From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ding Tianhong Subject: Re: [PATCH net-next v13 3/3] net: hisilicon: new hip04 ethernet driver Date: Thu, 15 Jan 2015 10:54:24 +0800 Message-ID: <54B72BE0.7040403@huawei.com> References: <1421217254-12008-1-git-send-email-dingtianhong@huawei.com> <1421217254-12008-4-git-send-email-dingtianhong@huawei.com> <3131780.HetDHI4Cfl@wuerfel> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: , , , , , , , , , , , To: Arnd Bergmann Return-path: Received: from szxga01-in.huawei.com ([119.145.14.64]:32066 "EHLO szxga01-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751183AbbAOCyt (ORCPT ); Wed, 14 Jan 2015 21:54:49 -0500 In-Reply-To: <3131780.HetDHI4Cfl@wuerfel> Sender: netdev-owner@vger.kernel.org List-ID: On 2015/1/14 16:53, Arnd Bergmann wrote: > On Wednesday 14 January 2015 14:34:14 Ding Tianhong wrote: >> +#define HIP04_MAX_TX_COALESCE_USECS 200 >> +#define HIP04_MIN_TX_COALESCE_USECS 100 >> +#define HIP04_MAX_TX_COALESCE_FRAMES 200 >> +#define HIP04_MIN_TX_COALESCE_FRAMES 100 > > It's not important, but in case you are creating another version of the > patch, maybe the allowed range can be extended somewhat. The example values > I picked when I sent my suggestion were really made up. It's great if > they work fine, but users might want to tune this far more depending on > their workloads, How about these > > #define HIP04_MAX_TX_COALESCE_USECS 100000 > #define HIP04_MIN_TX_COALESCE_USECS 1 > #define HIP04_MAX_TX_COALESCE_FRAMES (TX_DESC_NUM - 1) > #define HIP04_MIN_TX_COALESCE_FRAMES 1 > Is it really ok that the so big range may break the driver and hip04 could not work fine? I am not sure it is ok, I will fix it in next version. Ding > Arnd > > . >