Ah, yes - I see the issue now. I'll update that when I get back. And I
apologize for the top posting! My phone's email client doesn't do inline too
well...
Sent from my Verizon, Samsung Galaxy smartphone
-------- Original message --------
From: Sven Eckelmann <s...@narfation.org>
Date: 10/25/18 3:42 AM (GMT-07:00)
To: b.a.t.m.a.n@lists.open-mesh.org
Cc: Jonathan Haws <jh...@sdl.usu.edu>, guohuizou2...@sina.com
Subject: Re: [B.A.T.M.A.N.] [PATCH v2] alfred: Request MAC resolution for IPv4
address not in ARP cache
On Mittwoch, 24. Oktober 2018 13:21:28 CEST Jonathan Haws wrote:
[...]
>
> + while (retries-- && !(arpreq.arp_flags & ATF_COM)) {
> + ipv4_request_mac_resolve(addr);
> + usleep(200000);
> +
> + if (ioctl(interface->netsock, SIOCGARP, &arpreq) < 0)
> + return -1;
> + }
> +
> if (arpreq.arp_flags & ATF_COM) {
> memcpy(mac, arpreq.arp_ha.sa_data, sizeof(*mac));
> } else {
According to gary [1], this doesn't work because the ioctl fails for him and
then the function returns immediately (before the while loop).
Please adjust your patch - but please don't use his code - it looks rather
ugly and also doesn't work for multiple retries. You most likely want to
drop the if-ioctl completely and then put everything in your while loop:
while (ioctl(interface->netsock, SIOCGARP, &arpreq) < 0 ||
!(arpreq.arp_flags & ATF_COM)) {
if (retries-- <= 0)
break;
ipv4_request_mac_resolve(addr);
usleep(200000);
}
But feel to propose a different (cleaner) approach.
Kind regards,
Sven
[1] https://lists.open-mesh.org/pipermail/b.a.t.m.a.n/2018-October/018195.html
