[PATCH v2 3/3] powerpc/powernv/cpuidle: Add workaround to enable fastsleep

From: Shreyas B. Prabhu
Date: Wed Oct 01 2014 - 03:47:01 EST

From: Preeti U Murthy <preeti@xxxxxxxxxxxxxxxxxx>

Fast sleep is an idle state, where the core and the L1 and L2
caches are brought down to a threshold voltage. This also means that
the communication between L2 and L3 caches have to be fenced. However
the current P8 chips have a bug wherein this fencing between L2 and
L3 caches get delayed by a cpu cycle. This can delay L3 response to
the other cpus if they request for data during this time. Thus they
would fetch the same data from the memory which could lead to data
corruption if L3 cache is not flushed.

The cpu idle states save power at a core level and not at a thread level.
Hence powersavings is based on the shallowest idle state that a thread
of a core is in. The above issue in fastsleep will arise only when
all the threads in a core either enter fastsleep or some of them enter
any deeper idle states, with only a few being in fastsleep. This patch
therefore implements a workaround this bug by ensuring
that, each time a cpu goes to fastsleep, it checks if it is the last
thread in the core to enter fastsleep. If so, it needs to make an opal
call to get around the above mentioned fastsleep problem in the hardware
before issuing the sleep instruction.

Similarly when a thread in a core comes out of fastsleep, it needs
to verify if its the first thread in the core to come out of fastsleep
and issue the opal call to revert the changes made while entering
fastsleep.

For the same reason mentioned above we need to take care of offline threads
as well since we allow them to enter fastsleep and with support for
deep winkle soon coming in they can enter winkle as well. We therefore
ensure that even offline threads make the above mentioned opal calls
similarly, so that as long as the threads in a core are in and
idle state >= fastsleep, we have the workaround in place. Whenever a
thread comes out of either of these states, it needs to verify if the
opal call has been made and if so it will revert it. For now this patch
ensures that offline threads enter fastsleep.

We need to be able to synchronize the cpus in a core which are entering
and exiting fastsleep so as to ensure that the last thread in the core
to enter fastsleep and the first to exit fastsleep *only* issue the opal
call. To do so, we need a per-core lock and counter. The counter is
required to keep track of the number of threads in a core which are in
idle state >= fastsleep. To make the implementation of this simple, we
introduce a per-cpu lock and counter and every thread always takes the
primary thread's lock, modifies the primary thread's counter. This
effectively makes them per-core entities.

But the workaround is abstracted in the powernv core code and neither
the hotplug path nor the cpuidle driver need to bother about it. All
they need to know is if fastsleep, with error or no error is present as
an idle state.

Cc: linux-pm@xxxxxxxxxxxxxxx
Cc: linuxppc-dev@xxxxxxxxxxxxxxxx
Cc: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
Cc: Paul Mackerras <paulus@xxxxxxxxx>
Cc: Michael Ellerman <mpe@xxxxxxxxxxxxxx>
Cc: Rafael J. Wysocki <rjw@xxxxxxxxxxxxx>
Signed-off-by: Shreyas B. Prabhu <shreyas@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Preeti U Murthy <preeti@xxxxxxxxxxxxxxxxxx>
---
arch/powerpc/include/asm/machdep.h | 3 +
arch/powerpc/include/asm/opal.h | 3 +
arch/powerpc/include/asm/processor.h | 4 +-
arch/powerpc/kernel/idle.c | 19 ++++
arch/powerpc/kernel/idle_power7.S | 2 +-
arch/powerpc/platforms/powernv/opal-wrappers.S | 1 +
arch/powerpc/platforms/powernv/setup.c | 139 ++++++++++++++++++-------
drivers/cpuidle/cpuidle-powernv.c | 8 +-
8 files changed, 140 insertions(+), 39 deletions(-)

diff --git a/arch/powerpc/include/asm/machdep.h b/arch/powerpc/include/asm/machdep.h
index b125cea..f37014f 100644
--- a/arch/powerpc/include/asm/machdep.h
+++ b/arch/powerpc/include/asm/machdep.h
@@ -298,6 +298,9 @@ struct machdep_calls {
#ifdef CONFIG_MEMORY_HOTREMOVE
int (*remove_memory)(u64, u64);
#endif
+ /* Idle handlers */
+ void (*setup_idle)(void);
+ unsigned long (*power7_sleep)(void);
};

extern void e500_idle(void);
diff --git a/arch/powerpc/include/asm/opal.h b/arch/powerpc/include/asm/opal.h
index 28b8342..166d572 100644
--- a/arch/powerpc/include/asm/opal.h
+++ b/arch/powerpc/include/asm/opal.h
@@ -149,6 +149,7 @@ struct opal_sg_list {
#define OPAL_DUMP_INFO2 94
#define OPAL_PCI_EEH_FREEZE_SET 97
#define OPAL_HANDLE_HMI 98
+#define OPAL_CONFIG_IDLE_STATE 99
#define OPAL_REGISTER_DUMP_REGION 101
#define OPAL_UNREGISTER_DUMP_REGION 102

@@ -775,6 +776,7 @@ extern struct device_node *opal_node;
/* Flags used for idle state discovery from the device tree */
#define IDLE_INST_NAP 0x00010000 /* nap instruction can be used */
#define IDLE_INST_SLEEP 0x00020000 /* sleep instruction can be used */
+#define IDLE_INST_SLEEP_ER1 0x00080000 /* Use sleep with work around*/

/* API functions */
int64_t opal_invalid_call(void);
@@ -975,6 +977,7 @@ extern int opal_handle_hmi_exception(struct pt_regs *regs);

extern void opal_shutdown(void);
extern int opal_resync_timebase(void);
+int64_t opal_config_idle_state(uint64_t state, uint64_t enter);

extern void opal_lpc_init(void);

diff --git a/arch/powerpc/include/asm/processor.h b/arch/powerpc/include/asm/processor.h
index dda7ac4..41953cd 100644
--- a/arch/powerpc/include/asm/processor.h
+++ b/arch/powerpc/include/asm/processor.h
@@ -451,8 +451,10 @@ extern unsigned long cpuidle_disable;
enum idle_boot_override {IDLE_NO_OVERRIDE = 0, IDLE_POWERSAVE_OFF};

extern int powersave_nap; /* set if nap mode can be used in idle loop */
+extern void arch_setup_idle(void);
extern void power7_nap(int check_irq);
-extern void power7_sleep(void);
+extern unsigned long power7_sleep(void);
+extern unsigned long __power7_sleep(void);
extern void flush_instruction_cache(void);
extern void hard_reset_now(void);
extern void poweroff_now(void);
diff --git a/arch/powerpc/kernel/idle.c b/arch/powerpc/kernel/idle.c
index d7216c9..1f268e0 100644
--- a/arch/powerpc/kernel/idle.c
+++ b/arch/powerpc/kernel/idle.c
@@ -32,6 +32,9 @@
#include <asm/machdep.h>
#include <asm/runlatch.h>
#include <asm/smp.h>
+#include <asm/cputhreads.h>
+#include <asm/firmware.h>
+#include <asm/opal.h>


unsigned long cpuidle_disable = IDLE_NO_OVERRIDE;
@@ -78,6 +81,22 @@ void arch_cpu_idle(void)
HMT_medium();
ppc64_runlatch_on();
}
+void arch_setup_idle(void)
+{
+ if (ppc_md.setup_idle)
+ ppc_md.setup_idle();
+}
+
+unsigned long power7_sleep(void)
+{
+ unsigned long ret;
+
+ if (ppc_md.power7_sleep)
+ ret = ppc_md.power7_sleep();
+ else
+ ret = __power7_sleep();
+ return ret;
+}

int powersave_nap;

diff --git a/arch/powerpc/kernel/idle_power7.S b/arch/powerpc/kernel/idle_power7.S
index be05841..c3481c9 100644
--- a/arch/powerpc/kernel/idle_power7.S
+++ b/arch/powerpc/kernel/idle_power7.S
@@ -129,7 +129,7 @@ _GLOBAL(power7_nap)
b power7_powersave_common
/* No return */

-_GLOBAL(power7_sleep)
+_GLOBAL(__power7_sleep)
li r3,1
li r4,1
b power7_powersave_common
diff --git a/arch/powerpc/platforms/powernv/opal-wrappers.S b/arch/powerpc/platforms/powernv/opal-wrappers.S
index 2e6ce1b..8d1e724 100644
--- a/arch/powerpc/platforms/powernv/opal-wrappers.S
+++ b/arch/powerpc/platforms/powernv/opal-wrappers.S
@@ -245,5 +245,6 @@ OPAL_CALL(opal_sensor_read, OPAL_SENSOR_READ);
OPAL_CALL(opal_get_param, OPAL_GET_PARAM);
OPAL_CALL(opal_set_param, OPAL_SET_PARAM);
OPAL_CALL(opal_handle_hmi, OPAL_HANDLE_HMI);
+OPAL_CALL(opal_config_idle_state, OPAL_CONFIG_IDLE_STATE);
OPAL_CALL(opal_register_dump_region, OPAL_REGISTER_DUMP_REGION);
OPAL_CALL(opal_unregister_dump_region, OPAL_UNREGISTER_DUMP_REGION);
diff --git a/arch/powerpc/platforms/powernv/setup.c b/arch/powerpc/platforms/powernv/setup.c
index 2dca1d8..9d9a898 100644
--- a/arch/powerpc/platforms/powernv/setup.c
+++ b/arch/powerpc/platforms/powernv/setup.c
@@ -36,9 +36,20 @@
#include <asm/opal.h>
#include <asm/kexec.h>
#include <asm/smp.h>
+#include <asm/cputhreads.h>

#include "powernv.h"

+/* Per-cpu structures to keep track of cpus of a core that
+ * are in idle states >= fastsleep so as to call opal for
+ * sleep setup when the entire core is ready to go to fastsleep.
+ *
+ * We need sometihng similar to a per-core lock. For now we
+ * achieve this by taking the lock of the primary thread in the core.
+ */
+static DEFINE_PER_CPU(spinlock_t, fastsleep_override_lock);
+static DEFINE_PER_CPU(int, fastsleep_cnt);
+
static void __init pnv_setup_arch(void)
{
set_arch_panic_timeout(10, ARCH_PANIC_TIMEOUT);
@@ -254,35 +265,8 @@ static unsigned long pnv_memory_block_size(void)
}
#endif

-static void __init pnv_setup_machdep_opal(void)
-{
- ppc_md.get_boot_time = opal_get_boot_time;
- ppc_md.get_rtc_time = opal_get_rtc_time;
- ppc_md.set_rtc_time = opal_set_rtc_time;
- ppc_md.restart = pnv_restart;
- ppc_md.power_off = pnv_power_off;
- ppc_md.halt = pnv_halt;
- ppc_md.machine_check_exception = opal_machine_check;
- ppc_md.mce_check_early_recovery = opal_mce_check_early_recovery;
- ppc_md.hmi_exception_early = opal_hmi_exception_early;
- ppc_md.handle_hmi_exception = opal_handle_hmi_exception;
-}
-
-#ifdef CONFIG_PPC_POWERNV_RTAS
-static void __init pnv_setup_machdep_rtas(void)
-{
- if (rtas_token("get-time-of-day") != RTAS_UNKNOWN_SERVICE) {
- ppc_md.get_boot_time = rtas_get_boot_time;
- ppc_md.get_rtc_time = rtas_get_rtc_time;
- ppc_md.set_rtc_time = rtas_set_rtc_time;
- }
- ppc_md.restart = rtas_restart;
- ppc_md.power_off = rtas_power_off;
- ppc_md.halt = rtas_halt;
-}
-#endif /* CONFIG_PPC_POWERNV_RTAS */
-
static unsigned int supported_cpuidle_states;
+static int need_fastsleep_workaround;

unsigned int pnv_get_supported_cpuidle_states(void)
{
@@ -292,12 +276,13 @@ unsigned int pnv_get_supported_cpuidle_states(void)
static int __init pnv_probe_idle_states(void)
{
struct device_node *power_mgt;
- struct property *prop;
int dt_idle_states;
- u32 *flags;
+ const __be32 *idle_state_flags;
+ u32 len_flags, flags;
int i;

supported_cpuidle_states = 0;
+ need_fastsleep_workaround = 0;

if (cpuidle_disable != IDLE_NO_OVERRIDE)
return 0;
@@ -311,21 +296,28 @@ static int __init pnv_probe_idle_states(void)
return 0;
}

- prop = of_find_property(power_mgt, "ibm,cpu-idle-state-flags", NULL);
- if (!prop) {
+ idle_state_flags = of_get_property(power_mgt,
+ "ibm,cpu-idle-state-flags", &len_flags);
+ if (!idle_state_flags) {
pr_warn("DT-PowerMgmt: missing ibm,cpu-idle-state-flags\n");
return 0;
}

- dt_idle_states = prop->length / sizeof(u32);
- flags = (u32 *) prop->value;
+ dt_idle_states = len_flags / sizeof(u32);

for (i = 0; i < dt_idle_states; i++) {
- if (flags[i] & IDLE_INST_NAP)
+
+ flags = be32_to_cpu(idle_state_flags[i]);
+ if (flags & IDLE_INST_NAP)
supported_cpuidle_states |= IDLE_USE_NAP;

- if (flags[i] & IDLE_INST_SLEEP)
+ if (flags & IDLE_INST_SLEEP)
supported_cpuidle_states |= IDLE_USE_SLEEP;
+
+ if (flags & IDLE_INST_SLEEP_ER1) {
+ supported_cpuidle_states |= IDLE_USE_SLEEP;
+ need_fastsleep_workaround = 1;
+ }
}

return 0;
@@ -333,6 +325,81 @@ static int __init pnv_probe_idle_states(void)

subsys_initcall(pnv_probe_idle_states);

+static void pnv_setup_idle(void)
+{
+ int cpu;
+
+ for_each_possible_cpu(cpu) {
+ spin_lock_init(&per_cpu(fastsleep_override_lock, cpu));
+ per_cpu(fastsleep_cnt, cpu) = threads_per_core;
+ }
+}
+
+static void
+pnv_apply_fastsleep_workaround(bool enter_fastsleep, int primary_thread)
+{
+ if (enter_fastsleep) {
+ spin_lock(&per_cpu(fastsleep_override_lock, primary_thread));
+ if (--(per_cpu(fastsleep_cnt, primary_thread)) == 0)
+ opal_config_idle_state(1, 1);
+ spin_unlock(&per_cpu(fastsleep_override_lock, primary_thread));
+ } else {
+ spin_lock(&per_cpu(fastsleep_override_lock, primary_thread));
+ if ((per_cpu(fastsleep_cnt, primary_thread)) == 0)
+ opal_config_idle_state(1, 0);
+ per_cpu(fastsleep_cnt, primary_thread)++;
+ spin_unlock(&per_cpu(fastsleep_override_lock, primary_thread));
+ }
+}
+
+static unsigned long pnv_power7_sleep(void)
+{
+ int cpu, primary_thread;
+ unsigned long srr1;
+
+ cpu = smp_processor_id();
+ primary_thread = cpu_first_thread_sibling(cpu);
+
+ if (need_fastsleep_workaround) {
+ pnv_apply_fastsleep_workaround(1, primary_thread);
+ srr1 = __power7_sleep();
+ pnv_apply_fastsleep_workaround(0, primary_thread);
+ } else {
+ srr1 = __power7_sleep();
+ }
+ return srr1;
+}
+
+static void __init pnv_setup_machdep_opal(void)
+{
+ ppc_md.get_boot_time = opal_get_boot_time;
+ ppc_md.get_rtc_time = opal_get_rtc_time;
+ ppc_md.set_rtc_time = opal_set_rtc_time;
+ ppc_md.restart = pnv_restart;
+ ppc_md.power_off = pnv_power_off;
+ ppc_md.halt = pnv_halt;
+ ppc_md.machine_check_exception = opal_machine_check;
+ ppc_md.mce_check_early_recovery = opal_mce_check_early_recovery;
+ ppc_md.hmi_exception_early = opal_hmi_exception_early;
+ ppc_md.handle_hmi_exception = opal_handle_hmi_exception;
+ ppc_md.setup_idle = pnv_setup_idle;
+ ppc_md.power7_sleep = pnv_power7_sleep;
+}
+
+#ifdef CONFIG_PPC_POWERNV_RTAS
+static void __init pnv_setup_machdep_rtas(void)
+{
+ if (rtas_token("get-time-of-day") != RTAS_UNKNOWN_SERVICE) {
+ ppc_md.get_boot_time = rtas_get_boot_time;
+ ppc_md.get_rtc_time = rtas_get_rtc_time;
+ ppc_md.set_rtc_time = rtas_set_rtc_time;
+ }
+ ppc_md.restart = rtas_restart;
+ ppc_md.power_off = rtas_power_off;
+ ppc_md.halt = rtas_halt;
+}
+#endif /* CONFIG_PPC_POWERNV_RTAS */
+
static int __init pnv_probe(void)
{
unsigned long root = of_get_flat_dt_root();
diff --git a/drivers/cpuidle/cpuidle-powernv.c b/drivers/cpuidle/cpuidle-powernv.c
index 23d2743..8ad97a9 100644
--- a/drivers/cpuidle/cpuidle-powernv.c
+++ b/drivers/cpuidle/cpuidle-powernv.c
@@ -18,6 +18,7 @@
#include <asm/firmware.h>
#include <asm/opal.h>
#include <asm/runlatch.h>
+#include <asm/processor.h>

/* Flags and constants used in PowerNV platform */

@@ -195,7 +196,8 @@ static int powernv_add_idle_states(void)
nr_idle_states++;
}

- if (flags & IDLE_INST_SLEEP) {
+ if ((flags & IDLE_INST_SLEEP_ER1) ||
+ (flags & IDLE_INST_SLEEP)) {
/* Add FASTSLEEP state */
strcpy(powernv_states[nr_idle_states].name, "FastSleep");
strcpy(powernv_states[nr_idle_states].desc, "FastSleep");
@@ -247,6 +249,10 @@ static int __init powernv_processor_idle_init(void)

register_cpu_notifier(&setup_hotplug_notifier);
printk(KERN_DEBUG "powernv_idle_driver registered\n");
+
+ /* If any idle states require special
+ * initializations before cpuidle kicks in */
+ arch_setup_idle();
return 0;
}

--
1.9.3

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/