From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christophe Ricard Subject: [RFC] netlink_ack: send a capped message in case of error Date: Mon, 24 Aug 2015 00:06:16 +0200 Message-ID: <1440367577-15786-1-git-send-email-christophe-h.ricard@st.com> Cc: netdev@vger.kernel.org, davem@davemloft.net, Christophe Ricard To: pablo@netfilter.org, jbenc@redhat.com, sameo@linux.intel.com Return-path: Received: from mail-wi0-f169.google.com ([209.85.212.169]:35979 "EHLO mail-wi0-f169.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752759AbbHWWGX (ORCPT ); Sun, 23 Aug 2015 18:06:23 -0400 Received: by wicja10 with SMTP id ja10so56168659wic.1 for ; Sun, 23 Aug 2015 15:06:22 -0700 (PDT) Sender: netdev-owner@vger.kernel.org List-ID: Hi, I have found with netlink we can miss a ack message when exchanging large buffers. I came across this "bug ?" when doing some tests on the NFC subsytem by trying to send extended APDU to a secure element (up to 65K) or when sending specific device firmware data (up to 8K) (NFC_ATTR_SE_APDU & NFC_ATTR_VENDOR_DATA). After some search on the web, i found this topic was already threated by Jiri Benc in November 2013: - http://patchwork.ozlabs.org/patch/290976/ - http://www.spinics.net/lists/netdev/msg256938.html I haven't found any update or any fix in recent kernel. I liked the netlink message cap approach. Do you have any update ? Do you think a netlink message cap in netlink_ack in case of an error is an acceptable approach ? Best Regards Christophe Christophe Ricard (1): netlink: netlink_ack send a capped message in case of error net/netlink/af_netlink.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) -- 2.1.4