netdev - Re: [PATCH net-next] net: stmmac: add BQL support

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 PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

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

Date:	Tue, 30 Dec 2014 22:27:31 +0100
From:	Beniamino Galvani <b.galvani@...il.com>
To:	Florian Fainelli <f.fainelli@...il.com>
Cc:	"David S. Miller" <davem@...emloft.net>,
	Giuseppe Cavallaro <peppe.cavallaro@...com>,
	netdev <netdev@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH net-next] net: stmmac: add BQL support

On Sun, Dec 28, 2014 at 11:48:14PM -0800, Florian Fainelli wrote:
> 2014-12-28 6:57 GMT-08:00 Beniamino Galvani <b.galvani@...il.com>:
> > Add support for Byte Queue Limits to the STMicro MAC driver.
> >
> > Tested on a Amlogic S805 Cortex-A5 board, where the use of BQL
> > slightly decreases the ping latency from ~10ms to ~3ms when the
> > 100Mbps link is saturated by TCP streams. No difference is
> > observed at 1Gbps.
> >
> > Signed-off-by: Beniamino Galvani <b.galvani@...il.com>
> > ---
> 
> [snip]
> 
> >         priv->dev->stats.tx_errors++;
> > @@ -2049,6 +2057,7 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
> >                 skb_tx_timestamp(skb);
> >
> >         priv->hw->dma->enable_dma_transmission(priv->ioaddr);
> > +       netdev_sent_queue(dev, skb->len);
> 
> You are introducing a potential use after free here in case tx_lock is
> eliminated one day and your TX reclaim logic kicks in and frees the
> freshly transmitted SKB, it would be safer to just cache skb->len in a
> local variable, and use it here.

Ok, I will change this part; probably a simpler solution is to call
netdev_sent_queue() before enabling the DMA like in other drivers.

BTW, I'm not sure this lock is really needed, since it should be
possible to safely access the ring without locks from both the
transmit and the reclaim functions if the pointers are updated
carefully. So maybe it will be really removed one day.

Beniamino

> 
> >
> >         spin_unlock(&priv->tx_lock);
> >         return NETDEV_TX_OK;

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html