Re: [PATCH linux-wpan] 6lowpan: ensure header compression does not corrupt ipv6 header

[Alexander Aring <alex.aring@gmail.com>]

Hi Marcel,

good to see you. I think this is bug occurs also in 6LoWPAN bluetooth.

I cc Jukka here.

On Sat, Sep 20, 2014 at 04:55:27PM +0200, Marcel Holtmann wrote:
> Hi Simon,
> 
> I am missing a proper commit message with explanation here. One-line subject online commit message are really only ever acceptable if it is a dead trivial fix. And even then I prefer a proper commit message with detailed explanation what the patch changes and why.
> 
> > Signed-off-by: Simon Vincent <simon.vincent@xsilon.com>
> > ---
> > net/ieee802154/6lowpan_rtnl.c | 131 +++++++++++++++++++++++++++++++-----------
> > 1 file changed, 97 insertions(+), 35 deletions(-)
> > 
> > diff --git a/net/ieee802154/6lowpan_rtnl.c b/net/ieee802154/6lowpan_rtnl.c
> > index 6591d27..cb94504 100644
> > --- a/net/ieee802154/6lowpan_rtnl.c
> > +++ b/net/ieee802154/6lowpan_rtnl.c
> > @@ -71,6 +71,26 @@ struct lowpan_dev_record {
> > 	struct list_head list;
> > };
> > 
> > +union lowpan_addr_u {
> > +	/* IPv6 needs big endian here */
> > +	__be64 extended;
> > +	__be16 short_;
> 
> I do not like short_ as field name. I know why you are doing this, but why bother. You always access via u.short anyway. Then again, it seems it is not used at all.
> 
Sorry my fault. I used "short_" in the rework, I don't like it too, but
would to change it in future. Do you have a better name?

The not using of this parameter is... that it should be using but we
doesn't support it right now. :-)

> > +};
> > +
> > +/* don't save pan id, it's intra pan */
> > +struct lowpan_addr {
> > +	/* non converted address mode bits here
> > +	 * make this at first, improve memcmp on this struct
> > +	 */
> > +	__le16 mode;
> > +	union lowpan_addr_u u;
> 
> If you do not need the union declaration by itself then you can just add it into the struct directly. No need for these extra _u declaration.
> 

yea, he mainly grab code from other branch and I used the union for some
parameter. But of course for this patch he can change it.

> > +};
> > +
> > +struct lowpan_addr_info {
> > +	struct lowpan_addr daddr;
> > +	struct lowpan_addr saddr;
> > +};
> > +
> > static inline struct
> > lowpan_dev_info *lowpan_dev_info(const struct net_device *dev)
> > {
> > @@ -85,14 +104,21 @@ static inline void lowpan_address_flip(u8 *src, u8 *dest)
> > 		(dest)[IEEE802154_ADDR_LEN - i - 1] = (src)[i];
> > }
> > 
> > +static inline struct
> > +lowpan_addr_info *lowpan_skb_priv(const struct sk_buff *skb)
> > +{
> > +	WARN_ON_ONCE(skb_headroom(skb) < sizeof(struct lowpan_addr_info));
> > +	return (struct lowpan_addr_info *)(skb->data -
> > +			sizeof(struct lowpan_addr_info));
> > +}
> > +
> > static int lowpan_header_create(struct sk_buff *skb, struct net_device *dev,
> > 				unsigned short type, const void *_daddr,
> > 				const void *_saddr, unsigned int len)
> > {
> > 	const u8 *saddr = _saddr;
> > 	const u8 *daddr = _daddr;
> > -	struct ieee802154_addr sa, da;
> > -	struct ieee802154_mac_cb *cb = mac_cb_init(skb);
> > +	struct lowpan_addr_info *info;
> > 
> > 	/* TODO:
> > 	 * if this package isn't ipv6 one, where should it be routed?
> > @@ -106,41 +132,15 @@ static int lowpan_header_create(struct sk_buff *skb, struct net_device *dev,
> > 	raw_dump_inline(__func__, "saddr", (unsigned char *)saddr, 8);
> > 	raw_dump_inline(__func__, "daddr", (unsigned char *)daddr, 8);
> > 
> > -	lowpan_header_compress(skb, dev, type, daddr, saddr, len);
> > -
> > -	/* NOTE1: I'm still unsure about the fact that compression and WPAN
> > -	 * header are created here and not later in the xmit. So wait for
> > -	 * an opinion of net maintainers.
> > -	 */
> > -	/* NOTE2: to be absolutely correct, we must derive PANid information
> > -	 * from MAC subif of the 'dev' and 'real_dev' network devices, but
> > -	 * this isn't implemented in mainline yet, so currently we assign 0xff
> > -	 */
> > -	cb->type = IEEE802154_FC_TYPE_DATA;
> > -
> > -	/* prepare wpan address data */
> > -	sa.mode = IEEE802154_ADDR_LONG;
> > -	sa.pan_id = ieee802154_mlme_ops(dev)->get_pan_id(dev);
> > -	sa.extended_addr = ieee802154_devaddr_from_raw(saddr);
> > +	info = lowpan_skb_priv(skb);
> > 
> > -	/* intra-PAN communications */
> > -	da.pan_id = sa.pan_id;
> > +	info->daddr.mode = IEEE802154_ADDR_LONG;
> > +	memcpy(&info->daddr.u.extended, daddr, sizeof(info->daddr.u.extended));
> > 
> > -	/* if the destination address is the broadcast address, use the
> > -	 * corresponding short address
> > -	 */
> > -	if (lowpan_is_addr_broadcast(daddr)) {
> > -		da.mode = IEEE802154_ADDR_SHORT;
> > -		da.short_addr = cpu_to_le16(IEEE802154_ADDR_BROADCAST);
> > -	} else {
> > -		da.mode = IEEE802154_ADDR_LONG;
> > -		da.extended_addr = ieee802154_devaddr_from_raw(daddr);
> > -	}
> > +	info->saddr.mode = IEEE802154_ADDR_LONG;
> > +	memcpy(&info->saddr.u.extended, saddr, sizeof(info->daddr.u.extended));
> > 
> > -	cb->ackreq = !lowpan_is_addr_broadcast(daddr);
> > -
> > -	return dev_hard_header(skb, lowpan_dev_info(dev)->real_dev,
> > -			type, (void *)&da, (void *)&sa, 0);
> > +	return 0;
> > }
> > 
> > static int lowpan_give_skb_to_devices(struct sk_buff *skb,
> > @@ -338,13 +338,74 @@ err:
> > 	return rc;
> > }
> > 
> > +static int lowpan_header(struct sk_buff *skb, struct net_device *dev)
> > +{
> > +	struct ieee802154_addr sa, da;
> > +	struct ieee802154_mac_cb *cb = mac_cb_init(skb);
> > +	struct lowpan_addr_info info;
> > +	void *daddr, *saddr;
> > +
> > +	memcpy(&info, lowpan_skb_priv(skb), sizeof(info));
> > +
> > +	/* TODO complicated bug why we support extended_addr only */
> 
> This is not a good comment. Explain why since especially only extended addresses are supported, I have no idea on why we are adding all these struct addition in the first place.
> 

again my fault. Just a hint for me in rework branch to add some better
comment there, or fix the issue.

> > +	daddr = &info.daddr.u.extended;
> > +	saddr = &info.saddr.u.extended;
> > +
> > +	lowpan_header_compress(skb, dev, ETH_P_IPV6, daddr, saddr, skb->len);
> > +
> > +	/* NOTE1: I'm still unsure about the fact that compression and WPAN
> > +	 * header are created here and not later in the xmit. So wait for
> > +	 * an opinion of net maintainers.
> > +	 */
> 
> Now I wonder why this has not yet been answered and you are in the need of copying around. Maybe this should have been addressed by now.
> 

Very old comment, should be removed.

> > +	/* NOTE2: to be absolutely correct, we must derive PANid information
> > +	 * from MAC subif of the 'ldev' and 'wdev' network devices, but
> > +	 * this isn't implemented in mainline yet, so currently we assign 0xff
> > +	 */

same.

> > +	cb->type = IEEE802154_FC_TYPE_DATA;
> > +
> > +	/* prepare wpan address data */
> > +	sa.mode = IEEE802154_ADDR_LONG;
> > +	sa.pan_id = ieee802154_mlme_ops(dev)->get_pan_id(dev);
> > +	sa.extended_addr = ieee802154_devaddr_from_raw(saddr);
> > +
> > +	/* intra-PAN communications */
> > +	da.pan_id = sa.pan_id;
> > +
> > +	/* if the destination address is the broadcast address, use the
> > +	 * corresponding short address
> > +	 */
> > +	if (lowpan_is_addr_broadcast((const u8 *)daddr)) {
> > +		da.mode = IEEE802154_ADDR_SHORT;
> > +		da.short_addr = cpu_to_le16(IEEE802154_ADDR_BROADCAST);
> > +		cb->ackreq = false;
> > +	} else {
> > +		da.mode = IEEE802154_ADDR_LONG;
> > +		da.extended_addr = ieee802154_devaddr_from_raw(daddr);
> > +		cb->ackreq = true;
> > +	}
> > +
> > +	return dev_hard_header(skb, lowpan_dev_info(dev)->real_dev,
> > +			ETH_P_IPV6, (void *)&da, (void *)&sa, 0);
> > +}
> > +
> > static netdev_tx_t lowpan_xmit(struct sk_buff *skb, struct net_device *dev)
> > {
> > 	struct ieee802154_hdr wpan_hdr;
> > -	int max_single;
> > +	int max_single, ret;
> > 
> > 	pr_debug("package xmit\n");
> > 
> > +	/* We must take a copy of the skb before we modify/replace the ipv6
> > +	 * header as the header could be used elsewhere
> > +	 */
> > +	skb = skb_unshare(skb, GFP_KERNEL);
> > +

need to be GFP_ATOMIC, xmit callback has atomic context.

- Alex