#
…

2022-01-02

00:15Dcow_ has joined #asahi
00:21Dcow_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
00:21Dcow_ has joined #asahi
00:30jeffmiw has quit [Ping timeout: 480 seconds]
00:30jeffmiw has joined #asahi
00:43aleasto has quit [Quit: Konversation terminated!]
01:17wCPO has quit [Remote host closed the connection]
01:17wCPO has joined #asahi
01:19Dcow_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
01:23amwGlanzmann: I did find my MBA got very hot when I left it shut in my bag under Linux - even though it is barely warm at one with the lid open
01:24amwSo you were right
01:36Dcow_ has joined #asahi
02:17artemist0 has quit [Quit: artemist0]
02:17artemist has joined #asahi
02:51DarkShadow4444 has joined #asahi
02:51DarkShadow4444 has quit []
02:52DarkShadow4444 has joined #asahi
02:54DarkShadow4444 has quit []
02:58DarkShadow44 has quit [Ping timeout: 480 seconds]
03:02psykose has quit [Remote host closed the connection]
03:04psykose has joined #asahi
03:04psykose has quit [Remote host closed the connection]
03:05psykose has joined #asahi
03:06psykose has quit [Remote host closed the connection]
03:07psykose has joined #asahi
03:17DarkShadow44 has joined #asahi
03:23darkapex1 has joined #asahi
03:28darkapex has quit [Ping timeout: 480 seconds]
03:37darkapex2 has joined #asahi
03:42darkapex1 has quit [Ping timeout: 480 seconds]
03:42marcanI would not recommend using Linux in a "portable" scenario without shutting down the full system every time you store it, yet
03:43marcanmaking that work acceptably well needs more power management stuff that isn't done yet (CPU deep sleep, whole system sleep, both of which depend on the PSCI stuff)
03:44PhilippvK has joined #asahi
03:47phiologe has quit [Ping timeout: 480 seconds]
03:49jmr2 has joined #asahi
03:51jmr2Be especially cautious if you're using a tree that include the October DCP code. On shutdown, it does turns off the display, but the laptop isn't powered down, since shutdown isn't implemented yet.
03:51jmr2 has quit []
04:12marvin24 has joined #asahi
04:16marvin24_ has quit [Ping timeout: 480 seconds]
05:22burlmancer[m] has joined #asahi
05:29i509vcb[m] is now known as i509VCB
05:31i509VCB has quit [Quit: Reconnecting]
05:31i509VCB has joined #asahi
05:33i509VCB has quit []
05:33i509VCB has joined #asahi
07:02Dcow_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
07:05Dcow_ has joined #asahi
07:14vnogueira has quit [Ping timeout: 480 seconds]
07:19Dcow_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
07:54the_lanetly_052__ has joined #asahi
08:01the_lanetly_052__ has quit [Remote host closed the connection]
08:28ewrvvvvvvvvvvvd has joined #asahi
09:55n1c has joined #asahi
10:14jeffmiw has quit [Remote host closed the connection]
10:18jeffmiw has joined #asahi
10:24ewrvvvvvvvvvvvd has quit [Remote host closed the connection]
10:45aleasto has joined #asahi
10:51os0 has joined #asahi
10:51os has quit [Read error: Connection reset by peer]
10:59landscape15[m]marcan: Could it be possible to run Linux only with Icestorm cores as a power saving feature?
11:05amwPerhaps there's a magic register to toggle speed up or down - but I guess it's way more complex...
11:06j`eyamw: CPU freq is already worked out
11:06j`eyturning cores off, would require the PSCI thing, mwhich needs some figuring out
11:07amwj`ey: So it is a simple register write ... :-)
11:07j`eywell.. ish :P
11:08jannaulandscape15[m]: right now or when support has progressed more. dynamically or just during boot until it is powered down?
11:09amwI think if you turn cores off they lose all register / cache values and need to be re-started correctly - but perhaps changing frequency might be simpler
11:10landscape15[m]jannau: I mean dynamically. But it would also be good to support this with a “power saving” mode (as in macOS).
11:10amwIt's always simpler when you don't know what you are talking about :-D
11:11j`eyyou can turn cores of in linux with: echo 0 > /sys/devices/system/cpu/cpuN/online
11:11j`ey(wont work yet though)
11:12landscape15[m]j`ey: ok thanks. So it only needs power management optimization.
11:12amwAnything in /sys/devices/system/cpu/cpu0/power do anything yet ? wifi take5 code
11:15marcanwe actually don't know how to turn cores off, but we know how to put them into deep sleep, which should be the ~same
11:15marcanthey lose GPR register state except pc, sp, which means you need a little push/pop deal around the sleep
11:15marcanm1n1 already implements this
11:16marcanI don't think macOS attempts to actually power cores off at any time except when going into whole system sleep
11:16marcanmaybe not even then, logically speaking
11:17amwIs this done in m1n1/proxyclient/experiments/cpu_pstates.py ?
11:17marcanno, that's for cpufreq
11:17amwSo that's not deep enough sleep?
11:17marcancpufreq is not sleep
11:17marcancpufreq just changes the clock speed
11:18marcanit's a completely different thing
11:19amwdeep_wfi - is that the routine?
11:20marcanonce the PSCI stuff is implemented linux will call m1n1 (which will remain resident in RAM) to implement this, as well as full system sleep
11:20marcanso linux will never actually have this code itself
11:20amwok - like a bios or UEFI routine?
11:21marcanyes, similar mechanism to UEFI runtime services
11:21marcanbut unrelated, we will also have UEFI runtime services via u-boot
11:21marcanthe problem is this is normally done by a supervisor or hypervisor, but we don't have either of those on M1 (normally), so we need to invent/standardize a new method for doing it without crossing privilege levels
11:22amwWhy is it bad to just call that code directly in Linux - if it is so small
11:23marcanbecause every other ARM platform does it via PSCI and we don't want to make a mess of things
11:23j`eyamw: upstream wont accept it
11:23amwok - PSCI = Power State Coordination Interface (according google)
11:24marcanas a bonus, this gives us even *more* chances of things like sleep modes magically working on newer Apple CPUs without any kernel changes
11:24marcanpushing low level stuff into m1n1 is always a win from that point of view
11:26amwOK - but it does sound a little complex if we have m1n1 + u-boot remaining whilst also booting through grub - perhaps just go u-boot => linux?
11:27darkapex3 has joined #asahi
11:28mpsamw: you mean u-boot numeric menu
11:29mpsand extlinux.conf?
11:29amwmps: Just worried about how everyone doesn't step on the previous layer's code?
11:30amwPerhaps dtb points to reserved address that implements https://www.kernel.org/doc/Documentation/devicetree/bindings/arm/psci.txt
11:30mpsiiuc uboot and grub are removed from memory when they finish their job
11:32amwmps: In the boards I've worked I think so - but perhaps there is firmware routines that can be implemented by m1n1 + u-boot in a special reserved area denoted in the dtb that the kernel reads (from u-boot)
11:32darkapex2 has quit [Ping timeout: 480 seconds]
11:33amwok - according to https://www.thegoodpenguin.co.uk/blog/an-overview-of-psci/ this is new in 64bit ARM - i've only worked on 32bit
11:34j`eyarm32 can use PSCI too, but it's newer than a bunch of power related code in the arm32 kernel
11:40landscape15[m]Is it true that grub requires a separated partition and can’t be loaded as a payload with m1n1?
11:45mpstpw_rules: here is image of u-boot booting from nvme https://arvanta.net/2x3d1kfg/u-boot-m1.jpg
11:45mazit could, if you were to port grub to m1n1. but the amount of work required for that is absolutely silly, and that's exactly why we have an EFI layer in u-boot...
11:46mpsI tried with picocom attached to ttyACM1 on second machine but didn't get anything on it during boot
11:47mpskettenis: ^
11:49landscape15[m]maz: thanks. this means that you can install grub directly into the rootFS right?
11:51mazlandscape15[m]: grub needs to live somewhere where the EFI loader will find it, and that's in the EFI partition.
11:51mpsgrub on my usb disk is in rootFS partition
11:52mpson nvme, even if I put it in efi sys partition it can't be found
11:53mazmps: are you sure? you should have a file called bootaa64.efi in the EFI partition.
11:53mpsok, it is found but doesn't start for unknown to me reason, just puts 'me' in grub cli prompt, image is few msgs above
11:54mpsmaz: I have it
11:54mazmps: that's grub.
11:54landscape15[m]Well, running grub on USB seems easier (you can make an EFI partition or load the .efi from an EFI folder)
11:54mpsmaz: ok, that is one part of grub, rest is on rootFS
11:55mazmps: well, everything else is a module that can live anywhere wher grub can load it. the core grub has to live in the EFI partition.
11:56j`eymps: does your 'u-boot.cfg' has CONFIG_FS_FAT?
11:56mpsj`ey: I think so because it works fine from same setup on usb disk
11:56landscape15[m]maz: this means that you need an APFS Container, an ext4 partition, a FAT32 EFI partition. To me it seems a lot
11:56j`eyoh right, hm
11:57mazlandscape15[m]: a lot? have you ever seen an Android install, or a ChromeOS install?
11:57mpsif I break u-boot and set boot_order usb0 it works and boots from usb
11:58mazlandscape15[m]: and you want to consider some swap as well.
11:58j`eylandscape15[m]: having a EFI partition is normal
11:58os0 has quit [Ping timeout: 480 seconds]
11:58os has joined #asahi
11:59j`eyand an APFS container is a level higher than the partitions
11:59landscape15[m]Ok, but what if you have two or more Linux distributions installed…
11:59mazand it's not like partitioning is wasting much space either...
12:00j`eysame as with any other computer
12:00j`eyyou'd have more paritions
12:00mpsj`ey: APFS container look like it is ZFS, am I right?
12:00amwThe Quick Start Partitioning does say you want a FAT32 /boot + / ext4 : https://github.com/AsahiLinux/docs/wiki/Developer-Quickstart#step-1-partitioning
12:00j`eymps: I haven't used ZFS, so Im not sure
12:01mpsj`ey: also I didn't but that is what I concluded looking at it
12:01landscape15[m]amw: yes, but the EFI partition is not needed if Linux is loaded by a m1n1 > kernel payload
12:02j`eylandscape15[m]: m1n1+kernel payload is not easy to upgrade
12:02j`eythat's why you want m1n1+uboot, so the kernel can be upgraded without having to go through 1TR+kmutil
12:03amwj`ey: Can't we do m1n1+u-boot payload => ext4 / linux?
12:04j`eyamw: yes, but the point is to use the "normal" boot path from whatever distro, which probably involves grub
12:04j`eythe idea is to integrate with other distros, not make a new one!
12:04landscape15[m]j`ey: So the “standard” way will be m1n1 > uboot payload, which then search for grub, right?
12:04j`eylandscape15[m]: yes
12:04amwok - so debian et al won't support u-boot?
12:05j`eywell debian packages u-boot, but the intended path is to be more UEFI like
12:05mazamw: the standard way to use u-boot from a distro these days is to use u-boot's EFI interface. it makes things a lot easier for Debian (and everyone else), as nobody has to care about the various versions of u-boot and their level of brokenness across 2000 different platforms.
12:06landscape15[m]j`ey: thanks. Also Linux kernel updates will be easier and not dependent on the boot chain.
12:07j`eylandscape15[m]: exactly
12:07mazAPFS != APFS container.
12:07amwok - so mps problem is his 1. bad dtb and 2. grub doesn't understand the "default" boot partition
12:08mpsamw: dtb is good because it boots from usb
12:08amwWhat does "fdt_check_header(): FDT_ERR_BADMAGIC" mean?
12:10landscape15[m]mps: it would be amazing if Linux could be installed on an APFS volume :D
12:10mpsamw: yes, this msg is what I don't understand, and why it works with usb but not nvme, same m1n1+dtb+u-boot chain
12:11kettenisamw: that message is cause by the clunky way the u-boot bootscripts are written and it is harmless
12:11mpsj`ey: I recall that ZFS also have 'container' but I don't know what is it
12:11amwkettenis: ok
12:12kettenisthe script looks at $fdt_addr_r first just in case you loaded a dtb there
12:12kettenis(which is something you should never do on these systems)
12:13kettenisthis fails (and causes that message) and then the script falls back on the dtb passed by m1n1
12:13kettenismps: your jpg clearly shows grub running
12:14kettenisso that "error: unknown filesystem" is from grub
12:16landscape15[m]kettenis: why does it enter rescue mode?
12:16amwmps: I suppose you have no keyboard (even USB keyboard?) working to type commands into the grub rescue prompt?
12:17mpsamw: yes, keyboard works fine
12:17mpsexcept arrow keys
12:17amwCan you probe the disks it sees?
12:18mpsI mean, internal keyboard on mb pro
12:19mpsamw: I don't know grub cli commands, have to read some docs first (mostly use syslinux and u-boot on my machines for years)
12:20amwTry "set" and see what variables it sets?
12:21kettenislandscape15[m]: I don't know why grub is upset (and I don't really use grub as OpenBSD has its own EFI bootloader)
12:21ketteniscould be misconfiguration, or maybe it gets confused by the presence of the apple-specific partitions
12:22amwDoes "ls" see anything - perhaps it is defaulting to the wrong partition?
12:23mpslooks like it scans first 3 partitions (o, 1, 2) and then give up
12:23mazit looks like grub doesn't find its configuration file, or that the fonfig file doesn't match the system that grub sees (UUIDs, partition layout and such)
12:23amwgrub has a clunky (hdX,Y) syntax for partitions
12:23mpsand rootFS with grub config and modules are on partition 5
12:24amwDoes your grub support nvme devices? Can you do ls (hd0,5) or (hd1,5)
12:24mpsI have to boot chromebook and try again
12:25mpsamw: give me 5 minutes
12:25kettenisgrub doesn't need to care about if this is an nvme device or not
12:26amwIt kind of sounds like u-boot can see the device, loads grub which can't read any partitions and drops into rescue shell
12:26mazthere are *two* grub config files in most installs: one that is colocated with bootaa64.efi (used to botstrap the environment), and another one with the full config. obviously, the first has to know how to find the second one.
12:27kettenisbut even without a config file, grub should be able to recognize the partitions
12:28kettenisfun fact: the nvme uses 4k sectors
12:28mazyes, absolutely. it may not be able to use the FS though.
12:28amwUnder Linux my USB device is /dev/sda but my nvme is nvme0n1 - quite different driver?
12:28mpsamw: ls in grub shell shows all partitions
12:29amwThen you should be able to manual load the kernel / initrd
12:31mpsset shows this 'prefix=(hd1,gpt2)/boot/grub'
12:31amwDoes ls (hd1,gpt2)/boot/grub work?
12:32mpsamw: no
12:32kettenisgpt2 is almost certainly the APFS partition
12:32amwI think that is where grub will look for it's files (and kernel)
12:33amwPerhaps try changing to the ext4 or what ever linux partition?
12:33amwe.g. set prefix=(hdX,Y)/boot/grub sort of command
12:34mpsamw: what is grub command to try boot after changing prefix
12:35amwThere's alot of other stuff there that looks useful too
12:35mpshmm, uncrecognized
12:35amwDid the "ls /" show anything?
12:36amwIn particular those notes say "set prefix=(hdX,Y)/boot/grub => se the actual location of the grub folder"
12:36mpsamw: yes, 'unknown filesysystem'
12:37mpsoh, there is root variable
12:37amwTry ls (hd1,gpt5) - if your linux on 5th partition?
12:39mpsgot 'filesystem is ext2'
12:40mpssomething is missing with grub, help and boot are unrecognized
12:43amwmps: That's better than unknown filesystem :-}
12:44amwIf your grub is on the EFI - shouldn't the partition it wants be FAT32 ?
12:44amwIs that (hd1,gpt1) ?
12:44kettenis(hd1,gpt4) would be my guess
12:44mpsamw: no, grub is gpt5
12:45mpsefi is gpt4
12:46kettenisanyway, you should be able to list files on those partitions with
12:46mpskettenis: I tried to put complete grub on efi partition (gpt4) but still didn't worked
12:46kettenisls (h1,gpt4)/
12:46kettenisls (hd1,gtp5)/
12:47amwI'm afraid I have to go to bed - those notes say grub rescue> => grub can't find the grub folder
12:47mpsls (hd1,gpt4) => filesystem unknown
12:48mpsls (hd1,gpt5) => ext2
12:48mpsok, good night
12:48ketteniswell, if (hd1,gpt4) really is your efi filesystem, grub doesn't recognize it as such
12:49kettenisso either the partition type is wrong, or the fat32 filesystem on there is busted
12:50kettenisor you have a grub that doesn't include fat32 support, which would be silly
12:50mpshm, I tried to copy all this to usb and it works fine
12:51eroux has joined #asahi
12:51mpsand opposite, copied all from usb disk and it doesn't find /boot/grub on gpt5
13:04eroux has quit [Ping timeout: 480 seconds]
13:06marcanlandscape15[m]: GPT can handle up to 128 partitions, so I'm really not worried about typical users running out of partitions by doing multiple linux installs
13:07marcanbesides, "traditional" linux is at least root, boot, swap; here we'd have stub, EFI, root, maybe swap (no need for root these days) so it's not that different
13:07marcan*for boot
13:08marcanamw: subsequent stages do not step over each other if they shouldn't; there are multiple reserved memory mechanisms depending on how you're booting
13:08marcanm1n1 is actually *required* to stay in RAM for us to implement sleep, due to how the CPU RVBAR gets locked down
13:09marcanit is a hard requirement of the platform as it exists today that the computer will wake up at m1n1's base address, the way I have addresses aligned right now
13:09marcanso there has to be *something* there if nothing else
13:10marcanthat doesn't mean m1n1 needs to stay entirely resident; I'll probably constrain the PSCI code to not use any dynamic memory, so we can throw away the heap; that would keep the memory wastage to under a couple MBs max, and we could shrink it further if we really want to with some section trickery
13:15kettenisfwiw, u-boot does section trickery and has a PSCI implementation for some boards that don't have TF-A that fits in a few pages
13:15kettenispeanuts compared to what apple already steals for some of the firmware blobs ;)
13:20kettenis(that PSCI implementation wouldn't be very useful to us though)
13:21marcanmakes sense, yeah
13:21kettenisthe efi runtime in u-boot uses a similar mechanism
13:22marcanand we certainly could fit the PSCI code in a few kBs if we want, though that would involve forgoing all the m1n1 infra and mostly making it a separate codebase, which may or may not be what we want
13:22marcan(e.g. leaving the UART iodev active, at least if debug was active, is quite legitimate for debugging suspend/resume and will largely play fine with the kernel driver)
13:22kettenismaking it a separate codebase would probably be what I'd do
13:23marcandepends on how much machinery ends up going in there
13:23kettenisbut it would make it easier to audit the code
13:23marcanyeah, but it's not like there's much of an attack surface here, and m1n1 is the root of trust for our ecosystem anyway
13:24marcane.g. there's a chance we end up having to talk to SPMI or something for sleep, and if it comes to that I'm not sure I want to start over without our usual scaffolding :p
13:24marcanbut either way, this is all relatively easy to refactor post facto; m1n1 is small enough it's easy to do major surgery on it
13:25marcanso unless it turns out the PSCI code will be basically freestanding once we figure out what needs to go there (i.e. one C file), I'll probably just start by leaving all of m1n1 resident
13:25marcanand then if we want to slim it down in the future we certainly can
13:29NightRaven[m]is there any easy way to get a wifi dongle working ?
13:32mpsNightRaven[m]: which dongle, do you know its usb id
13:32GlanzmannNightRaven[m]: wifi already works.
13:32NightRaven[m]i have a realtek 811
13:32GlanzmannNightRaven[m]: And USB works, too. So you probably only have to plug it in and compile in the right modules into the kernel.
13:33GlanzmannNightRaven[m]: Why don't you use the wifi from marcan?
13:33NightRaven[m]i don't have a second machine
13:33NightRaven[m]to enable the firmware stuff
13:33GlanzmannNightRaven[m]: Oh, I see.
13:33mpsNightRaven[m]: type 'lsusb' and look at ID column
13:34GlanzmannNightRaven[m]: THan I would use mps kernel config and than add the kernel module for you wifi stick.
13:34NightRaven[m]Glanzmann: but how would you add a kernel module ?
13:34GlanzmannNightRaven[m]: Do you have a usb ethernet card?
13:35NightRaven[m]Glanzmann: i don't
13:35NightRaven[m]only a wifi dongle
13:35GlanzmannNightRaven[m]: Buy one, makes your life easier.
13:35NightRaven[m]hmm i will but for now i want try to get the dongle working so i build the kernel using mps's config
13:36mpsNightRaven[m]: tell me what lsusb says for your dongle and I will which module you need
13:36mpss/will/will look/
13:36NightRaven[m]```Bus 001 Device 003: ID 0bda:818b Realtek Semiconductor Corp. RTL8192EU 802.11b/g/n WLAN Adapter```
13:37mpsNightRaven[m]: ok
13:37NightRaven[m]Glanzmann: i already have a usb c adapter i can get this
13:39mpsNightRaven[m]: hmm, kernel have only RTL8192U but not RTL8192EU
13:40mpsNightRaven[m]: ah, you are lucky 'drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_core.c: * The RTL8723BU/RTL8192EU vendor driver use coexistence table type'
13:41NightRaven[m]so i will build the kernel with your config is this your config wget https://tg.st/u/.config
13:42eroux has joined #asahi
13:46___nick___ has joined #asahi
13:46___nick___ has quit []
13:47___nick___ has joined #asahi
13:48GlanzmannNightRaven[m]: Yes, that is mps config. And than you have to add the additional module.
13:48GlanzmannAlso don't forget about the firmware.
13:48GlanzmannNightRaven[m]: What Linux Distribution do you use?
13:48NightRaven[m]i want to use arch
13:48mpsNightRaven[m]: I think it is CONFIG_RTL8XXXU in kernel config, you have to enable this
13:49NightRaven[m]mps: there is not CONFIG_RTL8XXXU in the .config file
13:49NightRaven[m]or do i have to add it ?
13:51mpsNightRaven[m]: run 'make nconfig' or 'make menuconfig' and go to drivers/networking menu and look there
13:51NightRaven[m]oh ok thx i willl post any problems i have i will try it now
14:09landscape15[m]<mps> "NightRaven: type 'lsusb' and..." <- Do you need to do that also for a usb-c to Ethernet dongle? So that I can enable (if present) that card in my config.
14:11NightRaven[m]do i just press enter or do i have to add some stuff ?
14:11NightRaven[m]like enable on this part
14:13mpslandscape15[m]: yes, you need enable drivers for your hardware if they are not already enabled
14:25NightRaven[m]<mps> "NightRaven: run 'make nconfig..." <- i have found the realtek folder in drivers/net/ but it shows a makefile and some C files
14:28mpsNightRaven[m]: do you have any experience with building kernels
14:29NightRaven[m]mps: nope i have no experience building kernels
14:30mpsNightRaven[m]: if you want to build it my advice is to first read some guides about this, there are many on the internet
14:31NightRaven[m]Do you have any recommendations ?
14:31mpsor, I can build new version and upload somewhere
14:31NightRaven[m]mps: sure if that is fine with you
14:32mpsNightRaven[m]: not really, I learned this things in '92-93 years and didn't looked after this time
14:32landscape15[m]mps: thanks. However, internal Wi-Fi (with config enabled) should work OOB I suppose
14:33mpsNightRaven[m]: ok, will do when I finish some things with one arm32 SBC, in about hour or two
14:33landscape15[m]NightRaven: I always keep my Linux pocket guide in my jacket :)
14:33NightRaven[m]landscape15[m]: might get that : )
14:33NightRaven[m]mps: oh ok i am not in a rush take your time with it : )
14:34mpsI'll enable all usb wifis
14:46vnogueira has joined #asahi
14:54mpsI made a break of my hobby with arm32
14:56landscape15[m]mps: thanks. I didn’t try it yet. But I’m also considering to build mine, to get basic Wi-Fi stuff working.
14:58kettenisreworked the keyboard mapping in u-boot a bit to make the arrow keys work
14:59mpskettenis: thanks
14:59kettenistested with grub from the debian installer
15:00mpskettenis: it is OT question, but do you know how to increase character size in u-boot
15:00kettenisbooted from usb, but it does recognize my efi and openbsd partitions just fine
15:01mpsI tried with setting ttf font but got only blank screen (always used u-boot with serial consoles till now)
15:04NightRaven[m]<mps> "NightRaven: landscape15[m..." <- ok i will use that let me just payload that and boot from the .bin file
15:06kettenismps: no I have no idea
15:06mpsNightRaven[m]: you have kernel and dtbs (and config) in boot dir when you unpack tar file
15:07mpsI hope you have m1n1.macho already
15:07mpskettenis: np, will try to find on the net
15:10NightRaven[m]mps: wait so i have the .bin file
15:10NightRaven[m]but where do i put the stuff inside the tar file ?
15:11riker77 has quit [Quit: Quitting IRC - gone for good...]
15:14mpsNightRaven[m]: best is to mount you target rootFS somewhere (/mnt for example) and do 'tar xf linux-5.16.0-rc5-a15.tar.gz -C /mnt'
15:23kettenisfwiw, I just chainloaded the OpenBSD EFI bootloader from the grub that is part of the debian installer ISO
15:23tpw_rulesmps: that looks like grub's fault. u-boot loaded it fine
15:24mpstpw_rules: yes
15:25mpstpw_rules: only strange thing is it works fine from usb disk
15:25tpw_rulesis the USB disk plugged in at this point?
15:27tpw_rulesdid you compile grub yourself? mine has a whole bunch of modules.
15:27mpsdoesn't matter, same happens with it plugged or not
15:27tpw_rulesok, i've had weird problems before when i have a usb with the same filesystem UUIDs as the nvme plugged in
15:29mpslooks like i need to look at grub build options
15:30kettenistpw_rules: UUIDs that aren't quite unique? ;)
15:30tpw_rulesalso also nixos keeps its kernel and initrd on the ESP so it doesn't technically need to access ext4. not sure how your distro does it
15:30tpw_ruleskettenis: blame the abhorrent practice of SD card images!
15:30kettenisI'll gladly do that!
15:31kettenisoh, interesting
15:31kettenisDebian actually uses SHIM on aarch64
15:33tpw_rulesbtw random arcana: when i made my ESP, i had to pass `-s 1` to mkfs.vfat to stop it complaining about the data structure size on a 512MiB ESP. maybe if you don't it creates a not quite standard partition grub isn't happy with. the nvme is 4k hard sectored and your USB presumably isn't
15:38mpstpw_rules: I created both partitions with 'mkfs.fat -F 32'
15:38mpss/created/made FS/
15:40j`eykettenis: not anymore
15:40mps(time for lunch, will be afk)
16:00riker77 has joined #asahi
16:04NightRaven[m]Mps I have done the tar xf on the /mnt but now I get this
16:04NightRaven[m]I will try again
16:16landscape15[m]NightRaven: I used rsync to copy the rootFS onto the nvme partition. It worked but you may need to repair it with e2fsck
16:19kettenistpw_rules: applied the usb timer thingy to my branch
16:24aleasto has quit [Quit: Konversation terminated!]
16:25aleasto has joined #asahi
16:30NightRaven[m]<landscape15[m]> "NightRaven: I used rsync to copy..." <- Hmm I have formatted multiple times I have done mkfs.ext4 /dev/sda then I tar -xf in a folder with the rootFS the sudo cp -r * /mnt Then I did mp's command
16:47os5 has joined #asahi
16:47os has quit [Ping timeout: 480 seconds]
16:48bradfier has quit [Quit: Leaving...]
16:49bradfier has joined #asahi
16:51NightRaven[m]Ok I get the same problem with my other HDD could it be the image.gz ?
16:58landscape15[m]NightRaven: why don’t you simply cp -r from the extracted rootFS to the root USB device?
16:58NightRaven[m]That's what I do
16:59NightRaven[m]Oops I missed a sentence
16:59NightRaven[m]I am on my phone sorry I am not the best at typing at it
17:00NightRaven[m]But I have tried my different HDD and same problem so I am going to try a different image.gz
17:00landscape15[m]NightRaven: yeah it could be the kernel broken
17:03NightRaven[m]Same message
17:04landscape15[m]With my config I get keyboard and nvme working, but USB doesn’t work. Which options should I enable?
17:04NightRaven[m]I used Glanzmann's one
17:04NightRaven[m]I am not sure
17:05NightRaven[m]<NightRaven[m]> "so i will build the kernel..." <- Try
17:05NightRaven[m]That . config
17:06landscape15[m]NightRaven[m]: mmh, try to build it yourself with debootstrap. In my case it worked.
17:06NightRaven[m]With debootstrap ?
17:07landscape15[m]NightRaven: yeah it’s a tool that let you build a Linux rootFS with basic tools. Then under chroot you can install everything you want.
17:08NightRaven[m]Oh ok I will try that
17:08landscape15[m]NightRaven: ok let me know
17:09NightRaven[m]If it does not work I am just going to post a picture and wait for mps or someone
17:17landscape15[m]<landscape15[m]> "With my config I get keyboard..." <- I’m comparing my config with mps’s one to find why USB doesn’t work. I enabled all APPLE configs, but still doesn’t work.
17:25zimsneexh has quit [Read error: Connection reset by peer]
17:33tpw_rulesmps: maybe try with mkfs.vfat -F 32 -s 1 . how big is your ESP?
17:34tpw_ruleskettenis: cool. i saw your review email too, thank you! ( i received it twice though? )
17:34kettenisyeah I mistyped one of the addresses in the CC's a added
17:35kettenisso I sent it again
17:37tpw_rulesi assume you've made the internal keyboard poll not take 40ms
17:42kettenissomething around 1ms I think
17:43tpw_rulesok. i still think the 5us poll timer is kind of absurd but 1ms should only delay for a second or so i imagine.
17:44___nick___ has quit []
17:45___nick___ has joined #asahi
17:55mpstpw_rules: not sure this will help, looks like grub stop scanning partitions on number 2, i.e. scans 0, 1 and 2 and then stops
17:56NightRaven[m]yea i keep getting this
17:56NightRaven[m]i am not sure now
17:56NightRaven[m] * i am not sure why
17:58tpw_rulesmps: well partition 3 might be a not quite valid FAT
17:58tpw_rulesif it's the ESP
17:58tpw_rulesit should be a simple matter to copy files off the ESP then reformat it right
17:58landscape15[m]NightRaven: try to keep the default mount point for your USB and cp -rf directly to it.
17:58NightRaven[m]so ```sudo cp -rf * /mnt/debian```
17:59os has joined #asahi
17:59NightRaven[m]* is the rootFS
18:00mpstpw_rules: right, it is simple task, I can try this
18:00os5 has quit [Read error: Connection reset by peer]
18:01landscape15[m]NightRaven: so yes. But you should add the entire rootFS path, not only `*`
18:02NightRaven[m]See what i do is i put all the rootFS into a folder at home and then i cd into the folder and do ```sudo cp -rf * /mnt/debian``` i do ```*``` for everything in that folder
18:02NightRaven[m]so the rootFS
18:04landscape15[m]NightRaven: ok then. It should work. But have you moved the rootFS files with cp to that folder? Permissions could be broken.
18:05NightRaven[m]but i use sudo should i go to root
18:05mpsNightRaven[m]: from your image above looks like your root FS is broken
18:06landscape15[m]NightRaven: if you used sudo then no
18:06NightRaven[m]mps: but i do ```sudo mkfs.ext4 /dev/sda```
18:06NightRaven[m]i am cleaning the hdd
18:06NightRaven[m]then mounting it and putting the rootFS on too it
18:06NightRaven[m]<mps> "NightRaven: you have kernel..." <- and then your files
18:07mpswhy didn't you created partition tables and partitions on your disk
18:08NightRaven[m]i mean this worked before so i thought it will work
18:08NightRaven[m]i did not know you need it but it worked without it
18:09mpsit could work but is unusual
18:10mpsif you could 'follow' shell script I created to install alpine I can post it to use it as hints for your distro
18:10NightRaven[m]i am going to try landscape's sudo cp -rf
18:10NightRaven[m]instead of -r
18:10NightRaven[m]mps: sure i will try to change some stuff
18:12mpsit is for alpine and didn't tested much
18:12mpsbut basic things are there and if you adapt it for your use case it should work
18:12NightRaven[m]hmm i will take a look at it after the restart i do
18:13mpswhen we get SMC and wifi working without second machine I will write short howto about this
18:14mpsyes, this one is also useful
18:14tpw_rulesare there two people here with the same favorite distro? :)
18:15mpstpw_rules: :)
18:16kettenisand they make jokes about the BSDs...
18:16mpskettenis: I like BSD but don't use on workstations
18:16NightRaven[m]tpw_rules: lol i will take a look at it and try some stuff
18:19os has quit [Read error: Connection reset by peer]
18:24tpw_rulesif you like, you don't have to use that to install to the internal nvme. you can load the nixos installer to ram then reformat your usb flash drive
18:25NightRaven[m]Hmm yeah I might try nix os but never heard of it
18:27NightRaven[m]Btw mps I am not sure if it is the rootFS, Image.gz or .bin file but I am going to try something
18:28mpsNightRaven[m]: from your image it is seen that kernel trying to run /sbin/init but can't to run it for unknown reason
18:29mpssorry, /etc/init
18:29mpsit clearly says that it mounted sda as root fs
18:30NightRaven[m]Yea I think the linux.tar.gz corrupted it
18:31mpsno, it can't find init
18:34NightRaven[m]I FOUND THE ISSUE
18:36NightRaven[m]<mps> "NightRaven: best is to mount you..." <- it was your tar.gz when I did tar xf linux.tar.gz to my mount point
18:36NightRaven[m]The command broke it.
18:37eroux has quit [Ping timeout: 480 seconds]
18:37NightRaven[m]Hmm ok Then how do I use your . tar.gz
18:42mps'tar xf linux......tar.gz -C /rootFS_mountpoint' where you mounted your rootFS
18:43NightRaven[m]And that is why it was not wkeko
18:43mpslook at script which I posted how it creates payload.macho
18:45NightRaven[m]Oh so basically the .tar.gz Is like image.gz ?
18:46mpsNightRaven[m]: no, it contains modules kernel and dtbs
18:47NightRaven[m]I know how to make a .bin/macho file
18:47mpswhen you untar it properly it install drivers in proper directory on rootFS
18:47NightRaven[m]It was the command that messed it up
18:47eroux has joined #asahi
18:50NightRaven[m]i found the issue
19:54ewrvvvvvvvvvvvvvvvvvvvvvvvvvvd has joined #asahi
19:57n1c has quit [Quit: ZNC 1.8.2+deb1+focal2 - https://znc.in]
19:58n1c has joined #asahi
20:02n1c has quit []
20:04DarkShadow44 has quit [Quit: ZNC - https://znc.in]
20:04n1c has joined #asahi
20:16DarkShadow44 has joined #asahi
20:23mps_ has joined #asahi
20:26landscape15[m]How can I get a lower resolution?
20:26j`eyin menuconfig you can pick a bigger FONT
20:26j`eyassuming thats just the framebuffer console
20:27NightRaven[m]mps: I might just wait till it's easier to get the internal wifi working
20:27mps has quit [Ping timeout: 480 seconds]
20:32landscape15[m]j`ey: thanks. Does menuconfig automatically generate .config? Because then I have to use —set-value to change some configs.
20:33j`eylandscape15[m]: yes menuconfig reads your current .config, and then makes the modifications
20:34mps_NightRaven[m]: in next day I will work to make install u-boot on usb stick, it will be then simpler
20:34mps_if I find time, ofc
20:34mps_s/day/days/
20:37Dcow_ has joined #asahi
20:37eroux has quit [Ping timeout: 480 seconds]
20:38landscape15[m]j`ey: ok thanks
20:39Dcow_ has quit []
20:39aleasto has quit [Quit: Konversation terminated!]
20:40Dcow_ has joined #asahi
20:41j`eyi dont thinkn that ./script/config thing is very good, i dont think it sets dependencies?
20:46landscape15[m]j`ey: well, it only changes CONFIG values. But I can also manually change them (using a text editor).
20:46j`eyright, so you can maybe make some unsupported configuration
20:48landscape15[m]j`ey: yes. For this reason I can’t make USB working. I don’t find the right configs.
20:48j`eymaybe using menuconfig will help
20:50eroux has joined #asahi
20:50landscape15[m]But why does nvme work only by enabling the required config?
20:50landscape15[m]Maybe there are no dependences for that
20:51j`eywhat did you have to enable for that?
20:51j`eyBLK_DEV_NVME_APPLE is default ARCH_APPLE, so should have already been on
20:53landscape15[m]j`ey: In my case it was disabled. I also enabled APPLE_PROPERTIES
20:55landscape15[m]btw I’m browsing the .config with menuconfig
21:00eroux has quit [Ping timeout: 480 seconds]
21:02j`eyI just ran defconfig, BLK_DEV_NVME_APPLE was on by default
21:07os has joined #asahi
21:08___nick___ has quit [Ping timeout: 480 seconds]
21:13Dcow_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
21:16crabbedhaloablut has quit []
21:16crabbedhaloablut has joined #asahi
21:35chengsun_ has joined #asahi
21:42chengsun has quit [Ping timeout: 480 seconds]
21:55chengsun has joined #asahi
22:00chengsun_ has quit [Ping timeout: 480 seconds]
22:34XeR has quit []
22:40landscape15[m]I managed to run it with USB working using mps’s config. Is it possible to switch to cli once you are at GUI login screen.
22:40j`eytry ctrl-alt-f2.. but I just realised you have an mbp.. so not sure
22:41landscape15[m]j`ey: I can use an external keyboard
22:44XeR has joined #asahi
22:47Dcow_ has joined #asahi
22:50XeR has quit []
22:51mps_landscape15[m]: you can stop your display manager
22:51XeR has joined #asahi
22:52mps_that is how I switch between console and xorg
22:53mps_ is now known as mps
22:55eroux has joined #asahi
22:59thickrocks has quit [Remote host closed the connection]
23:00yrlf has quit [Quit: The Lounge - https://thelounge.chat]
23:03eroux has quit [Ping timeout: 480 seconds]
23:13landscape15[m]@oftc_mps:matrix.org: ok, I’ll get a keyboard and try. Thanks.
23:54eroux has joined #asahi
…