[RFC PATCH v2 0/8] accel/rocket: RK3576 NPU (RKNN) enablement

[Jiaxing Hu <gahing@gahingwoo.com>]

This adds RK3576 NPU support to accel/rocket.  RFC: the NPU brings up and
runs, but full inference is not correct yet, and I would rather get
feedback now than sit on it.

Tested on a Radxa ROCK 4D.  The NPU probes, powers on, brings up its
IOMMUs, and runs jobs to completion (rocket + Mesa Teflon).  Getting there
took a few fixes outside the driver too: a power-domain settle delay and
reset cycling (else a cold power-on throws an async SError), the full
clock set on the MMU nodes, clearing a stale firmware page fault, and both
NPU power domains on the core.

Not solved: only the first operation in a power session gives the right
answer.  Every later one engages and DMAs its input, but the compute
output never lands, so the readback is the zero-point.  I chased it
through the whole software surface (registers, IOMMU/TLB, cache,
power/clock/reset, dispatch) and it reproduces even with the driver
environment matched to the vendor's on the same kernel, so it looks to be
below software, in the fixed-function block.  Traces available.

Two asks: review of the bring-up and the shared pmdomain/iommu changes;
and if anyone knows the RK3576 NPU internals, whether that
per-power-session arm can be driven from software at all.

The series is split by subsystem (binding, pmdomain, iommu, driver, DT).
The board DT keeps vdd_npu_s0 always-on for now; happy to switch to
proper runtime control if preferred.

Changes in v2 (thanks to Chaoyi Chen and the Sashiko review bot):
  - dt-bindings: keep sram-supply required for rk3588 via allOf/if-then
    instead of dropping it from the shared required list
  - pmdomain: release the resets in rockchip_pm_remove_one_domain()
  - iommu: drop the redundant "err == 0" check in the clock probe
  - iommu: replace the two "skip orphaned-fault banks" patches with one
    that clears the stale firmware page fault (CMD_PAGE_FAULT_DONE) before
    enabling stall, so no bank needs special-casing -- Chaoyi asked
    whether dealing with the fault avoids the skip, and it does
  - accel/rocket: gate every RK3576-specific difference (extra clocks,
    single reset, multi power-domain attach, polled completion) behind
    of_device_id match data so the RK3588 path is byte-for-byte unchanged;
    drop the unused DPU register mapping (the driver did no pre-arm)
  - series is now 8 patches (was 9)

I used Claude Opus 4.8 to help trim this series out of my debugging tree
and generate the diffs.

Jiaxing Hu (8):
  dt-bindings: npu: rockchip: add rockchip,rk3576-rknn-core
  pmdomain/rockchip: add optional per-domain power-on settle delay
  pmdomain/rockchip: cycle optional power-domain resets on power-on
  iommu/rockchip: take all DT clocks
  iommu/rockchip: clear stale page faults before enabling stall
  accel/rocket: add RK3576 NPU (RKNN) support
  arm64: dts: rockchip: rk3576: add NPU (RKNN) nodes
  arm64: dts: rockchip: rk3576-rock-4d: enable NPU

 .../npu/rockchip,rk3588-rknn-core.yaml        |  12 +-
 .../boot/dts/rockchip/rk3576-rock-4d.dts      |  16 +++
 arch/arm64/boot/dts/rockchip/rk3576.dtsi      |  78 ++++++++++-
 drivers/accel/rocket/rocket_core.c            |  39 +++++-
 drivers/accel/rocket/rocket_core.h            |  22 ++-
 drivers/accel/rocket/rocket_device.c          |   4 +
 drivers/accel/rocket/rocket_drv.c             |  22 ++-
 drivers/accel/rocket/rocket_job.c             | 127 ++++++++++++++++--
 drivers/iommu/rockchip-iommu.c                |  41 +++---
 drivers/pmdomain/rockchip/pm-domains.c        |  69 ++++++----
 10 files changed, 371 insertions(+), 59 deletions(-)

-- 
2.43.0