2025-12-01 11:07:43 <mps> now when povervr (imagination) in kernel we should enable it also in mesa 25.3 series. some riscv boards have these gpus
2025-12-01 11:09:33 <mps> already built mesa 25.3.0 for riscv64 but can't test right now, hope tomorrow I will have time and peripherals 
2025-12-01 11:16:36 <mps> also, these should be enabled on aarch64, some arm64 have this gpus
2025-12-03 22:15:52 <ncopa> I wonder how well the 6.18 kernel works on riscv. With 6.12 only starfive boards has worked out of the box.
2025-12-03 22:16:36 <ncopa> Are there any new boards that works with 6.18 kernel?
2025-12-04 05:28:20 <iggy> I think the powerVR driver is the big deal for riscv in 6.18, but maybe the SiFive HiFive Premier P550 board?
2025-12-04 05:33:27 <iggy> Orange Pi RV2 maybe too
2025-12-04 05:33:59 <iggy> if you're talking about the delta between 6.12 and 6.18 for riscv64, that's pretty massive
2025-12-04 07:33:57 <mps> iggy: afaik starfive vf2 have imagination powervr 6 series gpu
2025-12-04 07:35:44 <ncopa> from what I read the Orange Pi RV2 boots, but no MMC support yet. (with 6.18-rc1)
2025-12-04 07:35:47 <mps> spacemit k1x have older powervr gpu which is not supported in mainline kernel
2025-12-04 07:36:43 <ncopa> ESWIN7700 (HiFive Premiere p550, starpr64) got various drivers but the ethernet driver is sceduled for 6.19. so no network there
2025-12-04 07:38:10 <ncopa> oh i think maybe the milk-v pioneer machines may work out of the box
2025-12-04 07:38:50 <mps> most board got improvements but afaik only jh7110 is full support in mainline, and some sifive
2025-12-04 07:40:55 <mps> starfive was most supportive to open source and they seriously worked to upstream all
2025-12-04 07:42:11 <mps> "next year is riscv year" TM :)
2025-12-04 14:00:02 <mps> i can't find mesa source anywhere for imagination gpu which is on bpi-f3
2025-12-04 14:01:23 <mps> it mentioned in latest mesa but looks like it doesn't work or i don't know how to use vulkan driver
2025-12-04 18:22:05 <mps> ncopa: I'm not sure that the milk-v pioneer is fully supported but you can try, maybe it will work
2025-12-04 18:36:06 <ddevault> anyone with a visionfive2 experience unreliable nvme issues?
2025-12-04 18:36:12 <ddevault> pcie-starfive 2c000000.pcie: error -ENODEV: failed to get valid pcie domain
2025-12-04 18:36:35 <ikke> ddevault: I had when I didn't provide enough power
2025-12-04 18:37:01 <ddevault> I suspected that but I am giving it a 20W PD supply
2025-12-04 18:37:04 <ikke> like a 3 amp power supply
2025-12-04 18:37:07 <ikke> hmm
2025-12-04 18:42:22 <mps> also I had this with some power supplies
2025-12-04 18:45:36 <ddevault> I'll try it with a different one
2025-12-04 18:46:33 <ddevault> I was able to follow your advice for u-boot and got it set up nicely with EFI on NVMe root btw mps, thanks again 
2025-12-04 18:47:07 <mps> nice
2025-12-04 18:47:23 <mps> and you are welcome
2025-12-04 18:48:19 <mps> now I know that it works, thanks for report
2025-12-04 18:48:49 <mps> dis you upgraded kernel to 6.18
2025-12-04 18:50:24 <mps> s/dis/did/
2025-12-04 19:07:12 <ddevault> I'm running everything from alpine edge upstream 
2025-12-04 19:07:37 <ddevault> I used your old sdcard image to run setup-alpine
2025-12-04 19:12:58 <mps> ah, then probably 6.18
2025-12-05 08:58:13 <ddevault> morning
2025-12-05 08:58:25 <ddevault> a day later and two USB PD power supplies and still have only booted to nvme once :/
2025-12-05 08:59:37 <ddevault> what's odd is that u-boot consistently loads the kernel/initrd from nvme
2025-12-05 08:59:41 <ddevault> it's only the kernel that fails to initialize it
2025-12-05 08:59:54 <ddevault> maybe it's the drive in question?
2025-12-05 09:19:00 <mps> ddevault: this reminds me that I saw such posts on #riscv channel (or/and here) from some people that they had problems with some nvmes types or vendors
2025-12-05 09:19:14 <ddevault> grabbed another one and it seems somewhat more reliable
2025-12-05 09:19:30 <mps> this was 2-3 years ago
2025-12-05 09:19:32 <ddevault> haven't finished installing alpine on it yet
2025-12-05 09:19:43 <ddevault> so we'll see
2025-12-05 09:19:54 <ddevault> this board is very picky about storage devices
2025-12-05 09:20:57 <mps> I had problem with one samsung evo till I found fix. later fix was added to mainline
2025-12-05 09:22:25 <ddevault> samsung 980 pro is the one that's having issues
2025-12-05 09:22:32 <ddevault> random nvme I stole out of a spare laptop is doing better
2025-12-05 09:25:28 <ddevault> yep, we're in business
2025-12-05 09:25:52 <ddevault> aside, does anyone know offhand how to configure chronyd to set the time properly on boot for devices without a persistent RTC
2025-12-05 09:27:21 <mps> usually i have this in config 'makestep 0.1 3'
2025-12-05 09:27:29 <ddevault> thanks, will try that
2025-12-05 09:28:29 <mps> full config https://tpaste.us/6Kw8
2025-12-05 09:30:14 <ddevault> thank you
2025-12-05 09:31:43 <ddevault> what's alpine's setup for risc-v builders at this time?
2025-12-05 09:33:20 <ncopa> We use milk-v pioneer
2025-12-05 09:33:32 <ddevault> nice
2025-12-05 09:34:01 <ncopa> I have been looking for a reliable solution for non-rtc setups
2025-12-05 09:35:01 <ddevault> great, as soon as I button up the board in its case it stops booting
2025-12-05 09:39:03 <ddevault> pcie failures reappeared. so weird
2025-12-05 09:39:20 <ddevault> only ever manage to boot it once
2025-12-05 09:40:13 <ddevault> comes up fine on 6.7.4.0-starfive
2025-12-05 09:46:45 <ddevault> mps: that chronyd config works great, thank you!
2025-12-05 09:46:48 <mps> works very stable in my case with 'Non-Volatile memory controller: Samsung Electronics Co Ltd NVMe SSD Controller SM981/PM981/PM983' (lspci shows)
2025-12-05 09:47:32 <ddevault> linux-stable has the same issue
2025-12-05 09:47:44 <ddevault> and linux-lts
2025-12-05 09:47:45 <ddevault> alas
2025-12-05 09:47:55 <mps> ddevault: you are welcome
2025-12-05 09:49:01 <mps> `uname -a` => Linux fx 6.17.8-0-starfive #1-Alpine SMP PREEMPT_DYNAMIC Thu, 20 Nov 2025 12:45:17 +0000 riscv64 Linux
2025-12-05 09:49:50 <mps> uptime => 10:49:12 up 14 days, 17:40,  0 user,  load average: 0.00, 0.00, 0.00
2025-12-05 09:50:34 <ddevault> I have one more drive I can try
2025-12-05 09:51:39 <mps> usual hackers day start :)
2025-12-05 09:51:49 <ddevault> quite :)
2025-12-05 09:55:46 <ddevault> nope, same on the third drive
2025-12-05 09:55:47 <ddevault> damn
2025-12-05 09:56:56 <ddevault> maybe I should just set this up to boot from SD
2025-12-05 09:58:24 <ddevault> I mean, with this error message it's probably _not_ the drive
2025-12-05 09:58:28 <ddevault> pcie-starfive 2c000000.pcie: error -ENODEV: failed to get valid pcie domain
2025-12-05 09:58:32 <ddevault> that sounds like an issue with the SoC
2025-12-05 09:59:51 <mps> ddevault: do you have version a or b of board
2025-12-05 09:59:59 <ddevault> not sure, how can I tell
2025-12-05 10:00:35 <mps> v1.2a or v1.3b
2025-12-05 10:01:11 <ddevault> b
2025-12-05 10:01:24 <mps> from u-boot cli
2025-12-05 10:02:08 <mps> hm, v1.3b should have all fixed
2025-12-05 10:02:45 <mps> i'm out of ideas
2025-12-05 10:03:01 <ddevault> there is some weird residue all over the bottom of my board
2025-12-05 10:03:05 <ddevault> maybe I should just clean it lol
2025-12-05 10:03:17 <mps> :)
2025-12-05 10:04:03 <mps> nvme screw is properly set
2025-12-05 10:05:00 <ddevault> yep
2025-12-05 10:05:11 <ddevault> or it was until I removed it to clean the board
2025-12-05 10:05:13 <ddevault> :)
2025-12-05 10:07:02 <mps> this reminds me that i had looooong ago few days to work at one big server. at the end discovered than one screw on disk was longer for one millimeter than it should be
2025-12-05 10:08:15 <mps> s/than/that/ for first than
2025-12-05 10:08:59 <ddevault> board is clean. hail mary, let's give it another go
2025-12-05 10:10:38 <ddevault> nope :(
2025-12-05 10:11:37 <ddevault> I should have a spare emmc module somewhere, I guess that's my last resort
2025-12-05 10:21:23 <ddevault> oh, for the love of god
2025-12-05 10:21:27 <ddevault> the emmc is unreliable too
2025-12-05 10:21:46 <ddevault> I guess I'm going with the microSD card
2025-12-05 10:22:06 <ddevault> why do I bother with SoCs again
2025-12-05 10:26:12 <mps> strange, ncopa, ikke and I use this board with nvme and emmc without serious problems
2025-12-05 10:27:46 <ddevault> it's probably a bad board
2025-12-05 10:27:58 <ddevault> maybe I'll order a new one, if they're available 
2025-12-05 10:34:30 <ddevault> what's weird is that if I boot up your old image, mps, I'm able to access the nvme no problem
2025-12-05 10:34:56 <ddevault> I should run updates on that and see if I still have issues on the latest kernel
2025-12-05 10:36:17 <ddevault> no, it works consitently when I boot up from microSD
2025-12-05 10:52:23 <ncopa> I have starfive vf2 but runs 6.12 kernel from alpine 3.22 IIRC. I think I boot it from sd an run from tmpfs root. I think I have an nvme.
2025-12-05 11:01:15 <mps> for me board works no matters from where booted
2025-12-05 11:02:04 <ddevault> thanks for testing that
2025-12-05 11:02:16 <ddevault> I have a reliable setup booting to SD, I guess I can live with that for now
2025-12-05 11:02:46 <mps> and that is for 3 years iirc, when I managed to resolve all big issues
2025-12-05 11:05:37 <mps> ddevault: anyway, try when you can with different power supplies, i tried two supplies from chromebooks, one bought in local shop and it worked properly with apple m1 power supply
2025-12-05 11:05:53 <ddevault> hm, alright
2025-12-05 11:05:55 <ddevault> I have a few more I can test out
2025-12-05 11:06:46 <mps> without nvme it worked on simple 5V supply
2025-12-05 17:35:21 <richhenn> I am trying to load 3.23 via serial ttl on a visionfive 2.  sdio mode fails (boot error ffffffff) and qspi mode gets as far as grub and dies with hart0 failures).  Is there any info out there about prerequisite states for switches, flash rom minimum levels. Can't find any related to alpiine.
2025-12-05 17:40:12 <richhenn> Would the debian trixie or ubuntu load instructions be relevant to alpine as things to try? They want flash rom updated to 1/25 levels to load, for example.  What levels have been tested?
2025-12-05 18:17:57 <mps> what kernel version and u-boot. what media
2025-12-05 18:50:34 <richhenn> media: micro sd card, boot sbi says it found device 1, then cannot set block device, input data exceeds 1048576 bytes - truncated, failed to load /boot/uEnv.txt; defaulting to text format; error boot2 not defined; switch to partition #0 OK, 
2025-12-05 18:53:32 <richhenn> scanning mmc 1:2..., libfdt fdt_check_header(): FDT_ERR_BAD_MAGIC; Unable to read ubootevi.var; failed to loac EFI variables; found EFI removable media binary efi/bootriscv64.efi; 1298432 bytes read in 115 ms (10.8 MiB/s)
2025-12-05 18:55:47 <richhenn> libfdt fdt_check-header() FDT_ERR_BADMAGIC; then Grub loads offering *Linux lts.  Then Booting 'linux lts'; then clk already disabled, then sbi_trap_error: hart0 trap handler failed (error -2) then a register dump
2025-12-05 18:56:44 <mps> did you upgraded it from 3.22 to 3.23 or new install
2025-12-05 19:00:53 <richhenn> attempting a brand new install.
2025-12-05 19:01:42 <mps> how, if i'm allowed to ask
2025-12-05 19:01:53 <richhenn> been using debian until now, now trying to switch to alpine
2025-12-05 19:02:29 <mps> and why grub over serial line
2025-12-05 19:03:23 <mps> here is alpine short guide
2025-12-05 19:03:27 <richhenn> serial line has been the least problematic for inital boots.
2025-12-05 19:03:33 <mps> https://arvanta.net/alpine/alpine-on-visionfive/
2025-12-05 19:04:21 <mps> right, serial is very good for test problems
2025-12-05 19:04:50 <richhenn> i do have usb kbd and mouse with hdmi.  x86_64 alpine works fine with those on another box, but nothing shows up at all on those for riscv. so serial is the only sign of life.
2025-12-05 19:05:28 <richhenn> thanks.  i'll check out the link.
2025-12-05 19:05:33 <mps> I think old u-boot don't have video
2025-12-05 19:06:26 <mps> here is how to update u-boot https://arvanta.net/alpine/u-boot-visionfive2-alpine/
2025-12-05 19:06:52 <richhenn> firmware is not the latest - suitable for debian image 69 from a year ago.  There is a newer flash image.  I can upgrade that.  Has anybody tested w visionfive 2, lately?
2025-12-05 19:07:52 <mps> but this old guide, after install you have to alpine upgrade to latest (or prefered) release
2025-12-05 19:08:40 <mps> yes, VF2 works fine for some of us
2025-12-05 19:10:31 <mps> yesterday and one user (well, not ordinary user but experienced man) installed alpine following above guides
2025-12-05 19:10:41 <richhenn> Great.  Probably, they have a newer flash rom software, so I'll upgrade rom and try again. No upgrade involved. I am starting out from scratch with 3.23 as first iso image.
2025-12-05 19:12:16 <richhenn> Thanks for the assistance so far. I'll give the arvanta posts a whirl and get back with results.
2025-12-05 19:13:00 <mps> rom can't be upgraded but you can flash/write u-boot to spi nor flash
2025-12-05 19:15:09 <richhenn> meaning...
2025-12-05 19:20:28 <mps> ?
2025-12-05 19:21:15 <mps> (i'm not native english speaker)
2025-12-05 19:44:55 <richhenn> not a problem, I was saying I simply accepted your correction of my terminology
2025-12-05 19:46:20 <mps> ah ok
2025-12-06 14:07:05 <ddevault> success https://redacted.moe/f/965c6c9f.png
2025-12-06 14:22:34 <mps> rootFS on nvme?
2025-12-06 14:23:24 <ddevault> no, still on microSD
2025-12-06 14:23:31 <ddevault> gave up on NVMe for the time being
2025-12-06 14:32:57 <mps> huh :(
2025-12-06 14:37:13 <mps> i'm struggling to boot from mmc on bpi-f3. boot from emmc and nvme works fine
2025-12-06 14:38:26 <mps> but it worked full year and more till now
2025-12-06 14:39:44 <mps> that's life with experimental hardware (and software)
2025-12-06 16:37:10 <mps> milk-v pioneer got improves in 6.19, maybe will work 'out of the box'
2025-12-06 16:39:27 <iggy> ddevault: dod you have the latest board firmwares (uboot, sbi, etc)... I think they made some PD improvements that helped that at one point... I ended up just throwing a dumb 20v power suppoly on the USB-C input so the craptacular PD chip would stop messing things up
2025-12-06 16:40:07 <ddevault> yeah, it's all the latest
2025-12-06 16:40:18 <ddevault> and I tried two 20W power supplies with PD, to no avail
2025-12-07 11:34:25 <mps> heh, interesting patch in mainline kernel b08ee4d666f216a6f9e7194a9b335147d4717f33
2025-12-07 11:35:58 <mps> I made this nearly 4 years ago same in aports/testing/linux-elm with 12x24 fonts
2025-12-07 11:36:08 <mps> funny
2025-12-08 15:42:14 <richhenn> I finally got the new riscv64 alpine live iso to boot in my visionfive2.  The issues, as expected, were related to older versions of the spl and u-boot in flash memory.  The latest version of these (6.0.0) turn out to be too big to burn into flash via flashcp.
2025-12-08 15:44:42 <richhenn> One could update via xmodem and make it work, or, following ubuntu's iso instructions, update it using their server image as a wedge.  Once I was on 6.0.0, Grub started and began the good ol' alpine install environment.
2025-12-08 15:52:43 <richhenn> My ethernet card wasn't found initially, but ifup and using eth1 instead of eth0 solved that.  The serial/ttl console and ssh work fine and the hdmi console keyboard and mouse function, but the screen stays off.
2025-12-08 15:59:45 <richhenn> Is anyone using the hdmi output yet? I'm not sure what drivers need to be present to just get console I/O working on the monitor.  Is this even working yet?
2025-12-08 16:00:56 <richhenn> I presume that the gpu is not amd, nvidia or intel, but for simple console i/o should that even be an issue?
2025-12-08 16:10:39 <richhenn> I did the install via the serial/ttl interface, if that makes a difference.
2025-12-08 16:34:34 <mps> iirc hdmi works
2025-12-08 16:34:53 <mps> gpu is imagination
2025-12-08 16:35:30 <mps> what is 6.0.0 version
2025-12-08 16:46:12 <richhenn> 6.0.0 is the sbl/u-boot pair that get loaded into flash memory for an spl boot of isos.  It is provided by starfive.  Alpine didn't load until I installed it.  Have others been using another technique to get the new alpine live iso up and running?
2025-12-08 16:47:44 <richhenn> are there packages I should install to support the imagination gpu talking to alpine - that didn't come in during an install using the serial/ttl console?
2025-12-08 16:48:16 <mps> alpine u-uboot-visionfive can be flashed
2025-12-08 16:49:05 <richhenn> is there a specific u-boot version that alpine is supposed to use booting from flash?
2025-12-08 16:49:10 <mps> we don't have but sdcard image
2025-12-08 16:50:00 <mps> https://arvanta.net/alpine/u-boot-visionfive2-alpine/
2025-12-08 16:50:51 <richhenn> I am only aware of the vendor supplied version.  The copy I used was the one supplied by ubuntu.  I could redo it with the vendor one now that the mdt has been expanded.
2025-12-08 16:53:10 <mps> vf2 have u-boot in mainline for 2 years and alpine build it
2025-12-08 16:55:02 <mps> I enabled gpu in latest linux-starfive kernel but didn't have time to test
2025-12-08 17:00:19 <richhenn> OK, so there is an alpine-specific u-boot I should be using.  Since that post was from 2023, I presumed it was outdated for an iso load. Mostly distributions have been supplying pre-installed images.  I was assuming that the live iso installs are new and needing a recent spl/u-boot combo.
2025-12-08 17:02:34 <richhenn> I read recently (debian site, I think) that mmc and sdcard boots are no longer supported and only uart and flash boots (which now detect where iso is: usb, which I used, or sdcard) should be used now. 
2025-12-08 17:04:26 <richhenn> Anyway, except for no vga-style console on my hdmi monitor, alpine riscv seems to be working great via serial and ssh.
2025-12-08 17:04:35 <mps> alpine is different, other distros doc doesn't apply
2025-12-08 17:05:32 <mps> alpine VF1 also works fine with kbd/display
2025-12-08 17:06:45 <richhenn> I worried about that.  So I should at some point go back, find and load an alpine supplied u-boot
2025-12-08 17:07:06 <mps> (send me board and it will with alpine :) )
2025-12-08 17:07:56 <mps> here is guide about u-boot https://arvanta.net/alpine/u-boot-visionfive2-alpine/
2025-12-08 17:08:33 <mps> but board can be booted fine without flash even
2025-12-08 17:10:09 <richhenn> Good, I am glad that kbd/display is expected to work.  Booted without flash? I tried that and got nowhere with the spl I was using at the time. I can go back that try that now that I have a newer spl.
2025-12-08 17:10:37 <richhenn> thanks for your help on this, btw.
2025-12-08 17:11:05 <mps> spl/u-boot can be read from sd card
2025-12-08 17:11:36 <mps> look at image I posted url above
2025-12-08 17:13:34 <mps> hm, i should update guide at last
2025-12-08 17:13:51 <mps> it is outdated
2025-12-08 17:19:17 <richhenn> Right.  Once you update it, it might be useful to post an url to your pages on the risc64 page in the wiki. At least until actual load requirements and instructions can be enterred into the wiki pages.
2025-12-08 17:21:14 <mps> i don't know to use wiki
2025-12-08 17:24:58 <richhenn> With your permission on the links, I could try to edit that page once the posts are up to day.  The riscv64 wiki page is currently very sparce with no installing info whatsoever.  As you''ve shown me, the load isn't quite the same as a normal alpne install.  The system needs to be in a specific state, first.
2025-12-08 17:26:38 <mps> help in writing docs will be appreciated
2025-12-08 17:27:44 <mps> i'm not native english speaker
2025-12-08 21:17:08 <richhenn> Glad to help where I can. I am a native English speaker, but I'll have to check out getting priviledges to add to the wiki.  Will explore.
2025-12-08 21:29:42 <richhenn> OK, I believe I can propose edits, now.  So let me know when your pages are brought up to date, and I'll put the links into the existing riscv64 page.  There is only one.  doc is meager to say the least. Once I am comfortable with the load enough to describe to others, I'd be happy to document that in the wiki.
2025-12-08 21:30:55 <richhenn> The main issue is getting the board in a state where it can receive the iso.
2025-12-11 09:52:53 <mps> ikke: forgot to tell you that i found meaning of these 3 unknown TLV eeprom codes, we talked earlier about them
2025-12-11 09:54:16 <mps> they are named in new source of bianbu u-boot for spacemit/bpi-f3
2025-12-11 11:22:20 <ncopa> clandmeter: I have looked at the kernel for pioneer box. Seems like most of the stuff is included in 6.18 kernel, but there are a handful patches that may be needed.
2025-12-11 11:22:41 <ncopa> clandmeter: are you interested in testing the 6.18 kernel on one of our pioneer boxes? if so when?
2025-12-11 11:34:07 <mps> ncopa: 6.19 will have more things
2025-12-11 11:34:28 <mps> maybe try with 6.19-rc
2025-12-11 11:40:08 <ncopa> I have 12 patches, which are expected to be pulled in for 6.19. They apply fine on th 6.18 kernel
2025-12-11 11:40:27 <ncopa> so I wonder if I should just cherry-pick them. they are non-intrusive
2025-12-11 11:41:49 <mps> `ping 172.16.30.101` => 5 packets transmitted, 0 packets received, 100% packet loss
2025-12-11 11:42:17 <mps> maybe clandmeter already upgraded it ;)
2025-12-11 12:08:21 <clandmeter> ncopa: sure
2025-12-11 12:08:25 <clandmeter> not in officen ow
2025-12-11 12:08:33 <clandmeter> can try tomorrow morning
2025-12-11 12:08:39 <clandmeter> or week after
2025-12-11 12:08:51 <clandmeter> i think the last patches go into .19
2025-12-11 12:09:13 <ncopa> thats what it looks like, yes. I am backporting them to 6.18
2025-12-11 12:39:42 <ncopa> I have installed the kernel on nld-bld-2 (.3). the apk is in /root
2025-12-11 12:39:58 <ncopa> but I think grub will default to the -sophgo kernel
2025-12-11 12:41:37 <ncopa> we should probalby try the .30.2 first, which seems to be more problematic
2025-12-11 12:48:14 <ncopa> it looks like the 6.18 kernel should work on spacemit k1? https://github.com/spacemit-com/linux/wiki
2025-12-11 13:19:53 <ncopa> kernel_comp_addr_r or kernel_comp_size is not provided!
2025-12-11 13:20:08 <ncopa> i did gunzip the kernel and the it boots to:
2025-12-11 13:20:11 <ncopa> https://tpaste.us/do05
2025-12-11 14:19:52 <achill> btw 6.18.1 will likely release the next days
2025-12-12 10:29:37 <mps> my riscv64 lxc (172.16.30.101) still not accessible? is the host down?
2025-12-12 12:33:03 <ncopa> i think the host is down
2025-12-15 08:43:46 <mps> riscv64 developer host is still down? or my lxc is not started?
2025-12-15 09:00:22 <ncopa> the host has been up 4 mins now
2025-12-15 09:02:00 <ncopa> iptables policy prevented the lxc to get ip address. it should be back now
2025-12-15 09:07:15 <mps> yes, it works now. thank you
2025-12-15 09:09:44 <mps> ncopa: btw, from changelog of 6.19-rc1 looks like it could work on milk-v pioneer
2025-12-15 09:17:20 <ncopa> ok
2025-12-15 09:17:50 <ncopa> yeah, i looked at what patches to packport to 6.18 last week. I think most of them went in to 6.19
2025-12-15 09:18:25 <ncopa> there may have been one or two patchsets that was still under review. Don't know if they went in to 6.19
2025-12-15 09:23:09 <clandmeter> hi, i just rebooted the machines
2025-12-15 09:23:12 <clandmeter> should be ok now
2025-12-15 09:25:26 <mps> and bananapi f3 is nearly ready, except some usb fixes
2025-12-15 10:27:42 <ncopa> looks like everything for pioneer went into 6.19. the changes are non-intrusive so I think I will simply backport them for our linux-lts-6.18 kernel
2025-12-15 10:31:47 <clandmeter> ok
2025-12-15 10:31:50 <clandmeter> dts also?
2025-12-15 10:32:00 <clandmeter> im looking at new disks
2025-12-15 10:32:10 <clandmeter> prices are going up rapidly 
2025-12-15 10:32:31 <clandmeter> memory is crazy...
2025-12-15 10:34:47 <ncopa> yes dts also
2025-12-15 10:35:11 <ncopa> i think we can try the 6.18 kernel with the nvme
2025-12-15 10:35:33 <ncopa> i am hoping it will solve the nvme issues we have had
2025-12-15 10:38:37 <ncopa> ouch.. i looked at a samsung 9100 pro in nov, around black friday
2025-12-15 10:38:38 <clandmeter> you will try it today? 
2025-12-15 10:38:45 <ncopa> prices has went up 50% since then
2025-12-15 10:39:03 <ncopa> yes, i can test it today if you are available to help
2025-12-15 10:39:14 <ncopa> im building it as we speak. will take 20 mins or so
2025-12-15 10:39:35 <ncopa> i regret i didnt by the disk in november
2025-12-15 10:40:48 <clandmeter> it will only get more crazy
2025-12-15 10:41:16 <clandmeter> expected to carry over to 2028
2025-12-15 10:42:08 <ncopa> alright, maybe I just buy it now then
2025-12-15 10:42:56 <ncopa> I'm also thinking of buying a milk-v titan: https://arace.tech/products/milk-v-titan-1?variant=44084141097140
2025-12-15 10:43:05 <ncopa> but need to buy RAM for it as well... 
2025-12-15 10:44:34 <ncopa> looks like lots of the things for spacemit k1, bananapi bpi-f3 and orangepi rv2 went into 6.19
2025-12-15 10:45:11 <ncopa> but it looks like significantly bigger changes than sophgo, so I think we just wait for linux-stable 6.19 for those
2025-12-15 10:47:59 <ncopa> clandmeter: i'm installing the 6.18.1 kernel now on 172.16.30.2
2025-12-15 10:48:12 <ncopa> clandmeter: are you available to help rescue if it goes south?
2025-12-15 10:49:07 <clandmeter> let me check if I have an usb
2025-12-15 10:50:53 <clandmeter> ncopa: looks like i have a boot usb
2025-12-15 10:51:01 <ncopa> ok good.
2025-12-15 10:51:11 <ncopa> it uses extlinux.conf, right?
2025-12-15 10:51:16 <clandmeter> yes
2025-12-15 10:51:27 <ncopa> I added an entry: LABEL lts
2025-12-15 10:51:39 <clandmeter> which box is it?
2025-12-15 10:51:42 <clandmeter> .2
2025-12-15 10:51:44 <ncopa> so I think we need to change it in the boot menu
2025-12-15 10:51:48 <ncopa> yes .2
2025-12-15 10:51:54 <clandmeter> let me see which one it is
2025-12-15 10:52:05 <clandmeter> so i plug in the kvm in the right one
2025-12-15 10:53:06 <clandmeter> .2 is on console now
2025-12-15 10:53:12 <clandmeter> so go ahead
2025-12-15 10:53:37 <ncopa> alright I reboot it and at boot you have to chose '2' or 'lts' at extlinux boot prompt
2025-12-15 10:54:00 <ncopa> rebooting
2025-12-15 10:54:03 <clandmeter> its a diff bootloader, but i get the point :)
2025-12-15 10:54:32 <ncopa> Connection to 172.16.30.2 closed.
2025-12-15 10:55:14 <clandmeter> Booting
2025-12-15 10:55:24 <ncopa> this is very exciting
2025-12-15 10:55:49 <clandmeter> No screen 
2025-12-15 10:56:09 <mps> spacemit pushing lot things to upstream
2025-12-15 10:56:14 <clandmeter> Did you fix the fw issue?
2025-12-15 10:56:25 <ncopa> the fw issue?
2025-12-15 10:56:38 <ncopa> mps: yes I saw. its very nice to see
2025-12-15 10:56:54 <clandmeter> Fw compression
2025-12-15 10:57:03 <clandmeter> I mention it last time
2025-12-15 10:57:14 <mps> openbi already merged spacemit in master
2025-12-15 10:57:27 <ncopa> clandmeter: I simply copied the entry in extlinux.conf
2025-12-15 10:57:40 <clandmeter> i mean in the kernel config
2025-12-15 10:57:48 <clandmeter> to support compressed firmware
2025-12-15 10:57:56 <ncopa> i dont remember exactly the fw issue. We lacked fw decompression in kernel?
2025-12-15 10:58:00 <ncopa> let me check
2025-12-15 10:58:04 <clandmeter> can you try to connect from serial? 
2025-12-15 10:58:09 <clandmeter> from the other boxc
2025-12-15 10:58:14 <ncopa> from .1?
2025-12-15 10:58:17 <clandmeter> nod
2025-12-15 10:59:04 <ncopa> what terminal emulator do we use there for serial?
2025-12-15 10:59:38 <mps> meh, s/openbi/opensbi/
2025-12-15 11:00:23 <mps> ime, best terminal for serial is screen
2025-12-15 11:01:07 <mps> screen -S rb -U -A -O -T screen-256color -fn /dev/${tty} 115200
2025-12-15 11:01:23 <clandmeter> i normally use screen
2025-12-15 11:01:30 <mps> usually works everywhere
2025-12-15 11:03:32 <ncopa> which serial console is it? apparently not ttyS0
2025-12-15 11:05:04 <ncopa> clandmeter: no I cannot connect to any serial console from .1
2025-12-15 11:05:20 <clandmeter> hmm
2025-12-15 11:05:25 <clandmeter> it does not list any dev?
2025-12-15 11:06:06 <ncopa> there are ttyS[0-3] but none of them appear to have anything responding
2025-12-15 11:06:13 <clandmeter> ah
2025-12-15 11:06:16 <clandmeter> no thats the router
2025-12-15 11:06:19 <clandmeter> you need the other box
2025-12-15 11:06:26 <clandmeter> .3 i assume
2025-12-15 11:06:42 <ncopa> alright
2025-12-15 11:06:46 <clandmeter> the 2 pioneer boxes are interconnected via usb
2025-12-15 11:06:55 <ncopa> and it appears to be ttyUSB0
2025-12-15 11:06:57 <ncopa> makes sense
2025-12-15 11:07:05 <ncopa> [11:06:37.225] Connected to /dev/ttyUSB0
2025-12-15 11:07:05 <ncopa> [11:06:37.219] tio 3.8
2025-12-15 11:07:05 <ncopa> [11:06:37.219] Press ctrl-t q to quit
2025-12-15 11:07:11 <ncopa> but no response on anything
2025-12-15 11:07:26 <ncopa> can you power cycle it?
2025-12-15 11:07:29 <clandmeter> yes 
2025-12-15 11:07:33 <clandmeter> thats what i was suggesting :)
2025-12-15 11:07:41 <clandmeter> keep connected and ill kick it
2025-12-15 11:08:03 <ncopa> I am connectec now to the ttyUSB0
2025-12-15 11:08:09 <ncopa> but no output
2025-12-15 11:09:18 <ncopa> its booting
2025-12-15 11:09:45 <ncopa> ok it booted the 6.1.80-sophgo kernel now
2025-12-15 11:09:52 <ncopa> i dont get the boot menu on serial
2025-12-15 11:10:28 <ncopa> lets have a break and continue in 30 mins?
2025-12-15 11:10:45 <clandmeter> make it 1h
2025-12-15 11:10:49 <ncopa> ok
2025-12-15 11:10:49 <clandmeter> i will go out for lunch
2025-12-15 11:10:52 <ncopa> sounds good
2025-12-15 11:11:01 <clandmeter> you can join if you want :)
2025-12-15 11:11:02 <clandmeter> KFC
2025-12-15 12:09:45 <ncopa> clandmeter: yeah you are right, I had forgot the compressed FW_LOADER thingy
2025-12-15 12:27:08 <ncopa> ok I now have both compressed and uncompressed firware available
2025-12-15 12:27:46 <ncopa> an I have updated kernel with CONFIG_FW_LOADER_COMPRESS_ZSTD=y
2025-12-15 12:29:04 <ncopa> clandmeter: I am ready to reboot it again
2025-12-15 12:29:20 <ncopa> but I dont get access to the boot menu from serial (I think)
2025-12-15 12:32:04 <ncopa> i suppose I could revert the default boot option so I only need your help to revert to old kernel
2025-12-15 12:57:31 <mps> strange that u-boot doesn't show on serial console
2025-12-15 13:04:22 <mps> and u-boot can be set if not load first kernel then it load default
2025-12-15 13:25:40 <ncopa> clandmeter: i changed it so lts is the default in extlinux.conf. I will need your help to recover it fails now
2025-12-15 13:25:59 <ncopa> rebooting now
2025-12-15 13:29:12 <ncopa> nothing?
2025-12-15 13:29:15 <ncopa> [   70.644637] ISOFS: Unable to identify CD-ROM format.
2025-12-15 13:31:17 <ncopa> nope :-(
2025-12-15 13:31:21 <ncopa> it does not boot
2025-12-15 13:31:26 <ncopa> clandmeter: help!
2025-12-15 13:38:26 <mps> this reminds that milkv titan have optional BMC
2025-12-15 14:01:28 <ncopa> clandmeter: whenever you get a chance, can you please reboot the .2 machine, and at the boot menu select '1' or 'sophgo'
2025-12-15 14:01:31 <ncopa> sorry
2025-12-15 14:01:32 <ncopa> thanks
2025-12-15 14:20:50 <clandmeter> ncopa: sorry was busy
2025-12-15 14:20:52 <clandmeter> whats up?
2025-12-15 14:25:22 <ncopa> i changed the extlinux.conf to boot the lts kernel as the default. and it didnt boot
2025-12-15 14:26:20 <ncopa> can you try boot it and select the 'sophgo' kernel at the boot menu
2025-12-15 14:26:33 <clandmeter> oh i just tried again
2025-12-15 14:26:43 <clandmeter> ok let me boot back to oringial
2025-12-15 14:26:56 <ncopa> i dont know what it was this time
2025-12-15 14:27:07 <ncopa> i would have expected serial console to work
2025-12-15 14:27:16 <ncopa> does the bootloader handle compressed kernels?
2025-12-15 14:28:48 <clandmeter> its booting sopho now
2025-12-15 14:28:55 <ncopa> i can see that. thanks
2025-12-15 14:29:02 <ncopa> i dont know why it fails with lts kernel
2025-12-15 14:29:11 <ncopa> did you see anything ouput?
2025-12-15 14:29:41 <ncopa> there comes nothing from the serial console, which is a bit worrying
2025-12-15 14:30:47 <clandmeter> no only bootloader
2025-12-15 14:30:57 <clandmeter> after the switching of kernel its gone
2025-12-15 14:31:01 <clandmeter> due to the fw issue
2025-12-15 14:31:16 <clandmeter> ssh is working
2025-12-15 14:31:21 <ncopa> yeah, im in
2025-12-15 14:31:38 <ncopa> i did upgrade the linux-firmware package, but I also extracted all the firmware files
2025-12-15 14:31:47 <ncopa> so now there should be both compressed and uncompressed
2025-12-15 14:32:01 <ncopa> I also updated the kernel config to support compressed firmware
2025-12-15 14:32:30 <ncopa> i dont know why it does not find the serial console at early boot
2025-12-15 14:32:43 <clandmeter> weird about the fw
2025-12-15 14:32:50 <clandmeter> i see you uncompressed them
2025-12-15 14:32:58 <clandmeter> but dmesg shows it cant find the fw
2025-12-15 14:33:39 <ncopa> maybe it needs older version of the firmware package?
2025-12-15 14:34:04 <clandmeter> yeah that could be it
2025-12-15 14:34:05 <ncopa> (11/22) Upgrading linux-firmware-none (20240610-r1 -> 20251125-r1)
2025-12-15 14:34:05 <ncopa> (12/22) Upgrading linux-firmware-amdgpu (20240610-r1 -> 20251125-r1)
2025-12-15 14:34:05 <ncopa> (13/22) Upgrading linux-firmware-radeon (20240610-r1 -> 20251125-r1)
2025-12-15 14:34:05 <ncopa> (14/22) Upgrading linux-firmware-rtl_nic (20240610-r1 -> 20251125-r1)
2025-12-15 14:34:11 <clandmeter> it doesnt say unable to find
2025-12-15 14:34:19 <ncopa> atleast the network works
2025-12-15 14:36:38 <ncopa> do we use the mmcblk1 for anything during boot? It looks like it only boots from nvme
2025-12-15 14:36:51 <ncopa> and it does support compressed kernel images
2025-12-15 15:10:41 <ncopa> will have to continue another day
2025-12-15 15:31:02 <clandmeter> sorry was on the phone
2025-12-15 15:35:09 <ncopa> i was busy as well
2025-12-15 18:48:31 <Anoduck[m]> ncopa: Not only because I am efficiently thrifty, but because so much is happening with RISC-V right now, I would hold off until a true market outlier has been established. Right now Risc-V is just playing catch-up to ARM and X86.
2025-12-15 19:10:57 <mps> heh, free/open source developers cannot wait but want to help early
2025-12-16 08:32:21 <mps> akms pkgs are not build binaries by builders? users have to build them locally?
2025-12-16 08:34:13 <clandmeter> ncopa: ping
2025-12-16 09:51:21 <ncopa> clandmeter: pong
2025-12-16 10:01:14 <clandmeter> you want to do some rv stuff? 
2025-12-16 10:01:18 <clandmeter> i mean the box
2025-12-16 10:05:08 <ikke> mps: the packages are built, but they just contain instructions how to build the kernel module against the current kernel 
2025-12-16 10:05:33 <ncopa> clandmeter: give me 10 mins
2025-12-16 10:19:54 <mps> ikke: thanks for explanation
2025-12-16 10:21:26 <mps> don't know how but think that should be better
2025-12-16 10:23:07 <mps> btw, in new gitlab how to restart pipeline on CI, can't find this 'button'
2025-12-16 10:25:48 <mps> oh, https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/94730/pipelines it running iiuc
2025-12-16 10:26:42 <ncopa> clandmeter: something broken with the network. I cannot copy the kernel to 172.16.30.2:
2025-12-16 10:26:48 <ncopa>  1%  255KB   0.7KB/s - stalled -
2025-12-16 10:27:00 <clandmeter> i have the same issue
2025-12-16 10:27:05 <clandmeter> not sure whats going on
2025-12-16 10:27:11 <clandmeter> also network to .1
2025-12-16 10:27:31 <ncopa> now it is progressing
2025-12-16 10:28:08 <clandmeter> network is really bad
2025-12-16 10:28:34 <clandmeter> it looks like its related to WG
2025-12-16 10:28:47 <clandmeter> you can direct connect to the public ip 
2025-12-16 10:29:05 <ncopa> I think I use the dmvpn connection
2025-12-16 10:29:10 <ikke> mps: the goal is to support out-of-tree modules without giving ncopa more work
2025-12-16 10:29:22 <clandmeter> then its an issue with dmvbpn
2025-12-16 10:29:30 <clandmeter> if i connect directly, its fast
2025-12-16 10:29:41 <ncopa> i was able to copy over the files and have updated the kernel
2025-12-16 10:29:56 <ncopa> i dont think it will boot, but it should hopefully give something on the serial console
2025-12-16 10:30:36 <ncopa> I have changed the default kernel to 'sophogo' so we always can reboot and get back to the old
2025-12-16 10:30:46 <clandmeter> i am not able to connect to .3
2025-12-16 10:31:08 <clandmeter> are you on the serial of .3?
2025-12-16 10:31:48 <ncopa> no, i got kicked out
2025-12-16 10:31:50 <ncopa> nld-bld-1 login: Read from remote host 172.16.30.3: Operation timed out
2025-12-16 10:31:50 <ncopa> Connection to 172.16.30.3 closed.
2025-12-16 10:31:55 <ncopa> i thikn the machine is dead again
2025-12-16 10:32:27 <ncopa> ssh: connect to host 172.16.30.3 port 22: Host is unreachable
2025-12-16 10:32:28 <ncopa> yup
2025-12-16 10:32:33 <ncopa> needs power cycle
2025-12-16 10:33:19 <clandmeter> lets reboot .3 first? 
2025-12-16 10:34:44 <ncopa> yes please
2025-12-16 10:36:52 <ncopa> i think we may need to enable the radeon and amdgpu drivers as well. they are currently disabled in riscv64 kernel
2025-12-16 10:37:59 <ncopa> the .3 is back
2025-12-16 10:38:45 <ncopa> [10:38:36.893] Warning: Could not open /dev/ttyUSB0 (No such file or directory)
2025-12-16 10:41:19 <mps> ttyUSB1?
2025-12-16 10:41:37 <ncopa> there are not ttyUSB*
2025-12-16 10:41:43 <mps> ah
2025-12-16 10:41:48 <ncopa> maybe its unplugged?
2025-12-16 10:42:17 <ncopa> and I lost the other pinoeer machine
2025-12-16 10:44:07 <clandmeter> i dont have screen on both
2025-12-16 10:44:10 <clandmeter> got a bit lost now
2025-12-16 10:44:29 <clandmeter> bld2 is online
2025-12-16 10:44:31 <ncopa> the .2 is back online
2025-12-16 10:44:42 <ncopa> but it has no /dev/ttyUSB* device
2025-12-16 10:44:53 <clandmeter> .2 is offline for me
2025-12-16 10:44:54 <ncopa> so I wonder if the usb serial cable was unplugged?
2025-12-16 10:44:57 <clandmeter> .3 is online
2025-12-16 10:45:03 <clandmeter> serial cable is online
2025-12-16 10:45:04 <ncopa> oh sorry
2025-12-16 10:45:15 <ncopa> yes. .3 is online
2025-12-16 10:45:18 <clandmeter> i just put stickers on which one is .3 and .2...
2025-12-16 10:45:20 <ncopa> the .2 is gone
2025-12-16 10:45:34 <clandmeter> let me reboot .2
2025-12-16 10:45:41 <ncopa> thanks
2025-12-16 10:45:47 <clandmeter> did you copyu the kernel to  both? 
2025-12-16 10:46:02 <clandmeter> cause i thought i could identify from boot menu, but it looks same.
2025-12-16 10:46:31 <ncopa> I have copied *a* kernel to both, but only the .2 has the current one. i think it is the only place it is installed
2025-12-16 10:46:40 <ncopa> was thinking getting it running before trying on .3
2025-12-16 10:46:58 <ncopa> .3 is the edge and 3.23 builder
2025-12-16 10:47:09 <ncopa> .2 has the older 3.2*-stable
2025-12-16 10:47:33 <ncopa> so I was thinking experiment with .2 only, til we got something that works
2025-12-16 10:48:03 <ncopa> correction: I have only copied the 6.18 kernel to .2
2025-12-16 10:49:44 <clandmeter> ok both online now
2025-12-16 10:49:49 <clandmeter> both on old kernel
2025-12-16 10:50:36 <ncopa> thanks
2025-12-16 10:50:49 <clandmeter> so both have .18 kernel?
2025-12-16 10:51:04 <ncopa> no. only .2 has 6.18 kernel
2025-12-16 10:51:13 <clandmeter> check boot
2025-12-16 10:51:16 <clandmeter> both have
2025-12-16 10:51:30 <clandmeter> 6,18.0 vs 1
2025-12-16 10:52:03 <ncopa> alrigh, that something old, let me remove that
2025-12-16 10:52:18 <ncopa> deleted it
2025-12-16 10:52:34 <ncopa> so we dont get confused
2025-12-16 10:52:50 <ncopa> but I dont have any /dev/ttyUSB* on .3
2025-12-16 10:52:50 <clandmeter> ok so bld1 has new kernel
2025-12-16 10:52:55 <clandmeter> and we serial from bld2
2025-12-16 10:53:09 <ncopa> yes, except that I dont have the serial device to connect to from bld2
2025-12-16 10:53:20 <clandmeter> weird
2025-12-16 10:53:23 <ncopa> the /dev/ttyUSB* is missing
2025-12-16 10:53:28 <clandmeter> let me replug
2025-12-16 10:54:08 <ncopa> [  347.625419] usb 1-1.4: USB disconnect, device number 4
2025-12-16 10:54:08 <ncopa> [  444.569233] dw-apb-uart 7040001000.serial: failed to request DMA
2025-12-16 10:55:10 <clandmeter> so bld1 has ok serial
2025-12-16 10:55:15 <clandmeter> but bld2 is erroring
2025-12-16 10:55:47 <ncopa> yup
2025-12-16 10:56:59 <ncopa> looks like bld2 boots with grub?
2025-12-16 10:57:31 <clandmeter> its linuxboot
2025-12-16 10:57:34 <clandmeter> there is no grub
2025-12-16 10:58:02 <ncopa> ok. where does linuxoot get its config from? there are no extlinux.conf
2025-12-16 10:58:55 <clandmeter> it can read from grub
2025-12-16 10:59:05 <ncopa> ok maybe that is what it does then
2025-12-16 10:59:26 <clandmeter> documentation of linuxboot is non existing afaik
2025-12-16 10:59:26 <ncopa> annoying that we dont have serial connection working
2025-12-16 10:59:37 <clandmeter> i think it was working before
2025-12-16 10:59:41 <ncopa> it was
2025-12-16 10:59:48 <clandmeter> im not sure what is causing it
2025-12-16 10:59:52 <ncopa> i dont know what happened
2025-12-16 10:59:56 <clandmeter> the BSD could also influece it
2025-12-16 11:00:09 <clandmeter> BSP
2025-12-16 11:00:22 <clandmeter> the BSP loads its own kernel with DTS
2025-12-16 11:00:43 <clandmeter> i read somehwere the DTS is not overwritten by keec
2025-12-16 11:01:09 <ncopa> oh, that could be related
2025-12-16 11:01:13 <clandmeter> so i wonder if we use a newer kernel, if the new dts is actually used
2025-12-16 11:01:27 <ncopa> can you try replug the usb serial device?
2025-12-16 11:01:31 <clandmeter> i was checking /dev/device-tree
2025-12-16 11:01:35 <ncopa> i removed the usbserial kernel module
2025-12-16 11:02:48 <clandmeter> replugged
2025-12-16 11:03:23 <ncopa> nothing
2025-12-16 11:04:37 <clandmeter> lets kick it one more time
2025-12-16 11:04:42 <clandmeter> maybe the reinit will help?
2025-12-16 11:04:46 <clandmeter> i mean boot
2025-12-16 11:04:56 <ncopa> unplug it, wait 10 sec and re plug it
2025-12-16 11:04:59 <clandmeter> these systems are so unstable
2025-12-16 11:05:02 <ncopa> yea
2025-12-16 11:05:08 <ncopa> maybe try other usb port too if possible
2025-12-16 11:05:28 <ncopa> really annoying when you cannot trust the hardware
2025-12-16 11:07:18 <clandmeter> its not reporting anything anymore
2025-12-16 11:09:02 <clandmeter> its back
2025-12-16 11:09:40 <ncopa> \o/
2025-12-16 11:09:46 <ncopa> great! what did you do?
2025-12-16 11:09:55 <clandmeter> move it to back usb 
2025-12-16 11:10:05 <ncopa> thanks
2025-12-16 11:10:17 <clandmeter> i guess the kernel disabled the front one 
2025-12-16 11:10:25 <ncopa> alright, lets try reboot the bld1 (.2) then 
2025-12-16 11:10:41 <ncopa> reboot it and select "lts" at boot menu
2025-12-16 11:10:49 <clandmeter> ok go for it
2025-12-16 11:10:57 <ncopa> rebooting
2025-12-16 11:12:31 <ncopa> the last thing I see on screen is:
2025-12-16 11:12:33 <ncopa> [   11.147357] ISOFS: Unable to identify CD-ROM format.
2025-12-16 11:12:33 <ncopa> [   11.164550] ISOFS: Unable to identify CD-ROM format.
2025-12-16 11:12:35 <ncopa> then nothing
2025-12-16 11:12:45 <clandmeter> thats the boot menu
2025-12-16 11:12:47 <clandmeter> linuxboot
2025-12-16 11:12:55 <clandmeter> i get the same on screen 
2025-12-16 11:13:04 <ncopa> did you boot the lts kernel?
2025-12-16 11:13:07 <clandmeter> nod
2025-12-16 11:13:10 <ncopa> ok
2025-12-16 11:13:17 <ncopa> so exactly nothign comes out
2025-12-16 11:13:21 <ncopa> on serial
2025-12-16 11:13:26 <ncopa> and it does not boot
2025-12-16 11:13:29 <clandmeter> i think something switches the kernel
2025-12-16 11:13:40 <clandmeter> err serial
2025-12-16 11:14:24 <ncopa> or the serial driver is not working
2025-12-16 11:14:33 <ncopa> or the kernel itself is completely broken
2025-12-16 11:14:46 <ncopa> alright
2025-12-16 11:14:52 <ncopa> back to the drawing board
2025-12-16 11:14:56 <clandmeter> after the isofs msg, it jumps to linuxboot boot menu
2025-12-16 11:15:02 <clandmeter> on the console
2025-12-16 11:15:09 <ncopa> can you please reboot it again? to sophgo
2025-12-16 11:15:27 <ncopa> it could be the DTS as you said. I pick another DTS
2025-12-16 11:15:30 <ncopa> FDT
2025-12-16 11:15:44 <ncopa> maybe we can try the FDT for the old kernel and see what happens
2025-12-16 11:16:47 <ncopa> can we continue on 30 mins?
2025-12-16 11:17:36 <ncopa> need to go out a bit
2025-12-16 11:55:19 <ncopa> and im back
2025-12-16 11:55:26 <ncopa> i think i need work on 3.23.1 release now
2025-12-16 12:02:05 <mps> i tried 6.19-rc1 on bpi-f3, it doesn't boot