[MIPS] SMP: Fix initialization order bug.

    
A recent change requires cpu_possible_map to be initialized before
smp_sched_init() but most MIPS platforms were initializing their
processors in the prom_prepare_cpus callback of smp_prepare_cpus.  The
simple fix of calling prom_prepare_cpus from one of the earlier SMP
initialization hooks doesn't work well either since IPIs may require
init_IRQ() to have completed, so bit the bullet and split
prom_prepare_cpus into two initialization functions, plat_smp_setup
which is called early from setup_arch and plat_prepare_cpus called where
prom_prepare_cpus used to be called.
    
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

1 files changed, 6 insertions, 1 deletions

diff --git a/arch/mips/sgi-ip27/ip27-smp.c b/arch/mips/sgi-ip27/ip27-smp.c
index dbef3f6..09fa7f5 100644
--- a/arch/mips/sgi-ip27/ip27-smp.c
+++ b/arch/mips/sgi-ip27/ip27-smp.c
@@ -140,7 +140,7 @@ static __init void intr_clear_all(nasid_t nasid)
 		REMOTE_HUB_CLR_INTR(nasid, i);
 }
 
-void __init prom_prepare_cpus(unsigned int max_cpus)
+void __init plat_smp_setup(void)
 {
 	cnodeid_t	cnode;
 
@@ -161,6 +161,11 @@ void __init prom_prepare_cpus(unsigned int max_cpus)
 	alloc_cpupda(0, 0);
 }
 
+void __init plat_prepare_cpus(unsigned int max_cpus)
+{
+	/* We already did everything necessary earlier */
+}
+
 /*
  * Launch a slave into smp_bootstrap().  It doesn't take an argument, and we
  * set sp to the kernel stack of the newly created idle process, gp to the proc