Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 20 Jul 2020 10:24:43 -0700
From:   Stephen Hemminger <stephen@networkplumber.org>
To:     Chi Song <chisong@linux.microsoft.com>
Cc:     "K. Y. Srinivasan" <kys@microsoft.com>,
        Haiyang Zhang <haiyangz@microsoft.com>,
        Stephen Hemminger <sthemmin@microsoft.com>,
        Wei Liu <wei.liu@kernel.org>,
        "David S. Miller" <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>, linux-hyperv@vger.kernel.org,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 net-next] net: hyperv: Add attributes to show TX
 indirection table
Message-ID: <20200720102443.63d8ddf2@hermes.lan>
In-Reply-To: <alpine.LRH.2.23.451.2007192357400.30908@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.inter>
References: <alpine.LRH.2.23.451.2007192357400.30908@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.inter>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, 20 Jul 2020 00:12:10 -0700 (PDT)
Chi Song <chisong@linux.microsoft.com> wrote:

> An imbalanced TX indirection table causes netvsc to have low
> performance. This table is created and managed during runtime. To help
> better diagnose performance issues caused by imbalanced tables, add
> device attributes to show the content of TX indirection tables.
> 
> Signed-off-by: Chi Song <chisong@microsoft.com>
> ---


> v2: remove RX as it's in ethtool already, show single value in each file,
>  and update description.
> 
> Thank you for comments. Let me know, if I miss something.
> 
> ---
>  drivers/net/hyperv/netvsc_drv.c | 53 +++++++++++++++++++++++++++++++++
>  1 file changed, 53 insertions(+)
> 
> diff --git a/drivers/net/hyperv/netvsc_drv.c
> b/drivers/net/hyperv/netvsc_drv.c
> index 6267f706e8ee..222c2fad9300 100644
> --- a/drivers/net/hyperv/netvsc_drv.c
> +++ b/drivers/net/hyperv/netvsc_drv.c
> @@ -2370,6 +2370,55 @@ static int netvsc_unregister_vf(struct net_device
> *vf_netdev)
>  	return NOTIFY_OK;
>  }
> 
> +static struct device_attribute
> dev_attr_netvsc_dev_attrs[VRSS_SEND_TAB_SIZE];
> +static struct attribute *netvsc_dev_attrs[VRSS_SEND_TAB_SIZE + 1];
> +
> +const struct attribute_group netvsc_dev_group = {
> +	.name = NULL,
> +	.attrs = netvsc_dev_attrs,
> +};
> +
> +static ssize_t tx_indirection_table_show(struct device *dev,
> +					 struct device_attribute
> *dev_attr,
> +					 char *buf)
> +{
> +	struct net_device *ndev = to_net_dev(dev);
> +	struct net_device_context *ndc = netdev_priv(ndev);
> +	ssize_t offset = 0;

useless initialization

> +	int index = dev_attr - dev_attr_netvsc_dev_attrs;
> +
> +	offset = sprintf(buf, "%u\n", ndc->tx_table[index]);
> +
> +	return offset;
why not just
	return sprintf(buf, "%u\n", ndc->tx_table[index]);
> +}
> +
> +static void netvsc_attrs_init(void)
> +{
> +	int i;
> +	char buffer[32];
> +
> +	for (i = 0; i < VRSS_SEND_TAB_SIZE; i++) {
> +		sprintf(buffer, "tx_indirection_table_%02u", i);

Although this has one value per file it leads to a mess.
Why not put it in a separate directory (/sys/class/net/eth0/tx_indirection/N)?