From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Duyck Subject: [net PATCH 0/3] netpoll fixes for Intel drivers Date: Tue, 22 Sep 2015 14:35:22 -0700 Message-ID: <20150922213002.2774.37924.stgit@ahduyck-vm-fedora22> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Cc: jeffrey.t.kirsher@intel.com To: netdev@vger.kernel.org, intel-wired-lan@lists.osuosl.org Return-path: Received: from mail-pa0-f51.google.com ([209.85.220.51]:34341 "EHLO mail-pa0-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S934184AbbIVVfY (ORCPT ); Tue, 22 Sep 2015 17:35:24 -0400 Received: by padhy16 with SMTP id hy16so20144424pad.1 for ; Tue, 22 Sep 2015 14:35:24 -0700 (PDT) Sender: netdev-owner@vger.kernel.org List-ID: While reviewing some other patches related to netpoll I started looking at how the drivers were handling their budgets. It turns out that the ixgbe, fm10k, and i40e drivers contained cases where they would increase the budget from 0 and as a result it was possible for them to clean Rx frames inside of netpoll context. This patch set corrects that by adding a check before the Rx portion of the polling routines for a budget less than or equal to 0. If the value is in that range we simply exit and leave ourselves on the polling list rather than attempt to clean any Rx frames. --- Alexander Duyck (3): i40e: Fix handling of napi budget fm10k: Fix handling of napi budget when multiple queues are enabled per vector ixgbe: Fix handling of napi budget when multiple queues are enabled per vector drivers/net/ethernet/intel/fm10k/fm10k_main.c | 4 ++++ drivers/net/ethernet/intel/i40e/i40e_txrx.c | 5 +++++ drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 3 ++- 3 files changed, 11 insertions(+), 1 deletion(-) --