| Andi Kleen | 7eb903f | 2006-01-11 22:42:39 +0100 | [diff] [blame] | 1 | Firmware support for CPU hotplug under Linux/x86-64 |
| 2 | --------------------------------------------------- |
| 3 | |
| 4 | Linux/x86-64 supports CPU hotplug now. For various reasons Linux wants to |
| Randy Dunlap | 57d3077 | 2007-02-13 13:26:23 +0100 | [diff] [blame] | 5 | know in advance of boot time the maximum number of CPUs that could be plugged |
| Andi Kleen | 7eb903f | 2006-01-11 22:42:39 +0100 | [diff] [blame] | 6 | into the system. ACPI 3.0 currently has no official way to supply |
| 7 | this information from the firmware to the operating system. |
| 8 | |
| 9 | In ACPI each CPU needs an LAPIC object in the MADT table (5.2.11.5 in the |
| 10 | ACPI 3.0 specification). ACPI already has the concept of disabled LAPIC |
| 11 | objects by setting the Enabled bit in the LAPIC object to zero. |
| 12 | |
| 13 | For CPU hotplug Linux/x86-64 expects now that any possible future hotpluggable |
| 14 | CPU is already available in the MADT. If the CPU is not available yet |
| 15 | it should have its LAPIC Enabled bit set to 0. Linux will use the number |
| 16 | of disabled LAPICs to compute the maximum number of future CPUs. |
| 17 | |
| 18 | In the worst case the user can overwrite this choice using a command line |
| 19 | option (additional_cpus=...), but it is recommended to supply the correct |
| 20 | number (or a reasonable approximation of it, with erring towards more not less) |
| 21 | in the MADT to avoid manual configuration. |