[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZC52VRfUOOObx2fw@gondor.apana.org.au>
Date: Thu, 6 Apr 2023 15:35:49 +0800
From: Herbert Xu <herbert@...dor.apana.org.au>
To: Jakub Kicinski <kuba@...nel.org>
Cc: davem@...emloft.net, netdev@...r.kernel.org, edumazet@...gle.com,
pabeni@...hat.com, alexander.duyck@...il.com, hkallweit1@...il.com,
andrew@...n.ch, willemb@...gle.com, michael.chan@...adcom.com,
jesse.brandeburg@...el.com, anthony.l.nguyen@...el.com
Subject: Re: [PATCH net-next v3 7/7] net: piggy back on the memory barrier in
bql when waking queues
On Wed, Apr 05, 2023 at 03:31:34PM -0700, Jakub Kicinski wrote:
> \
> @@ -82,10 +82,26 @@
> _res; \
> }) \
>
> +/* Variant of netdev_tx_completed_queue() which guarantees smp_mb() if
> + * @bytes != 0, regardless of kernel config.
> + */
> +static inline void
> +netdev_txq_completed_mb(struct netdev_queue *dev_queue,
> + unsigned int pkts, unsigned int bytes)
> +{
> +#ifdef CONFIG_BQL
> + netdev_tx_completed_queue(dev_queue, pkts, bytes);
> +#else
> + if (bytes)
> + smp_mb();
> +#endif
> +}
Minor nit, I would write this as
if (IS_ENABLED(CONFIG_BQL))
netdev_tx_completed_queue(dev_queue, pkts, bytes);
else if (bytes)
smp_mb();
Actually, why is this checking bytes while the caller is checking
pkts? Do we need to check them at all? If pkts/bytes is commonly
non-zero, then we should just do a barrier unconditionally and make
the uncommon path pay the penalty.
Thanks,
--
Email: Herbert Xu <herbert@...dor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
Powered by blists - more mailing lists