On Wed, Apr 01, 2020 at 12:34:05PM -0700, Paul E. McKenney wrote: > On Wed, Apr 01, 2020 at 09:05:48PM +0200, Uladzislau Rezki wrote: > > On Wed, Apr 01, 2020 at 11:54:39AM -0700, Paul E. McKenney wrote: > > > On Wed, Apr 01, 2020 at 08:37:45PM +0200, Uladzislau Rezki wrote: > > > > On Wed, Apr 01, 2020 at 11:26:15AM -0700, Paul E. McKenney wrote: > > > > > On Wed, Apr 01, 2020 at 08:16:01PM +0200, Uladzislau Rezki wrote: > > > > > > > > > > > > > > > > > > Right. Per discussion with Paul, we discussed that it is better if we > > > > > > > > > pre-allocate N number of array blocks per-CPU and use it for the cache. > > > > > > > > > Default for N being 1 and tunable with a boot parameter. I agree with this. > > > > > > > > > > > > > > > > > As discussed before, we can make use of memory pool API for such > > > > > > > > purpose. But i am not sure if it should be one pool per CPU or > > > > > > > > one pool per NR_CPUS, that would contain NR_CPUS * N pre-allocated > > > > > > > > blocks. > > > > > > > > > > > > > > There are advantages and disadvantages either way. The advantage of the > > > > > > > per-CPU pool is that you don't have to worry about something like lock > > > > > > > contention causing even more pain during an OOM event. One potential > > > > > > > problem wtih the per-CPU pool can happen when callbacks are offloaded, > > > > > > > in which case the CPUs needing the memory might never be getting it, > > > > > > > because in the offloaded case (RCU_NOCB_CPU=y) the CPU posting callbacks > > > > > > > might never be invoking them. > > > > > > > > > > > > > > But from what I know now, systems built with CONFIG_RCU_NOCB_CPU=y > > > > > > > either don't have heavy callback loads (HPC systems) or are carefully > > > > > > > configured (real-time systems). Plus large systems would probably end > > > > > > > up needing something pretty close to a slab allocator to keep from dying > > > > > > > from lock contention, and it is hard to justify that level of complexity > > > > > > > at this point. > > > > > > > > > > > > > > Or is there some way to mark a specific slab allocator instance as being > > > > > > > able to keep some amount of memory no matter what the OOM conditions are? > > > > > > > If not, the current per-CPU pre-allocated cache is a better choice in the > > > > > > > near term. > > > > > > > > > > > > > As for mempool API: > > > > > > > > > > > > mempool_alloc() just tries to make regular allocation taking into > > > > > > account passed gfp_t bitmask. If it fails due to memory pressure, > > > > > > it uses reserved preallocated pool that consists of number of > > > > > > desirable elements(preallocated when a pool is created). > > > > > > > > > > > > mempoll_free() returns an element to to pool, if it detects that > > > > > > current reserved elements are lower then minimum allowed elements, > > > > > > it will add an element to reserved pool, i.e. refill it. Otherwise > > > > > > just call kfree() or whatever we define as "element-freeing function." > > > > > > > > > > Unless I am missing something, mempool_alloc() acquires a per-mempool > > > > > lock on each invocation under OOM conditions. For our purposes, this > > > > > is essentially a global lock. This will not be at all acceptable on a > > > > > large system. > > > > > > > > > It uses pool->lock to access to reserved objects, so if we have one memory > > > > pool per one CPU then it would be serialized. > > > > > > I am having difficulty parsing your sentence. It looks like your thought > > > is to invoke mempool_create() for each CPU, so that the locking would be > > > on a per-CPU basis, as in 128 invocations of mempool_init() on a system > > > having 128 hardware threads. Is that your intent? > > > > > In order to serialize it, you need to have it per CPU. So if you have 128 > > cpus, it means: > > > > <snip> > > for_each_possible_cpu(...) > > cpu_pool = mempool_create(); > > <snip> > > > > but please keep in mind that it is not my intention, but i had a though > > about mempool API. Because it has pre-reserve logic inside. > > OK, fair point on use of mempool API, but my guess is that extending > the current kfree_rcu() logic will be simpler. > Agree :) -- Vlad Rezki
