Re: [PATCH] for some issues in tcp window tracking patch

[Patrick McHardy <kaber@trash.net>]

Henrik Nordstrom wrote:
> On Fri, 21 May 2004, Pablo Neira wrote:
> 
> 
>>b) We don't need to set assured bit all the time, I think that I told 
>>you, well if so, sorry for repeating me.
> 
> 
>>         } else if ((old_state == TCP_CONNTRACK_SYN_RECV
>>                    || old_state == TCP_CONNTRACK_ESTABLISHED)
>>-                   && new_state == TCP_CONNTRACK_ESTABLISHED)
>>+                 && new_state == TCP_CONNTRACK_ESTABLISHED
>>+                 && !test_bit(IPS_ASSURED_BIT, &conntrack->status)) {
> 
> 
> Are you sure this really is an optimization in the current code? Is 
> test_bit really faster than set_bit when the bit is already set?

It should be, on x86 set_bit locks the bus:

static __inline__ void set_bit(int nr, volatile unsigned long * addr)
{
         __asm__ __volatile__( LOCK_PREFIX
                 "btsl %1,%0"
                 :"=m" (ADDR)
                 :"Ir" (nr));
}

but test_bit doesn't:

static __inline__ int variable_test_bit(int nr, const volatile unsigned 
long * addr)
{
         int oldbit;
 

         __asm__ __volatile__(
                 "btl %2,%1\n\tsbbl %0,%0"
                 :"=r" (oldbit)
                 :"m" (ADDR),"Ir" (nr));
         return oldbit;
}

Regards
Patrick