On Friday 16 January 2004 05:01 pm, Dag-Erling Smørgrav wrote: > John Baldwin <jhb_at_FreeBSD.org> writes: > > You could only lock newfdp #ifdef INVARIANTS perhaps since that is the > > only reason you are doing it. That doesn't pessimize production kernels > > while still letting your assertions work ok. > > now that's an idea (though it's a bit of a hack) > > > You could also perhaps tweak the > > mtx_assert to somehow check the state of the fd pointer to see if it is a > > new table (refcount of 0 or some such) > > that would pessimize the common case... Well, only for kernels with INVARIANTS in them. :) If you do 'if (foo) mtx_assert(...)' then gcc should optimize the whole thing out without invariants. -- John Baldwin <jhb_at_FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve" = http://www.FreeBSD.orgReceived on Fri Jan 16 2004 - 13:33:27 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:38 UTC