From: Jakub Kicinski <kuba@kernel.org>
To: Heiner Kallweit <hkallweit1@gmail.com>
Cc: davem@davemloft.net, netdev@vger.kernel.org, edumazet@google.com,
pabeni@redhat.com
Subject: Re: [PATCH net-next 1/3] net: provide macros for commonly copied lockless queue stop/wake code
Date: Sat, 1 Apr 2023 11:03:41 -0700 [thread overview]
Message-ID: <20230401110341.7213900c@kernel.org> (raw)
In-Reply-To: <5b997cab-d5ee-9772-7555-8d7e8eaccb16@gmail.com>
On Sat, 1 Apr 2023 17:04:17 +0200 Heiner Kallweit wrote:
> > +#define netif_tx_queue_try_stop(txq, get_desc, start_thrs) \
> > + ({ \
> > + int _res; \
> > + \
> > + netif_tx_stop_queue(txq); \
> > + \
> > + smp_mb(); \
>
> Wouldn't a smp_mb__after_atomic() be sufficient here, because netif_tx_stop_queue()
> includes a set_bit()? At least on X86 this would result in a no-op.
Yup, good point, __after_atomic() should be perfectly fine. I didn't
think too much about the smp_mb() 'cause it's what existing code was
using and it's a slow path. I'll fix in v3.
> > + \
> > + /* We need to check again in a case another \
> > + * CPU has just made room available. \
> > + */ \
> > + _res = 0; \
> > + if (unlikely(get_desc >= start_thrs)) { \
> > + netif_tx_start_queue(txq); \
> > + _res = -1; \
> > + } \
> > + _res; \
> > + }) \
> > +
> > +/**
> > + * netif_tx_queue_maybe_stop() - locklessly stop a Tx queue, if needed
> > + * @txq: struct netdev_queue to stop/start
> > + * @get_desc: get current number of free descriptors (see requirements below!)
> > + * @stop_thrs: minimal number of available descriptors for queue to be left
> > + * enabled
> > + * @start_thrs: minimal number of descriptors to re-enable the queue, can be
> > + * equal to @stop_thrs or higher to avoid frequent waking
> > + *
> > + * All arguments may be evaluated multiple times, beware of side effects.
> > + * @get_desc must be a formula or a function call, it must always
> > + * return up-to-date information when evaluated!
> > + * Expected to be used from ndo_start_xmit, see the comment on top of the file.
> > + *
> > + * Returns:
> > + * 0 if the queue was stopped
> > + * 1 if the queue was left enabled
> > + * -1 if the queue was re-enabled (raced with waking)
> > + */
> > +#define netif_tx_queue_maybe_stop(txq, get_desc, stop_thrs, start_thrs) \
> > + ({ \
> > + int _res; \
> > + \
> > + _res = 1; \
> > + if (unlikely(get_desc < stop_thrs)) \
> > + _res = netif_tx_queue_try_stop(txq, get_desc, \
> > + start_thrs); \
> > + _res; \
> > + }) \
> > +
> > +#define __netif_tx_queue_try_wake(txq, get_desc, start_thrs, down_cond) \
>
> Maybe I miss something, but: Why the get_desc and start_thrs parameters
> if they aren't used?
Copy'n'paste fail, will fix :(
next prev parent reply other threads:[~2023-04-01 18:03 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-01 5:12 [PATCH net-next 0/3] net: provide macros for commonly copied lockless queue stop/wake code Jakub Kicinski
2023-04-01 5:12 ` [PATCH net-next 1/3] " Jakub Kicinski
2023-04-01 15:04 ` Heiner Kallweit
2023-04-01 18:03 ` Jakub Kicinski [this message]
2023-04-01 15:18 ` Heiner Kallweit
2023-04-01 18:58 ` Jakub Kicinski
2023-04-01 20:41 ` Heiner Kallweit
2023-04-03 15:18 ` Alexander Duyck
2023-04-03 15:56 ` Jakub Kicinski
2023-04-03 18:11 ` Alexander Duyck
2023-04-03 19:03 ` Jakub Kicinski
2023-04-03 20:27 ` Alexander Duyck
2023-04-05 22:20 ` Paul E. McKenney
2023-04-06 5:15 ` Herbert Xu
2023-04-06 14:17 ` Paul E. McKenney
2023-04-06 14:46 ` Jakub Kicinski
2023-04-06 15:45 ` Paul E. McKenney
2023-04-06 15:56 ` Jakub Kicinski
2023-04-06 16:25 ` Paul E. McKenney
2023-04-07 0:58 ` Herbert Xu
2023-04-07 1:03 ` Jakub Kicinski
2023-04-07 1:14 ` Herbert Xu
2023-04-07 1:21 ` Jakub Kicinski
2023-04-04 6:39 ` Herbert Xu
2023-04-04 22:36 ` Jakub Kicinski
2023-04-01 5:12 ` [PATCH net-next 2/3] ixgbe: use new queue try_stop/try_wake macros Jakub Kicinski
2023-04-01 5:12 ` [PATCH net-next 3/3] bnxt: " Jakub Kicinski
2023-04-01 18:35 ` Michael Chan
-- strict thread matches above, loose matches on Subject: below --
2023-03-22 23:30 [PATCH net-next 1/3] net: provide macros for commonly copied lockless queue stop/wake code Jakub Kicinski
2023-03-23 0:35 ` Andrew Lunn
2023-03-23 1:04 ` Jakub Kicinski
2023-03-23 21:02 ` Andrew Lunn
2023-03-23 22:46 ` Jakub Kicinski
2023-03-23 3:05 ` Yunsheng Lin
2023-03-23 3:27 ` Jakub Kicinski
2023-03-23 4:53 ` Pavan Chebbi
2023-03-23 5:08 ` Jakub Kicinski
2023-03-23 16:05 ` Alexander H Duyck
2023-03-24 3:09 ` Jakub Kicinski
2023-03-24 15:45 ` Alexander Duyck
2023-03-24 21:28 ` Jakub Kicinski
2023-03-26 21:23 ` Alexander Duyck
2023-03-29 0:56 ` Jakub Kicinski
2023-03-30 14:56 ` Paolo Abeni
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230401110341.7213900c@kernel.org \
--to=kuba@kernel.org \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=hkallweit1@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).