netdev - Re: [PATCH v5 bpf 2/2] veth: update mem type in xdp

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: <aQSvuMWDMyMYRI+W@boxer>
Date: Fri, 31 Oct 2025 13:46:48 +0100
From: Maciej Fijalkowski <maciej.fijalkowski@...el.com>
To: <bot+bpf-ci@...nel.org>
CC: <bpf@...r.kernel.org>, <ast@...nel.org>, <daniel@...earbox.net>,
	<hawk@...nel.org>, <netdev@...r.kernel.org>, <magnus.karlsson@...el.com>,
	<aleksander.lobakin@...el.com>, <ilias.apalodimas@...aro.org>,
	<toke@...hat.com>, <lorenzo@...nel.org>, <kuba@...nel.org>,
	<andrii@...nel.org>, <martin.lau@...nel.org>, <eddyz87@...il.com>,
	<yonghong.song@...ux.dev>, <clm@...a.com>, <ihor.solodrai@...ux.dev>
Subject: Re: [PATCH v5 bpf 2/2] veth: update mem type in xdp_buff

On Wed, Oct 29, 2025 at 10:59:36PM +0000, bot+bpf-ci@...nel.org wrote:
> 
> ```
> commit 96d6e22337a81e8606de33d22173ff80ab961bc8
> Author: Maciej Fijalkowski <maciej.fijalkowski@...el.com>
> 
> veth: update mem type in xdp_buff
> 
> The commit fixes a page_pool leak by properly detecting page_pool memory
> and using appropriate reference counting APIs. The xdp_buff mem type is
> now dynamically set based on skb->pp_recycle instead of always using
> MEM_TYPE_PAGE_SHARED.
> 
> Closes: https://lore.kernel.org/bpf/CAADnVQ+bBofJDfieyOYzSmSujSfJwDTQhiz3aJw7hE+4E2_iPA@mail.gmail.com/
> 
> > diff --git a/drivers/net/veth.c b/drivers/net/veth.c
> > index a3046142c..187f30e2c 100644
> > --- a/drivers/net/veth.c
> > +++ b/drivers/net/veth.c
> 
> [ ... ]
> 
> > @@ -733,7 +733,7 @@ static void veth_xdp_rcv_bulk_skb(struct veth_rq *rq, void **frames,
> >  	}
> >  }
> >
> > -static void veth_xdp_get(struct xdp_buff *xdp)
> > +static void veth_xdp_get_shared(struct xdp_buff *xdp)
> >  {
> >  	struct skb_shared_info *sinfo = xdp_get_shared_info_from_buff(xdp);
> >  	int i;
> > @@ -746,12 +746,33 @@ static void veth_xdp_get(struct xdp_buff *xdp)
> >  		__skb_frag_ref(&sinfo->frags[i]);
> >  }
> >
> > +static void veth_xdp_get_pp(struct xdp_buff *xdp)
> > +{
> > +	struct skb_shared_info *sinfo = xdp_get_shared_info_from_buff(xdp);
> > +	int i;
> > +
> > +	page_pool_ref_page(virt_to_page(xdp->data));
> > +	if (likely(!xdp_buff_has_frags(xdp)))
> > +		return;
> > +
> > +	for (i = 0; i < sinfo->nr_frags; i++) {
> > +		skb_frag_t *frag = &sinfo->frags[i];
> > +
> > +		page_pool_ref_page(netmem_to_page(frag->netmem));
> > +	}
> > +}
> > +
> > +static void veth_xdp_get(struct xdp_buff *xdp)
> > +{
> > +	xdp->rxq->mem.type == MEM_TYPE_PAGE_POOL ?
> > +		veth_xdp_get_pp(xdp) : veth_xdp_get_shared(xdp);
> > +}
> > +
> 
> [ ... ]
> 
> > @@ -822,24 +831,24 @@ static struct sk_buff *veth_xdp_rcv_skb(struct veth_rq *rq,
> >  	case XDP_TX:
> >  		veth_xdp_get(xdp);
> >  		consume_skb(skb);
> > -		xdp->rxq->mem = rq->xdp_mem;
> >  		if (unlikely(veth_xdp_tx(rq, xdp, bq) < 0)) {
> >  			trace_xdp_exception(rq->dev, xdp_prog, act);
> >  			stats->rx_drops++;
> >  			goto err_xdp;
> >  		}
> 
> Can veth_xdp_rcv_skb() leak page references in the XDP_TX error path?
> When veth_xdp_tx() fails, the code jumps to err_xdp without releasing
> the page references acquired by veth_xdp_get(). The skb was already
> consumed, so the pages still have the extra references from either
> veth_xdp_get_pp() or veth_xdp_get_shared().

You're missing xdp_return_buff() on err_xdp path.

> 
> veth_xdp_tx() can return -EOVERFLOW when xdp_convert_buff_to_frame()
> returns NULL, which is a reachable error path.
> 
> >  		stats->xdp_tx++;
> > +		rq->xdp_rxq.mem = rq->xdp_mem;
> >  		rcu_read_unlock();
> >  		goto xdp_xmit;
> >  	case XDP_REDIRECT:
> >  		veth_xdp_get(xdp);
> >  		consume_skb(skb);
> > -		xdp->rxq->mem = rq->xdp_mem;
> >  		if (xdp_do_redirect(rq->dev, xdp, xdp_prog)) {
> >  			stats->rx_drops++;
> >  			goto err_xdp;
> >  		}
> 
> Similar question for the XDP_REDIRECT case - when xdp_do_redirect()
> fails, does this also leak the page references taken by veth_xdp_get()?
> 
> >  		stats->xdp_redirect++;
> > +		rq->xdp_rxq.mem = rq->xdp_mem;
> >  		rcu_read_unlock();
> >  		goto xdp_xmit;
> >  	default:
> 
> 
> ```
> 
> ---
> AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
> See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
> 
> In-Reply-To-Subject: `veth: update mem type in xdp_buff`
> CI run summary: https://github.com/kernel-patches/bpf/actions/runs/18923812102