netdev - Re: [PATCH ethtool 2/2] netlink: Print and return an error when features weren't changed

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Sun, 23 Aug 2020 18:03:20 +0200
From:   Michal Kubecek <mkubecek@...e.cz>
To:     Maxim Mikityanskiy <maximmi@...lanox.com>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Andrew Lunn <andrew@...n.ch>, Jakub Kicinski <kuba@...nel.org>,
        netdev@...r.kernel.org
Subject: Re: [PATCH ethtool 2/2] netlink: Print and return an error when
 features weren't changed

On Fri, Aug 14, 2020 at 04:17:45PM +0300, Maxim Mikityanskiy wrote:
> The legacy ethtool prints an error message and returns 1 if no features
> were changed as requested. Port this behavior to ethtool-netlink.
> req_mask is compared to wanted_mask to detect if any feature was
> changed. If these masks are equal, it means that the kernel hasn't
> changed anything, and all bits got to wanted.
> 
> Signed-off-by: Maxim Mikityanskiy <maximmi@...lanox.com>
> ---
>  netlink/features.c | 25 ++++++++++++++++++-------
>  1 file changed, 18 insertions(+), 7 deletions(-)
> 
> diff --git a/netlink/features.c b/netlink/features.c
> index 133529d..4f63fa2 100644
> --- a/netlink/features.c
> +++ b/netlink/features.c
[...]
> @@ -471,8 +480,10 @@ int sfeatures_reply_cb(const struct nlmsghdr *nlhdr, void *data)
>  		return MNL_CB_OK;
>  	}
>  
> -	show_feature_changes(nlctx, tb);
> -	return MNL_CB_OK;
> +	if (show_feature_changes(nlctx, tb))
> +		return MNL_CB_OK;
> +	else
> +		return MNL_CB_ERROR;

I agree with the general change and the code aboved detecting the
condition but this kind of error is IMHO not so critical that it would
justify bailing out and completely ignoring the final NLMSG_ERROR with
kernel return code and possible extack (error/warning message).

I would rather suggest to set a flag (e.g. in sfctx) when "no requested
change performed" result is detected and leave displaying the error
message and setting the exit code after the whole message queue is
processed. What do you think?

Michal

>  }
>  
>  int nl_sfeatures(struct cmd_context *ctx)
> -- 
> 2.21.0
> 

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)