netdev - Re: [PATCH net-next v3 1/7] dsa: move call to driver port

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]

Date: Mon, 1 Apr 2024 18:53:27 +0300
From: Vladimir Oltean <olteanv@...il.com>
To: Andrew Lunn <andrew@...n.ch>
Cc: Florian Fainelli <f.fainelli@...il.com>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	Russell King <linux@...linux.org.uk>,
	Gregory Clement <gregory.clement@...tlin.com>,
	netdev@...r.kernel.org
Subject: Re: [PATCH net-next v3 1/7] dsa: move call to driver port_setup
 after creation of netdev.

Title: "net: dsa:" prefix, no "."

On Mon, Apr 01, 2024 at 08:35:46AM -0500, Andrew Lunn wrote:
> The drivers call port_setup() is a good place to add the LEDs of a
> port to the netdev representing the port. However, when port_setup()
> is called in dsa_port_devlink_setup() the netdev does not exist
> yet. That only happens in dsa_user_create() which is latter in

later

> dsa_port_setup().
> 
> Move the call to port_setup() out of dsa_port_devlink_setup() and to
> the end of dsa_port_setup() where the netdev will exist.
> 
> Signed-off-by: Andrew Lunn <andrew@...n.ch>
> ---
>  net/dsa/devlink.c | 17 +----------------
>  net/dsa/dsa.c     |  3 +++
>  2 files changed, 4 insertions(+), 16 deletions(-)
> 
> diff --git a/net/dsa/devlink.c b/net/dsa/devlink.c
> index 431bf52290a1..9c3dc6319269 100644
> --- a/net/dsa/devlink.c
> +++ b/net/dsa/devlink.c
> @@ -294,20 +294,12 @@ int dsa_port_devlink_setup(struct dsa_port *dp)
>  	struct dsa_switch_tree *dst = dp->ds->dst;
>  	struct devlink_port_attrs attrs = {};
>  	struct devlink *dl = dp->ds->devlink;
> -	struct dsa_switch *ds = dp->ds;
>  	const unsigned char *id;
>  	unsigned char len;
> -	int err;
>  
>  	memset(dlp, 0, sizeof(*dlp));
>  	devlink_port_init(dl, dlp);
>  
> -	if (ds->ops->port_setup) {
> -		err = ds->ops->port_setup(ds, dp->index);
> -		if (err)
> -			return err;
> -	}
> -
>  	id = (const unsigned char *)&dst->index;
>  	len = sizeof(dst->index);
>  
> @@ -331,14 +323,7 @@ int dsa_port_devlink_setup(struct dsa_port *dp)
>  	}
>  
>  	devlink_port_attrs_set(dlp, &attrs);
> -	err = devlink_port_register(dl, dlp, dp->index);
> -	if (err) {
> -		if (ds->ops->port_teardown)
> -			ds->ops->port_teardown(ds, dp->index);
> -		return err;
> -	}
> -
> -	return 0;
> +	return devlink_port_register(dl, dlp, dp->index);
>  }
>  
>  void dsa_port_devlink_teardown(struct dsa_port *dp)
> diff --git a/net/dsa/dsa.c b/net/dsa/dsa.c
> index 09d2f5d4b3dd..6ffee2a7de94 100644
> --- a/net/dsa/dsa.c
> +++ b/net/dsa/dsa.c
> @@ -520,6 +520,9 @@ static int dsa_port_setup(struct dsa_port *dp)
>  		break;
>  	}
>  
> +	if (ds->ops->port_setup)
> +		err = ds->ops->port_setup(ds, dp->index);
> +

This overwrites the not-yet-checked "err", masking the dsa_user_create()
return code, and breaking the error handling logic below. Not to
mention, if ds->ops->port_setup() fails for a user port, we should call
dsa_user_destroy().

>  	if (err && dsa_port_enabled)
>  		dsa_port_disable(dp);
>  	if (err && dsa_port_link_registered)
> 
> -- 
> 2.43.0
> 

It would have been good for the API, if we want the netdev to be
available for user ports at port_setup() time, for it to be available at
port_teardown() time as well. So dsa_port_devlink_teardown() needs
changing too.

Additionally, for CPU and DSA ports, this change will make
ds->ops->port_enable() be visible from the driver API earlier than
ds->ops->port_setup(), which isn't exactly intuitive or great or better
than before.

In fact, I think it's very difficult not to make mistakes changing the
code in its current form. These 3 patches I've prepared - which replace
this patch - should help (see attached).

View attachment "0001-net-dsa-consolidate-setup-and-teardown-for-shared-po.patch" of type "text/x-diff" (1842 bytes)

View attachment "0002-net-dsa-break-out-port-setup-and-teardown-code-per-p.patch" of type "text/x-diff" (3824 bytes)

View attachment "0003-net-dsa-move-call-to-driver-port_setup-after-creatio.patch" of type "text/x-diff" (4536 bytes)