netdev - Re: [PATCH net] ionic: linearize tso skb with too many frags

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: <20210316145431.232672d5@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date:   Tue, 16 Mar 2021 14:54:31 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Shannon Nelson <snelson@...sando.io>
Cc:     netdev@...r.kernel.org, davem@...emloft.net, drivers@...sando.io
Subject: Re: [PATCH net] ionic: linearize tso skb with too many frags

On Tue, 16 Mar 2021 11:52:43 -0700 Shannon Nelson wrote:
> We were linearizing non-TSO skbs that had too many frags, but
> we weren't checking number of frags on TSO skbs.  This could
> lead to a bad page reference when we received a TSO skb with
> more frags than the Tx descriptor could support.
> 
> Fixes: 0f3154e6bcb3 ("ionic: Add Tx and Rx handling")
> Signed-off-by: Shannon Nelson <snelson@...sando.io>
> ---
>  .../net/ethernet/pensando/ionic/ionic_txrx.c  | 28 ++++++++++---------
>  1 file changed, 15 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/net/ethernet/pensando/ionic/ionic_txrx.c b/drivers/net/ethernet/pensando/ionic/ionic_txrx.c
> index 162a1ff1e9d2..462b0d106be4 100644
> --- a/drivers/net/ethernet/pensando/ionic/ionic_txrx.c
> +++ b/drivers/net/ethernet/pensando/ionic/ionic_txrx.c
> @@ -1079,25 +1079,27 @@ static int ionic_tx_descs_needed(struct ionic_queue *q, struct sk_buff *skb)
>  {
>  	int sg_elems = q->lif->qtype_info[IONIC_QTYPE_TXQ].max_sg_elems;
>  	struct ionic_tx_stats *stats = q_to_tx_stats(q);
> +	int ndescs;
>  	int err;
>  
> -	/* If TSO, need roundup(skb->len/mss) descs */
> +	/* If TSO, need roundup(skb->len/mss) descs
> +	 * If non-TSO, just need 1 desc and nr_frags sg elems
> +	 */
>  	if (skb_is_gso(skb))
> -		return (skb->len / skb_shinfo(skb)->gso_size) + 1;
> +		ndescs = (skb->len / skb_shinfo(skb)->gso_size) + 1;

Slightly unrelated but why not gso_segs? len / gso_size + 1 could be
over counting, not to mention that div is expensive.

Are you segmenting in the driver? Why do you need #segs descriptors?

> +	else
> +		ndescs = 1;
>  
> -	/* If non-TSO, just need 1 desc and nr_frags sg elems */
> -	if (skb_shinfo(skb)->nr_frags <= sg_elems)
> -		return 1;
> +	/* If too many frags, linearize */
> +	if (skb_shinfo(skb)->nr_frags > sg_elems) {
> +		err = skb_linearize(skb);
> +		if (err)
> +			return err;
>  
> -	/* Too many frags, so linearize */
> -	err = skb_linearize(skb);
> -	if (err)
> -		return err;
> -
> -	stats->linearize++;
> +		stats->linearize++;
> +	}
>  
> -	/* Need 1 desc and zero sg elems */
> -	return 1;
> +	return ndescs;

I'd be tempted to push back on the refactoring here, you could've 
just replaced return 1;s with return ndescs;s without changing 
the indentation.. this will give all backporters a pause. But 
not the end of the world, I guess.

>  }
>  
>  static int ionic_maybe_stop_tx(struct ionic_queue *q, int ndescs)