From mboxrd@z Thu Jan 1 00:00:00 1970 From: Josh Hunt Date: Tue, 22 Sep 2015 11:01:04 -0500 Subject: [Intel-wired-lan] e1000e tx queue timeout with small mss values Message-ID: <56017B40.90308@akamai.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: intel-wired-lan@osuosl.org List-ID: We are seeing transmit queue timeouts with the 82574L even with the following patch applied, e1000e: DoS while TSO enabled caused by link partner with small MSS, commit d821a4c4d11ad160925dab2bb009b8444beff484 upstream. We're running upstream 3.14.43. We've noticed an increase in the frequency of hitting the timeouts when we have tcp mtu probing enabled. This is allowing some connections to get MSS values of 36. There was some discussion during the original reporting of this issue on trying to reproduce it in a lab setting, possibly setting the advmss via ip route. I've tried this in our lab, but am unable to trigger the problem. I was wondering if anyone there knows how to possibly trigger this issue in a lab setting? Thanks for any help you can provide. Josh