Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 4 Dec 2019 10:39:55 -0800
From:   Jakub Kicinski <jakub.kicinski@netronome.com>
To:     Navid Emamdoost <navid.emamdoost@gmail.com>
Cc:     emamd001@umn.edu, smccaman@umn.edu, kjlu@umn.edu,
        "David S. Miller" <davem@davemloft.net>,
        Pablo Neira Ayuso <pablo@netfilter.org>,
        John Hurley <john.hurley@netronome.com>,
        Colin Ian King <colin.king@canonical.com>,
        oss-drivers@netronome.com, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3] nfp: abm: fix memory leak in
 nfp_abm_u32_knode_replace
Message-ID: <20191204103955.63c4d9af@cakuba.netronome.com>
In-Reply-To: <20190927015157.20070-1-navid.emamdoost@gmail.com>
References: <20190925215314.10cf291d@cakuba.netronome.com>
        <20190927015157.20070-1-navid.emamdoost@gmail.com>
Organization: Netronome Systems, Ltd.
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 Thu, 26 Sep 2019 20:51:46 -0500, Navid Emamdoost wrote:
> In nfp_abm_u32_knode_replace if the allocation for match fails it should
> go to the error handling instead of returning. Updated other gotos to
> have correct errno returned, too.
> 
> Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
> ---
> Changes in v2:
> 	- Reused err variable for erorr value returning.
> Changes in v3:
> 	- Fix the err value in the first goto, and fix subject prefix.

Ugh damn this. Apparently this "fix" has made the news:

https://news.softpedia.com/news/canonical-releases-major-kernel-security-update-for-ubuntu-19-10-and-18-04-lts-528433.shtml

https://nvd.nist.gov/vuln/detail/CVE-2019-19076

and (a) it would be a damn control path, root-only memory leak, but
also (b) upon closer inspection there is no leak here at all!

We don't need to delete the entry if we failed to allocate it...
The delete path is in case the entry for the handle is changed, but 
if we're trying to allocate one anew there can't be any on the list.

Congratulations to whoever classified this as a security fix.

I will send a revert, and go ask for the CVE to be marked invalid.
What a waste of time. I should have paid more attention :/

> diff --git a/drivers/net/ethernet/netronome/nfp/abm/cls.c b/drivers/net/ethernet/netronome/nfp/abm/cls.c
> index 23ebddfb9532..9f8a1f69c0c4 100644
> --- a/drivers/net/ethernet/netronome/nfp/abm/cls.c
> +++ b/drivers/net/ethernet/netronome/nfp/abm/cls.c
> @@ -176,8 +176,10 @@ nfp_abm_u32_knode_replace(struct nfp_abm_link *alink,
>  	u8 mask, val;
>  	int err;
>  
> -	if (!nfp_abm_u32_check_knode(alink->abm, knode, proto, extack))
> +	if (!nfp_abm_u32_check_knode(alink->abm, knode, proto, extack)) {
> +		err = -EOPNOTSUPP;
>  		goto err_delete;
> +	}
>  
>  	tos_off = proto == htons(ETH_P_IP) ? 16 : 20;
>  
> @@ -198,14 +200,18 @@ nfp_abm_u32_knode_replace(struct nfp_abm_link *alink,
>  		if ((iter->val & cmask) == (val & cmask) &&
>  		    iter->band != knode->res->classid) {
>  			NL_SET_ERR_MSG_MOD(extack, "conflict with already offloaded filter");
> +			err = -EOPNOTSUPP;
>  			goto err_delete;
>  		}
>  	}
>  
>  	if (!match) {
>  		match = kzalloc(sizeof(*match), GFP_KERNEL);
> -		if (!match)
> -			return -ENOMEM;
> +		if (!match) {
> +			err = -ENOMEM;
> +			goto err_delete;
> +		}
> +
>  		list_add(&match->list, &alink->dscp_map);
>  	}
>  	match->handle = knode->handle;
> @@ -221,7 +227,7 @@ nfp_abm_u32_knode_replace(struct nfp_abm_link *alink,
>  
>  err_delete:
>  	nfp_abm_u32_knode_delete(alink, knode);
> -	return -EOPNOTSUPP;
> +	return err;
>  }
>  
>  static int nfp_abm_setup_tc_block_cb(enum tc_setup_type type,