On 26/11/14 10:59, Liviu Dudau wrote:

On Wed, Nov 26, 2014 at 10:30:24AM +0000, Marc Zyngier wrote:

quoted

Hi Liviu,

On 11/11/14 17:32, Liviu Dudau wrote:

quoted

This adds support for ARM's Juno development board (rev 0).
It enables most of the board peripherals: UART, I2C, USB, MMC and
100Mb ethernet. There is no support at the moment for clock setting
and HDLCD driver which depends on it.

Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
---
 arch/arm64/boot/dts/arm/Makefile              |   1 +
 arch/arm64/boot/dts/arm/juno-clocks.dtsi      |  45 ++++++
 arch/arm64/boot/dts/arm/juno-motherboard.dtsi | 129 +++++++++++++++
 arch/arm64/boot/dts/arm/juno.dts              | 218 ++++++++++++++++++++++++++
 4 files changed, 393 insertions(+)
 create mode 100644 arch/arm64/boot/dts/arm/juno-clocks.dtsi
 create mode 100644 arch/arm64/boot/dts/arm/juno-motherboard.dtsi
 create mode 100644 arch/arm64/boot/dts/arm/juno.dts

[...]

quoted

diff --git a/arch/arm64/boot/dts/arm/juno.dts b/arch/arm64/boot/dts/arm/juno.dts
new file mode 100644
index 0000000..097ecc4
--- /dev/null
+++ b/arch/arm64/boot/dts/arm/juno.dts

[...]

quoted

+       timer {
+               compatible = "arm,armv8-timer";
+               interrupts = <GIC_PPI 13 (GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_EDGE_RISING)>,
+                            <GIC_PPI 14 (GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_EDGE_RISING)>,
+                            <GIC_PPI 11 (GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_EDGE_RISING)>,
+                            <GIC_PPI 10 (GIC_CPU_MASK_SIMPLE(6) | IRQ_TYPE_EDGE_RISING)>;
+       };

Sorry, I should have spotted this earlier: all these interrupts are
*level*, not edge. This happens to work because the GIC's config
register is RO for PPIs on Juno, but still...

Yes, I've seen your other email to Suravee regarding interrupt triggering and I was trying to
find the relevant bits in Juno to tell me why I've put the info in DT this way. Maybe I *did*
copy-paste this part from FVP DT though.

I will send a patch to Olof to update.

Yeah, looks like most (if not all) DTs in the tree are broken. Any
chance you could write a sweeping patch to fix them all (at least for
the ARM implementations)?

The APM and Cavium implementations also carry the same values, and I
suppose this is a bug too, but someone with access to documentation
should figure this out.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...