linux-kernel - Re: [PATCH 4.19 625/639] packet: fix data-race in fanout_flow_is

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: <20200126182917.GA26911@amd>
Date:   Sun, 26 Jan 2020 19:29:18 +0100
From:   Pavel Machek <pavel@...x.de>
To:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc:     linux-kernel@...r.kernel.org, stable@...r.kernel.org,
        Eric Dumazet <edumazet@...gle.com>,
        Willem de Bruijn <willemb@...gle.com>,
        "David S. Miller" <davem@...emloft.net>,
        Sasha Levin <sashal@...nel.org>
Subject: Re: [PATCH 4.19 625/639] packet: fix data-race in
 fanout_flow_is_huge()

Hi!

> From: Eric Dumazet <edumazet@...gle.com>
> 
> [ Upstream commit b756ad928d98e5ef0b74af7546a6a31a8dadde00 ]
> 
> KCSAN reported the following data-race [1]
> 
> Adding a couple of READ_ONCE()/WRITE_ONCE() should silence it.
> 
> Since the report hinted about multiple cpus using the history
> concurrently, I added a test avoiding writing on it if the
> victim slot already contains the desired value.

>  static bool fanout_flow_is_huge(struct packet_sock *po, struct sk_buff *skb)
>  {
> -	u32 rxhash;
> +	u32 *history = po->rollover->history;
> +	u32 victim, rxhash;
>  	int i, count = 0;
>  
>  	rxhash = skb_get_hash(skb);
>  	for (i = 0; i < ROLLOVER_HLEN; i++)
> -		if (po->rollover->history[i] == rxhash)
> +		if (READ_ONCE(history[i]) == rxhash)
>  			count++;
>  
> -	po->rollover->history[prandom_u32() % ROLLOVER_HLEN] = rxhash;
> +	victim = prandom_u32() % ROLLOVER_HLEN;
> +
> +	/* Avoid dirtying the cache line if possible */
> +	if (READ_ONCE(history[victim]) != rxhash)
> +		WRITE_ONCE(history[victim], rxhash);
> +

Replacing simple asignment with if() is ... not nice and with all the
"volatile" magic in _ONCE macros may not be win for
everyone. [Actually, I don't think this is win here. This is not
exactly hot path, is it? Is it likely that array aready contains
required value?]

If this is going to get more common, should we get
WRITE_ONCE_NONDIRTY() macro hiding the uglyness?

Best regards,
								Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Download attachment "signature.asc" of type "application/pgp-signature" (182 bytes)