Skip to contentSkip to conversation

xen: Add NUMA support to Xen

PATCH Superseded Ryan Harper 1 participant 1 message May 1, 2006

lore
Share & tools
lore
#0 lore raw

cover xen: Add NUMA support to Xen

The patchset will add basic NUMA support to Xen (hypervisor only). We borrowed from Linux support for NUMA SRAT table parsing, discontiguous memory tracking (memchunk_t), and cpu support (node_to_cpumask etc).

The hypervisor parses the SRAT tables and constructs mappings for each node such as node to cpu mappings and memory range to node mappings.

Using this information, we also modified the page allocator to provide a simple NUMA-aware API. The allocator is initialized on a per-node basis. Upon bootup, memory is inserted into the proper per-node allocator based
on SRAT table information. When requesting pages, the allocator will
attempt to supply pages from a local node, but will use other non-local node memory to satisfy a particular zone or order request.

We also modified Xen's increase_reservation memory op to balance memory distribution across the vcpus in use by a domain. Relying on previous patches which have already been committed to xen-unstable, a guest can be constructed such that its entire memory is contained within a specific NUMA node.

We export NUMA system information via the physinfo hypercall. This information provides cpu/memory topology and configuration information gleaned from the SRAT tables to userspace applications. Currently, xend doesn't leverage any of the information automatically but we intend to do so in the future.

There is a fix for some incorrect assumptions about the symmetry of NUMA systems w.r.t the sockets_per_node value. Details are available in a later email with the patch.

We've added a new tool, xen_numastat, which dumps information about the current heap free page status, as well as probing where a domain's memory is from.

These patches have been tested on several IBM NUMA and non-NUMA systems:

NUMA-aware systems:

IBM Dual Opteron:  2 Node,  2 CPU,  4GB 
IBM x445        :  4 Node, 32 CPU, 32GB 
IBM x460        :  1 Node,  8 CPU, 16GB
IBM x460        :  2 Node, 32 CPU, 32GB
IBM x460        :  4 Node, 32 CPU, 24GB

Non NUMA-aware systems:

IBM Dual Xeon   :  1 Node,  2 CPU,  2GB 
IBM P4          :  1 Node,  1 CPU,  1GB

We look forward to your review of the patches for acceptance.

Signature · 5 lines
Ryan Harper
Software Engineer; Linux Technology Center
IBM Corp., Austin, Tx
(512) 838-9253   T/L: 678-9253
[email protected]