[PATCH] x86_64: New heuristics to find out hotpluggable CPUs.

With a NR_CPUS==128 kernel with CPU hotplug enabled we would waste 4MB
on per CPU data of all possible CPUs.  The reason was that HOTPLUG
always set up possible map to NR_CPUS cpus and then we need to allocate
that much (each per CPU data is roughly ~32k now)

The underlying problem is that ACPI didn't tell us how many hotplug CPUs
the platform supports.  So the old code just assumed all, which would
lead to this memory wastage.

This implements some new heuristics:

 - If the BIOS specified disabled CPUs in the ACPI/mptables assume they
   can be enabled later (this is bending the ACPI specification a bit,
   but seems like a obvious extension)
 - The user can overwrite it with a new additionals_cpus=NUM option
 - Otherwise use half of the available CPUs or 2, whatever is more.

Cc: ashok.raj@intel.com
Cc: len.brown@intel.com
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/Documentation/x86_64/boot-options.txt b/Documentation/x86_64/boot-options.txt
index ffe1c06..a831396 100644 (file)
--- a/Documentation/x86_64/boot-options.txt
+++ b/Documentation/x86_64/boot-options.txt
@@ -122,6 +122,9 @@ SMP
 
   cpumask=MASK   only use cpus with bits set in mask
 
+  additional_cpus=NUM Allow NUM more CPUs for hotplug
+                (defaults are specified by the BIOS or half the available CPUs)
+
 NUMA
 
   numa=off     Only set up a single NUMA node spanning all memory.

diff --git a/arch/x86_64/kernel/mpparse.c b/arch/x86_64/kernel/mpparse.c
index 4a581d1..1d61f10 100644 (file)
--- a/arch/x86_64/kernel/mpparse.c
+++ b/arch/x86_64/kernel/mpparse.c
@@ -65,7 +65,9 @@ unsigned long mp_lapic_addr = 0;
 /* Processor that is doing the boot up */
 unsigned int boot_cpu_id = -1U;
 /* Internal processor count */
-static unsigned int num_processors = 0;
+unsigned int num_processors __initdata = 0;
+
+unsigned disabled_cpus __initdata;
 
 /* Bitmask of physically existing CPUs */
 physid_mask_t phys_cpu_present_map = PHYSID_MASK_NONE;
@@ -109,8 +111,10 @@ static void __init MP_processor_info (struct mpc_config_processor *m)
        int ver, cpu;
        static int found_bsp=0;
 
-       if (!(m->mpc_cpuflag & CPU_ENABLED))
+       if (!(m->mpc_cpuflag & CPU_ENABLED)) {
+               disabled_cpus++;
                return;
+       }
 
        printk(KERN_INFO "Processor #%d %d:%d APIC version %d\n",
                m->mpc_apicid,

diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c
index 3393fc0..f74319a 100644 (file)
--- a/arch/x86_64/kernel/smpboot.c
+++ b/arch/x86_64/kernel/smpboot.c
@@ -880,6 +880,9 @@ static __init void disable_smp(void)
 }
 
 #ifdef CONFIG_HOTPLUG_CPU
+
+int additional_cpus __initdata = -1;
+
 /*
  * cpu_possible_map should be static, it cannot change as cpu's
  * are onlined, or offlined. The reason is per-cpu data-structures
@@ -888,14 +891,38 @@ static __init void disable_smp(void)
  * cpu_present_map on the other hand can change dynamically.
  * In case when cpu_hotplug is not compiled, then we resort to current
  * behaviour, which is cpu_possible == cpu_present.
- * If cpu-hotplug is supported, then we need to preallocate for all
- * those NR_CPUS, hence cpu_possible_map represents entire NR_CPUS range.
  * - Ashok Raj
+ *
+ * Three ways to find out the number of additional hotplug CPUs:
+ * - If the BIOS specified disabled CPUs in ACPI/mptables use that.
+ * - otherwise use half of the available CPUs or 2, whatever is more.
+ * - The user can overwrite it with additional_cpus=NUM
+ * We do this because additional CPUs waste a lot of memory.
+ * -AK
  */
 __init void prefill_possible_map(void)
 {
        int i;
-       for (i = 0; i < NR_CPUS; i++)
+       int possible;
+
+       if (additional_cpus == -1) {
+               if (disabled_cpus > 0) {
+                       additional_cpus = disabled_cpus;
+               } else {
+                       additional_cpus = num_processors / 2;
+                       if (additional_cpus == 0)
+                               additional_cpus = 2;
+               }
+       }
+       possible = num_processors + additional_cpus;
+       if (possible > NR_CPUS) 
+               possible = NR_CPUS;
+
+       printk(KERN_INFO "SMP: Allowing %d CPUs, %d hotplug CPUs\n",
+               possible,
+               max_t(int, possible - num_processors, 0));
+
+       for (i = 0; i < possible; i++)
                cpu_set(i, cpu_possible_map);
 }
 #endif
@@ -1151,6 +1178,12 @@ void __cpu_die(unsigned int cpu)
        printk(KERN_ERR "CPU %u didn't die...\n", cpu);
 }
 
+static __init int setup_additional_cpus(char *s)
+{
+       return get_option(&s, &additional_cpus);
+}
+__setup("additional_cpus=", setup_additional_cpus);
+
 #else /* ... !CONFIG_HOTPLUG_CPU */
 
 int __cpu_disable(void)

diff --git a/include/asm-x86_64/smp.h b/include/asm-x86_64/smp.h
index 592161e..cf8f969 100644 (file)
--- a/include/asm-x86_64/smp.h
+++ b/include/asm-x86_64/smp.h
@@ -81,6 +81,8 @@ extern int safe_smp_processor_id(void);
 extern int __cpu_disable(void);
 extern void __cpu_die(unsigned int cpu);
 extern void prefill_possible_map(void);
+extern unsigned num_processors;
+extern unsigned disabled_cpus;
 
 #endif /* !ASSEMBLY */