Allwinner H6 is a new SoC with Cortex-A53 cores from Allwinner, with its
memory map fully reworked and some high-speed peripherals (PCIe, USB
3.0) introduced.
This commit adds the basical DTSI file of it, including the clock
support and UART support.
Signed-off-by: Icenowy Zheng
Pine H64 is an Allwinner H6-based SBC from Pine64, with the following
features:
- 1GiB/2GiB/4GiB LPDDR3 DRAM (in 4GiB situation only 3GiB is
accessible)
- AXP805 PMIC
- Raspberry-Pi-compatible GPIO header, "Euler" GPIO header (not
compatible with the "Euler" on Pine A64) and "Expansion" pin
The Allwinner H6 SoC has a CCU which has been largely rearranged.
Add support for it in the sunxi-ng CCU framework.
Signed-off-by: Icenowy Zheng
---
Changes in v2:
- Exported APB1 bus clock for PIO.
- Switch to SPDX license identifier.
On the new Allwinner H6 SoC, multiple PLL's are NMP style clocks
(modelled as NKMP with no K) and have fixed post-dividers.
Add fixed post divider support to the NKMP style clocks.
Signed-off-by: Icenowy Zheng
---
No changes in v2.
drivers/clk/sunxi-ng/ccu_nkmp.c | 20
The Allwinner H6 SoC has two pin controllers, one main controller
(called CPUX-PORT in user manual) and one controller in CPUs power
domain (called CPUS-PORT in user manual).
This commit introduces support for the main pin controller on H6.
The pin bank A and B are not wired out and hidden from
The Allwinner H6 SoC have its pin controllers with the first IRQ-capable
GPIO bank at IRQ bank 1 and the second bank at IRQ bank 5. This
situation cannot be processed with the current pinctrl IRQ code, as it
only expects a offset to all IRQ banks.
Update the code to use a logical IRQ bank to
This patchset adds initial support for the Allwinner H6 SoC.
It's quite different from earlier Allwinner SoCs. For example, the
memory map is refactored, and the CCU is rearranged. It's also the first
Allwinner SoC with PCI Express interface, and the second one with USB
3.0 (the first one is
This reverts commit 7daa213700758b5b08fc0daab09bb139dd334165.
The original commit has several problems:
- vdd-cpus and aldo3 (AVCC of the SoC) are not set to always-on, which
leads to system hang when disabling unused regulators.
- GMAC (which uses dldo1 and aldo2) and Wi-Fi (which uses aldo1)
Hello!
First of all I would like to thank you all for your amazing contribution! More
I dive into understanding the kernel, more I appreciate your work and more I
learn.
As Marc Zyngier said:
"Now, if someone could try and run guests on this turd and report whether
this works correctly or