netdev - Re: [PATCH net-next] r8169: Add 64bit statistics

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: <1321534921.2751.31.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC>
Date:	Thu, 17 Nov 2011 14:02:01 +0100
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Junchang Wang <junchangwang@...il.com>
Cc:	Stephen Hemminger <stephen.hemminger@...tta.com>,
	netdev@...r.kernel.org, romieu@...zoreil.com,
	nic swsd <nic_swsd@...ltek.com>
Subject: Re: [PATCH net-next] r8169: Add 64bit statistics

Le jeudi 17 novembre 2011 à 18:59 +0800, Junchang Wang a écrit :
> > Yes, look at sky2.c for a template
> >
> > drivers/net/ethernet/marvell/sky2.c contains code like that
> > (different syncp for rx/tx)
> >
> > TX path:
> >                        u64_stats_update_begin(&sky2->tx_stats.syncp);
> >                        ++sky2->tx_stats.packets;
> >                        sky2->tx_stats.bytes += skb->len;
> >                        u64_stats_update_end(&sky2->tx_stats.syncp);
> >
> >
> > RX path:
> >
> >        u64_stats_update_begin(&sky2->rx_stats.syncp);
> >        sky2->rx_stats.packets += packets;
> >        sky2->rx_stats.bytes += bytes;
> >        u64_stats_update_end(&sky2->rx_stats.syncp);
> >
> 
> Thanks, Eric.
> 
> I'm still confused about why we need two sync entries. Please correct
> me if I'm wrong.
> 
> Take r8169 for example, All statistic entries are updated in
> rtl8169_rx_interrupt() or rtl8169_tx_interrupt(). Those two functions
> are called in rtl8169_poll().
> As far as I know, rtl8169_poll() is protected by NAPI_STATE_SCHED bit
> to run on a single core at one moment. So there is not compulsory to
> use two sync entries.
> 
> One benefit from two sync is that readers can avoid many retries.
> 

Oh well...

Dont try to optimize all this stuff, I spent some time on it already,
just use the infrastructure and forget about races and bugs.

The short answer is :

CPU1                    CPU2                  
TX path                 RX path               


since CPU1 & CPU2 will do the update_begin(), and the memory increment
is not atomic, we might lose one increment, and block a stat reader
forever.




--
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