On Wed, May 4, 2022 at 7:55 PM Rex-BC Chen <rex-bc.chen@xxxxxxxxxxxx> wrote: > > This series is for supporting devapc implementation in MT8186. > > V2: > - Add a patch to separate register offsets from mtk_devapc_data. > > V1: > - Add dt-binding and add devapc data for MT8186. > > Rex-BC Chen (3): > dt-bindings: soc: mediatek: devapc: Add bindings for MT8186 > soc: mediatek: devapc: Separate register offsets from mtk_devapc_data > soc: mediatek: devapc: Add support for MT8186 Applied this on next-20220602 (with a few fixes for other section mismatch errors), booted and got: [ 1.948483] mtk-devapc 10207000.devapc: Read Violation [ 1.948488] mtk-devapc 10207000.devapc: Bus ID:0x100, Dom ID:0x0, Vio Addr:0x13000000 [ 1.948520] mtk-devapc 10207000.devapc: Read Violation [ 1.948523] mtk-devapc 10207000.devapc: Bus ID:0x200, Dom ID:0x0, Vio Addr:0x140001a0 [ 1.948537] mtk-devapc 10207000.devapc: Read Violation [ 1.948540] mtk-devapc 10207000.devapc: Bus ID:0x200, Dom ID:0x0, Vio Addr:0x140001a0 [ 1.948555] mtk-devapc 10207000.devapc: Read Violation [ 1.948558] mtk-devapc 10207000.devapc: Bus ID:0x200, Dom ID:0x0, Vio Addr:0x14000100 [ 1.948573] mtk-devapc 10207000.devapc: Read Violation [ 1.948576] mtk-devapc 10207000.devapc: Bus ID:0x200, Dom ID:0x0, Vio Addr:0x14000100 [ 1.948590] mtk-devapc 10207000.devapc: Read Violation [ 1.948593] mtk-devapc 10207000.devapc: Bus ID:0x200, Dom ID:0x0, Vio Addr:0x14000100 [ 1.948607] mtk-devapc 10207000.devapc: Read Violation [ 1.948610] mtk-devapc 10207000.devapc: Bus ID:0x200, Dom ID:0x0, Vio Addr:0x14000100 [ 1.948624] mtk-devapc 10207000.devapc: Read Violation [ 1.948627] mtk-devapc 10207000.devapc: Bus ID:0x200, Dom ID:0x0, Vio Addr:0x14000100 during the boot process. So I think this works well. Manually reading a known secure address also triggers it: root@hayato:~# busybox devmem 0x1000e000 0x00000000 [ 135.069121] mtk-devapc 10207000.devapc: Read Violation [ 135.069132] mtk-devapc 10207000.devapc: Bus ID:0x482, Dom ID:0x0, Vio Addr:0x1000e000 So, Tested-by: Chen-Yu Tsai <wenst@xxxxxxxxxxxx> BTW, looks like MT8186 support in mainline is shaping up real good. SCP firmware is still missing, so video codec stuff won't work.
