Re: spinlock random exponential backoff

[Matthew Dillon <dillon@xxxxxxxxxxxxxxxxxxxx>]

:Hi,
:
:I have problem under specific test case, if I use spinlock on ifq (got
:many "indefinite wait" logging).  With following patch things improve
:greatly, there is almost no "indefinite wait"!.  Please review it:
:http://leaf.dragonflybsd.org/~sephe/kern_spinlock.c.diff1
:
:Best Regards,
:sephe

    Looks good.  We might want to also change that cpu_nop() into a
    cpu_pause(), or perhaps even call both.

					-Matt
					Matthew Dillon 
					<dillon@backplane.com>