Received-SPF: pass (google.com: domain of linux-kernel+bounces-58488-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Precedence: bulk
MIME-Version: 1.0
References: <20240205210453.11301-1-jdamato@fastly.com> <20240205210453.11301-2-jdamato@fastly.com>
In-Reply-To: <20240205210453.11301-2-jdamato@fastly.com>
From: Eric Dumazet <edumazet@google.com>
Date: Thu, 8 Feb 2024 18:46:25 +0100
Message-ID: <CANn89iJY8mTn3PViBvTh_DewUKWjc0z3cvJvr8AcQgcbWC4G0Q@mail.gmail.com>
Subject: Re: [PATCH net-next v6 1/4] eventpoll: support busy poll per epoll instance
To: Joe Damato <jdamato@fastly.com>
Cc: linux-kernel@vger.kernel.org, netdev@vger.kernel.org, 
	chuck.lever@oracle.com, jlayton@kernel.org, linux-api@vger.kernel.org, 
	brauner@kernel.org, davem@davemloft.net, alexander.duyck@gmail.com, 
	sridhar.samudrala@intel.com, kuba@kernel.org, willemdebruijn.kernel@gmail.com, 
	weiwan@google.com, David.Laight@aculab.com, arnd@arndb.de, sdf@google.com, 
	amritha.nambiar@intel.com, Alexander Viro <viro@zeniv.linux.org.uk>, 
	Jan Kara <jack@suse.cz>, 
	"open list:FILESYSTEMS (VFS and infrastructure)" <linux-fsdevel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Mon, Feb 5, 2024 at 10:05=E2=80=AFPM Joe Damato <jdamato@fastly.com> wro=
te:
>
> Allow busy polling on a per-epoll context basis. The per-epoll context
> usec timeout value is preferred, but the pre-existing system wide sysctl
> value is still supported if it specified.
>
> Signed-off-by: Joe Damato <jdamato@fastly.com>
> ---
>  fs/eventpoll.c | 49 +++++++++++++++++++++++++++++++++++++++++++++----
>  1 file changed, 45 insertions(+), 4 deletions(-)
>
> diff --git a/fs/eventpoll.c b/fs/eventpoll.c
> index 3534d36a1474..ce75189d46df 100644
> --- a/fs/eventpoll.c
> +++ b/fs/eventpoll.c
> @@ -227,6 +227,8 @@ struct eventpoll {
>  #ifdef CONFIG_NET_RX_BUSY_POLL
>         /* used to track busy poll napi_id */
>         unsigned int napi_id;
> +       /* busy poll timeout */
> +       u64 busy_poll_usecs;
>  #endif
>
>  #ifdef CONFIG_DEBUG_LOCK_ALLOC
> @@ -386,12 +388,44 @@ static inline int ep_events_available(struct eventp=
oll *ep)
>                 READ_ONCE(ep->ovflist) !=3D EP_UNACTIVE_PTR;
>  }
>
> +/**
> + * busy_loop_ep_timeout - check if busy poll has timed out. The timeout =
value
> + * from the epoll instance ep is preferred, but if it is not set fallbac=
k to
> + * the system-wide global via busy_loop_timeout.
> + *
> + * @start_time: The start time used to compute the remaining time until =
timeout.
> + * @ep: Pointer to the eventpoll context.
> + *
> + * Return: true if the timeout has expired, false otherwise.
> + */
> +static inline bool busy_loop_ep_timeout(unsigned long start_time, struct=
 eventpoll *ep)
> +{
> +#ifdef CONFIG_NET_RX_BUSY_POLL

It seems this local helper is only called from code compiled when
CONFIG_NET_RX_BUSY_POLL
is set.

Not sure why you need an #ifdef here.

> +       unsigned long bp_usec =3D READ_ONCE(ep->busy_poll_usecs);
> +
> +       if (bp_usec) {
> +               unsigned long end_time =3D start_time + bp_usec;
> +               unsigned long now =3D busy_loop_current_time();
> +
> +               return time_after(now, end_time);
> +       } else {
> +               return busy_loop_timeout(start_time);
> +       }
> +#endif
> +       return true;
> +}
> +
>  #ifdef CONFIG_NET_RX_BUSY_POLL
> +static bool ep_busy_loop_on(struct eventpoll *ep)
> +{
> +       return !!ep->busy_poll_usecs || net_busy_loop_on();
> +}
> +
>  static bool ep_busy_loop_end(void *p, unsigned long start_time)
>  {
>         struct eventpoll *ep =3D p;
>
> -       return ep_events_available(ep) || busy_loop_timeout(start_time);
> +       return ep_events_available(ep) || busy_loop_ep_timeout(start_time=
, ep);
>  }
>
>  /*
> @@ -404,7 +438,7 @@ static bool ep_busy_loop(struct eventpoll *ep, int no=
nblock)
>  {
>         unsigned int napi_id =3D READ_ONCE(ep->napi_id);
>
> -       if ((napi_id >=3D MIN_NAPI_ID) && net_busy_loop_on()) {
> +       if ((napi_id >=3D MIN_NAPI_ID) && ep_busy_loop_on(ep)) {
>                 napi_busy_loop(napi_id, nonblock ? NULL : ep_busy_loop_en=
d, ep, false,
>                                BUSY_POLL_BUDGET);
>                 if (ep_events_available(ep))
> @@ -430,7 +464,8 @@ static inline void ep_set_busy_poll_napi_id(struct ep=
item *epi)
>         struct socket *sock;
>         struct sock *sk;
>
> -       if (!net_busy_loop_on())
> +       ep =3D epi->ep;
> +       if (!ep_busy_loop_on(ep))
>                 return;
>
>         sock =3D sock_from_file(epi->ffd.file);
> @@ -442,7 +477,6 @@ static inline void ep_set_busy_poll_napi_id(struct ep=
item *epi)
>                 return;
>
>         napi_id =3D READ_ONCE(sk->sk_napi_id);
> -       ep =3D epi->ep;
>
>         /* Non-NAPI IDs can be rejected
>          *      or
> @@ -466,6 +500,10 @@ static inline void ep_set_busy_poll_napi_id(struct e=
pitem *epi)
>  {
>  }
>
> +static inline bool ep_busy_loop_on(struct eventpoll *ep)
> +{
> +       return false;
> +}
>  #endif /* CONFIG_NET_RX_BUSY_POLL */
>
>  /*
> @@ -2058,6 +2096,9 @@ static int do_epoll_create(int flags)
>                 error =3D PTR_ERR(file);
>                 goto out_free_fd;
>         }
> +#ifdef CONFIG_NET_RX_BUSY_POLL
> +       ep->busy_poll_usecs =3D 0;
> +#endif
>         ep->file =3D file;
>         fd_install(fd, file);
>         return fd;
> --
> 2.25.1
>