s390/numa: add emulation support

NUMA emulation (aka fake NUMA) distributes the available memory to nodes
without using real topology information about the physical memory of the
machine.

Splitting the system memory into nodes replicates the memory management
structures for each node. Particularly each node has its own "mm locks"
and its own "kswapd" task.

For large systems, under certain conditions, this results in improved
system performance and/or latency based on reduced pressure on the mm
locks and the kswapd tasks.

NUMA emulation distributes CPUs to nodes while respecting the original
machine topology information. This is done by trying to avoid to separate
CPUs which reside on the same book or even on the same MC. Because the
current Linux scheduler code requires a stable cpu to node mapping, cores
are pinned to nodes when the first CPU thread is set online.

This patch is based on the initial implementation from Philipp Hachtmann.

Signed-off-by: Michael Holzheu <holzheu@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

1 files changed, 37 insertions, 0 deletions

diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index 25510adb07d3..cb418dcc2d45 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -423,6 +423,43 @@ config NODES_SHIFT
 	  Specify the maximum number of NUMA nodes available on the target
 	  system. Increases memory reserved to accommodate various tables.
 
+menu "Select NUMA modes"
+	depends on NUMA
+
+config NUMA_EMU
+	bool "NUMA emulation"
+	default y
+	help
+	  Numa emulation mode will split the available system memory into
+	  equal chunks which then are distributed over the configured number
+	  of nodes in a round-robin manner.
+
+	  The number of fake nodes is limited by the number of available memory
+	  chunks (i.e. memory size / fake size) and the number of supported
+	  nodes in the kernel.
+
+	  The CPUs are assigned to the nodes in a way that partially respects
+	  the original machine topology (if supported by the machine).
+	  Fair distribution of the CPUs is not guaranteed.
+
+config EMU_SIZE
+	hex "NUMA emulation memory chunk size"
+	default 0x10000000
+	range 0x400000 0x100000000
+	depends on NUMA_EMU
+	help
+	  Select the default size by which the memory is chopped and then
+	  assigned to emulated NUMA nodes.
+
+	  This can be overridden by specifying
+
+	  emu_size=<n>
+
+	  on the kernel command line where also suffixes K, M, G, and T are
+	  supported.
+
+endmenu
+
 config SCHED_MC
 	def_bool n