1 files changed, 43 insertions, 6 deletions

diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c
index 12e656ffe60e..cb64a6e1dc51 100644
--- a/arch/powerpc/kernel/prom_init.c
+++ b/arch/powerpc/kernel/prom_init.c
@@ -196,6 +196,8 @@ static int __initdata mem_reserve_cnt;
 
 static cell_t __initdata regbuf[1024];
 
+static bool rtas_has_query_cpu_stopped;
+
 
 /*
  * Error results ... some OF calls will return "-1" on error, some
@@ -856,7 +858,8 @@ static void __init prom_send_capabilities(void)
 {
         ihandle root;
         prom_arg_t ret;
-        __be32 *cores;
+        u32 cores;
+        unsigned char *ptcores;
 
         root = call_prom("open", 1, 1, ADDR("/"));
         if (root != 0) {
@@ -866,15 +869,30 @@ static void __init prom_send_capabilities(void)
                  * (we assume this is the same for all cores) and use it to
                  * divide NR_CPUS.
                  */
-                cores = (__be32 *)&ibm_architecture_vec[IBM_ARCH_VEC_NRCORES_OFFSET];
-                if (be32_to_cpup(cores) != NR_CPUS) {
+
+                /* The core value may start at an odd address. If such a word
+                 * access is made at a cache line boundary, this leads to an
+                 * exception which may not be handled at this time.
+                 * Forcing a per byte access to avoid exception.
+                 */
+                ptcores = &ibm_architecture_vec[IBM_ARCH_VEC_NRCORES_OFFSET];
+                cores = 0;
+                cores |= ptcores[0] << 24;
+                cores |= ptcores[1] << 16;
+                cores |= ptcores[2] << 8;
+                cores |= ptcores[3];
+                if (cores != NR_CPUS) {
                         prom_printf("WARNING ! "
                                     "ibm_architecture_vec structure inconsistent: %lu!\n",
-                                    be32_to_cpup(cores));
+                                    cores);
                 } else {
-                        *cores = cpu_to_be32(DIV_ROUND_UP(NR_CPUS, prom_count_smt_threads()));
+                        cores = DIV_ROUND_UP(NR_CPUS, prom_count_smt_threads());
                         prom_printf("Max number of cores passed to firmware: %lu (NR_CPUS = %lu)\n",
-                                    be32_to_cpup(cores), NR_CPUS);
+                                    cores, NR_CPUS);
+                        ptcores[0] = (cores >> 24) & 0xff;
+                        ptcores[1] = (cores >> 16) & 0xff;
+                        ptcores[2] = (cores >> 8) & 0xff;
+                        ptcores[3] = cores & 0xff;
                 }
 
                 /* try calling the ibm,client-architecture-support method */
@@ -1574,6 +1592,11 @@ static void __init prom_instantiate_rtas(void)
         prom_setprop(rtas_node, "/rtas", "linux,rtas-entry",
                      &val, sizeof(val));
 
+        /* Check if it supports "query-cpu-stopped-state" */
+        if (prom_getprop(rtas_node, "query-cpu-stopped-state",
+                         &val, sizeof(val)) != PROM_ERROR)
+                rtas_has_query_cpu_stopped = true;
+
 #if defined(CONFIG_PPC_POWERNV) && defined(__BIG_ENDIAN__)
         /* PowerVN takeover hack */
         prom_rtas_data = base;
@@ -1815,6 +1838,18 @@ static void __init prom_hold_cpus(void)
                 = (void *) LOW_ADDR(__secondary_hold_acknowledge);
         unsigned long secondary_hold = LOW_ADDR(__secondary_hold);
 
+        /*
+         * On pseries, if RTAS supports "query-cpu-stopped-state",
+         * we skip this stage, the CPUs will be started by the
+         * kernel using RTAS.
+         */
+        if ((of_platform == PLATFORM_PSERIES ||
+             of_platform == PLATFORM_PSERIES_LPAR) &&
+            rtas_has_query_cpu_stopped) {
+                prom_printf("prom_hold_cpus: skipped\n");
+                return;
+        }
+
         prom_debug("prom_hold_cpus: start...\n");
         prom_debug("    1) spinloop       = 0x%x\n", (unsigned long)spinloop);
         prom_debug("    1) *spinloop      = 0x%x\n", *spinloop);
@@ -3011,6 +3046,8 @@ unsigned long __init prom_init(unsigned long r3, unsigned long r4,
          * On non-powermacs, put all CPUs in spin-loops.
          *
          * PowerMacs use a different mechanism to spin CPUs
+         *
+         * (This must be done after instanciating RTAS)
          */
         if (of_platform != PLATFORM_POWERMAC &&
             of_platform != PLATFORM_OPAL)