diff options
| author |   Ravikiran Thirumalai <kiran@scalex86.org> | 2006-01-11 22:45:36 +0100 |
|---|---|---|
| committer |   Linus Torvalds <torvalds@g5.osdl.org> | 2006-01-11 19:04:59 -0800 |
| commit | 05b3cbd8bb98736387df8a2e1efe311b1fb4d2ad (patch) | |
| tree | 46ca01cfe38d2df4c6f93ba0325a3f2a1e97ea71 /arch | |
| parent | [PATCH] x86_64: Fix up white space in time.c (diff) | |
| download | linux-dev-05b3cbd8bb98736387df8a2e1efe311b1fb4d2ad.tar.xz linux-dev-05b3cbd8bb98736387df8a2e1efe311b1fb4d2ad.zip | |
[PATCH] x86_64: Early initialization of cpu_to_node
Patch enables early intialization of cpu_to_node.
apicid_to_node is built by reading the SRAT table, from acpi_numa_init with
ACPI_NUMA and k8_scan_nodes with K8_NUMA.
x86_cpu_to_apicid is built by parsing the ACPI MADT table, from acpi_boot_init.
We combine these two tables and setup cpu_to_node.
Early intialization helps the static per_cpu_areas in getting pages from
correct node.
Change since last release:
Do not initialize early init_cpu_to_node for faking node cases.
Patch tested on TYAN dual core 4P board with K8 only, ACPI_NUMA.
Tested on EM64T NUMA. Also tested with numa=off, numa=fake, and running
a kernel compiled with NUMA on a regular EM64 2 way SMP.
Signed-off-by: Alok N Kataria <alokk@calsoftinc.com>
Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Shai Fultheim <shai@scalex86.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'arch')
| -rw-r--r-- | arch/x86_64/kernel/setup.c | 2 | ||||
| -rw-r--r-- | arch/x86_64/mm/numa.c | 25 |
2 files changed, 27 insertions, 0 deletions
diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c index 46bf556efc30..173bdc55113b 100644 --- a/arch/x86_64/kernel/setup.c +++ b/arch/x86_64/kernel/setup.c @@ -713,6 +713,8 @@ void __init setup_arch(char **cmdline_p) acpi_boot_init(); #endif + init_cpu_to_node(); + #ifdef CONFIG_X86_LOCAL_APIC /* * get boot-time SMP configuration: diff --git a/arch/x86_64/mm/numa.c b/arch/x86_64/mm/numa.c index 42ec1f4f491a..876e34374412 100644 --- a/arch/x86_64/mm/numa.c +++ b/arch/x86_64/mm/numa.c @@ -330,6 +330,31 @@ __init int numa_setup(char *opt) return 1; } +/* + * Setup early cpu_to_node. + * + * Populate cpu_to_node[] only if x86_cpu_to_apicid[], + * and apicid_to_node[] tables have valid entries for a CPU. + * This means we skip cpu_to_node[] initialisation for NUMA + * emulation and faking node case (when running a kernel compiled + * for NUMA on a non NUMA box), which is OK as cpu_to_node[] + * is already initialized in a round robin manner at numa_init_array, + * prior to this call, and this initialization is good enough + * for the fake NUMA cases. + */ +void __init init_cpu_to_node(void) +{ + int i; + for (i = 0; i < NR_CPUS; i++) { + u8 apicid = x86_cpu_to_apicid[i]; + if (apicid == BAD_APICID) + continue; + if (apicid_to_node[apicid] == NUMA_NO_NODE) + continue; + cpu_to_node[i] = apicid_to_node[apicid]; + } +} + EXPORT_SYMBOL(cpu_to_node); EXPORT_SYMBOL(node_to_cpumask); EXPORT_SYMBOL(memnode_shift); |