program loading: add prog_run() function

The prog_run() function abstracts away what is required
for running a given program. Within it, there are 2
calls: 1. platform_prog_run() and 2. arch_prog_run().
The platform_prog_run() allows for a chipset to intercept
a program that will be run. This allows for CPU switching
as currently needed in t124 and t132.

Change-Id: I22a5dd5bfb1018e7e46475e47ac993a0941e2a8c
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/8846
Tested-by: build bot (Jenkins)
Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com>
Reviewed-by: Marc Jones <marc.jones@se-eng.com>

src/arch/x86/boot/Makefile.inc

@@ -2,6 +2,7 @@
 ifeq ($(CONFIG_ARCH_ROMSTAGE_X86_32),y)
 
 romstage-y += cbmem.c
+romstage-y += boot.c
 
 endif # CONFIG_ARCH_ROMSTAGE_X86_32
 

src/arch/x86/boot/boot.c

@@ -123,7 +123,7 @@ static void jmp_payload(void *entry, unsigned long buffer, unsigned long size)
 		);
 }
 
-void arch_payload_run(const struct payload *payload)
+void arch_payload_run(struct payload *payload)
 {
 	if (IS_ENABLED(CONFIG_RELOCATABLE_RAMSTAGE))
 		jmp_payload_no_bounce_buffer(prog_entry(&payload->prog));
@@ -132,3 +132,11 @@ void arch_payload_run(const struct payload *payload)
 				(uintptr_t)payload->bounce.data,
 				payload->bounce.size);
 }
+
+void arch_prog_run(struct prog *prog)
+{
+	__asm__ volatile (
+		"jmp  *%%edi\n"
+		:: "D"(prog_entry(prog))
+	);
+}