On Sat, Apr 02, 2011 at 11:36:20AM +0200, Lars Heidieker wrote: > >> being moved to kmem any way the point doesn't seem to be very strong > to me. > > > > Why ? A leak is still a leak > > > Sure but even now we wound find a leak that happens if the memory > comes from the kmem allocator, therefore we need a detector there > anyway. malloc currently covers only a small part of allocations.
sure, but it's better than nothing. This and pools allows to track a fair number of alloc/free in the kernel these days. > >> But tracing who is doing the allocation is definitely something worse > >> having, tracing this with dtrace should be possible as Andrew pointed > >> out end january, when I made the patch public. > > > > AFAIK dtrace doesn't work on non-modular kernels ... > I haven't checked but if so it should be made usable on > modular-kernels as well. You meant "non-modular" I guess ? -- Manuel Bouyer <[email protected]> NetBSD: 26 ans d'experience feront toujours la difference --
