Re: [PATCH net-next v12 12/13] net: ethtool: tsinfo: Add support for hwtstamp provider and get/set hwtstamp config

[Simon Horman <horms@kernel.org>]

On Tue, Apr 30, 2024 at 05:49:55PM +0200, Kory Maincent wrote:
> Enhance 'get' command to retrieve tsinfo of hwtstamp providers within a
> network topology and read current hwtstamp configuration.
> 
> Introduce support for ETHTOOL_MSG_TSINFO_SET ethtool netlink socket to
> configure hwtstamp of a PHC provider. Note that simultaneous hwtstamp
> isn't supported; configuring a new one disables the previous setting.
> 
> Also, add support for a specific dump command to retrieve all hwtstamp
> providers within the network topology, with added functionality for
> filtered dump to target a single interface.
> 
> Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>

...

> diff --git a/net/core/dev_ioctl.c b/net/core/dev_ioctl.c
> index acb0cadb7512..ec6dd81a5add 100644
> --- a/net/core/dev_ioctl.c
> +++ b/net/core/dev_ioctl.c
> @@ -270,10 +270,13 @@ static int dev_eth_ioctl(struct net_device *dev,
>  int dev_get_hwtstamp_phylib(struct net_device *dev,
>  			    struct kernel_hwtstamp_config *cfg)
>  {
> -	if (dev->hwtstamp) {
> -		struct ptp_clock *ptp = dev->hwtstamp->ptp;
> +	struct hwtstamp_provider *hwtstamp;
>  
> -		cfg->qualifier = dev->hwtstamp->qualifier;
> +	hwtstamp = rtnl_dereference(dev->hwtstamp);
> +	if (hwtstamp) {
> +		struct ptp_clock *ptp = hwtstamp->ptp;
> +
> +		cfg->qualifier = hwtstamp->qualifier;
>  		if (ptp_clock_from_phylib(ptp))
>  			return phy_hwtstamp_get(ptp_clock_phydev(ptp), cfg);
>  
> @@ -340,13 +343,15 @@ int dev_set_hwtstamp_phylib(struct net_device *dev,
>  {
>  	const struct net_device_ops *ops = dev->netdev_ops;
>  	struct kernel_hwtstamp_config old_cfg = {};
> +	struct hwtstamp_provider *hwtstamp;
>  	struct phy_device *phydev;
>  	bool changed = false;
>  	bool phy_ts;
>  	int err;
>  
> -	if (dev->hwtstamp) {
> -		struct ptp_clock *ptp = dev->hwtstamp->ptp;
> +	hwtstamp = rtnl_dereference(dev->hwtstamp);
> +	if (hwtstamp) {
> +		struct ptp_clock *ptp = hwtstamp->ptp;
>  
>  		if (ptp_clock_from_phylib(ptp)) {
>  			phy_ts = true;

Hi Kory,

A few lines beyond this hunk, within the "if (hwtstamp)" block,
is the following:

		cfg->qualifier = dev->hwtstamp->qualifier;

Now that dev->hwtstamp is managed using RCU, I don't think it is correct
to dereference it directly like this. Rather, the hwtstamp local variable,
which has rcu_dereference'd this pointer should be used:

		 cfg->qualifier = hwtstamp->qualifier;

Flagged by Sparse.

...

> diff --git a/net/ethtool/tsinfo.c b/net/ethtool/tsinfo.c

...

> +static int ethnl_tsinfo_dump_one_dev(struct sk_buff *skb, struct net_device *dev,
> +				     struct netlink_callback *cb)
> +{
> +	struct ethnl_tsinfo_dump_ctx *ctx = (void *)cb->ctx;
> +	struct ptp_clock *ptp;
> +	int ret;
> +
> +	netdev_for_each_ptp_clock_start(dev, ctx->pos_phcindex, ptp,
> +					ctx->pos_phcindex) {
> +		ret = ethnl_tsinfo_dump_one_ptp(skb, dev, cb, ptp);
> +		if (ret < 0 && ret != -EOPNOTSUPP)
> +			break;
> +		ctx->pos_phcqualifier = HWTSTAMP_PROVIDER_QUALIFIER_PRECISE;
> +	}
> +
> +	return ret;

Perhaps it is not possible, but if the loop iterates zero times then
ret will be used uninitialised here.

Flagged by Smatch.

> +}

...