[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220309134341.6037bb1d@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date: Wed, 9 Mar 2022 13:43:41 -0800
From: Jakub Kicinski <kuba@...nel.org>
To: Eric Dumazet <eric.dumazet@...il.com>
Cc: "David S . Miller" <davem@...emloft.net>,
netdev <netdev@...r.kernel.org>,
Eric Dumazet <edumazet@...gle.com>,
jeffreyji <jeffreyji@...gle.com>,
Brian Vazquez <brianvv@...gle.com>
Subject: Re: [PATCH net-next] net: add per-cpu storage and net->core_stats
On Wed, 9 Mar 2022 13:18:08 -0800 Eric Dumazet wrote:
> From: Eric Dumazet <edumazet@...gle.com>
>
> Before adding yet another possibly contended atomic_long_t,
> it is time to add per-cpu storage for existing ones:
> dev->tx_dropped, dev->rx_dropped, and dev->rx_nohandler
>
> Because many devices do not have to increment such counters,
> allocate the per-cpu storage on demand, so that dev_get_stats()
> does not have to spend considerable time folding zero counters.
>
> Note that some drivers have abused these counters which
> were supposed to be only used by core networking stack.
>
> Signed-off-by: Eric Dumazet <edumazet@...gle.com>
> Cc: jeffreyji <jeffreyji@...gle.com>
> Cc: Brian Vazquez <brianvv@...gle.com>
Reviewed-by: Jakub Kicinski <kuba@...nel.org>
Things to mention in passing, in case there is a v2:
- we could use a different header for this, netdevice.h is a monster
- netdev_core_stats_alloc() could have the last if inverted to keep
success path unindented
Powered by blists - more mailing lists