From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Michael S. Tsirkin" Subject: Re: [PATCH] tun: Use netif_receive_skb instead of netif_rx Date: Tue, 29 Nov 2016 18:20:22 +0200 Message-ID: <20161129181631-mutt-send-email-mst@kernel.org> References: <1480433136-7922-1-git-send-email-andreyknvl@google.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Herbert Xu , "David S . Miller" , Jason Wang , Eric Dumazet , Paolo Abeni , Soheil Hassas Yeganeh , Markus Elfring , Mike Rapoport , netdev@vger.kernel.org, linux-kernel@vger.kernel.org, Dmitry Vyukov , Kostya Serebryany , syzkaller@googlegroups.com To: Andrey Konovalov Return-path: Received: from mx1.redhat.com ([209.132.183.28]:35262 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753188AbcK2QUZ (ORCPT ); Tue, 29 Nov 2016 11:20:25 -0500 Content-Disposition: inline In-Reply-To: <1480433136-7922-1-git-send-email-andreyknvl@google.com> Sender: netdev-owner@vger.kernel.org List-ID: On Tue, Nov 29, 2016 at 04:25:36PM +0100, Andrey Konovalov wrote: > This patch changes tun.c to call netif_receive_skb instead of netif_rx > when a packet is received. The difference between the two is that netif_rx > queues the packet into the backlog, and netif_receive_skb proccesses the > packet in the current context. > > This patch is required for syzkaller [1] to collect coverage from packet > receive paths, when a packet being received through tun (syzkaller collects > coverage per process in the process context). > > A similar patch was introduced back in 2010 [2, 3], but the author found > out that the patch doesn't help with the task he had in mind (for cgroups > to shape network traffic based on the original process) and decided not to > go further with it. The main concern back then was about possible stack > exhaustion with 4K stacks, but CONFIG_4KSTACKS was removed and stacks are > 8K now. > > [1] https://github.com/google/syzkaller > > [2] https://www.spinics.net/lists/netdev/thrd440.html#130570 > > [3] https://www.spinics.net/lists/netdev/msg130570.html > > Signed-off-by: Andrey Konovalov This was on my list of things to investigate ever since 8k stack default went in. Thanks for looking into this! I note that there are still seem to exit 3 architectures that do have CONFIG_4KSTACKS. How about a wrapper that does netif_rx_ni with CONFIG_4KSTACKS. > --- > drivers/net/tun.c | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > diff --git a/drivers/net/tun.c b/drivers/net/tun.c > index 8093e39..4b56e91 100644 > --- a/drivers/net/tun.c > +++ b/drivers/net/tun.c > @@ -1304,7 +1304,9 @@ static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile, > skb_probe_transport_header(skb, 0); > > rxhash = skb_get_hash(skb); > - netif_rx_ni(skb); > + local_bh_disable(); > + netif_receive_skb(skb); > + local_bh_enable(); > > stats = get_cpu_ptr(tun->pcpu_stats); > u64_stats_update_begin(&stats->syncp); > -- > 2.8.0.rc3.226.g39d4020