From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eyal Birger Subject: Re: [PATCH net-next v3 3/4] net: use skb->priority for overloading skb->dropcount and skb->reserved_tailroom instead of skb->mark Date: Tue, 24 Feb 2015 12:10:40 +0200 Message-ID: References: <1424713924-6821-2-git-send-email-eyal.birger@gmail.com> <1424713924-6821-4-git-send-email-eyal.birger@gmail.com> <20150223205633.7fbd08d4@halley> <20150223.164801.842537100812023841.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Cc: Shmulik Ladkani , Eric Dumazet , netdev@vger.kernel.org To: David Miller Return-path: Received: from mail-yh0-f44.google.com ([209.85.213.44]:44944 "EHLO mail-yh0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751314AbbBXKKl (ORCPT ); Tue, 24 Feb 2015 05:10:41 -0500 Received: by yhaf73 with SMTP id f73so13638209yha.11 for ; Tue, 24 Feb 2015 02:10:41 -0800 (PST) In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Tue, Feb 24, 2015 at 8:58 AM, Eyal Birger wrote: > On Mon, Feb 23, 2015 at 11:48 PM, David Miller wrote: >> From: Shmulik Ladkani >> Date: Mon, 23 Feb 2015 20:56:33 +0200 >> >>> May I suggest to unalias 'mark' out of the union? >> >> Only in exchange for finding a way to remove completely another object >> of the same size from sk_buff. >> >> I refuse to let it grow any more, it's already way too big. > > I agree that aliasing skb->priority with skb->dropcount and > skb->reserved_tailroom is > less than optimal. > > I can suggest an alternative patch following Eric Dumazet's suggestion in which > skb->dropcount is removed and skb->cb[] is used instead for keeping dropcount. > > That would require protocol family implementations to access skb->cb[] > via a common > accessor or use a common offset #define from the start of skb->cb[]. > Well, gave it a shot... it looks like several protocol families (packet, rxrpc, bluetooth) do not have room in skb->cb[] for the dropcount - at least on my 64 bit machine.