RE: [hyperv] BUG at drivers/hv/channel.c:462 while changing MTU

[Haiyang Zhang <haiyangz@microsoft.com>]

> -----Original Message-----
> From: Sitsofe Wheeler [mailto:sitsofe@gmail.com]
> Sent: Tuesday, August 19, 2014 8:37 AM
> To: KY Srinivasan
> Cc: Greg Kroah-Hartman; Haiyang Zhang; devel@linuxdriverproject.org;
> linux-kernel@vger.kernel.org; Jason Wang; Daniel Borkmann; David S.
> Miller; netdev@vger.kernel.org
> Subject: [hyperv] BUG at drivers/hv/channel.c:462 while changing MTU
> 
> Now we get to the issue I've been meaning to report: changing the MTU of
> a Hyper-V network interface while traffic is flowing through it triggers
> a BUG. I've seen this happen on a variety of kernels but the trace below
> is from 3.17.0-rc1.
> 
> Steps to reproduce:
> (Guests were customised Fedora 20 cloud images)
> 1.On Hyper-V VM 1 run
> iperf -s
> 2. On Hyper-V VM 2 run
> iperf -c <Hyper-V VM 1 host address> -t 200
> 3. On Hyper-V VM 2 run
> ip link set dev eth0 mtu 9000
> 
> Expected results:
> MTU of eth0 to be changed, packets to keep flowing or stop, interface to
> be at alive enough to be reconfigured.
> 
> Actual results:
> Error messages on the console, eth0 winds up dead.
> 

Thanks for the report.
I tried changing mtu while iperf is running, but unable to reproduce it.
I repeated the changing MTU test 10+ times, the link is stable, and passing 
traffic at ~8Gbps.

Could you try a Distro without customization?
I just want to narrow down the possible causes for it.

Thanks,
- Haiyang