00:00
<
samueldr >
uh legacy bootable in gpt, bootable in mbr
00:00
<
samueldr >
also make sure your u-boot can do usb
00:00
<
cole-h >
How would I make sure it can read from my partition? Any easy way to test if it supports ZFS? Or am I gonna have to go for ext4?
00:02
<
samueldr >
noneucat: what is the best method to test and confirm crust works, in your opinion?
00:02
<
samueldr >
it almost looks like crust doesn't need any more downstream patches for the pinephone
00:03
<
samueldr >
all of the patches in pine64's tree seems to target h5 and h6 AFAICT
00:04
<
samueldr >
>> U-Boot: upstream support for loading Crust into SRAM was merged in commit 18261b855223, and is present in all releases starting with v2021.01-rc1.
00:04
<
samueldr >
it's in the readme of the project
00:04
* samueldr
deletes patch
00:05
<
samueldr >
also explains how it went from ~17 patches to 0
00:05
dstzd has joined #nixos-aarch64
00:11
dstzd has joined #nixos-aarch64
00:17
dstzd has joined #nixos-aarch64
00:18
<
samueldr >
I'm too lazy to assemble a working sd card manually, so I started an image cross-compilation for examples/hello
00:20
<
cole-h >
samueldr: Does it matter if I do gpt vs mbr?
00:21
dstzd has quit [Client Quit]
00:22
dstzd has joined #nixos-aarch64
00:25
dstzd has quit [Client Quit]
00:38
<
samueldr >
yes and no
00:38
<
samueldr >
it matters because you need to know how to do the legacy boot flag in gpt
00:38
<
samueldr >
but it should work
00:38
<
samueldr >
u-boot knows about gpt
00:41
<
flokli >
samueldr: any opinions on more cross-NixOS stuff in cache.nixos.org?
00:41
<
samueldr >
yes please?
00:41
<
flokli >
It seems nice garbage collected again
00:41
<
flokli >
s/nice/nix/
00:41
<
flokli >
In my Workstation
00:41
<
samueldr >
I think it's more about ressources than anything else
00:42
<
samueldr >
though if we do it, we probably want a specific jobset, and maybe do the trick like with mobile nixos
00:42
<
flokli >
damn autocomplete
00:42
<
samueldr >
where it only makes a new eval on stable updates
00:42
<
samueldr >
not sure it's
*that* needed
00:42
<
samueldr >
if it can handle on unstable updates, which should be daily, it should be fine otherwise
00:43
<
flokli >
I mean, we have parts of the cross toolchain already there anyways
00:43
<
samueldr >
(should be daily... the dream of a healthy channel)
00:43
<
samueldr >
yes, but only as a side-effect of mobile nixos
00:43
<
flokli >
I wonder how much more a minimal sd image toplevel would be
00:43
<
samueldr >
but yes, that's less impact
00:43
<
samueldr >
probably negligeible considering what mobile nixos builds I think
00:44
<
samueldr >
so yeah, I already had a quick discussion about that in the past, that it would be nice to have
00:44
<
samueldr >
now it's about actually doing it I guess
00:44
<
flokli >
so we need to expose this somewhere as an attribute, and create a jobset or other hydra voodoo for it?
00:44
<
samueldr >
figuring out if the nixos.org foundation thinks it's okay to do it ;)
00:44
<
samueldr >
I would argue new jobset
00:45
<
samueldr >
where not only the sd image is exposed, but... more?
00:45
<
samueldr >
I don't know what, or even if we could do all-packages?
00:46
<
flokli >
oof, I assume a lot of these would currently fail, and waste some build time in realizing this over and over again
00:47
<
flokli >
something like unstableSmall, but cross could be a good starter
00:47
<
samueldr >
what would be really helpful is having all the attributes used in the sd image exposed in the jobset
00:47
<
samueldr >
*and* a manually triggered jobset with all-packages
00:47
<
samueldr >
so we can send it off, look at what's the status
00:49
<
flokli >
I mean, the more and more mobile-nixos focus shifts to the userland components it'd become interesting anyways
00:49
<
flokli >
maybe it can also just be some more attributes in that jobset
00:50
<
flokli >
I'm not sure. Maybe you can just ask the foundation in what they think?
00:55
<
samueldr >
I don't really have any more direct access than you do
01:02
dstzd has joined #nixos-aarch64
01:07
dstzd has quit [Client Quit]
01:07
dstzd has joined #nixos-aarch64
01:13
rajivr has joined #nixos-aarch64
01:17
dstzd has joined #nixos-aarch64
01:20
dstzd has quit [Client Quit]
01:20
dstzd has joined #nixos-aarch64
01:21
<
cole-h >
Meh, I'm satisfied with having my rock64 living on the SD card (for now).
01:21
<
cole-h >
Maybe I'll investigate further in the future (read: once this SD card inexplicably dies).
01:21
<
cole-h >
Thanks for your pointers samueldr++.
01:21
<
{^_^} >
samueldr's karma got increased to -2147483648
01:40
dstzd has joined #nixos-aarch64
01:42
dstzd has quit [Client Quit]
01:43
dstzd has joined #nixos-aarch64
01:44
dstzd has quit [Remote host closed the connection]
01:45
dstzd has joined #nixos-aarch64
01:49
dstzd has quit [Client Quit]
02:03
<
{^_^} >
mobile-nixos#283 (by samueldr, 9 minutes ago, open): pine64-pinephone: u-boot: 2020.07 -> 2021.01
02:06
h0m1 has quit [Ping timeout: 272 seconds]
02:07
h0m1 has joined #nixos-aarch64
02:07
dstzd has joined #nixos-aarch64
03:13
<
pbb >
I love how we can now boot all 64-bit capable Raspberry Pis with the sd-image-aarch64-new-kernel
03:14
<
samueldr >
mainline now works?
03:15
<
samueldr >
at least enough so that a user can rebuild with the pi foundation kernel, or build without u-boot?
03:15
<
samueldr >
[on the pi 4]
03:15
<
samueldr >
relatedly: nobody tested the raspberry pi 2 with the 64 bit chip
03:15
<
samueldr >
it
*should* work, AFAIUI it's just like a pi 3, with less RAM
03:17
<
samueldr >
I should hit the canakit peeps with a mail to ask if they have that pi 2 variant on hand
03:23
<
hexa- >
rpi 2 can't do 64 bit, can it?
03:23
<
samueldr >
not all of them
03:24
<
samueldr >
the later revision used another CPU
03:24
<
samueldr >
officially speaking it's treated as a 32 bit board still
03:26
<
samueldr >
and it can boot in 64 bit mode^[citation needed]
03:49
<
pbb >
samueldr: mainline worked fine when i tested your uboot update pr
03:54
<
pbb >
When the next hydra evaluation runs I'll post the link to the sd-image-aarch64-new-kernel artifact in that issue about booting mainline kernels
04:00
<
samueldr >
cool, I wasn't aware someone checked
04:00
<
samueldr >
that means we can delete the pi4-specific image
04:01
<
samueldr >
as long as someone can
*then* build their own image, there is no need for a non-mainline image
04:12
orivej has quit [Ping timeout: 256 seconds]
04:34
<
samueldr >
huh, booting the aarch64 image with aarch64 emulation on qemu now works just fine
04:34
<
samueldr >
and trivially
04:35
<
samueldr >
it's slow (as expected)
05:08
srk has quit [Remote host closed the connection]
05:09
srk has joined #nixos-aarch64
05:23
craige[m]1 has joined #nixos-aarch64
05:25
<
Ke >
sphalerite: since you are further along now, what available modules/config did you use?
05:28
<
craige[m]1 >
Thanks for your 2020 wrap up samueldr !
05:28
<
samueldr >
thanks for reading it
05:29
<
craige[m]1 >
All the way to the RSS recommendation, which is exactly how I digest it 🤣
07:05
dstzd has joined #nixos-aarch64
07:09
dstzd has quit [Client Quit]
07:10
orivej has joined #nixos-aarch64
07:21
orivej has quit [Ping timeout: 264 seconds]
07:22
orivej has joined #nixos-aarch64
07:26
dstzd has joined #nixos-aarch64
07:28
dstzd has quit [Client Quit]
07:29
dstzd has joined #nixos-aarch64
07:31
dstzd has quit [Client Quit]
07:43
dstzd has joined #nixos-aarch64
07:44
dstzd has quit [Client Quit]
07:47
cole-h has quit [Quit: zzz]
07:53
<
Ke >
I swithed to defaults indeed already, but now wondering, if using the same config for pbp is fine
07:58
<
LinuxHackerman >
Only one way to find out :D
07:58
<
Ke >
linus.heckemann: did you enable some extra modules, do you have amdgpu?
07:59
<
LinuxHackerman >
no, my honeycomb is headless
07:59
<
LinuxHackerman >
But this should have basically all the modules since autoModules is enabled
08:00
<
LinuxHackerman >
yep, modprobe amdgpu works
08:00
<
Ke >
I need it on initrd
08:00
<
LinuxHackerman >
right, well then go ahead :p
08:01
<
LinuxHackerman >
I have no explicit initrd modules
08:01
<
Ke >
I also need the firmware there
08:01
<
Ke >
how is that even possible?
08:01
<
Ke >
you need at least rootfs block device support?
08:01
<
LinuxHackerman >
zfs is in there implicitly, because / is zfs
08:01
<
LinuxHackerman >
emmc is builtin
08:02
<
Ke >
but it should be module =o(
08:02
<
LinuxHackerman >
why?
08:02
<
LinuxHackerman >
I think having emmc built-in makes sense on ARM
08:02
<
Ke >
I guess defconfig's gonna defconfig
08:02
<
LinuxHackerman >
hehe
08:03
<
Ke >
modules are compressed, image is not
08:03
dstzd has joined #nixos-aarch64
08:04
<
LinuxHackerman >
ah, fair
08:05
<
LinuxHackerman >
ok yeah 41M is a fairly big image
08:06
<
Ke >
I guess LTO should help with that, if nixos ever switches to LLVM here
08:07
dstzd has quit [Client Quit]
08:13
cfinch has quit [Read error: No route to host]
08:13
cfinch has joined #nixos-aarch64
08:15
ky0ko has quit [Remote host closed the connection]
08:16
cfinch has quit [Read error: Connection reset by peer]
08:18
cfinch has joined #nixos-aarch64
08:29
dstzd has joined #nixos-aarch64
08:30
dstzd has quit [Remote host closed the connection]
08:30
dstzd has joined #nixos-aarch64
08:31
dstzd has quit [Client Quit]
08:44
dstzd has joined #nixos-aarch64
08:45
dstzd has quit [Remote host closed the connection]
08:45
sorki has joined #nixos-aarch64
08:46
srk has quit [Ping timeout: 240 seconds]
08:48
sorki is now known as srk
08:51
cfinch has quit [Read error: Connection reset by peer]
08:51
cfinch has joined #nixos-aarch64
08:58
cfinch has quit [Ping timeout: 256 seconds]
08:58
cfinch has joined #nixos-aarch64
09:04
dstzd has joined #nixos-aarch64
09:45
noonien5 has joined #nixos-aarch64
09:46
noonien has quit [Ping timeout: 264 seconds]
09:46
noonien5 is now known as noonien
09:58
cfinch has quit [Ping timeout: 272 seconds]
10:36
noonien9 has joined #nixos-aarch64
10:37
noonien has quit [Ping timeout: 265 seconds]
10:37
noonien9 is now known as noonien
10:49
monk has left #nixos-aarch64 ["Error from remote client"]
10:50
monk has joined #nixos-aarch64
10:51
zarel has joined #nixos-aarch64
11:02
monk has left #nixos-aarch64 ["Error from remote client"]
11:03
pbb has joined #nixos-aarch64
11:07
monk has joined #nixos-aarch64
11:11
asbachb has joined #nixos-aarch64
11:14
<
asbachb >
samueldr: Doe you have any precompiled image for that PR. I'm unable to build that stuff on my machine.
11:19
<
noneucat >
samueldr: u-boot prints that the spci is available at boot through uart if crust is present
11:20
zupo has joined #nixos-aarch64
11:23
<
noneucat >
other than that i'm not sure other than actually going into suspend and checking power draw
11:27
ib07 has quit [Ping timeout: 256 seconds]
11:27
ib07 has joined #nixos-aarch64
11:39
<
Ke >
ok, pbp still works with the defconfig
11:39
<
Ke >
no need to have separate pbp defconfig
11:44
<
Ke >
if only the manjaro patches were less messy
11:52
dstzd has joined #nixos-aarch64
11:57
dstzd has quit [Client Quit]
12:00
asbachb has quit [Ping timeout: 248 seconds]
12:33
dstzd has joined #nixos-aarch64
12:40
tilpner has quit [Remote host closed the connection]
12:40
tilpner has joined #nixos-aarch64
13:09
orivej has quit [Ping timeout: 272 seconds]
13:12
FRidh has joined #nixos-aarch64
13:35
<
LinuxHackerman >
I've got a successful UEFI-based boot (with systemd-boot) on the honeycomb as well now :)
13:35
<
LinuxHackerman >
Didn't build the firmware myself though :(
14:06
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
14:15
orivej has joined #nixos-aarch64
14:41
dstzd has joined #nixos-aarch64
14:42
dstzd has quit [Remote host closed the connection]
14:43
zupo has joined #nixos-aarch64
15:04
dstzd has joined #nixos-aarch64
15:06
dstzd has quit [Client Quit]
15:07
WilliButz has quit [Ping timeout: 256 seconds]
15:18
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
15:19
WilliButz has joined #nixos-aarch64
15:28
zupo has joined #nixos-aarch64
15:32
zupo has quit [Client Quit]
15:42
dstzd has joined #nixos-aarch64
15:48
zupo has joined #nixos-aarch64
16:00
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
16:16
dstzd has joined #nixos-aarch64
16:17
dstzd has quit [Client Quit]
16:18
dstzd has joined #nixos-aarch64
16:19
dstzd has quit [Client Quit]
16:33
orivej has quit [Ping timeout: 272 seconds]
16:43
yorick has quit [Ping timeout: 260 seconds]
16:43
yorick has joined #nixos-aarch64
17:06
dstzd has joined #nixos-aarch64
17:07
dstzd has quit [Client Quit]
17:21
Raito_Bezarius has quit [Ping timeout: 260 seconds]
17:25
dstzd has joined #nixos-aarch64
17:26
dstzd has quit [Client Quit]
17:26
dstzd has joined #nixos-aarch64
17:28
dstzd has quit [Client Quit]
17:39
FRidh has quit [Ping timeout: 260 seconds]
17:47
dstzd has joined #nixos-aarch64
17:48
dstzd has quit [Client Quit]
18:06
dstzd has joined #nixos-aarch64
18:08
dstzd has quit [Client Quit]
18:12
rajivr has quit [Quit: Connection closed for inactivity]
18:19
cole-h has joined #nixos-aarch64
18:22
monk has left #nixos-aarch64 ["Error from remote client"]
18:26
monk has joined #nixos-aarch64
18:26
ib07 has quit [Ping timeout: 260 seconds]
18:27
dstzd has joined #nixos-aarch64
18:29
dstzd has quit [Client Quit]
18:30
zupo has joined #nixos-aarch64
18:30
dstzd has joined #nixos-aarch64
18:31
dstzd has quit [Client Quit]
18:35
dstzd has joined #nixos-aarch64
18:37
dstzd has quit [Client Quit]
18:37
dstzd has joined #nixos-aarch64
18:37
dstzd has quit [Client Quit]
18:38
dstzd has joined #nixos-aarch64
18:41
Raito_Bezarius has joined #nixos-aarch64
18:42
asbachb has joined #nixos-aarch64
18:42
asbachb has left #nixos-aarch64 [#nixos-aarch64]
18:51
dstzd has joined #nixos-aarch64
18:52
dstzd has quit [Remote host closed the connection]
18:53
dstzd has joined #nixos-aarch64
18:53
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
19:04
cole-h has quit [Quit: Goodbye]
19:05
cole-h has joined #nixos-aarch64
19:06
ib07 has joined #nixos-aarch64
19:14
tilpner has quit [Remote host closed the connection]
19:14
tilpner has joined #nixos-aarch64
19:25
veleiro has joined #nixos-aarch64
19:51
zupo has joined #nixos-aarch64
19:55
dstzd has joined #nixos-aarch64
19:57
dstzd has quit [Read error: Connection reset by peer]
20:08
jumper149 has joined #nixos-aarch64
20:14
dstzd has joined #nixos-aarch64
20:15
dstzd has quit [Client Quit]
20:36
<
veleiro >
i just realized I wont be able to make a nixOS lxc container on my router
20:36
dstzd has joined #nixos-aarch64
20:37
<
veleiro >
damn armv7, i can hardly get armv7 to work with nixos on other platforms
20:37
<
veleiro >
mostly because of the lack of caches, and building against custom nixos configs...
20:38
<
veleiro >
ur on ur own
20:48
orivej has joined #nixos-aarch64
21:39
dstzd has joined #nixos-aarch64
21:45
<
pbb >
samueldr: Just to be sure, I didn't check if linuxPackages (5.4 lts) boots on the rpi4. I only checked the _latest (5.10).
21:45
<
samueldr >
5.4 AFAIK doesn't
21:45
<
pbb >
yeah, probably not.
21:46
<
samueldr >
since 5.10 is the next LTS, it's great news
21:46
dstzd has joined #nixos-aarch64
21:46
<
samueldr >
5.9 is even already EOL'd
21:46
<
samueldr >
this is where you can see the LTS versions once decided
21:47
<
samueldr >
there is no strict schedule
21:47
<
pbb >
whatever, I still need the rpi-specific kernel to make my hifiberrys work :(
21:47
<
pbb >
why can't they upstream things :(
21:47
<
samueldr >
yeah, that's another whole issue, at the very least you could theoretically start yourself up with an aarch64 build with mainline :)
21:48
<
pbb >
I always use the qemu-user binfmt thing
21:48
<
pbb >
even with the rpi 4 everything is
_so_ slow
21:48
<
samueldr >
since there's not much to actually build, qemu-user here should work fine enough
21:48
<
samueldr >
but still, it's quite YMMV
21:48
<
samueldr >
any failures under qemu-user has to be validated under actually-native
21:49
<
pbb >
have seen things fail to build with qemu-user only
21:50
dstzd has quit [Client Quit]
21:50
<
pbb >
but the current deviceTree module looks like a convoluted mess and I kind of don't want to touch it
21:51
<
pbb >
for reference, this is how my module is currently used:
21:51
dstzd has joined #nixos-aarch64
21:51
<
samueldr >
I really haven't been paying enough attention to the deviceTree modules
21:51
dstzd has quit [Client Quit]
21:52
<
samueldr >
(which was partly because of my initial lack of complete understanding)
21:52
dstzd has joined #nixos-aarch64
21:52
<
samueldr >
I should try again to make that touchscreen display work
21:52
<
samueldr >
as an exercise
21:52
<
samueldr >
but uh... time flies
21:52
<
pbb >
1/24 of of 2021 is already over :D
21:53
dstzd has quit [Client Quit]
21:53
dstzd has joined #nixos-aarch64
21:54
dstzd has quit [Client Quit]
21:55
dstzd has joined #nixos-aarch64
22:01
<
sphalerite >
veleiro: "just" buy a honeycomb lx2k and run a 32-bit VM on it? :p
22:05
<
samueldr >
is the CPU 32 bit aware? :)
22:05
<
samueldr >
(I would guess it is)
22:06
<
ar >
speaking of honeycomb
22:06
<
ar >
i should do something with mine
22:10
<
sphalerite >
samueldr: yep, cortex a72 \o/
22:10
<
sphalerite >
(I've tested it to make sure already)
22:10
jumper149 has quit [Quit: WeeChat 3.0]
22:13
<
sphalerite >
I even used a 32-bit cpuburn to test power consumption with heavy load x)
22:13
<
sphalerite >
(because apparently that's more effective at generating heat)
22:16
zupo has quit [Ping timeout: 256 seconds]
22:16
zupo has joined #nixos-aarch64
22:24
dstzd has joined #nixos-aarch64
22:33
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
22:42
zupo has joined #nixos-aarch64
22:42
zupo_ has joined #nixos-aarch64
22:46
zupo has quit [Ping timeout: 246 seconds]
23:27
ib07 has quit [Ping timeout: 264 seconds]
23:33
* colemickens
gets ready to do the u-boot mainline rpi4 thang
23:34
<
samueldr >
colemickens: I don't think unstable updated yet, but 2021.01 is in master
23:34
<
samueldr >
don't know if it matters
23:34
<
samueldr >
I guess not if pbb started from the hydra-built sd images
23:34
<
samueldr >
(unless selecting a fresh enough image)
23:37
<
pbb >
I tested it by cherry-picking 5ff4a674125abbe8586884db62c66a00a8b4f665, a315a1503ab94caf281c3f584fd621109ba2cbb4 and 3d3f22e7bee920b8b3f0e67b3619e932b37fe986 on top of nixos-unstable
23:38
<
pbb >
Because when I tested it, master was completely broken
23:39
<
pbb >
But I think rpi4 booting linux_latest was already possible with 2020.10
23:39
<
pbb >
It just needed that patch (that I still need to send upstream, duh)
23:50
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
23:52
dstzd has joined #nixos-aarch64