2019-05-27 16:23:30

by Miles Chen

Subject: [PATCH] arm64: mm: make CONFIG_ZONE_DMA32 configurable

This change makes CONFIG_ZONE_DMA32 defuly y and allows users
to overwrite it.

For the SoCs that do not need CONFIG_ZONE_DMA32, this is the
first step to manage all available memory by a single
zone(normal zone) to reduce the overhead of multiple zones.

The change also fixes a build error when CONFIG_NUMA=y and
CONFIG_ZONE_DMA32=n.

arch/arm64/mm/init.c:195:17: error: use of undeclared identifier 'ZONE_DMA32'
max_zone_pfns[ZONE_DMA32] = PFN_DOWN(max_zone_dma_phys());

Signed-off-by: Miles Chen <[email protected]>
---
arch/arm64/Kconfig | 3 ++-
arch/arm64/mm/init.c | 2 ++
2 files changed, 4 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 76f6e4765f49..9d20a736d1d1 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -260,7 +260,8 @@ config GENERIC_CALIBRATE_DELAY
def_bool y

config ZONE_DMA32
- def_bool y
+ bool "Support DMA32 zone"
+ default y

config HAVE_GENERIC_GUP
def_bool y
diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c
index d2adffb81b5d..96829ce21f99 100644
--- a/arch/arm64/mm/init.c
+++ b/arch/arm64/mm/init.c
@@ -191,8 +191,10 @@ static void __init zone_sizes_init(unsigned long min, unsigned long max)
{
unsigned long max_zone_pfns[MAX_NR_ZONES] = {0};

+#ifdef CONFIG_ZONE_DMA32
if (IS_ENABLED(CONFIG_ZONE_DMA32))
max_zone_pfns[ZONE_DMA32] = PFN_DOWN(max_zone_dma_phys());
+#endif
max_zone_pfns[ZONE_NORMAL] = max;

free_area_init_nodes(max_zone_pfns);
--
2.18.0

2019-05-28 10:44:57

by Robin Murphy

Subject: Re: [PATCH] arm64: mm: make CONFIG_ZONE_DMA32 configurable

On 27/05/2019 17:08, Miles Chen wrote:
> This change makes CONFIG_ZONE_DMA32 defuly y and allows users
> to overwrite it.
>
> For the SoCs that do not need CONFIG_ZONE_DMA32, this is the
> first step to manage all available memory by a single
> zone(normal zone) to reduce the overhead of multiple zones.
>
> The change also fixes a build error when CONFIG_NUMA=y and
> CONFIG_ZONE_DMA32=n.
>
> arch/arm64/mm/init.c:195:17: error: use of undeclared identifier 'ZONE_DMA32'
> max_zone_pfns[ZONE_DMA32] = PFN_DOWN(max_zone_dma_phys());
>
> Signed-off-by: Miles Chen <[email protected]>
> ---
> arch/arm64/Kconfig | 3 ++-
> arch/arm64/mm/init.c | 2 ++
> 2 files changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> index 76f6e4765f49..9d20a736d1d1 100644
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -260,7 +260,8 @@ config GENERIC_CALIBRATE_DELAY
> def_bool y
>
> config ZONE_DMA32
> - def_bool y
> + bool "Support DMA32 zone"

This probably warrants an "if EMBEDDED" or "if EXPERT", since turning it
off produces a kernel which won't work at all on certain systems (I've
played around with this before...)

> + default y
>
> config HAVE_GENERIC_GUP
> def_bool y
> diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c
> index d2adffb81b5d..96829ce21f99 100644
> --- a/arch/arm64/mm/init.c
> +++ b/arch/arm64/mm/init.c
> @@ -191,8 +191,10 @@ static void __init zone_sizes_init(unsigned long min, unsigned long max)
> {
> unsigned long max_zone_pfns[MAX_NR_ZONES] = {0};
>
> +#ifdef CONFIG_ZONE_DMA32
> if (IS_ENABLED(CONFIG_ZONE_DMA32))

There's no point keeping the IS_ENABLED() check when it's entirely
redundant with the #ifdefs.

Robin.

> max_zone_pfns[ZONE_DMA32] = PFN_DOWN(max_zone_dma_phys());
> +#endif
> max_zone_pfns[ZONE_NORMAL] = max;
>
> free_area_init_nodes(max_zone_pfns);
>

2019-05-28 15:49:09

by Miles Chen

Subject: Re: [PATCH] arm64: mm: make CONFIG_ZONE_DMA32 configurable

On Tue, 2019-05-28 at 11:43 +0100, Robin Murphy wrote:
> On 27/05/2019 17:08, Miles Chen wrote:
> > This change makes CONFIG_ZONE_DMA32 defuly y and allows users
> > to overwrite it.
> >
> > For the SoCs that do not need CONFIG_ZONE_DMA32, this is the
> > first step to manage all available memory by a single
> > zone(normal zone) to reduce the overhead of multiple zones.
> >
> > The change also fixes a build error when CONFIG_NUMA=y and
> > CONFIG_ZONE_DMA32=n.
> >
> > arch/arm64/mm/init.c:195:17: error: use of undeclared identifier 'ZONE_DMA32'
> > max_zone_pfns[ZONE_DMA32] = PFN_DOWN(max_zone_dma_phys());
> >
> > Signed-off-by: Miles Chen <[email protected]>
> > ---
> > arch/arm64/Kconfig | 3 ++-
> > arch/arm64/mm/init.c | 2 ++
> > 2 files changed, 4 insertions(+), 1 deletion(-)
> >
> > diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> > index 76f6e4765f49..9d20a736d1d1 100644
> > --- a/arch/arm64/Kconfig
> > +++ b/arch/arm64/Kconfig
> > @@ -260,7 +260,8 @@ config GENERIC_CALIBRATE_DELAY
> > def_bool y
> >
> > config ZONE_DMA32
> > - def_bool y
> > + bool "Support DMA32 zone"
>
> This probably warrants an "if EMBEDDED" or "if EXPERT", since turning it
> off produces a kernel which won't work at all on certain systems (I've
> played around with this before...)

Thanks for your comment.
I'll put a "if EXPERT" here to avoid this case.

>
> > + default y
> >
> > config HAVE_GENERIC_GUP
> > def_bool y
> > diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c
> > index d2adffb81b5d..96829ce21f99 100644
> > --- a/arch/arm64/mm/init.c
> > +++ b/arch/arm64/mm/init.c
> > @@ -191,8 +191,10 @@ static void __init zone_sizes_init(unsigned long min, unsigned long max)
> > {
> > unsigned long max_zone_pfns[MAX_NR_ZONES] = {0};
> >
> > +#ifdef CONFIG_ZONE_DMA32
> > if (IS_ENABLED(CONFIG_ZONE_DMA32))
>
> There's no point keeping the IS_ENABLED() check when it's entirely
> redundant with the #ifdefs.

I'll remove the IS_ENABLE() code in next patch.

-Miles
>
> Robin.
>
> > max_zone_pfns[ZONE_DMA32] = PFN_DOWN(max_zone_dma_phys());
> > +#endif
> > max_zone_pfns[ZONE_NORMAL] = max;
> >
> > free_area_init_nodes(max_zone_pfns);
> >