powerpc: Fix RCU idle and hcall tracing

Tracepoints should not be called inside an rcu_idle_enter/rcu_idle_exit
region. Since pSeries calls H_CEDE in the idle loop, we were violating
this rule.

commit a7b152d5342c (powerpc: Tell RCU about idle after hcall tracing)
tried to work around it by delaying the rcu_idle_enter until after we
called the hcall tracepoint, but there are a number of issues with it.

The hcall tracepoint trampoline code is called conditionally when the
tracepoint is enabled. If the tracepoint is not enabled we never call
rcu_idle_enter. The idle_uses_rcu check was also done at compile time
which breaks multiplatform builds.

The simple fix is to avoid tracing H_CEDE and rely on other tracepoints
and the hypervisor dispatch trace log to work out if we called H_CEDE.

This fixes a hang during boot on pSeries.

Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

2 files changed, 12 insertions, 14 deletions

diff --git a/arch/powerpc/kernel/idle.c b/arch/powerpc/kernel/idle.c
index 7c66ce13da89..0a48bf5db6c8 100644
--- a/arch/powerpc/kernel/idle.c
+++ b/arch/powerpc/kernel/idle.c
@@ -50,12 +50,6 @@ static int __init powersave_off(char *arg)
 }
 __setup("powersave=off", powersave_off);
 
-#if defined(CONFIG_PPC_PSERIES) && defined(CONFIG_TRACEPOINTS)
-static const bool idle_uses_rcu = 1;
-#else
-static const bool idle_uses_rcu;
-#endif
-
 /*
  * The body of the idle task.
  */
@@ -67,8 +61,7 @@ void cpu_idle(void)
         set_thread_flag(TIF_POLLING_NRFLAG);
         while (1) {
                 tick_nohz_idle_enter();
-                if (!idle_uses_rcu)
-                        rcu_idle_enter();
+                rcu_idle_enter();
 
                 while (!need_resched() && !cpu_should_die()) {
                         ppc64_runlatch_off();
@@ -106,8 +99,7 @@ void cpu_idle(void)
 
                 HMT_medium();
                 ppc64_runlatch_on();
-                if (!idle_uses_rcu)
-                        rcu_idle_exit();
+                rcu_idle_exit();
                 tick_nohz_idle_exit();
                 preempt_enable_no_resched();
                 if (cpu_should_die())
diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c
index 948e0e3b3547..7bc73af6c7b9 100644
--- a/arch/powerpc/platforms/pseries/lpar.c
+++ b/arch/powerpc/platforms/pseries/lpar.c
@@ -546,6 +546,13 @@ void __trace_hcall_entry(unsigned long opcode, unsigned long *args)
         unsigned long flags;
         unsigned int *depth;
 
+        /*
+         * We cannot call tracepoints inside RCU idle regions which
+         * means we must not trace H_CEDE.
+         */
+        if (opcode == H_CEDE)
+                return;
+
         local_irq_save(flags);
 
         depth = &__get_cpu_var(hcall_trace_depth);
@@ -556,8 +563,6 @@ void __trace_hcall_entry(unsigned long opcode, unsigned long *args)
         (*depth)++;
         preempt_disable();
         trace_hcall_entry(opcode, args);
-        if (opcode == H_CEDE)
-                rcu_idle_enter();
         (*depth)--;
 
 out:
@@ -570,6 +575,9 @@ void __trace_hcall_exit(long opcode, unsigned long retval,
         unsigned long flags;
         unsigned int *depth;
 
+        if (opcode == H_CEDE)
+                return;
+
         local_irq_save(flags);
 
         depth = &__get_cpu_var(hcall_trace_depth);
@@ -578,8 +586,6 @@ void __trace_hcall_exit(long opcode, unsigned long retval,
                 goto out;
 
         (*depth)++;
-        if (opcode == H_CEDE)
-                rcu_idle_exit();
         trace_hcall_exit(opcode, retval, retbuf);
         preempt_enable();
         (*depth)--;