2020-10-04 05:56:22

by Phil Chang

[permalink] [raw]
Subject: [PATCH] [PATCH] of_reserved_mem: Increase the number of reserved regions

Certain SoCs need to support large amount of reserved memory
regions, especially to follow the GKI rules from Google.
In MTK new SoC requires more than 68 regions of reserved memory
for each IP's usage, such as load firmware to specific sapce,
so that need to reserve more regisions

Signed-off-by: Joe Liu <[email protected]>
Signed-off-by: YJ Chiang <[email protected]>
Signed-off-by: Alix Wu <[email protected]>
Signed-off-by: Phil Chang <[email protected]>
---
drivers/of/of_reserved_mem.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
index 46b9371c8a33..595f0741dcef 100644
--- a/drivers/of/of_reserved_mem.c
+++ b/drivers/of/of_reserved_mem.c
@@ -22,7 +22,7 @@
#include <linux/slab.h>
#include <linux/memblock.h>

-#define MAX_RESERVED_REGIONS 64
+#define MAX_RESERVED_REGIONS 128
static struct reserved_mem reserved_mem[MAX_RESERVED_REGIONS];
static int reserved_mem_count;

--
2.18.0


2020-10-04 13:07:01

by Chun-Kuang Hu

[permalink] [raw]
Subject: Re: [PATCH] [PATCH] of_reserved_mem: Increase the number of reserved regions

Hi, Phil:

Phil Chang <[email protected]> 於 2020年10月4日 週日 下午1:51寫道:
>
> Certain SoCs need to support large amount of reserved memory
> regions, especially to follow the GKI rules from Google.
> In MTK new SoC requires more than 68 regions of reserved memory
> for each IP's usage, such as load firmware to specific sapce,

space

> so that need to reserve more regisions

regions.

I guess this requirement is from Mediatek SoC, but I find below device
tree and just find one reserved memory region,

arch/arm64/boot/dts/mediatek/mt7622.dtsi
arch/arm64/boot/dts/mediatek/mt8173.dtsi
arch/arm64/boot/dts/mediatek/mt8516.dtsi

Could you show me the 68 regions?

Regards,
Chun-Kuang.

>
> Signed-off-by: Joe Liu <[email protected]>
> Signed-off-by: YJ Chiang <[email protected]>
> Signed-off-by: Alix Wu <[email protected]>
> Signed-off-by: Phil Chang <[email protected]>
> ---
> drivers/of/of_reserved_mem.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> index 46b9371c8a33..595f0741dcef 100644
> --- a/drivers/of/of_reserved_mem.c
> +++ b/drivers/of/of_reserved_mem.c
> @@ -22,7 +22,7 @@
> #include <linux/slab.h>
> #include <linux/memblock.h>
>
> -#define MAX_RESERVED_REGIONS 64
> +#define MAX_RESERVED_REGIONS 128
> static struct reserved_mem reserved_mem[MAX_RESERVED_REGIONS];
> static int reserved_mem_count;
>
> --
> 2.18.0
> _______________________________________________
> Linux-mediatek mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-mediatek

2020-10-05 13:48:46

by Rob Herring

[permalink] [raw]
Subject: Re: [PATCH] [PATCH] of_reserved_mem: Increase the number of reserved regions

On Sun, Oct 4, 2020 at 12:50 AM Phil Chang <[email protected]> wrote:
>
> Certain SoCs need to support large amount of reserved memory
> regions, especially to follow the GKI rules from Google.
> In MTK new SoC requires more than 68 regions of reserved memory
> for each IP's usage, such as load firmware to specific sapce,

typo

> so that need to reserve more regisions

typo. Missing punctuation.

>
> Signed-off-by: Joe Liu <[email protected]>
> Signed-off-by: YJ Chiang <[email protected]>
> Signed-off-by: Alix Wu <[email protected]>
> Signed-off-by: Phil Chang <[email protected]>
> ---
> drivers/of/of_reserved_mem.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> index 46b9371c8a33..595f0741dcef 100644
> --- a/drivers/of/of_reserved_mem.c
> +++ b/drivers/of/of_reserved_mem.c
> @@ -22,7 +22,7 @@
> #include <linux/slab.h>
> #include <linux/memblock.h>
>
> -#define MAX_RESERVED_REGIONS 64
> +#define MAX_RESERVED_REGIONS 128

At some point, this starts to feel like abuse of reserved regions.
Please provide details on what the regions are.

Also, this probably just needs to be dynamic. I think we're at that point.

Rob

2020-10-05 15:15:00

by Chun-Kuang Hu

[permalink] [raw]
Subject: Re: [PATCH] [PATCH] of_reserved_mem: Increase the number of reserved regions

Hi, Rob:

Rob Herring <[email protected]> 於 2020年10月5日 週一 下午9:45寫道:
>
> On Sun, Oct 4, 2020 at 12:50 AM Phil Chang <[email protected]> wrote:
> >
> > Certain SoCs need to support large amount of reserved memory
> > regions, especially to follow the GKI rules from Google.
> > In MTK new SoC requires more than 68 regions of reserved memory
> > for each IP's usage, such as load firmware to specific sapce,
>
> typo
>
> > so that need to reserve more regisions
>
> typo. Missing punctuation.
>
> >
> > Signed-off-by: Joe Liu <[email protected]>
> > Signed-off-by: YJ Chiang <[email protected]>
> > Signed-off-by: Alix Wu <[email protected]>
> > Signed-off-by: Phil Chang <[email protected]>
> > ---
> > drivers/of/of_reserved_mem.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> > index 46b9371c8a33..595f0741dcef 100644
> > --- a/drivers/of/of_reserved_mem.c
> > +++ b/drivers/of/of_reserved_mem.c
> > @@ -22,7 +22,7 @@
> > #include <linux/slab.h>
> > #include <linux/memblock.h>
> >
> > -#define MAX_RESERVED_REGIONS 64
> > +#define MAX_RESERVED_REGIONS 128
>
> At some point, this starts to feel like abuse of reserved regions.
> Please provide details on what the regions are.
>
> Also, this probably just needs to be dynamic. I think we're at that point.

How about using a config like DRM_FBDEV_OVERALLOC [1] ?

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/gpu/drm/Kconfig?h=v5.9-rc8#n125

Regards,
Chun-Kuang.

>
> Rob
>
> _______________________________________________
> Linux-mediatek mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-mediatek

2020-10-05 17:20:39

by Phil Chang

[permalink] [raw]
Subject: Re: [PATCH] [PATCH] of_reserved_mem: Increase the number of reserved regions

Hi Chun-Kuang

Sorry for typo. In fact, the dts of new SoC is not upstream yet. I'm so
sorry for couldn't show the detail now.

How about the configurable MAX_RESERVED_REGIONS size like this patch?
https://patchwork.kernel.org/patch/10692101/


On Sun, 2020-10-04 at 21:05 +0800, Chun-Kuang Hu wrote:
> Hi, Phil:
>
> Phil Chang <[email protected]> 於 2020年10月4日 週日 下午1:51寫道:
> >
> > Certain SoCs need to support large amount of reserved memory
> > regions, especially to follow the GKI rules from Google.
> > In MTK new SoC requires more than 68 regions of reserved memory
> > for each IP's usage, such as load firmware to specific sapce,
>
> space
>
> > so that need to reserve more regisions
>
> regions.
>
> I guess this requirement is from Mediatek SoC, but I find below device
> tree and just find one reserved memory region,
>
> arch/arm64/boot/dts/mediatek/mt7622.dtsi
> arch/arm64/boot/dts/mediatek/mt8173.dtsi
> arch/arm64/boot/dts/mediatek/mt8516.dtsi
>
> Could you show me the 68 regions?
>
> Regards,
> Chun-Kuang.
>
> >
> > Signed-off-by: Joe Liu <[email protected]>
> > Signed-off-by: YJ Chiang <[email protected]>
> > Signed-off-by: Alix Wu <[email protected]>
> > Signed-off-by: Phil Chang <[email protected]>
> > ---
> > drivers/of/of_reserved_mem.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> > index 46b9371c8a33..595f0741dcef 100644
> > --- a/drivers/of/of_reserved_mem.c
> > +++ b/drivers/of/of_reserved_mem.c
> > @@ -22,7 +22,7 @@
> > #include <linux/slab.h>
> > #include <linux/memblock.h>
> >
> > -#define MAX_RESERVED_REGIONS 64
> > +#define MAX_RESERVED_REGIONS 128
> > static struct reserved_mem reserved_mem[MAX_RESERVED_REGIONS];
> > static int reserved_mem_count;
> >
> > --
> > 2.18.0
> > _______________________________________________
> > Linux-mediatek mailing list
> > [email protected]
> > http://lists.infradead.org/mailman/listinfo/linux-mediatek

2020-10-05 21:11:22

by Rob Herring

[permalink] [raw]
Subject: Re: [PATCH] [PATCH] of_reserved_mem: Increase the number of reserved regions

On Mon, Oct 5, 2020 at 12:17 PM Phil Chang <[email protected]> wrote:
>
> Hi Chun-Kuang

Please don't top post to the lists.

> Sorry for typo. In fact, the dts of new SoC is not upstream yet. I'm so
> sorry for couldn't show the detail now.

Don't have to have the dts upstream. Can you point to a downstream dts
or post a snippet of the reserved memory?

> How about the configurable MAX_RESERVED_REGIONS size like this patch?
> https://patchwork.kernel.org/patch/10692101/

No, as I already said in that patch. But glad you found what's needed
to make it dynamic. But even for dynamic, I want to understand the
use.

Rob

>
>
> On Sun, 2020-10-04 at 21:05 +0800, Chun-Kuang Hu wrote:
> > Hi, Phil:
> >
> > Phil Chang <[email protected]> 於 2020年10月4日 週日 下午1:51寫道:
> > >
> > > Certain SoCs need to support large amount of reserved memory
> > > regions, especially to follow the GKI rules from Google.
> > > In MTK new SoC requires more than 68 regions of reserved memory
> > > for each IP's usage, such as load firmware to specific sapce,
> >
> > space
> >
> > > so that need to reserve more regisions
> >
> > regions.
> >
> > I guess this requirement is from Mediatek SoC, but I find below device
> > tree and just find one reserved memory region,
> >
> > arch/arm64/boot/dts/mediatek/mt7622.dtsi
> > arch/arm64/boot/dts/mediatek/mt8173.dtsi
> > arch/arm64/boot/dts/mediatek/mt8516.dtsi
> >
> > Could you show me the 68 regions?
> >
> > Regards,
> > Chun-Kuang.
> >
> > >
> > > Signed-off-by: Joe Liu <[email protected]>
> > > Signed-off-by: YJ Chiang <[email protected]>
> > > Signed-off-by: Alix Wu <[email protected]>
> > > Signed-off-by: Phil Chang <[email protected]>
> > > ---
> > > drivers/of/of_reserved_mem.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> > > index 46b9371c8a33..595f0741dcef 100644
> > > --- a/drivers/of/of_reserved_mem.c
> > > +++ b/drivers/of/of_reserved_mem.c
> > > @@ -22,7 +22,7 @@
> > > #include <linux/slab.h>
> > > #include <linux/memblock.h>
> > >
> > > -#define MAX_RESERVED_REGIONS 64
> > > +#define MAX_RESERVED_REGIONS 128
> > > static struct reserved_mem reserved_mem[MAX_RESERVED_REGIONS];
> > > static int reserved_mem_count;
> > >
> > > --
> > > 2.18.0
> > > _______________________________________________
> > > Linux-mediatek mailing list
> > > [email protected]
> > > http://lists.infradead.org/mailman/listinfo/linux-mediatek
>

2020-10-05 23:58:57

by Chun-Kuang Hu

[permalink] [raw]
Subject: Re: [PATCH] [PATCH] of_reserved_mem: Increase the number of reserved regions

Rob Herring <[email protected]> 於 2020年10月6日 週二 上午2:57寫道:
>
> On Mon, Oct 5, 2020 at 12:17 PM Phil Chang <[email protected]> wrote:
> >
> > Hi Chun-Kuang
>
> Please don't top post to the lists.
>
> > Sorry for typo. In fact, the dts of new SoC is not upstream yet. I'm so
> > sorry for couldn't show the detail now.
>
> Don't have to have the dts upstream. Can you point to a downstream dts
> or post a snippet of the reserved memory?
>
> > How about the configurable MAX_RESERVED_REGIONS size like this patch?
> > https://patchwork.kernel.org/patch/10692101/
>
> No, as I already said in that patch. But glad you found what's needed
> to make it dynamic. But even for dynamic, I want to understand the
> use.

I get the point. We should prevent too many config. If this
information could be get from dts, we should not get it from config.
Agree that WHY of this patch is more important than HOW of this patch.

Regards,
Chun-Kuang.

>
> Rob
>
> >
> >
> > On Sun, 2020-10-04 at 21:05 +0800, Chun-Kuang Hu wrote:
> > > Hi, Phil:
> > >
> > > Phil Chang <[email protected]> 於 2020年10月4日 週日 下午1:51寫道:
> > > >
> > > > Certain SoCs need to support large amount of reserved memory
> > > > regions, especially to follow the GKI rules from Google.
> > > > In MTK new SoC requires more than 68 regions of reserved memory
> > > > for each IP's usage, such as load firmware to specific sapce,
> > >
> > > space
> > >
> > > > so that need to reserve more regisions
> > >
> > > regions.
> > >
> > > I guess this requirement is from Mediatek SoC, but I find below device
> > > tree and just find one reserved memory region,
> > >
> > > arch/arm64/boot/dts/mediatek/mt7622.dtsi
> > > arch/arm64/boot/dts/mediatek/mt8173.dtsi
> > > arch/arm64/boot/dts/mediatek/mt8516.dtsi
> > >
> > > Could you show me the 68 regions?
> > >
> > > Regards,
> > > Chun-Kuang.
> > >
> > > >
> > > > Signed-off-by: Joe Liu <[email protected]>
> > > > Signed-off-by: YJ Chiang <[email protected]>
> > > > Signed-off-by: Alix Wu <[email protected]>
> > > > Signed-off-by: Phil Chang <[email protected]>
> > > > ---
> > > > drivers/of/of_reserved_mem.c | 2 +-
> > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/of/of_reserved_mem.c b/drivers/of/of_reserved_mem.c
> > > > index 46b9371c8a33..595f0741dcef 100644
> > > > --- a/drivers/of/of_reserved_mem.c
> > > > +++ b/drivers/of/of_reserved_mem.c
> > > > @@ -22,7 +22,7 @@
> > > > #include <linux/slab.h>
> > > > #include <linux/memblock.h>
> > > >
> > > > -#define MAX_RESERVED_REGIONS 64
> > > > +#define MAX_RESERVED_REGIONS 128
> > > > static struct reserved_mem reserved_mem[MAX_RESERVED_REGIONS];
> > > > static int reserved_mem_count;
> > > >
> > > > --
> > > > 2.18.0
> > > > _______________________________________________
> > > > Linux-mediatek mailing list
> > > > [email protected]
> > > > http://lists.infradead.org/mailman/listinfo/linux-mediatek
> >