Re: [PATCH] net: fix race in the receive/select

From: Davide Libenzi <redacted>
Date: Thu, 25 Jun 2009 19:04:04 -0700 (PDT)

On Fri, 26 Jun 2009, Eric Dumazet wrote:

quoted

Adding a plain smp_mb() in tcp_poll() for example would slowdown select()/poll() with NULL
timeout.

Do you think of it as good design adding an MB on a subsystem, because of 
the special locking logic of another one?
The (eventual) slowdown, IMO can be argued sideways, by saying that 
non-socket users will pay the price for their polls.

Perhaps every performance argument is moot since spin_unlock() used to
have the barrier :-)