[PATCH] drivers: firmware: psci: add __init mark to psci_dt_cpu_init_idle

[PATCH] drivers: firmware: psci: add __init mark to psci_dt_cpu_init_idle

[Jisheng Zhang]

psci_dt_cpu_init_idle() and psci_cpu_init_idle() are not needed after
booting, so mark them as __init.

Signed-off-by: Jisheng Zhang <jszhang@marvell.com>
---
 drivers/firmware/psci.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/firmware/psci.c b/drivers/firmware/psci.c
index 7d52186..284413f 100644
--- a/drivers/firmware/psci.c
+++ b/drivers/firmware/psci.c
@@ -250,7 +250,7 @@ static int __init psci_features(u32 psci_func_id)
 #ifdef CONFIG_CPU_IDLE
 static DEFINE_PER_CPU_READ_MOSTLY(u32 *, psci_power_state);
 
-static int psci_dt_cpu_init_idle(struct device_node *cpu_node, int cpu)
+static int __init psci_dt_cpu_init_idle(struct device_node *cpu_node, int cpu)
 {
 	int i, ret, count = 0;
 	u32 *psci_states;
@@ -310,7 +310,7 @@ free_mem:
 	return ret;
 }
 
-int psci_cpu_init_idle(unsigned int cpu)
+int __init psci_cpu_init_idle(unsigned int cpu)
 {
 	struct device_node *cpu_node;
 	int ret;
-- 
2.8.0.rc3

[PATCH] drivers: firmware: psci: add __init mark to psci_dt_cpu_init_idle

[Lorenzo Pieralisi]

Hi Jisheng,

On Tue, Mar 22, 2016 at 10:35:29PM +0800, Jisheng Zhang wrote:
> psci_dt_cpu_init_idle() and psci_cpu_init_idle() are not needed after
> booting, so mark them as __init.
> 
> Signed-off-by: Jisheng Zhang <jszhang@marvell.com>
> ---

I've slightly changed the $SUBJECT, patch below FYI.

Lorenzo

-- >8 --
Subject: [PATCH] drivers: firmware: psci: mark idle init functions __init

psci_dt_cpu_init_idle() and psci_cpu_init_idle() are not needed after
booting, so mark them as __init.

Signed-off-by: Jisheng Zhang <jszhang@marvell.com>
[lpieralisi: updated patch subject]
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
---
 drivers/firmware/psci.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/firmware/psci.c b/drivers/firmware/psci.c
index 6d86881..4a6a7de 100644
--- a/drivers/firmware/psci.c
+++ b/drivers/firmware/psci.c
@@ -250,7 +250,7 @@ static int __init psci_features(u32 psci_func_id)
 #ifdef CONFIG_CPU_IDLE
 static DEFINE_PER_CPU_READ_MOSTLY(u32 *, psci_power_state);
 
-static int psci_dt_cpu_init_idle(struct device_node *cpu_node, int cpu)
+static int __init psci_dt_cpu_init_idle(struct device_node *cpu_node, int cpu)
 {
 	int i, ret, count = 0;
 	u32 *psci_states;
@@ -310,7 +310,7 @@ free_mem:
 	return ret;
 }
 
-int psci_cpu_init_idle(unsigned int cpu)
+int __init psci_cpu_init_idle(unsigned int cpu)
 {
 	struct device_node *cpu_node;
 	int ret;

[PATCH] drivers: firmware: psci: add __init mark to psci_dt_cpu_init_idle

[Sudeep Holla]

On 19/04/16 16:05, Lorenzo Pieralisi wrote:
> Hi Jisheng,
>
> On Tue, Mar 22, 2016 at 10:35:29PM +0800, Jisheng Zhang wrote:
>> psci_dt_cpu_init_idle() and psci_cpu_init_idle() are not needed after
>> booting, so mark them as __init.
>>

On the other hand, when I was trying to reuse arm_cpuidle_init to
support ACPI LPIs, I found that we may need to remove __init tag on it
as it gets called by hotplug notifiers in ACPI context.

I am not objecting this change, just mentioning with the hope to get
solution to the above issue.

-- 
Regards,
Sudeep

[PATCH] drivers: firmware: psci: add __init mark to psci_dt_cpu_init_idle

[Lorenzo Pieralisi]

On Tue, Apr 19, 2016 at 04:20:38PM +0100, Sudeep Holla wrote:
> 
> 
> On 19/04/16 16:05, Lorenzo Pieralisi wrote:
> >Hi Jisheng,
> >
> >On Tue, Mar 22, 2016 at 10:35:29PM +0800, Jisheng Zhang wrote:
> >>psci_dt_cpu_init_idle() and psci_cpu_init_idle() are not needed after
> >>booting, so mark them as __init.
> >>
> 
> On the other hand, when I was trying to reuse arm_cpuidle_init to
> support ACPI LPIs, I found that we may need to remove __init tag on it
> as it gets called by hotplug notifiers in ACPI context.
> 
> I am not objecting this change, just mentioning with the hope to get
> solution to the above issue.

Yep, good point, it is not strictly necessary to merge it either
and I would avoid churning out changes given that we know we will
require to remove those __init tags anyway, so patch dropped.

Lorenzo