linux-kernel - Re: [PATCH v3 2/2] net-sysfs: Ensure begin/complete are called in speed_show() and duplex

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <425AF2D4-1FEE-437B-8520-452F818F7DEE@canonical.com>
Date:   Thu, 20 Feb 2020 13:45:16 +0800
From:   Kai-Heng Feng <kai.heng.feng@...onical.com>
To:     David Miller <davem@...emloft.net>,
        Michal Kubecek <mkubecek@...e.cz>,
        Jeff Kirsher <jeffrey.t.kirsher@...el.com>
Cc:     Jakub Kicinski <kuba@...nel.org>,
        Maxime Chevallier <maxime.chevallier@...tlin.com>,
        Heiner Kallweit <hkallweit1@...il.com>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Jouni Hogander <jouni.hogander@...kie.com>,
        Eric Dumazet <edumazet@...gle.com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Wang Hai <wanghai26@...wei.com>,
        YueHaibing <yuehaibing@...wei.com>,
        Kimberly Brown <kimbrownkd@...il.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Florian Fainelli <f.fainelli@...il.com>,
        Jiri Pirko <jiri@...lanox.com>, Andrew Lunn <andrew@...n.ch>,
        Li RongQing <lirongqing@...du.com>,
        open list <linux-kernel@...r.kernel.org>,
        "open list:NETWORKING [GENERAL]" <netdev@...r.kernel.org>
Subject: Re: [PATCH v3 2/2] net-sysfs: Ensure begin/complete are called in
 speed_show() and duplex_show()


> On Feb 7, 2020, at 18:10, Kai-Heng Feng <kai.heng.feng@...onical.com> wrote:
> 
> Device like igb gets runtime suspended when there's no link partner. We
> can't get correct speed under that state:
> $ cat /sys/class/net/enp3s0/speed
> 1000
> 
> In addition to that, an error can also be spotted in dmesg:
> [  385.991957] igb 0000:03:00.0 enp3s0: PCIe link lost
> 
> It's because the igb device doesn't get runtime resumed before calling
> get_link_ksettings().
> 
> So let's use a new helper to call begin() and complete() like what
> dev_ethtool() does, to runtime resume/suspend or power up/down the
> device properly.
> 
> Once this fix is in place, igb can show the speed correctly without link
> partner:
> $ cat /sys/class/net/enp3s0/speed
> -1
> 
> -1 here means SPEED_UNKNOWN, which is the correct value when igb is
> runtime suspended.
> 
> Signed-off-by: Kai-Heng Feng <kai.heng.feng@...onical.com>

A gentle ping...

Kai-Heng

> ---
> v3:
> - Specify -1 means SPEED_UNKNOWN.
> v2:
> - Add a new helper with begin/complete and use it in net-sysfs.
> 
> include/linux/ethtool.h |  4 ++++
> net/core/net-sysfs.c    |  4 ++--
> net/ethtool/ioctl.c     | 33 ++++++++++++++++++++++++++++++++-
> 3 files changed, 38 insertions(+), 3 deletions(-)
> 
> diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h
> index 95991e4300bf..785ec1921417 100644
> --- a/include/linux/ethtool.h
> +++ b/include/linux/ethtool.h
> @@ -160,6 +160,10 @@ extern int
> __ethtool_get_link_ksettings(struct net_device *dev,
> 			     struct ethtool_link_ksettings *link_ksettings);
> 
> +extern int
> +__ethtool_get_link_ksettings_full(struct net_device *dev,
> +				  struct ethtool_link_ksettings *link_ksettings);
> +
> /**
>  * ethtool_intersect_link_masks - Given two link masks, AND them together
>  * @dst: first mask and where result is stored
> diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
> index 4c826b8bf9b1..a199e15a080f 100644
> --- a/net/core/net-sysfs.c
> +++ b/net/core/net-sysfs.c
> @@ -201,7 +201,7 @@ static ssize_t speed_show(struct device *dev,
> 	if (netif_running(netdev)) {
> 		struct ethtool_link_ksettings cmd;
> 
> -		if (!__ethtool_get_link_ksettings(netdev, &cmd))
> +		if (!__ethtool_get_link_ksettings_full(netdev, &cmd))
> 			ret = sprintf(buf, fmt_dec, cmd.base.speed);
> 	}
> 	rtnl_unlock();
> @@ -221,7 +221,7 @@ static ssize_t duplex_show(struct device *dev,
> 	if (netif_running(netdev)) {
> 		struct ethtool_link_ksettings cmd;
> 
> -		if (!__ethtool_get_link_ksettings(netdev, &cmd)) {
> +		if (!__ethtool_get_link_ksettings_full(netdev, &cmd)) {
> 			const char *duplex;
> 
> 			switch (cmd.base.duplex) {
> diff --git a/net/ethtool/ioctl.c b/net/ethtool/ioctl.c
> index b987052d91ef..faeba247c1fb 100644
> --- a/net/ethtool/ioctl.c
> +++ b/net/ethtool/ioctl.c
> @@ -420,7 +420,9 @@ struct ethtool_link_usettings {
> 	} link_modes;
> };
> 
> -/* Internal kernel helper to query a device ethtool_link_settings. */
> +/* Internal kernel helper to query a device ethtool_link_settings. To be called
> + * inside begin/complete block.
> + */
> int __ethtool_get_link_ksettings(struct net_device *dev,
> 				 struct ethtool_link_ksettings *link_ksettings)
> {
> @@ -434,6 +436,35 @@ int __ethtool_get_link_ksettings(struct net_device *dev,
> }
> EXPORT_SYMBOL(__ethtool_get_link_ksettings);
> 
> +/* Internal kernel helper to query a device ethtool_link_settings. To be called
> + * outside of begin/complete block.
> + */
> +int __ethtool_get_link_ksettings_full(struct net_device *dev,
> +				      struct ethtool_link_ksettings *link_ksettings)
> +{
> +	int rc;
> +
> +	ASSERT_RTNL();
> +
> +	if (!dev->ethtool_ops->get_link_ksettings)
> +		return -EOPNOTSUPP;
> +
> +	if (dev->ethtool_ops->begin) {
> +		rc = dev->ethtool_ops->begin(dev);
> +		if (rc  < 0)
> +			return rc;
> +	}
> +
> +	memset(link_ksettings, 0, sizeof(*link_ksettings));
> +	rc = dev->ethtool_ops->get_link_ksettings(dev, link_ksettings);
> +
> +	if (dev->ethtool_ops->complete)
> +		dev->ethtool_ops->complete(dev);
> +
> +	return rc;
> +}
> +EXPORT_SYMBOL(__ethtool_get_link_ksettings_full);
> +
> /* convert ethtool_link_usettings in user space to a kernel internal
>  * ethtool_link_ksettings. return 0 on success, errno on error.
>  */
> -- 
> 2.17.1
>