ARM: dts: bcm5301x: Add TWD and I2C
cover ARM: dts: bcm5301x: Add TWD and I2C
These seemed to have been lost sometime last year. See
https://lkml.org/lkml/2016/5/11/953
I am dropping the syscon reboot from the series, as I believe that was the sticking point, and resending the TWD and i2c patches.
Jon Mason (2):
ARM: dts: bcm5301x: Add TWD WD Support to DT
ARM: dts: bcm5301x: Add I2C support to the DT
arch/arm/boot/dts/bcm5301x.dtsi | 24 +++++++++++++++++++++--- 1 file changed, 21 insertions(+), 3 deletions(-)
Signature · 1 line
2.7.4
PATCH 1/2 ARM: dts: bcm5301x: Add TWD WD Support to DT
From: Jon Mason <[email protected]>
Add support for the ARM TWD Watchdog to the bcm5301x device tree. The ARM TWD timer allocated the register space for the WDT, so this patch necessitated shrinking that. Also, the GIC masks were added for these.
--- arch/arm/boot/dts/bcm5301x.dtsi | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-)
arch/arm/boot/dts/bcm5301x.dtsi +12 -3
@@ -70,10 +70,19 @@ clocks = <&periph_clk>; }; - local-timer at 20600 { + timer at 20600 { compatible = "arm,cortex-a9-twd-timer"; - reg = <0x20600 0x100>; - interrupts = <GIC_PPI 13 IRQ_TYPE_LEVEL_HIGH>; + reg = <0x20600 0x20>; + interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(2) | + IRQ_TYPE_LEVEL_HIGH)>; + clocks = <&periph_clk>; + }; + + watchdog at 20620 { + compatible = "arm,cortex-a9-twd-wdt"; + reg = <0x20620 0x20>; + interrupts = <GIC_PPI 14 (GIC_CPU_MASK_SIMPLE(2) | + IRQ_TYPE_LEVEL_HIGH)>; clocks = <&periph_clk>; };
Signature · 1 line
2.7.4
Re: [PATCH 1/2] ARM: dts: bcm5301x: Add TWD WD Support to DT
arch/arm/boot/dts/bcm5301x.dtsi line 73
Quoted 6 lines · Jon Mason
arch/arm/boot/dts/bcm5301x.dtsi +1
-1@@ -70,10 +70,19 @@ clocks = <&periph_clk>; }; - local-timer at 20600 { + timer at 20600 {
Is there a reason for this rename? I don't really have preference for any name, it's just we shouldn't keep renaming this depending of personal preferences.
It also doesn't seem like adjusting to any schema:bcm63138.dtsi uses local-timer at 1e600 bcm-nsp.dtsi uses twd-timer at 20600 berlin2cd.dtsi uses local-timer at ad0600 berlin2.dtsi uses local-timer at ad0600 berlin2q.dtsi uses local-timer at ad0600
arch/arm/boot/dts/bcm5301x.dtsi line 79
Quoted 8 lines
arch/arm/boot/dts/bcm5301x.dtsi +5
-2compatible = "arm,cortex-a9-twd-timer"; - reg = <0x20600 0x100>; - interrupts = <GIC_PPI 13 IRQ_TYPE_LEVEL_HIGH>; + reg = <0x20600 0x20>; + interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(2) | + IRQ_TYPE_LEVEL_HIGH)>; + clocks = <&periph_clk>; + };
The rest looks OK, thanks! :)
Signature · 1 line
Rafa?
Re: [PATCH 1/2] ARM: dts: bcm5301x: Add TWD WD Support to DT
Quoted 12 lines · Rafa? Mi?ecki
arch/arm/boot/dts/bcm5301x.dtsi line 73
Quoted 6 lines · Jon Mason
arch/arm/boot/dts/bcm5301x.dtsi +1
-1@@ -70,10 +70,19 @@ clocks = <&periph_clk>; }; - local-timer at 20600 { + timer at 20600 {Is there a reason for this rename? I don't really have preference for any name, it's just we shouldn't keep renaming this depending of personal preferences.
Quoted 6 lines
It also doesn't seem like adjusting to any schema:
bcm63138.dtsiuses local-timer at 1e600bcm-nsp.dtsiuses twd-timer at 20600berlin2cd.dtsiuses local-timer at ad0600berlin2.dtsiuses local-timer at ad0600berlin2q.dtsiuses local-timer at ad0600
This was a suggestion from Sergei Shtylyov
http://lists.infradead.org/pipermail/linux-arm-kernel/2015-December/395160.html
Quoted 14 lines
arch/arm/boot/dts/bcm5301x.dtsi line 79
Quoted 9 lines
arch/arm/boot/dts/bcm5301x.dtsi +5
-2compatible = "arm,cortex-a9-twd-timer"; - reg = <0x20600 0x100>; - interrupts = <GIC_PPI 13 IRQ_TYPE_LEVEL_HIGH>; + reg = <0x20600 0x20>; + interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(2) | + IRQ_TYPE_LEVEL_HIGH)>; + clocks = <&periph_clk>; + };The rest looks OK, thanks! :)
--
Rafa?
Re: [PATCH 1/2] ARM: dts: bcm5301x: Add TWD WD Support to DT
Quoted 20 lines · Jon Mason
Quoted 12 lines · Rafa? Mi?ecki
arch/arm/boot/dts/bcm5301x.dtsi line 73
Quoted 6 lines · Jon Mason
arch/arm/boot/dts/bcm5301x.dtsi +1
-1@@ -70,10 +70,19 @@ clocks = <&periph_clk>; }; - local-timer at 20600 { + timer at 20600 {Is there a reason for this rename? I don't really have preference for any name, it's just we shouldn't keep renaming this depending of personal preferences.
Quoted 6 lines
It also doesn't seem like adjusting to any schema:
bcm63138.dtsiuses local-timer at 1e600bcm-nsp.dtsiuses twd-timer at 20600berlin2cd.dtsiuses local-timer at ad0600berlin2.dtsiuses local-timer at ad0600berlin2q.dtsiuses local-timer at ad0600
Quoted 2 lines
This was a suggestion from Sergei Shtylyov
http://lists.infradead.org/pipermail/linux-arm-kernel/2015-December/395160.html
Good to know, thanks!
Signature · 1 line
Rafa?
Re: [PATCH 1/2] ARM: dts: bcm5301x: Add TWD WD Support to DT
In reply to Jon MasonQuoted 19 lines · Jon Mason
From: Jon Mason <[email protected]>
Add support for the ARM TWD Watchdog to the bcm5301x device tree. The ARM TWD timer allocated the register space for the WDT, so this patch necessitated shrinking that. Also, the GIC masks were added for these.
Signed-off-by Jon Mason <[email protected]>--- arch/arm/boot/dts/bcm5301x.dtsi | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-)arch/arm/boot/dts/bcm5301x.dtsi +0
-0@@ -70,10 +70,19 @@ clocks = <&periph_clk>; };
arch/arm/boot/dts/bcm5301x.dtsi line 79
Quoted 10 lines
arch/arm/boot/dts/bcm5301x.dtsi +6
-3- local-timer at 20600 { + timer at 20600 { compatible = "arm,cortex-a9-twd-timer"; - reg = <0x20600 0x100>; - interrupts = <GIC_PPI 13 IRQ_TYPE_LEVEL_HIGH>; + reg = <0x20600 0x20>; + interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(2) | + IRQ_TYPE_LEVEL_HIGH)>; + clocks = <&periph_clk>; + };
If you follow my recent e-mail thread:
BCM5301X: GIC: PPI11 is secure or misconfigured (same for PPI13) you'll see IRQ_TYPE_LEVEL_HIGH type isn't correct. It should be IRQ_TYPE_EDGE_RISING.
I believe patch switching to IRQ_TYPE_EDGE_RISING should be sent with Cc stable for kernels 4.8+.
The same change is needed for "arm,cortex-a9-global-timer".
Would you find time to revise this patch?
Re: [PATCH 1/2] ARM: dts: bcm5301x: Add TWD WD Support to DT
Quoted 39 lines · Rafał Miłecki
Quoted 21 lines · Jon Mason
From: Jon Mason <[email protected]>
Add support for the ARM TWD Watchdog to the bcm5301x device tree. The ARM TWD timer allocated the register space for the WDT, so this patch necessitated shrinking that. Also, the GIC masks were added for these.
Signed-off-by Jon Mason <[email protected]>--- arch/arm/boot/dts/bcm5301x.dtsi | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-)arch/arm/boot/dts/bcm5301x.dtsi +0
-1@@ -70,10 +70,19 @@ clocks = <&periph_clk>; }; - local-timer@20600 {arch/arm/boot/dts/bcm5301x.dtsi line 80
Quoted 10 lines
arch/arm/boot/dts/bcm5301x.dtsi +6
-2+ timer@20600 { compatible = "arm,cortex-a9-twd-timer"; - reg = <0x20600 0x100>; - interrupts = <GIC_PPI 13 IRQ_TYPE_LEVEL_HIGH>; + reg = <0x20600 0x20>; + interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(2) | + IRQ_TYPE_LEVEL_HIGH)>; + clocks = <&periph_clk>; + };If you follow my recent e-mail thread:
BCM5301X: GIC: PPI11 is secure or misconfigured (same for PPI13) you'll seeIRQ_TYPE_LEVEL_HIGHtype isn't correct. It should beIRQ_TYPE_EDGE_RISING.
Quoted 8 lines
I believe patch switching to
IRQ_TYPE_EDGE_RISINGshould be sent with Cc stable
for kernels 4.8+.The same change is needed for "arm,cortex-a9-global-timer".
Would you find time to revise this patch?
I'll do 2 patches. One to revise this one and one to address the issue you've discovered. Will that be okay for you?
Re: [PATCH 1/2] ARM: dts: bcm5301x: Add TWD WD Support to DT
Quoted 46 lines · Jon Mason
Quoted 45 lines · Rafa? Mi?ecki
Quoted 29 lines · Jon Mason
From: Jon Mason <[email protected]>
Add support for the ARM TWD Watchdog to the bcm5301x device tree. The ARM TWD timer allocated the register space for the WDT, so this patch necessitated shrinking that. Also, the GIC masks were added for these.
Signed-off-by Jon Mason <[email protected]>--- arch/arm/boot/dts/bcm5301x.dtsi | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-)arch/arm/boot/dts/bcm5301x.dtsi +3
-3@@ -70,10 +70,19 @@ clocks = <&periph_clk>; }; - local-timer at 20600 { + timer at 20600 { compatible = "arm,cortex-a9-twd-timer"; - reg = <0x20600 0x100>; - interrupts = <GIC_PPI 13 IRQ_TYPE_LEVEL_HIGH>; + reg = <0x20600 0x20>; + interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(2)|
arch/arm/boot/dts/bcm5301x.dtsi line 79
Quoted 3 lines
arch/arm/boot/dts/bcm5301x.dtsi +3
-0+ IRQ_TYPE_LEVEL_HIGH)>; + clocks = <&periph_clk>; + };If you follow my recent e-mail thread:
BCM5301X: GIC: PPI11 is secure or misconfigured (same for PPI13) you'll seeIRQ_TYPE_LEVEL_HIGHtype isn't correct. It should beIRQ_TYPE_EDGE_RISING.I believe patch switching to
IRQ_TYPE_EDGE_RISINGshould be sent with Cc stable
for kernels 4.8+.
Quoted 7 lines
Quoted 3 lines
The same change is needed for "arm,cortex-a9-global-timer".
Would you find time to revise this patch?
I'll do 2 patches. One to revise this one and one to address the issue you've discovered. Will that be okay for you?
Sure. Please make sure you switch to IRQ_TYPE_EDGE_RISING before doing other changes, so it can be applied cleanly to the stable kernels.
Signature · 1 line
Rafa?
PATCH 2/2 ARM: dts: bcm5301x: Add I2C support to the DT
In reply to Jon MasonFrom: Jon Mason <[email protected]>
Add I2C support to the bcm5301x Device Tree. Since no driver changes are needed to enable this hardware, only the device tree changes are required to make this functional.
--- arch/arm/boot/dts/bcm5301x.dtsi | 9 +++++++++ 1 file changed, 9 insertions(+)
arch/arm/boot/dts/bcm5301x.dtsi +9 -0
@@ -338,6 +338,15 @@ }; }; + i2c0: i2c at 18009000 { + compatible = "brcm,iproc-i2c"; + reg = <0x18009000 0x50>; + interrupts = <GIC_SPI 121 IRQ_TYPE_NONE>; + #address-cells = <1>; + #size-cells = <0>; + clock-frequency = <100000>; + }; + lcpll0: lcpll0 at 1800c100 { #clock-cells = <1>; compatible = "brcm,nsp-lcpll0";
Signature · 1 line
2.7.4