On 07/05/20 12:39 pm, Michal Hocko wrote: > On Wed 06-05-20 17:50:28, Vlastimil Babka wrote: >> [...] >> >> How about something like this: >> >> diff --git a/Documentation/admin-guide/numastat.rst b/Documentation/admin-guide/numastat.rst >> index aaf1667489f8..08ec2c2bdce3 100644 >> --- a/Documentation/admin-guide/numastat.rst >> +++ b/Documentation/admin-guide/numastat.rst >> @@ -6,6 +6,21 @@ Numa policy hit/miss statistics >> >> All units are pages. Hugepages have separate counters. >> >> +The numa_hit, numa_miss and numa_foreign counters reflect how well processes >> +are able to allocate memory from nodes they prefer. If they succeed, numa_hit >> +is incremented on the preferred node, otherwise numa_foreign is incremented on >> +the preferred node and numa_miss on the node where allocation succeeded. >> + >> +Usually preferred node is the one local to the CPU where the process executes, >> +but restrictions such as mempolicies can change that, so there are also two >> +counters based on CPU local node. local_node is similar to numa_hit and is >> +incremented on allocation from a node by CPU on the same node. other_node is >> +similar to numa_miss and is incremented on the node where allocation succeeds >> +from a CPU from a different node. Note there is no counter analogical to >> +numa_foreign. >> + >> +In more detail: >> + >> =============== ============================================================ >> numa_hit A process wanted to allocate memory from this node, >> and succeeded. >> @@ -14,11 +29,13 @@ numa_miss A process wanted to allocate memory from another node, >> but ended up with memory from this node. >> >> numa_foreign A process wanted to allocate on this node, >> - but ended up with memory from another one. >> + but ended up with memory from another node. >> >> -local_node A process ran on this node and got memory from it. >> +local_node A process ran on this node's CPU, >> + and got memory from this node. >> >> -other_node A process ran on this node and got memory from another node. >> +other_node A process ran on a different node's CPU >> + and got memory from this node. >> >> interleave_hit Interleaving wanted to allocate from this node >> and succeeded. >> @@ -28,3 +45,11 @@ For easier reading you can use the numastat utility from the numactl package >> (http://oss.sgi.com/projects/libnuma/). Note that it only works >> well right now on machines with a small number of CPUs. >> >> +Note that on systems with memoryless nodes (where a node has CPUs but no >> +memory) the numa_hit, numa_miss and numa_foreign statistics can be skewed >> +heavily. In the current kernel implementation, if a process prefers a >> +memoryless node (i.e. because it is running on one of its local CPU), the >> +implementation actually treats one of the nearest nodes with memory as the >> +preferred node. As a result, such allocation will not increase the numa_foreign >> +counter on the memoryless node, and will skew the numa_hit, numa_miss and >> +numa_foreign statistics of the nearest node. > > This is certainly an improvement. Thanks! The question whether we can > identify where bogus numbers came from would be interesting as well. > Maybe those are not worth fixing but it would be great to understand > them at least. I have to say that the explanation via boot_pageset is > not really clear to me. > The documentation update will definitely help. Thanks for that. I did collect some stack traces on a ppc64 guest for calls to zone_statistics() in case of zones that are using the boot_pageset and most of them originate from kmem_cache_init() with eventual calls to allocate_slab(). [ 0.000000] [c00000000282b690] [c000000000402d98] zone_statistics+0x138/0x1d0 [ 0.000000] [c00000000282b740] [c000000000401190] rmqueue_pcplist+0xf0/0x120 [ 0.000000] [c00000000282b7d0] [c00000000040b178] get_page_from_freelist+0x2f8/0x2100 [ 0.000000] [c00000000282bb30] [c000000000401ae0] __alloc_pages_nodemask+0x1a0/0x2d0 [ 0.000000] [c00000000282bbc0] [c00000000044b040] alloc_slab_page+0x70/0x580 [ 0.000000] [c00000000282bc20] [c00000000044b5f8] allocate_slab+0xa8/0x610 ... In the remaining cases, the sources are ftrace_init() and early_trace_init(). Unless they are useful, can we avoid incrementing stats for zones using boot_pageset inside zone_statistics()? - Sandipan
