[Xen-devel] [PATCH v3 1/3] x86/numa: Allow arbitrary value of PXM in PXM<->node mapping

[Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>]

ACPI defines proximity domain identifier as a 32-bit integer. While
in most cases the values will be zero-based this is not guaranteed,
making current pxm2node[256] mapping structure not appropriate.

We will instead use MAX_NUMNODES-sized array of struct pxm2node to
store PXM-to-node mapping. To accommodate common case of zero-based
and contiguios PXMs we will, whenever possible, try to use PXM as
index into this array array for fast lookups.

Signed-off-by: Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>
---

Changes in v3:
* Renamed p2n to pxm2node (and struct pxm2node to struct pxm_to_node)
* Changed node type in pxm_to_node to u8
* pxm_to_node(), setup_node() and node_to_pxm()) use unsigned as node
  argument
* Reinitialized pxm2node in abd_srat();
* Added BUILD_BUG_ON(MAX_NUMNODES >= NUMA_NO_NODE);
* Made changes to K&R style (which is what this file uses)

 xen/arch/x86/srat.c        |   96 ++++++++++++++++++++++++++++++++------------
 xen/include/asm-x86/numa.h |    4 +-
 2 files changed, 72 insertions(+), 28 deletions(-)

diff --git a/xen/arch/x86/srat.c b/xen/arch/x86/srat.c
index 29fc724..1664393 100644
--- a/xen/arch/x86/srat.c
+++ b/xen/arch/x86/srat.c
@@ -27,35 +27,78 @@ static nodemask_t memory_nodes_parsed __initdata;
 static nodemask_t processor_nodes_parsed __initdata;
 static nodemask_t nodes_found __initdata;
 static struct node nodes[MAX_NUMNODES] __initdata;
-static u8 __read_mostly pxm2node[256] = { [0 ... 255] = NUMA_NO_NODE };
 
+struct pxm_to_node {
+       unsigned pxm;
+       u8 node;
+};
+static struct pxm_to_node __read_mostly pxm2node[MAX_NUMNODES] =
+       { [0 ... MAX_NUMNODES - 1] = {.node = NUMA_NO_NODE} };
+
+static int node_to_pxm(unsigned n);
 
 static int num_node_memblks;
 static struct node node_memblk_range[NR_NODE_MEMBLKS];
 static int memblk_nodeid[NR_NODE_MEMBLKS];
 
+static inline bool_t node_found(unsigned idx, unsigned pxm)
+{
+       return ((pxm2node[idx].pxm == pxm) &&
+               (pxm2node[idx].node != NUMA_NO_NODE));
+}
 
-static int node_to_pxm(int n);
-
-int pxm_to_node(int pxm)
+int pxm_to_node(unsigned pxm)
 {
-       if ((unsigned)pxm >= 256)
-               return -1;
+       unsigned i;
+
+       if ((pxm < MAX_NUMNODES) && node_found(pxm, pxm))
+               return pxm2node[pxm].node;
+
+       for (i = 0; i < MAX_NUMNODES; i++)
+               if (node_found(i, pxm))
+                       return pxm2node[i].node;
+
        /* Extend 0xff to (int)-1 */
-       return (signed char)pxm2node[pxm];
+       return (signed char)NUMA_NO_NODE;
 }
 
-__devinit int setup_node(int pxm)
+__devinit int setup_node(unsigned pxm)
 {
-       unsigned node = pxm2node[pxm];
-       if (node == 0xff) {
-               if (nodes_weight(nodes_found) >= MAX_NUMNODES)
-                       return -1;
-               node = first_unset_node(nodes_found); 
-               node_set(node, nodes_found);
-               pxm2node[pxm] = node;
+       int node;
+       unsigned idx;
+       static bool_t warned;
+
+       BUILD_BUG_ON(MAX_NUMNODES >= NUMA_NO_NODE);
+
+       if (pxm < MAX_NUMNODES) {
+               if (node_found(pxm, pxm))
+                       return pxm2node[pxm].node;
+
+               /* Try to maintain indexing of pxm2node by pxm */
+               if (pxm2node[pxm].node == NUMA_NO_NODE) {
+                       idx = pxm;
+                       goto finish;
+               }
+       }
+
+       for (idx = 0; idx < MAX_NUMNODES; idx++)
+               if (pxm2node[idx].node == NUMA_NO_NODE)
+                       goto finish;
+
+       if (!warned) {
+               printk(XENLOG_WARNING "More PXMs than available nodes\n");
+               warned = 1;
        }
-       return pxm2node[pxm];
+
+       return (signed char)NUMA_NO_NODE;
+
+ finish:
+       node = first_unset_node(nodes_found);
+       node_set(node, nodes_found);
+       pxm2node[idx].pxm = pxm;
+       pxm2node[idx].node = node;
+
+       return node;
 }
 
 int valid_numa_range(u64 start, u64 end, int node)
@@ -111,8 +154,8 @@ static __init void bad_srat(void)
        acpi_numa = -1;
        for (i = 0; i < MAX_LOCAL_APIC; i++)
                apicid_to_node[i] = NUMA_NO_NODE;
-       for (i = 0; i < ARRAY_SIZE(pxm2node); i++)
-               pxm2node[i] = NUMA_NO_NODE;
+       for (i = 0; i < MAX_NUMNODES; i++)
+               pxm2node[i].node = NUMA_NO_NODE;
        mem_hotplug = 0;
 }
 
@@ -438,15 +481,16 @@ int __init acpi_scan_nodes(u64 start, u64 end)
        return 0;
 }
 
-static int node_to_pxm(int n)
+static int node_to_pxm(unsigned n)
 {
-       int i;
-       if (pxm2node[n] == n)
-               return n;
-       for (i = 0; i < 256; i++)
-               if (pxm2node[i] == n)
-                       return i;
-       return 0;
+       unsigned i;
+
+       if ((n <  MAX_NUMNODES) && (pxm2node[n].node == n))
+               return pxm2node[n].pxm;
+       for (i = 0; i < MAX_NUMNODES; i++)
+               if (pxm2node[i].node == n)
+                       return pxm2node[i].pxm;
+       return 0;
 }
 
 int __node_distance(int a, int b)
diff --git a/xen/include/asm-x86/numa.h b/xen/include/asm-x86/numa.h
index 5959860..e491a9e 100644
--- a/xen/include/asm-x86/numa.h
+++ b/xen/include/asm-x86/numa.h
@@ -21,7 +21,7 @@ struct node {
 
 extern int compute_hash_shift(struct node *nodes, int numnodes,
                              int *nodeids);
-extern int pxm_to_node(int nid);
+extern int pxm_to_node(unsigned pxm);
 
 #define ZONE_ALIGN (1UL << (MAX_ORDER+PAGE_SHIFT))
 #define VIRTUAL_BUG_ON(x) 
@@ -33,7 +33,7 @@ extern int numa_off;
 
 extern int srat_disabled(void);
 extern void numa_set_node(int cpu, int node);
-extern int setup_node(int pxm);
+extern int setup_node(unsigned pxm);
 extern void srat_detect_node(int cpu);
 
 extern void setup_node_bootmem(int nodeid, u64 start, u64 end);
-- 
1.7.1


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel