public inbox for io-uring@vger.kernel.org
 help / color / mirror / Atom feed
From: David Wei <dw@davidwei.uk>
To: io-uring@vger.kernel.org, netdev@vger.kernel.org
Cc: David Wei <dw@davidwei.uk>, Jens Axboe <axboe@kernel.dk>,
	Pavel Begunkov <asml.silence@gmail.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jesper Dangaard Brouer <hawk@kernel.org>,
	David Ahern <dsahern@kernel.org>,
	Mina Almasry <almasrymina@google.com>
Subject: [PATCH v1 08/15] net: add helper executing custom callback from napi
Date: Mon,  7 Oct 2024 15:15:56 -0700	[thread overview]
Message-ID: <20241007221603.1703699-9-dw@davidwei.uk> (raw)
In-Reply-To: <20241007221603.1703699-1-dw@davidwei.uk>

From: Pavel Begunkov <asml.silence@gmail.com>

It's useful to have napi private bits and pieces like page pool's fast
allocating cache, so that the hot allocation path doesn't have to do any
additional synchronisation. In case of io_uring memory provider
introduced in following patches, we keep the consumer end of the
io_uring's refill queue private to napi as it's a hot path.

However, from time to time we need to synchronise with the napi, for
example to add more user memory or allocate fallback buffers. Add a
helper function napi_execute that allows to run a custom callback from
under napi context so that it can access and modify napi protected
parts of io_uring. It works similar to busy polling and stops napi from
running in the meantime, so it's supposed to be a slow control path.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: David Wei <dw@davidwei.uk>
---
 include/net/busy_poll.h |  6 +++++
 net/core/dev.c          | 53 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 59 insertions(+)

diff --git a/include/net/busy_poll.h b/include/net/busy_poll.h
index f03040baaefd..3fd9e65731e9 100644
--- a/include/net/busy_poll.h
+++ b/include/net/busy_poll.h
@@ -47,6 +47,7 @@ bool sk_busy_loop_end(void *p, unsigned long start_time);
 void napi_busy_loop(unsigned int napi_id,
 		    bool (*loop_end)(void *, unsigned long),
 		    void *loop_end_arg, bool prefer_busy_poll, u16 budget);
+void napi_execute(unsigned napi_id, void (*cb)(void *), void *cb_arg);
 
 void napi_busy_loop_rcu(unsigned int napi_id,
 			bool (*loop_end)(void *, unsigned long),
@@ -63,6 +64,11 @@ static inline bool sk_can_busy_loop(struct sock *sk)
 	return false;
 }
 
+static inline void napi_execute(unsigned napi_id,
+				void (*cb)(void *), void *cb_arg)
+{
+}
+
 #endif /* CONFIG_NET_RX_BUSY_POLL */
 
 static inline unsigned long busy_loop_current_time(void)
diff --git a/net/core/dev.c b/net/core/dev.c
index 1e740faf9e78..ba2f43cf5517 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -6497,6 +6497,59 @@ void napi_busy_loop(unsigned int napi_id,
 }
 EXPORT_SYMBOL(napi_busy_loop);
 
+void napi_execute(unsigned napi_id,
+		  void (*cb)(void *), void *cb_arg)
+{
+	struct napi_struct *napi;
+	bool done = false;
+	unsigned long val;
+	void *have_poll_lock = NULL;
+
+	rcu_read_lock();
+
+	napi = napi_by_id(napi_id);
+	if (!napi) {
+		rcu_read_unlock();
+		return;
+	}
+
+	if (!IS_ENABLED(CONFIG_PREEMPT_RT))
+		preempt_disable();
+	for (;;) {
+		local_bh_disable();
+		val = READ_ONCE(napi->state);
+
+		/* If multiple threads are competing for this napi,
+		* we avoid dirtying napi->state as much as we can.
+		*/
+		if (val & (NAPIF_STATE_DISABLE | NAPIF_STATE_SCHED |
+			  NAPIF_STATE_IN_BUSY_POLL))
+			goto restart;
+
+		if (cmpxchg(&napi->state, val,
+			   val | NAPIF_STATE_IN_BUSY_POLL |
+				 NAPIF_STATE_SCHED) != val)
+			goto restart;
+
+		have_poll_lock = netpoll_poll_lock(napi);
+		cb(cb_arg);
+		done = true;
+		gro_normal_list(napi);
+		local_bh_enable();
+		break;
+restart:
+		local_bh_enable();
+		if (unlikely(need_resched()))
+			break;
+		cpu_relax();
+	}
+	if (done)
+		busy_poll_stop(napi, have_poll_lock, false, 1);
+	if (!IS_ENABLED(CONFIG_PREEMPT_RT))
+		preempt_enable();
+	rcu_read_unlock();
+}
+
 #endif /* CONFIG_NET_RX_BUSY_POLL */
 
 static void napi_hash_add(struct napi_struct *napi)
-- 
2.43.5


  parent reply	other threads:[~2024-10-07 22:16 UTC|newest]

Thread overview: 126+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-10-07 22:15 [PATCH v1 00/15] io_uring zero copy rx David Wei
2024-10-07 22:15 ` [PATCH v1 01/15] net: devmem: pull struct definitions out of ifdef David Wei
2024-10-09 20:17   ` Mina Almasry
2024-10-09 23:16     ` Pavel Begunkov
2024-10-10 18:01       ` Mina Almasry
2024-10-10 18:57         ` Pavel Begunkov
2024-10-13 22:38           ` Pavel Begunkov
2024-10-07 22:15 ` [PATCH v1 02/15] net: prefix devmem specific helpers David Wei
2024-10-09 20:19   ` Mina Almasry
2024-10-07 22:15 ` [PATCH v1 03/15] net: generalise net_iov chunk owners David Wei
2024-10-08 15:46   ` Stanislav Fomichev
2024-10-08 16:34     ` Pavel Begunkov
2024-10-09 16:28       ` Stanislav Fomichev
2024-10-11 18:44         ` David Wei
2024-10-11 22:02           ` Pavel Begunkov
2024-10-11 22:25             ` Mina Almasry
2024-10-11 23:12               ` Pavel Begunkov
2024-10-09 20:44   ` Mina Almasry
2024-10-09 22:13     ` Pavel Begunkov
2024-10-09 22:19     ` Pavel Begunkov
2024-10-07 22:15 ` [PATCH v1 04/15] net: page_pool: create hooks for custom page providers David Wei
2024-10-09 20:49   ` Mina Almasry
2024-10-09 22:02     ` Pavel Begunkov
2024-10-07 22:15 ` [PATCH v1 05/15] net: prepare for non devmem TCP memory providers David Wei
2024-10-09 20:56   ` Mina Almasry
2024-10-09 21:45     ` Pavel Begunkov
2024-10-13 22:33       ` Pavel Begunkov
2024-10-07 22:15 ` [PATCH v1 06/15] net: page_pool: add ->scrub mem provider callback David Wei
2024-10-09 21:00   ` Mina Almasry
2024-10-09 21:59     ` Pavel Begunkov
2024-10-10 17:54       ` Mina Almasry
2024-10-13 17:25         ` David Wei
2024-10-14 13:37           ` Pavel Begunkov
2024-10-14 22:58           ` Mina Almasry
2024-10-16 17:42             ` Pavel Begunkov
2024-11-01 17:18               ` Mina Almasry
2024-11-01 18:35                 ` Pavel Begunkov
2024-11-01 19:24                   ` Mina Almasry
2024-11-01 21:38                     ` Pavel Begunkov
2024-11-04 20:42                       ` Mina Almasry
2024-11-04 21:27                         ` Pavel Begunkov
2024-10-07 22:15 ` [PATCH v1 07/15] net: page pool: add helper creating area from pages David Wei
2024-10-09 21:11   ` Mina Almasry
2024-10-09 21:34     ` Pavel Begunkov
2024-10-07 22:15 ` David Wei [this message]
2024-10-08 22:25   ` [PATCH v1 08/15] net: add helper executing custom callback from napi Joe Damato
2024-10-09 15:09     ` Pavel Begunkov
2024-10-09 16:13       ` Joe Damato
2024-10-09 19:12         ` Pavel Begunkov
2024-10-07 22:15 ` [PATCH v1 09/15] io_uring/zcrx: add interface queue and refill queue David Wei
2024-10-09 17:50   ` Jens Axboe
2024-10-09 18:09     ` Jens Axboe
2024-10-09 19:08     ` Pavel Begunkov
2024-10-11 22:11     ` Pavel Begunkov
2024-10-13 17:32     ` David Wei
2024-10-07 22:15 ` [PATCH v1 10/15] io_uring/zcrx: add io_zcrx_area David Wei
2024-10-09 18:02   ` Jens Axboe
2024-10-09 19:05     ` Pavel Begunkov
2024-10-09 19:06       ` Jens Axboe
2024-10-09 21:29   ` Mina Almasry
2024-10-07 22:15 ` [PATCH v1 11/15] io_uring/zcrx: implement zerocopy receive pp memory provider David Wei
2024-10-09 18:10   ` Jens Axboe
2024-10-09 22:01   ` Mina Almasry
2024-10-09 22:58     ` Pavel Begunkov
2024-10-10 18:19       ` Mina Almasry
2024-10-10 20:26         ` Pavel Begunkov
2024-10-10 20:53           ` Mina Almasry
2024-10-10 20:58             ` Mina Almasry
2024-10-10 21:22             ` Pavel Begunkov
2024-10-11  0:32               ` Mina Almasry
2024-10-11  1:49                 ` Pavel Begunkov
2024-10-07 22:16 ` [PATCH v1 12/15] io_uring/zcrx: add io_recvzc request David Wei
2024-10-09 18:28   ` Jens Axboe
2024-10-09 18:51     ` Pavel Begunkov
2024-10-09 19:01       ` Jens Axboe
2024-10-09 19:27         ` Pavel Begunkov
2024-10-09 19:42           ` Jens Axboe
2024-10-09 19:47             ` Pavel Begunkov
2024-10-09 19:50               ` Jens Axboe
2024-10-07 22:16 ` [PATCH v1 13/15] io_uring/zcrx: add copy fallback David Wei
2024-10-08 15:58   ` Stanislav Fomichev
2024-10-08 16:39     ` Pavel Begunkov
2024-10-08 16:40     ` David Wei
2024-10-09 16:30       ` Stanislav Fomichev
2024-10-09 23:05         ` Pavel Begunkov
2024-10-11  6:22           ` David Wei
2024-10-11 14:43             ` Stanislav Fomichev
2024-10-09 18:38   ` Jens Axboe
2024-10-07 22:16 ` [PATCH v1 14/15] io_uring/zcrx: set pp memory provider for an rx queue David Wei
2024-10-09 18:42   ` Jens Axboe
2024-10-10 13:09     ` Pavel Begunkov
2024-10-10 13:19       ` Jens Axboe
2024-10-07 22:16 ` [PATCH v1 15/15] io_uring/zcrx: throttle receive requests David Wei
2024-10-09 18:43   ` Jens Axboe
2024-10-07 22:20 ` [PATCH v1 00/15] io_uring zero copy rx David Wei
2024-10-08 23:10 ` Joe Damato
2024-10-09 15:07   ` Pavel Begunkov
2024-10-09 16:10     ` Joe Damato
2024-10-09 16:12     ` Jens Axboe
2024-10-11  6:15       ` David Wei
2024-10-09 15:27 ` Jens Axboe
2024-10-09 15:38   ` David Ahern
2024-10-09 15:43     ` Jens Axboe
2024-10-09 15:49       ` Pavel Begunkov
2024-10-09 15:50         ` Jens Axboe
2024-10-09 16:35       ` David Ahern
2024-10-09 16:50         ` Jens Axboe
2024-10-09 16:53           ` Jens Axboe
2024-10-09 17:12             ` Jens Axboe
2024-10-10 14:21               ` Jens Axboe
2024-10-10 15:03                 ` David Ahern
2024-10-10 15:15                   ` Jens Axboe
2024-10-10 18:11                 ` Jens Axboe
2024-10-14  8:42                   ` David Laight
2024-10-09 16:55 ` Mina Almasry
2024-10-09 16:57   ` Jens Axboe
2024-10-09 19:32     ` Mina Almasry
2024-10-09 19:43       ` Pavel Begunkov
2024-10-09 19:47       ` Jens Axboe
2024-10-09 17:19   ` David Ahern
2024-10-09 18:21   ` Pedro Tammela
2024-10-10 13:19     ` Pavel Begunkov
2024-10-11  0:35     ` David Wei
2024-10-11 14:28       ` Pedro Tammela
2024-10-11  0:29   ` David Wei
2024-10-11 19:43     ` Mina Almasry

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=20241007221603.1703699-9-dw@davidwei.uk \
    --to=dw@davidwei.uk \
    --cc=almasrymina@google.com \
    --cc=asml.silence@gmail.com \
    --cc=axboe@kernel.dk \
    --cc=davem@davemloft.net \
    --cc=dsahern@kernel.org \
    --cc=edumazet@google.com \
    --cc=hawk@kernel.org \
    --cc=io-uring@vger.kernel.org \
    --cc=kuba@kernel.org \
    --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