netdev - Re: r8169 regression: UDP packets dropped intermittantly

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1447935754.22599.194.camel@edumazet-glaptop2.roam.corp.google.com>
Date:	Thu, 19 Nov 2015 04:22:34 -0800
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Jonathan Woithe <jwoithe@...ad.com.au>
Cc:	Francois Romieu <romieu@...zoreil.com>, netdev@...r.kernel.org
Subject: Re: r8169 regression: UDP packets dropped intermittantly

On Thu, 2015-11-19 at 17:27 +1030, Jonathan Woithe wrote:
> On Thu, Nov 19, 2015 at 01:56:08AM +0100, Francois Romieu wrote:
> > Jonathan Woithe <jwoithe@...ad.com.au> :
> > [...]
> > > I could try.  What's the most reliable way to determine this?  Use regular
> > > ethtool queries about the time of the failures?
> > 
> > If you are neither space nor cpu constrained:
> > 
> > while : ; do
> > 	echo $(date +%s.%N):$(ethtool -d eth0 raw on | base64 -w 0):$(ethtool -S eth0 | base64 -w 0)
> > 	sleep 1
> > done > gloubi
> > 
> > Expect 2.5~3Mo per hour.
> 
> We're certainly not disc constrained.  The high speed acquisition takes a
> fair bit of CPU but there should still be sufficient free to at least give
> the above a go.  I'll run some tests tomorrow and see what transpires.

This looks like a race when/if RX happens very shortly after one
transmit.

Francois, are we really saving lot of cpu cycles testing status ?

Following patch would probably reduce the window.
(And doing TX completions before RX might help as well)


diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 79ef799f88ab1fc2518736163c5674aec360c065..ed3d0322643600c8c3521aa0ac7b935c929527aa 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -7541,17 +7541,14 @@ static int rtl8169_poll(struct napi_struct *napi, int budget)
 	struct rtl8169_private *tp = container_of(napi, struct rtl8169_private, napi);
 	struct net_device *dev = tp->dev;
 	u16 enable_mask = RTL_EVENT_NAPI | tp->event_slow;
-	int work_done= 0;
+	int work_done;
 	u16 status;
 
 	status = rtl_get_events(tp);
 	rtl_ack_events(tp, status & ~tp->event_slow);
 
-	if (status & RTL_EVENT_NAPI_RX)
-		work_done = rtl_rx(dev, tp, (u32) budget);
-
-	if (status & RTL_EVENT_NAPI_TX)
-		rtl_tx(dev, tp);
+	rtl_tx(dev, tp);
+	work_done = rtl_rx(dev, tp, (u32) budget);
 
 	if (status & tp->event_slow) {
 		enable_mask &= ~tp->event_slow;


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html