Use broadcast SIPI to startup siblings

The current code for initializing AP cpus has several shortcomings:

- it assumes APIC IDs are sequential
- it uses only the BSP for determining the AP count, which is bad if
  there's more than one physical CPU, and CPUs are of different type

Note that the new code call cpu->ops->init() in parallel, and therefore
some CPU code needs to be changed to address that. One example are old
Intel HT enabled CPUs which can't do microcode update in parallel.

Change-Id: Ic48a1ebab6a7c52aa76765f497268af09fa38c25
Signed-off-by: Sven Schnelle <svens@stackframe.org>
Reviewed-on: http://review.coreboot.org/1139
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>

1 files changed, 12 insertions, 0 deletions

diff --git a/src/cpu/x86/pae/pgtbl.c b/src/cpu/x86/pae/pgtbl.c
index 814c5f161f..7aa17c220d 100644
--- a/src/cpu/x86/pae/pgtbl.c
+++ b/src/cpu/x86/pae/pgtbl.c
@@ -3,8 +3,10 @@
 */
 
 #include <console/console.h>
+#include <device/device.h>
 #include <cpu/cpu.h>
 #include <cpu/x86/pae.h>
+#include <cpu/x86/lapic.h>
 #include <string.h>
 
 static void paging_off(void)
@@ -43,6 +45,14 @@ static void paging_on(void *pdp)
 		);
 }
 
+static int cpu_index(void)
+{
+	device_t dev = dev_find_lapic(lapicid());
+	if (!dev)
+		return -1;
+	return dev->path.apic.index;
+}
+
 void *map_2M_page(unsigned long page)
 {
 	struct pde {
@@ -60,7 +70,9 @@ void *map_2M_page(unsigned long page)
 	unsigned long window;
 	void *result;
 	int i;
+
 	index = cpu_index();
+
 	if ((index < 0) || (index >= CONFIG_MAX_CPUS)) {
 		return MAPPING_ERROR;
 	}