* [PATCH 1/4] arm64: dts: amlogic: s7: fix the GIC CPU interface region size
2026-10-22 18:30 [PATCH 0/4] Fix and complete the GIC description of Amlogic S7 and S7D Carlo Caione
@ 2026-10-22 18:30 ` Carlo Caione
2026-10-22 18:30 ` [PATCH 2/4] arm64: dts: amlogic: s7d: " Carlo Caione
` (3 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Carlo Caione @ 2026-10-22 18:30 UTC (permalink / raw)
To: Neil Armstrong, Kevin Hilman, Jerome Brunet, Martin Blumenstingl,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Xianwei Zhao
Cc: linux-arm-kernel, linux-amlogic, devicetree, linux-kernel, Carlo Caione
The CPU interface of the GIC-400 spans 8 KiB, but the S7 device tree
describes only its first 256 bytes. The GIC driver then disables the
split priority drop and deactivation mode, which needs the GICC_DIR
register in the second 4 KiB page, and warns at boot:
[ 0.000000] GIC: GICv2 detected, but range too small and irqchip.gicv2_force_probe not set
Describe the whole 8 KiB region, as the other Amlogic SoCs with a
GIC-400 at the same address do.
Fixes: 1b753fcfcff8 ("arm64: dts: add support for S7 based Amlogic BP201")
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
---
arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi b/arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi
index 64440df4f542..89b75bec40f0 100644
--- a/arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi
+++ b/arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi
@@ -121,7 +121,7 @@ gic: interrupt-controller@fff01000 {
#address-cells = <0>;
interrupt-controller;
reg = <0x0 0xfff01000 0 0x1000>,
- <0x0 0xfff02000 0 0x0100>;
+ <0x0 0xfff02000 0 0x2000>;
interrupts = <GIC_PPI 9 0xf04>;
};
--
2.56.0
_______________________________________________
linux-amlogic mailing list
linux-amlogic@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-amlogic
^ permalink raw reply [flat|nested] 6+ messages in thread* [PATCH 2/4] arm64: dts: amlogic: s7d: fix the GIC CPU interface region size
2026-10-22 18:30 [PATCH 0/4] Fix and complete the GIC description of Amlogic S7 and S7D Carlo Caione
2026-10-22 18:30 ` [PATCH 1/4] arm64: dts: amlogic: s7: fix the GIC CPU interface region size Carlo Caione
@ 2026-10-22 18:30 ` Carlo Caione
2026-10-22 18:30 ` [PATCH 3/4] arm64: dts: amlogic: s7: describe the GIC virtualization regions Carlo Caione
` (2 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Carlo Caione @ 2026-10-22 18:30 UTC (permalink / raw)
To: Neil Armstrong, Kevin Hilman, Jerome Brunet, Martin Blumenstingl,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Xianwei Zhao
Cc: linux-arm-kernel, linux-amlogic, devicetree, linux-kernel, Carlo Caione
The CPU interface of the GIC-400 spans 8 KiB, but the S7D device tree
describes only its first 256 bytes. The GIC driver then disables the
split priority drop and deactivation mode, which needs the GICC_DIR
register in the second 4 KiB page, and warns at boot that the range is
too small.
Describe the whole 8 KiB region, as the other Amlogic SoCs with a
GIC-400 at the same address do.
Fixes: 1a30661c20de ("arm64: dts: add support for S7D based Amlogic BM202")
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
---
arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi b/arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi
index 5593ff9e07eb..059a1bf37584 100644
--- a/arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi
+++ b/arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi
@@ -85,7 +85,7 @@ gic: interrupt-controller@fff01000 {
#address-cells = <0>;
interrupt-controller;
reg = <0x0 0xfff01000 0 0x1000>,
- <0x0 0xfff02000 0 0x0100>;
+ <0x0 0xfff02000 0 0x2000>;
interrupts = <GIC_PPI 9 0xf04>;
};
--
2.56.0
_______________________________________________
linux-amlogic mailing list
linux-amlogic@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-amlogic
^ permalink raw reply [flat|nested] 6+ messages in thread* [PATCH 3/4] arm64: dts: amlogic: s7: describe the GIC virtualization regions
2026-10-22 18:30 [PATCH 0/4] Fix and complete the GIC description of Amlogic S7 and S7D Carlo Caione
2026-10-22 18:30 ` [PATCH 1/4] arm64: dts: amlogic: s7: fix the GIC CPU interface region size Carlo Caione
2026-10-22 18:30 ` [PATCH 2/4] arm64: dts: amlogic: s7d: " Carlo Caione
@ 2026-10-22 18:30 ` Carlo Caione
2026-10-22 18:30 ` [PATCH 4/4] arm64: dts: amlogic: s7d: " Carlo Caione
2026-10-23 7:23 ` [PATCH 0/4] Fix and complete the GIC description of Amlogic S7 and S7D Neil Armstrong
4 siblings, 0 replies; 6+ messages in thread
From: Carlo Caione @ 2026-10-22 18:30 UTC (permalink / raw)
To: Neil Armstrong, Kevin Hilman, Jerome Brunet, Martin Blumenstingl,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Xianwei Zhao
Cc: linux-arm-kernel, linux-amlogic, devicetree, linux-kernel, Carlo Caione
The S7 device tree describes the distributor and the CPU interface of
the GIC-400, but not its virtual interface control (GICH) and virtual
CPU interface (GICV) regions. Without them the GIC driver does not hand
the GIC over to KVM, which then runs without the in-kernel vGIC.
Describe both regions at their GIC-400 offsets, as S4 does for its
GIC-400 at the same address.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
---
arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi b/arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi
index 89b75bec40f0..a337783c7783 100644
--- a/arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi
+++ b/arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi
@@ -121,7 +121,9 @@ gic: interrupt-controller@fff01000 {
#address-cells = <0>;
interrupt-controller;
reg = <0x0 0xfff01000 0 0x1000>,
- <0x0 0xfff02000 0 0x2000>;
+ <0x0 0xfff02000 0 0x2000>,
+ <0x0 0xfff04000 0 0x2000>,
+ <0x0 0xfff06000 0 0x2000>;
interrupts = <GIC_PPI 9 0xf04>;
};
--
2.56.0
_______________________________________________
linux-amlogic mailing list
linux-amlogic@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-amlogic
^ permalink raw reply [flat|nested] 6+ messages in thread* [PATCH 4/4] arm64: dts: amlogic: s7d: describe the GIC virtualization regions
2026-10-22 18:30 [PATCH 0/4] Fix and complete the GIC description of Amlogic S7 and S7D Carlo Caione
` (2 preceding siblings ...)
2026-10-22 18:30 ` [PATCH 3/4] arm64: dts: amlogic: s7: describe the GIC virtualization regions Carlo Caione
@ 2026-10-22 18:30 ` Carlo Caione
2026-10-23 7:23 ` [PATCH 0/4] Fix and complete the GIC description of Amlogic S7 and S7D Neil Armstrong
4 siblings, 0 replies; 6+ messages in thread
From: Carlo Caione @ 2026-10-22 18:30 UTC (permalink / raw)
To: Neil Armstrong, Kevin Hilman, Jerome Brunet, Martin Blumenstingl,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Xianwei Zhao
Cc: linux-arm-kernel, linux-amlogic, devicetree, linux-kernel, Carlo Caione
The S7D device tree describes the distributor and the CPU interface of
the GIC-400, but not its virtual interface control (GICH) and virtual
CPU interface (GICV) regions. Without them the GIC driver does not hand
the GIC over to KVM, which then runs without the in-kernel vGIC.
Describe both regions at their GIC-400 offsets, as S4 does for its
GIC-400 at the same address.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
---
arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi b/arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi
index 059a1bf37584..b77ac8898d16 100644
--- a/arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi
+++ b/arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi
@@ -85,7 +85,9 @@ gic: interrupt-controller@fff01000 {
#address-cells = <0>;
interrupt-controller;
reg = <0x0 0xfff01000 0 0x1000>,
- <0x0 0xfff02000 0 0x2000>;
+ <0x0 0xfff02000 0 0x2000>,
+ <0x0 0xfff04000 0 0x2000>,
+ <0x0 0xfff06000 0 0x2000>;
interrupts = <GIC_PPI 9 0xf04>;
};
--
2.56.0
_______________________________________________
linux-amlogic mailing list
linux-amlogic@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-amlogic
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH 0/4] Fix and complete the GIC description of Amlogic S7 and S7D
2026-10-22 18:30 [PATCH 0/4] Fix and complete the GIC description of Amlogic S7 and S7D Carlo Caione
` (3 preceding siblings ...)
2026-10-22 18:30 ` [PATCH 4/4] arm64: dts: amlogic: s7d: " Carlo Caione
@ 2026-10-23 7:23 ` Neil Armstrong
4 siblings, 0 replies; 6+ messages in thread
From: Neil Armstrong @ 2026-10-23 7:23 UTC (permalink / raw)
To: Carlo Caione, Kevin Hilman, Jerome Brunet, Martin Blumenstingl,
Rob Herring, Krzysztof Kozlowski, Conor Dooley, Xianwei Zhao
Cc: linux-arm-kernel, linux-amlogic, devicetree, linux-kernel
On 10/22/26 20:30, Carlo Caione wrote:
> The S7 and S7D device trees describe only the first 256 bytes of the
> GIC-400 CPU interface, which spans 8 KiB, and leave out its virtual
> interface control (GICH) and virtual CPU interface (GICV) regions.
>
> With the short CPU interface, the GIC driver disables the split priority
> drop and deactivation mode and warns at boot (patches 1 and 2). Without
> GICH and GICV, it cannot hand the GIC over to KVM (patches 3 and 4). It
> needs the split mode for that too, so KVM gets its in-kernel vGIC only
> with the whole series.
>
> Signed-off-by: Carlo Caione <ccaione@baylibre.com>
> ---
> Carlo Caione (4):
> arm64: dts: amlogic: s7: fix the GIC CPU interface region size
> arm64: dts: amlogic: s7d: fix the GIC CPU interface region size
> arm64: dts: amlogic: s7: describe the GIC virtualization regions
> arm64: dts: amlogic: s7d: describe the GIC virtualization regions
>
> arch/arm64/boot/dts/amlogic/amlogic-s7.dtsi | 4 +++-
> arch/arm64/boot/dts/amlogic/amlogic-s7d.dtsi | 4 +++-
> 2 files changed, 6 insertions(+), 2 deletions(-)
> ---
> base-commit: 8610d31587e139380db9cb8708f4263833b482d0
> change-id: 20261022-upstream-s7-s7d-gicc-size-db94cb656d36
>
> Best regards,
> --
> Carlo Caione <ccaione@baylibre.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Thanks,
Neil
_______________________________________________
linux-amlogic mailing list
linux-amlogic@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-amlogic
^ permalink raw reply [flat|nested] 6+ messages in thread