From mboxrd@z Thu Jan  1 00:00:00 1970
From: Stephen Hemminger <stephen@networkplumber.org>
Subject: Re: [iproute2 net-next] ip route: Add RTM_F_LOOKUP_TABLE flag and
 show table id
Date: Mon, 21 Sep 2015 14:19:29 -0700
Message-ID: <20150921141929.681ebf69@urahara>
References: <1442859588-14831-1-git-send-email-dsa@cumulusnetworks.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Cc: netdev@vger.kernel.org
To: David Ahern <dsa@cumulusnetworks.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-pa0-f53.google.com ([209.85.220.53]:36498 "EHLO
	mail-pa0-f53.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1752845AbbIUVTR (ORCPT
	<rfc822;netdev@vger.kernel.org>); Mon, 21 Sep 2015 17:19:17 -0400
Received: by padbj2 with SMTP id bj2so1966726pad.3
        for <netdev@vger.kernel.org>; Mon, 21 Sep 2015 14:19:16 -0700 (PDT)
In-Reply-To: <1442859588-14831-1-git-send-email-dsa@cumulusnetworks.com>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On Mon, 21 Sep 2015 11:19:48 -0700
David Ahern <dsa@cumulusnetworks.com> wrote:

> Currently 'ip route get' does not show the table the lookup result comes
> from and prior to kernel commit c36ba6603a11 the response from the kernel
> was hardcoded to the main table. From the discussion this appears to be
> a leftover from the route cache where the cached entry lost the table id
> and so the result was hardcoded to main table.
> 
> c36ba6603a11 added the RTM_F_LOOKUP_TABLE flag to maintain that behavior
> but to allow new tools to ask for the actual table id for the lookup.
> This patch adds that flag to ip route get request and if the result is
> not the main table shows the table id.
> 
> Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
> ---
>  ip/iproute.c              | 6 ++++--
>  1 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/ip/iproute.c b/ip/iproute.c
> index 8f49e6289003..bae43d5d8fb6 100644
> --- a/ip/iproute.c
> +++ b/ip/iproute.c
> @@ -421,9 +421,9 @@ int print_route(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
>  	if (tb[RTA_OIF] && filter.oifmask != -1)
>  		fprintf(fp, "dev %s ", ll_index_to_name(*(int*)RTA_DATA(tb[RTA_OIF])));
>  
> +	if (table && (table != RT_TABLE_MAIN || show_details > 0) && !filter.tb)
> +		fprintf(fp, " table %s ", rtnl_rttable_n2a(table, b1, sizeof(b1)));
>  	if (!(r->rtm_flags&RTM_F_CLONED)) {
> -		if ((table != RT_TABLE_MAIN || show_details > 0) && !filter.tb)
> -			fprintf(fp, " table %s ", rtnl_rttable_n2a(table, b1, sizeof(b1)));
>  		if ((r->rtm_protocol != RTPROT_BOOT || show_details > 0) && filter.protocolmask != -1)
>  			fprintf(fp, " proto %s ", rtnl_rtprot_n2a(r->rtm_protocol, b1, sizeof(b1)));
>  		if ((r->rtm_scope != RT_SCOPE_UNIVERSE || show_details > 0) && filter.scopemask != -1)
> @@ -1638,6 +1638,8 @@ static int iproute_get(int argc, char **argv)
>  	if (req.r.rtm_family == AF_UNSPEC)
>  		req.r.rtm_family = AF_INET;
>  
> +	req.r.rtm_flags |= RTM_F_LOOKUP_TABLE;
> +
>  	if (rtnl_talk(&rth, &req.n, &req.n, sizeof(req)) < 0)
>  		exit(2);
>  

How will this work (or not) on older kernels?