[PATCH v5] ARM: mvebu: use dt_fixup to provide fallback for enable-method

[PATCH v5] ARM: mvebu: use dt_fixup to provide fallback for enable-method

[Chris Packham]

We need to maintain backwards compatibility with device trees that don't
define an enable method. At the same time we want the device tree to be
able to specify an enable-method and have it stick.

Previously by having smp assigned in the DT_MACHINE definition this
would be picked up by setup_arch() and override whatever
arm_dt_init_cpu_maps() had configured. Now we move the initial
assignment of default smp_ops to a dt_fixup and let
arm_dt_init_cpu_maps() override that if the device tree defines an
enable-method.

Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
---
Pervious versions

v1: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/300182.html
v2: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/300480.html
v3: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/302945.html
v4: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/303899.html

Changes since v4:
- drop "RFC"

 arch/arm/mach-mvebu/board-v7.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/arch/arm/mach-mvebu/board-v7.c b/arch/arm/mach-mvebu/board-v7.c
index ccca95173e17..5bbde5e5258e 100644
--- a/arch/arm/mach-mvebu/board-v7.c
+++ b/arch/arm/mach-mvebu/board-v7.c
@@ -145,6 +145,11 @@ static void __init mvebu_dt_init(void)
 		i2c_quirk();
 }
 
+static void __init armada_370_xp_dt_fixup(void)
+{
+	smp_set_ops(smp_ops(armada_xp_smp_ops));
+}
+
 static const char * const armada_370_xp_dt_compat[] __initconst = {
 	"marvell,armada-370-xp",
 	NULL,
@@ -153,17 +158,12 @@ static const char * const armada_370_xp_dt_compat[] __initconst = {
 DT_MACHINE_START(ARMADA_370_XP_DT, "Marvell Armada 370/XP (Device Tree)")
 	.l2c_aux_val	= 0,
 	.l2c_aux_mask	= ~0,
-/*
- * The following field (.smp) is still needed to ensure backward
- * compatibility with old Device Trees that were not specifying the
- * cpus enable-method property.
- */
-	.smp		= smp_ops(armada_xp_smp_ops),
 	.init_machine	= mvebu_dt_init,
 	.init_irq       = mvebu_init_irq,
 	.restart	= mvebu_restart,
 	.reserve        = mvebu_memblock_reserve,
 	.dt_compat	= armada_370_xp_dt_compat,
+	.dt_fixup	= armada_370_xp_dt_fixup,
 MACHINE_END
 
 static const char * const armada_375_dt_compat[] __initconst = {
-- 
2.18.0

Re: [PATCH v5] ARM: mvebu: use dt_fixup to provide fallback for enable-method

[Gregory CLEMENT]

Hi Chris,
 
 On jeu., juil. 26 2018, Chris Packham <chris.packham@alliedtelesis.co.nz> wrote:

> We need to maintain backwards compatibility with device trees that don't
> define an enable method. At the same time we want the device tree to be
> able to specify an enable-method and have it stick.
>
> Previously by having smp assigned in the DT_MACHINE definition this
> would be picked up by setup_arch() and override whatever
> arm_dt_init_cpu_maps() had configured. Now we move the initial
> assignment of default smp_ops to a dt_fixup and let
> arm_dt_init_cpu_maps() override that if the device tree defines an
> enable-method.
>
> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>

I made several tests on an Armada XP based board: OpenBlock AX3: I
modify the enable-method in the decvice tree, and I confirm that without
your patch it is not taken into account whereas with this patch the
enable-method is applied form the device tree. I also didn't see any
regression with the original dtb.

So I added my:
Tested-by: Gregory CLEMENT <gregory.clement@bootlin.com>

and applied on mvebu/soc

Thanks,

Gregory

> ---
> Pervious versions
>
> v1: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/300182.html
> v2: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/300480.html
> v3: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/302945.html
> v4: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-November/303899.html
>
> Changes since v4:
> - drop "RFC"
>
>  arch/arm/mach-mvebu/board-v7.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)
>
> diff --git a/arch/arm/mach-mvebu/board-v7.c b/arch/arm/mach-mvebu/board-v7.c
> index ccca95173e17..5bbde5e5258e 100644
> --- a/arch/arm/mach-mvebu/board-v7.c
> +++ b/arch/arm/mach-mvebu/board-v7.c
> @@ -145,6 +145,11 @@ static void __init mvebu_dt_init(void)
>  		i2c_quirk();
>  }
>  
> +static void __init armada_370_xp_dt_fixup(void)
> +{
> +	smp_set_ops(smp_ops(armada_xp_smp_ops));
> +}
> +
>  static const char * const armada_370_xp_dt_compat[] __initconst = {
>  	"marvell,armada-370-xp",
>  	NULL,
> @@ -153,17 +158,12 @@ static const char * const armada_370_xp_dt_compat[] __initconst = {
>  DT_MACHINE_START(ARMADA_370_XP_DT, "Marvell Armada 370/XP (Device Tree)")
>  	.l2c_aux_val	= 0,
>  	.l2c_aux_mask	= ~0,
> -/*
> - * The following field (.smp) is still needed to ensure backward
> - * compatibility with old Device Trees that were not specifying the
> - * cpus enable-method property.
> - */
> -	.smp		= smp_ops(armada_xp_smp_ops),
>  	.init_machine	= mvebu_dt_init,
>  	.init_irq       = mvebu_init_irq,
>  	.restart	= mvebu_restart,
>  	.reserve        = mvebu_memblock_reserve,
>  	.dt_compat	= armada_370_xp_dt_compat,
> +	.dt_fixup	= armada_370_xp_dt_fixup,
>  MACHINE_END
>  
>  static const char * const armada_375_dt_compat[] __initconst = {
> -- 
> 2.18.0
>

-- 
Gregory Clement, Bootlin
Embedded Linux and Kernel engineering
http://bootlin.com

Re: [PATCH v5] ARM: mvebu: use dt_fixup to provide fallback for enable-method

[Olof Johansson]

On Fri, Sep 21, 2018 at 12:05:48PM +0200, Gregory CLEMENT wrote:
> Hi Chris,
>  
>  On jeu., juil. 26 2018, Chris Packham <chris.packham@alliedtelesis.co.nz> wrote:
> 
> > We need to maintain backwards compatibility with device trees that don't
> > define an enable method. At the same time we want the device tree to be
> > able to specify an enable-method and have it stick.
> >
> > Previously by having smp assigned in the DT_MACHINE definition this
> > would be picked up by setup_arch() and override whatever
> > arm_dt_init_cpu_maps() had configured. Now we move the initial
> > assignment of default smp_ops to a dt_fixup and let
> > arm_dt_init_cpu_maps() override that if the device tree defines an
> > enable-method.
> >
> > Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
> 
> I made several tests on an Armada XP based board: OpenBlock AX3: I
> modify the enable-method in the decvice tree, and I confirm that without
> your patch it is not taken into account whereas with this patch the
> enable-method is applied form the device tree. I also didn't see any
> regression with the original dtb.
> 
> So I added my:
> Tested-by: Gregory CLEMENT <gregory.clement@bootlin.com>
> 
> and applied on mvebu/soc

Hi,

Looks like this broke non-SMP. Not a huge deal, but please apply this as
closely as possible on top of the previous patch (or squash it in).


- 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< -



From 3190d9502607995c7aecce79beec36714574d494 Mon Sep 17 00:00:00 2001
From: Olof Johansson <olof@lixom.net>
Date: Mon, 24 Sep 2018 02:37:31 -0700
Subject: [PATCH] ARM: mvebu: fix !SMP build

Wrap set_smp_ops() in CONFIG_SMP.

Fixes: d6ec59de9a0a8 ("ARM: mvebu: use dt_fixup to provide fallback for enable-method")
Cc: Chris Packham <chris.packham@alliedtelesis.co.nz>
Signed-off-by: Olof Johansson <olof@lixom.net>
---
 arch/arm/mach-mvebu/board-v7.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/arm/mach-mvebu/board-v7.c b/arch/arm/mach-mvebu/board-v7.c
index 5bbde5e..0b10acd 100644
--- a/arch/arm/mach-mvebu/board-v7.c
+++ b/arch/arm/mach-mvebu/board-v7.c
@@ -147,7 +147,9 @@ static void __init mvebu_dt_init(void)
 
 static void __init armada_370_xp_dt_fixup(void)
 {
+#ifdef CONFIG_SMP
 	smp_set_ops(smp_ops(armada_xp_smp_ops));
+#endif
 }
 
 static const char * const armada_370_xp_dt_compat[] __initconst = {
-- 
2.8.6

Re: [PATCH v5] ARM: mvebu: use dt_fixup to provide fallback for enable-method

[Chris Packham]

On 24/09/18 21:54, Olof Johansson wrote:
> On Fri, Sep 21, 2018 at 12:05:48PM +0200, Gregory CLEMENT wrote:
>> Hi Chris,
>>   
>>   On jeu., juil. 26 2018, Chris Packham <chris.packham@alliedtelesis.co.nz> wrote:
>>
>>> We need to maintain backwards compatibility with device trees that don't
>>> define an enable method. At the same time we want the device tree to be
>>> able to specify an enable-method and have it stick.
>>>
>>> Previously by having smp assigned in the DT_MACHINE definition this
>>> would be picked up by setup_arch() and override whatever
>>> arm_dt_init_cpu_maps() had configured. Now we move the initial
>>> assignment of default smp_ops to a dt_fixup and let
>>> arm_dt_init_cpu_maps() override that if the device tree defines an
>>> enable-method.
>>>
>>> Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
>>
>> I made several tests on an Armada XP based board: OpenBlock AX3: I
>> modify the enable-method in the decvice tree, and I confirm that without
>> your patch it is not taken into account whereas with this patch the
>> enable-method is applied form the device tree. I also didn't see any
>> regression with the original dtb.
>>
>> So I added my:
>> Tested-by: Gregory CLEMENT <gregory.clement@bootlin.com>
>>
>> and applied on mvebu/soc
> 
> Hi,
> 
> Looks like this broke non-SMP. Not a huge deal, but please apply this as
> closely as possible on top of the previous patch (or squash it in).
> 
> 
> - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< - 8< -
> 
> 
> 
>  From 3190d9502607995c7aecce79beec36714574d494 Mon Sep 17 00:00:00 2001
> From: Olof Johansson <olof@lixom.net>
> Date: Mon, 24 Sep 2018 02:37:31 -0700
> Subject: [PATCH] ARM: mvebu: fix !SMP build
> 
> Wrap set_smp_ops() in CONFIG_SMP.
> 
> Fixes: d6ec59de9a0a8 ("ARM: mvebu: use dt_fixup to provide fallback for enable-method")
> Cc: Chris Packham <chris.packham@alliedtelesis.co.nz>
> Signed-off-by: Olof Johansson <olof@lixom.net>
> ---
>   arch/arm/mach-mvebu/board-v7.c | 2 ++
>   1 file changed, 2 insertions(+)
> 
> diff --git a/arch/arm/mach-mvebu/board-v7.c b/arch/arm/mach-mvebu/board-v7.c
> index 5bbde5e..0b10acd 100644
> --- a/arch/arm/mach-mvebu/board-v7.c
> +++ b/arch/arm/mach-mvebu/board-v7.c
> @@ -147,7 +147,9 @@ static void __init mvebu_dt_init(void)
>   
>   static void __init armada_370_xp_dt_fixup(void)
>   {
> +#ifdef CONFIG_SMP
>   	smp_set_ops(smp_ops(armada_xp_smp_ops));
> +#endif
>   }
>   
>   static const char * const armada_370_xp_dt_compat[] __initconst = {
> 

Makes sense to me.

Gregory, do you want me to send a v6 or are you able to squash this in?