mm, slub: better heuristic for number of cpus when calculating slab order

When creating a new kmem cache, SLUB determines how large the slab pages
will based on number of inputs, including the number of CPUs in the
system.  Larger slab pages mean that more objects can be allocated/free
from per-cpu slabs before accessing shared structures, but also
potentially more memory can be wasted due to low slab usage and
fragmentation.  The rough idea of using number of CPUs is that larger
systems will be more likely to benefit from reduced contention, and also
should have enough memory to spare.

Number of CPUs used to be determined as nr_cpu_ids, which is number of
possible cpus, but on some systems many will never be onlined, thus
commit a3579c3e856c ("mm/slub: let number of online CPUs determine the
slub page order") changed it to nr_online_cpus().  However, for kmem
caches created early before CPUs are onlined, this may lead to
permamently low slab page sizes.

Vincent reports a regression [1] of hackbench on arm64 systems:

  "I'm facing significant performances regression on a large arm64
   server system (224 CPUs). Regressions is also present on small arm64
   system (8 CPUs) but in a far smaller order of magnitude

   On 224 CPUs system : 9 iterations of hackbench -l 16000 -g 16
   v5.11-rc4 : 9.135sec (+/- 0.45%)
   v5.11-rc4 + revert this patch: 3.173sec (+/- 0.48%)
   v5.10: 3.136sec (+/- 0.40%)"

Mel reports a regression [2] of hackbench on x86_64, with lockstat suggesting
page allocator contention:

  "i.e. the patch incurs a 7% to 32% performance penalty. This bisected
   cleanly yesterday when I was looking for the regression and then
   found the thread.

   Numerous caches change size. For example, kmalloc-512 goes from
   order-0 (vanilla) to order-2 with the revert.

   So mostly this is down to the number of times SLUB calls into the
   page allocator which only caches order-0 pages on a per-cpu basis"

Clearly num_online_cpus() doesn't work too early in bootup.  We could
change the order dynamically in a memory hotplug callback, but runtime
order changing for existing kmem caches has been already shown as
dangerous, and removed in 5eb9277981cd ("mm, slub: remove runtime
allocation order changes").

It could be resurrected in a safe manner with some effort, but to fix
the regression we need something simpler.

We could use num_present_cpus() that should be the number of physically
present CPUs even before they are onlined.  That would work for PowerPC
[3], which triggered the original commit, but that still doesn't work on
arm64 [4] as explained in [5].

So this patch tries to determine the best available value without
specific arch knowledge.

 - num_present_cpus() if the number is larger than 1, as that means the
   arch is likely setting it properly

 - nr_cpu_ids otherwise

This should fix the reported regressions while also keeping the effect
of a3579c3e856c for PowerPC systems.  It's possible there are
configurations where num_present_cpus() is 1 during boot while
nr_cpu_ids is at the same time bloated, so these (if they exist) would
keep the large orders based on nr_cpu_ids as was before a3579c3e856c.

[1] https://lore.kernel.org/linux-mm/CAKfTPtA_JgMf_+zdFbcb_V9rM7JBWNPjAz9irgwFj7Rou=xzZg@mail.gmail.com/
[2] https://lore.kernel.org/linux-mm/20210128134512.GF3592@techsingularity.net/
[3] https://lore.kernel.org/linux-mm/20210123051607.GC2587010@in.ibm.com/
[4] https://lore.kernel.org/linux-mm/CAKfTPtAjyVmS5VYvU6DBxg4-JEo5bdmWbngf-03YsY18cmWv_g@mail.gmail.com/
[5] https://lore.kernel.org/linux-mm/20210126230305.GD30941@willie-the-truck/

Link: https://lkml.kernel.org/r/20210208134108.22286-1-vbabka@suse.cz
Fixes: a3579c3e856c ("mm/slub: let number of online CPUs determine the slub page order")
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reported-by: Vincent Guittot <vincent.guittot@linaro.org>
Reported-by: Mel Gorman <mgorman@techsingularity.net>
Tested-by: Mel Gorman <mgorman@techsingularity.net>
Tested-by: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Cc: Bharata B Rao <bharata@linux.ibm.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Jann Horn <jannh@google.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Will Deacon <will@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/mm/slub.c b/mm/slub.c
index 7ecbbbe5bc0c1fe65e3fc52b881205b18189b771..b22a4b101c846ea651e34b271280803a45fee867 100644 (file)
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -3423,6 +3423,7 @@ static inline int calculate_order(unsigned int size)
        unsigned int order;
        unsigned int min_objects;
        unsigned int max_objects;
+       unsigned int nr_cpus;
 
        /*
         * Attempt to find best configuration for a slab. This
@@ -3433,8 +3434,21 @@ static inline int calculate_order(unsigned int size)
         * we reduce the minimum objects required in a slab.
         */
        min_objects = slub_min_objects;
-       if (!min_objects)
-               min_objects = 4 * (fls(num_online_cpus()) + 1);
+       if (!min_objects) {
+               /*
+                * Some architectures will only update present cpus when
+                * onlining them, so don't trust the number if it's just 1. But
+                * we also don't want to use nr_cpu_ids always, as on some other
+                * architectures, there can be many possible cpus, but never
+                * onlined. Here we compromise between trying to avoid too high
+                * order on systems that appear larger than they are, and too
+                * low order on systems that appear smaller than they are.
+                */
+               nr_cpus = num_present_cpus();
+               if (nr_cpus <= 1)
+                       nr_cpus = nr_cpu_ids;
+               min_objects = 4 * (fls(nr_cpus) + 1);
+       }
        max_objects = order_objects(slub_max_order, size);
        min_objects = min(min_objects, max_objects);