#asahi-dev · 2021-09-21 11:30
kettenispower-domain-name would be most logical, but that is dangerously close to power-domain-names that is used in the client interface#asahi-dev · 2021-09-21 10:06
kettenisI suppose you want something unique to name the domain#asahi-dev · 2021-09-21 09:55
kettenismarcan: also, the domain-output-names property is a bit strange#asahi-dev · 2021-09-20 16:17
kettenisthe "apple,t8103-pmgr-pstate" compatible is only there as a safety in case the hardware isn't as generic as we think it is#asahi-dev · 2021-09-20 16:16
kettenismarcan: I think your pmgr driver should only match on "apple,pmgr-pstate"#asahi-dev · 2021-09-20 15:08
kettenisheh, then forget what I said#asahi-dev · 2021-09-20 14:57
kettenisyou might be able to get away with keeping the clock stuff and just adding the power-domain stuff if you want to have a DT that works with older kernels as well#asahi-dev · 2021-09-20 12:28
kettenisthis also means that "clock" properties of devices such as NVMe and PCIe need to be replaced with appropriate "power-domains" properties#asahi-dev · 2021-09-20 12:27
kettenisyes#asahi-dev · 2021-09-19 19:31
kettenislooks like u-boot handles powerdomains automagically is well#asahi-dev · 2021-09-11 18:10
kettenisI mean, hotplug in m1n1 isn't really essential#asahi-dev · 2021-09-11 18:10
kettenisalyssa: the linux driver is where this matters isn't it?#asahi-dev · 2021-08-26 10:48
kettenisoh, this is where folks are hanging out today#asahi-dev · 2021-07-13 08:24
kettenisbut this seems to argue for dma-coherent properties on the usb, nvme and pcie "devices" instead of putting a single one on the /soc simple-bus node#asahi-dev · 2021-07-13 08:22
kettenisI think I've seen SoCs with both coherent and incoherent devices on the same bus#asahi-dev · 2021-07-13 08:20
kettenisarnd: hmm, the device tree specification talks about devices instead of busses#asahi-dev · 2021-07-12 17:11
kettenisthat one tends to go on the root node though#asahi-dev · 2021-07-12 17:10
kettenisit probably makes sense to move the "interrupt-parent" property up as well to avoid repeating it for every device#asahi-dev · 2021-07-12 17:06
kettenisso we can probably put it on the "soc" node next to the "nonposted-mmio" property#asahi-dev · 2021-07-12 16:42
ketteniswhich means we need to add som "dma-coherent" properties to the device tree I think#asahi-dev · 2021-07-12 16:42
kettenisso as far as I can tell all devices that do DMA are cache coherent#asahi-dev · 2021-07-12 15:05
kettenisAnd I doubt all that infrastructure runs on MacOS ;)#asahi-dev · 2021-07-12 15:04
kettenisApple runs a lot of infrastructure and they might be interested in running that on ARM platforms#asahi-dev · 2021-06-18 15:32
kettenisnot clear to me if tony's voice carries any weight in that discussion though#asahi-dev · 2021-06-18 15:29
ketteniswe realy need "approved" clock bindings such that we can iron this out properly#asahi-dev · 2021-06-18 15:29
kettenisand u-boot may rely on on m1n1 enabling some of the clocks#asahi-dev · 2021-06-18 15:28
kettenisright now, the kernel may rely on u-boot enabling some of the clocks#asahi-dev · 2021-06-18 15:27
kettenistypically this means a clock hans't been enabled#asahi-dev · 2021-06-12 13:27
kettenisit is obviously running a custom firmware#asahi-dev · 2021-06-12 13:27
kettenisjannau: several other registers have different length as well#asahi-dev · 2021-06-11 19:44
kettenisis that 45 decimal?#asahi-dev · 2021-05-30 09:47
kettenisit'll trust whatever m1n1 passes as the framebuffer#asahi-dev · 2021-05-30 09:43
kettenis#define CONFIG_SYS_MALLOC_LEN SZ_64M#asahi-dev · 2021-05-30 09:42
kettenissven: you could try bumping the following #define in include/configs/apple.h#asahi-dev · 2021-05-29 18:58
kettenisI have two arm boards, a breadboard, a breakoutboard and two usb-to-serial converters (one acting as a power supply) just to get serial out of the mini#asahi-dev · 2021-05-29 18:57
kettenismuch cleaner than mine ;)#asahi-dev · 2021-05-29 17:49
kettenisthe presence of all those apple partitions seems to confuse the code that autosizes the partitions#asahi-dev · 2021-05-29 17:49
kettenisok, teaching the OpenBSD/arm64 installer about GPT partitions helped, but something went wrong as I ended up with 2G /home#asahi-dev · 2021-05-29 17:37
kettenisour device tree doesn't do this because frankly it is just confusing#asahi-dev · 2021-05-29 17:37
kettenisthe /arm-io node to be specefic#asahi-dev · 2021-05-29 17:36
kettenisso the ranges property of onr of the toplevel nodes defines an address-space translation just like it does for normal device trees#asahi-dev · 2021-05-29 17:32
kettenisranges#asahi-dev · 2021-05-29 17:25
kettenisbloom: biggest gotcha is that 64-bit addresses/sizes are "little endian" instead of "big endian"#asahi-dev · 2021-05-29 16:27
kettenisI simply commented out the reset_input_buffer() method#asahi-dev · 2021-05-29 16:25
ketteniswell, I have rk3399 boards as well#asahi-dev · 2021-05-29 16:25
kettenisno, pserial 3.4 issues#asahi-dev · 2021-05-29 16:25
kettenismarcan: me ;)#asahi-dev · 2021-05-27 19:07
kettenisheh, I wasn't complaining ;)#asahi-dev · 2021-05-27 19:06
ketteniss/set/said/#asahi-dev · 2021-05-27 19:06
kettenisthat set, getting clocks and dart upstreamed should probably be the priority