#asahi-dev · 2021-05-27 19:05
kettenisproposed bindings have been submitted but probably still need a little bit of back and forth with the relevant maintainers#asahi-dev · 2021-05-27 19:04
ketteniswithin the linux framework that is; the hardware registers involved have been figured out#asahi-dev · 2021-05-27 19:04
kettenisstandard ecam, so the only thing that really needs to be figured out is how to do MSIs#asahi-dev · 2021-05-27 19:02
kettenissomebody just needs to write a simple linux apcie driver; that gives you ethernet + usb storage on the mini#asahi-dev · 2021-03-14 19:15
kettenisthanks for the quick fix#asahi-dev · 2021-03-14 19:07
kettenisalso pushed that patch to the apple-m1-m1n1 branch of my u-boot fork#asahi-dev · 2021-03-14 19:06
kettenispatch sent; let's see what they say#asahi-dev · 2021-03-14 18:49
ketteniswell, I suppose they're mostly concerned about linux,cma-default and linux,dma-default nodes#asahi-dev · 2021-03-14 18:42
kettenisI'll send them a patch#asahi-dev · 2021-03-14 18:41
kettenisu-boot is buggy#asahi-dev · 2021-03-14 18:40
kettenisAnd the "Memory Reservation Block and UEFI" sections is crystal clear#asahi-dev · 2021-03-14 18:40
kettenisWhich points back at the device tree spec ;)#asahi-dev · 2021-03-14 18:35
kettenisand the clarification it claims to be made in the EBBR is completely absent as far as I can tell#asahi-dev · 2021-03-14 18:30
kettenisthis was changed not too long ago and the change references the EBBR#asahi-dev · 2021-03-14 18:04
kettenisI don't necessarily disagree with you, but it seems that u-boot only considers entries under /reserved-memory that have the no-map property to be permanently off-limits#asahi-dev · 2021-03-14 17:52
kettenisneed to read some docs to see what is supposed to happen#asahi-dev · 2021-03-14 17:52
kettenisand kernel thinks it can re-use this block of memory#asahi-dev · 2021-03-14 17:51
kettenisu-boot ends up marking that memory as EfiBootServicesData#asahi-dev · 2021-03-14 17:51
kettenishmm, that doesn't quite work#asahi-dev · 2021-03-14 17:09
kettenisand when my kernel lands on top of it, the secondary cores don't spin up#asahi-dev · 2021-03-14 17:09
kettenisby the looks of it you reserve none of it#asahi-dev · 2021-03-14 15:26
kettenismarcan: seems the memory where the secondar CPUs are spinning isn't reserved?#asahi-dev · 2021-03-14 13:43
kettenisregarding my earlier message: the secondary CPUs defenitely spin up with the MMU (and caches) disabled#asahi-dev · 2021-03-13 19:30
kettenishmm, maybe I'm just misinterpreting what I'm seeing#asahi-dev · 2021-03-13 19:05
kettenishmm, the secundary CPUs spin up with the MMU enabled?#asahi-dev · 2021-03-12 22:05
kettenisand in this case the copy in the kernel tree should just be there for reference#asahi-dev · 2021-03-12 22:04
kettenisbut a well-designed device tree doesn't need to change#asahi-dev · 2021-03-12 22:03
kettenisunavoidable at this stage of development#asahi-dev · 2021-03-12 21:48
kettenisso what I did to test grub is to simply dd the debian debian-10.8.0-arm64-netinst.iso to a USB stick and plug that in#asahi-dev · 2021-03-12 21:45
kettenisthat one is out of date#asahi-dev · 2021-03-12 21:31
kettenisthat should be a convenient place#asahi-dev · 2021-03-12 21:31
kettenisok, put the device tree in the u-boot repo#asahi-dev · 2021-03-12 21:11
ketteniswith that hack loading a linux kernel and initrd seems to work#asahi-dev · 2021-03-12 21:10
kettenispushed a hack for that#asahi-dev · 2021-03-12 20:31
kettenisboots grub, but the grub hits a snag loading a linux kernel#asahi-dev · 2021-03-12 20:16
kettenisframebuffer is black on blueish because u-boot doesn't really support 30-bit color#asahi-dev · 2021-03-12 20:15
ketteniswithout the device tree diff you should still be able to get a prompt on serial and framebuffer#asahi-dev · 2021-03-12 20:10
ketteniscloning a full linux kernel tree for that is a bit excessive#asahi-dev · 2021-03-12 20:09
ketteniswhat would be a good way to share the updated t8103.dtsi?#asahi-dev · 2021-03-12 20:07
kettenisnote that u-boot uses an 115200 serial console#asahi-dev · 2021-03-12 20:06
kettenisbut potentially model-specific#asahi-dev · 2021-03-12 20:06
kettenisseems the tunables are not machine specific#asahi-dev · 2021-03-12 20:06
kettenisincluding the tunables#asahi-dev · 2021-03-12 20:05
kettenisI have a devicetree into which a transplanted the nodes I need to enable PCIe#asahi-dev · 2021-03-12 19:57
kettenishttps://github.com/kettenis/u-boot/tree/apple-m1-m1n1#asahi-dev · 2021-03-12 19:55
kettenisuploading u-boot seems to mostly work#asahi-dev · 2021-03-12 19:44
kettenisin other words, not a particular reliable setup#asahi-dev · 2021-03-12 19:38
kettenisand then there is a breakout of the PD signals that goes to another OpenBSD arm64 box with a hacked kernel that sends the vendor-specific PD commands to turn on the serial ;)#asahi-dev · 2021-03-12 19:37
kettenisthis is a USB serial dongle that connects to a breadboard with some resistors to connect to the m1#asahi-dev · 2021-03-12 19:36
kettenis(an odroid-n2, but I doubt that matters)