From: Jens Axboe <[email protected]>
To: [email protected], [email protected]
Cc: [email protected], [email protected],
Stefan Roesch <[email protected]>, Jens Axboe <[email protected]>
Subject: [PATCH 1/7] net: split off __napi_busy_poll from napi_busy_poll
Date: Tue, 6 Feb 2024 09:30:03 -0700 [thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>
From: Stefan Roesch <[email protected]>
This splits off the key part of the napi_busy_poll function into its own
function, __napi_busy_poll, and changes the prefer_busy_poll bool to be
flag based to allow passing in more flags in the future.
This is done in preparation for an additional napi_busy_poll() function,
that doesn't take the rcu_read_lock(). The new function is introduced
in the next patch.
Signed-off-by: Stefan Roesch <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jens Axboe <[email protected]>
---
net/core/dev.c | 42 ++++++++++++++++++++++++++++--------------
1 file changed, 28 insertions(+), 14 deletions(-)
diff --git a/net/core/dev.c b/net/core/dev.c
index cb2dab0feee0..1eaed657f2c2 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -6177,8 +6177,12 @@ static void __busy_poll_stop(struct napi_struct *napi, bool skip_schedule)
clear_bit(NAPI_STATE_SCHED, &napi->state);
}
-static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, bool prefer_busy_poll,
- u16 budget)
+enum {
+ NAPI_F_PREFER_BUSY_POLL = 1,
+};
+
+static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock,
+ unsigned flags, u16 budget)
{
bool skip_schedule = false;
unsigned long timeout;
@@ -6198,7 +6202,7 @@ static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, bool
local_bh_disable();
- if (prefer_busy_poll) {
+ if (flags & NAPI_F_PREFER_BUSY_POLL) {
napi->defer_hard_irqs_count = READ_ONCE(napi->dev->napi_defer_hard_irqs);
timeout = READ_ONCE(napi->dev->gro_flush_timeout);
if (napi->defer_hard_irqs_count && timeout) {
@@ -6222,23 +6226,23 @@ static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, bool
local_bh_enable();
}
-void napi_busy_loop(unsigned int napi_id,
- bool (*loop_end)(void *, unsigned long),
- void *loop_end_arg, bool prefer_busy_poll, u16 budget)
+static void __napi_busy_loop(unsigned int napi_id,
+ bool (*loop_end)(void *, unsigned long),
+ void *loop_end_arg, unsigned flags, u16 budget)
{
unsigned long start_time = loop_end ? busy_loop_current_time() : 0;
int (*napi_poll)(struct napi_struct *napi, int budget);
void *have_poll_lock = NULL;
struct napi_struct *napi;
+ WARN_ON_ONCE(!rcu_read_lock_held());
+
restart:
napi_poll = NULL;
- rcu_read_lock();
-
napi = napi_by_id(napi_id);
if (!napi)
- goto out;
+ return;
if (!IS_ENABLED(CONFIG_PREEMPT_RT))
preempt_disable();
@@ -6254,14 +6258,14 @@ void napi_busy_loop(unsigned int napi_id,
*/
if (val & (NAPIF_STATE_DISABLE | NAPIF_STATE_SCHED |
NAPIF_STATE_IN_BUSY_POLL)) {
- if (prefer_busy_poll)
+ if (flags & NAPI_F_PREFER_BUSY_POLL)
set_bit(NAPI_STATE_PREFER_BUSY_POLL, &napi->state);
goto count;
}
if (cmpxchg(&napi->state, val,
val | NAPIF_STATE_IN_BUSY_POLL |
NAPIF_STATE_SCHED) != val) {
- if (prefer_busy_poll)
+ if (flags & NAPI_F_PREFER_BUSY_POLL)
set_bit(NAPI_STATE_PREFER_BUSY_POLL, &napi->state);
goto count;
}
@@ -6282,11 +6286,12 @@ void napi_busy_loop(unsigned int napi_id,
if (unlikely(need_resched())) {
if (napi_poll)
- busy_poll_stop(napi, have_poll_lock, prefer_busy_poll, budget);
+ busy_poll_stop(napi, have_poll_lock, flags, budget);
if (!IS_ENABLED(CONFIG_PREEMPT_RT))
preempt_enable();
rcu_read_unlock();
cond_resched();
+ rcu_read_lock();
if (loop_end(loop_end_arg, start_time))
return;
goto restart;
@@ -6294,10 +6299,19 @@ void napi_busy_loop(unsigned int napi_id,
cpu_relax();
}
if (napi_poll)
- busy_poll_stop(napi, have_poll_lock, prefer_busy_poll, budget);
+ busy_poll_stop(napi, have_poll_lock, flags, budget);
if (!IS_ENABLED(CONFIG_PREEMPT_RT))
preempt_enable();
-out:
+}
+
+void napi_busy_loop(unsigned int napi_id,
+ bool (*loop_end)(void *, unsigned long),
+ void *loop_end_arg, bool prefer_busy_poll, u16 budget)
+{
+ unsigned flags = prefer_busy_poll ? NAPI_F_PREFER_BUSY_POLL : 0;
+
+ rcu_read_lock();
+ __napi_busy_loop(napi_id, loop_end, loop_end_arg, flags, budget);
rcu_read_unlock();
}
EXPORT_SYMBOL(napi_busy_loop);
--
2.43.0
next prev parent reply other threads:[~2024-02-06 16:34 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-02-06 16:30 [PATCHSET v16 0/7] io_uring: add napi busy polling support Jens Axboe
2024-02-06 16:30 ` Jens Axboe [this message]
2024-02-06 16:30 ` [PATCH 2/7] net: add napi_busy_loop_rcu() Jens Axboe
2024-02-06 16:30 ` [PATCH 3/7] io-uring: move io_wait_queue definition to header file Jens Axboe
2024-02-06 16:30 ` [PATCH 4/7] io-uring: add napi busy poll support Jens Axboe
2024-02-06 16:30 ` [PATCH 5/7] io-uring: add sqpoll support for napi busy poll Jens Axboe
2024-02-06 16:30 ` [PATCH 6/7] io_uring: add register/unregister napi function Jens Axboe
2024-02-06 16:30 ` [PATCH 7/7] io_uring: add prefer busy poll to register and unregister napi api Jens Axboe
2024-02-09 18:50 ` [PATCHSET v16 0/7] io_uring: add napi busy polling support patchwork-bot+netdevbpf
2024-02-09 18:51 ` Jakub Kicinski
2024-02-09 18:59 ` Jens Axboe
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 \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
/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