super-quick (compile tested
> only) alternative, is
> this what you were advocating for?
> {
> + sum = csum_partial(saddr, sizeof(saddr->in6_u.u6_addr32), sum);
> + sum = csum_partial(saddr, sizeof(daddr->in6_u.u6_addr32), sum);
I'm pretty sure your 'saddr' and 'daddr' are adjacent.
Whether you can prove that is another matter.
> + sum = csum_add(sum, (__force __wsum)htonl(len));
> + sum = csum_add(sum, (__force __wsum)htons(proto));
> +
> + return sum;
> +}
On 64 bit systems you probably want to end up with something akin to:
__u64 a, b, c, d;
a = saddr[0] + saddr[1];
b = saddr[2] + saddr[3];
c = daddr[0] + daddr[1];
d = daddr[2] + daddr[3];
a += b;
c += d;
a += c;
and then collapse down the 64bit value.
However if you write the above in proper C, gcc will probably
convert it to long dependence chain.
An architecture specific csum_partial() probably manages to DTRT.
David
--
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