On 2020-11-12 15:36, Eric Dumazet wrote:
On Thu, Nov 12, 2020 at 12:41 PM Björn Töpel <bjorn.to...@gmail.com> wrote:

From: Björn Töpel <bjorn.to...@intel.com>

This option lets a user set a per socket NAPI budget for
busy-polling. If the options is not set, it will use the default of 8.

Signed-off-by: Björn Töpel <bjorn.to...@intel.com>
---


...

  #else /* CONFIG_NET_RX_BUSY_POLL */
  static inline unsigned long net_busy_loop_on(void)
@@ -106,7 +108,8 @@ static inline void sk_busy_loop(struct sock *sk, int 
nonblock)

         if (napi_id >= MIN_NAPI_ID)
                 napi_busy_loop(napi_id, nonblock ? NULL : sk_busy_loop_end, sk,
-                              READ_ONCE(sk->sk_prefer_busy_poll));
+                              READ_ONCE(sk->sk_prefer_busy_poll),
+                              sk->sk_busy_poll_budget ?: BUSY_POLL_BUDGET);

Please use :

        READ_ONCE(sk->sk_busy_poll_budget) ?: BUSY_POLL_BUDGET

Because sk_busy_loop() is usually called without socket lock being held.

This will prevent yet another KCSAN report.

  #endif
  }


...

--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1165,6 +1165,16 @@ int sock_setsockopt(struct socket *sock, int level, int 
optname,
                 else
                         sk->sk_prefer_busy_poll = valbool;
                 break;
+       case SO_BUSY_POLL_BUDGET:
+               if (val > sk->sk_busy_poll_budget && !capable(CAP_NET_ADMIN)) {
+                       ret = -EPERM;
+               } else {
+                       if (val < 0)

                if (val < 0 || val > (u16)~0)

+                               ret = -EINVAL;
+                       else
+                               sk->sk_busy_poll_budget = val;


                                WRITE_ONCE(sk->sk_busy_poll_budget, val);


Thanks for the review! I'll address it all.


Reply via email to