Date: Mon, 13 Apr 2015 12:16:26 +0200 From: Andre Albsmeier <Andre.Albsmeier@siemens.com> To: freebsd-net@freebsd.org Subject: Intel em (82574L and 82573L) problems: stopping on high network and cpu load (Watchdog timeout) Message-ID: <20150413101626.GA47411@bali>
next in thread | raw e-mail | index | archive | help
Hi, em0 = '82574L Gigabit Network Connection': em0: <Intel(R) PRO/1000 Network Connection 7.4.2> port 0xd000-0xd01f mem 0xf0500000-0xf051ffff,0xf052 0000-0xf0523fff irq 17 at device 0.0 on pci5 em0: Using MSIX interrupts with 3 vectors OS = 9.3-STABLE #2: Wed Apr 1 07:20:47 CEST 2015 Sometimes em0 freezes and comes back about 1 minute later. We see the infamous "Watchdog timeout -- resetting" message: em0: Watchdog timeout -- resetting em0: Queue(0) tdh = 284, hw tdt = 239 em0: TX(0) desc avail = 31,Next TX to Clean = 270 em0: link state changed to DOWN em0: link state changed to UP I don't want to say that I can easily reproduce it but chances are quite good when I transfer a lot of stuff over the network _AND_ the CPU is heavily loaded. The CPU is a CPU: Intel(R) Core(TM) i7-3770K CPU @ 3.50GHz (3492.14-MHz 686-class CPU) running 2 avidemux encoding processes (each with 8 threads) all the time. With just 1 avidemux task I didn't see it happen so far. It does not matter if the heavy network traffic is inbound or outbound. BTW, this happens on two identical machines but on two completely different networks. It also happens on a machine using a 82573L card. It never occured on a Marvell based machine (msk driver) running similar tasks... Any ideas?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20150413101626.GA47411>