2024-02-01 19:42:59 <mps> linux-starfive in testing repo is upgraded to 6.7.3
2024-02-01 19:43:54 <ikke> thanks
2024-02-01 19:43:55 <mps> clandmeter: did I told you that sophgo is enabled in linux-edge 6.7 but it only support milkv-duo
2024-02-01 19:44:09 <mps> not pioneer
2024-02-01 19:45:08 <mps> also I tested 6.8-rc1 on VF1 without patches, it boots and mmc works but most other drivers doesn't
2024-02-01 19:45:28 <ikke> Ok, I'll hold off upgrading the kernel on my VF1 then
2024-02-01 19:45:36 <mps> 6.8-rc1 can't boot on VF2
2024-02-01 19:46:27 <mps> ikke: I will try to upgrade kernel for VF1 with patches on weekend
2024-02-01 19:47:46 <mps> Esmil rebased it on 6.8-rc2
2024-02-01 19:47:54 <ikke> trying to upgrade my vf2
2024-02-01 19:50:18 <mps> lets hope it will work, I didn't tested 6.7.3
2024-02-01 19:54:07 <ikke> cannot reach it yet, checking with serial cable
2024-02-01 19:56:11 <ikke> It did boot
2024-02-01 19:56:24 <ikke> But no nics
2024-02-01 19:57:00 <mps> hm, give me some time to check on my board
2024-02-01 19:57:33 <ikke> Might be wrong dts
2024-02-01 19:57:59 <mps> oh, did you upgrade u-boot for it
2024-02-01 19:58:20 <ikke> No, I have not
2024-02-01 19:59:00 <ikke> Forgot, how should I do that again?
2024-02-01 19:59:50 <mps> ikke: https://arvanta.net/alpine/u-boot-visionfive2-alpine/
2024-02-01 20:00:20 <mps> but don't use starfive repo. use our main repo
2024-02-01 20:01:00 <mps> iirc apk add u-boot-jh7100
2024-02-01 20:02:25 <ikke> u-boot-starfive?
2024-02-01 20:02:37 <mps> yes, could be
2024-02-01 20:02:50 <mps> I forgot exact name
2024-02-01 20:02:56 <ikke> that's the one that's available
2024-02-01 20:03:05 <mps> then it is
2024-02-01 20:03:38 <mps> I have to go to my 'workshop' to check all
2024-02-01 20:04:35 <ikke> You mention a different one on your page
2024-02-01 20:04:46 <ikke> u-boot-jh7110-0_git20231016-r0.apk
2024-02-01 20:05:24 <mps> yes. I didn't updated this page when I moved all to aports
2024-02-01 20:05:36 <mps> sorry
2024-02-01 20:05:39 <ikke> no worry
2024-02-01 20:06:05 <mps> I planned to update this when finish everything, this weekend I hope
2024-02-01 20:07:35 <mps> I'm working rn on one arm64 machine kernel
2024-02-01 20:12:11 <ikke> So if I understand it correctly, I only need to run the flashcp commands?
2024-02-01 20:16:44 <mps> yes
2024-02-01 20:18:33 <mps> 2 files from /usr/share/u-boot/starfive_visionfive2/
2024-02-01 20:21:51 <mps> uhm, cryptsetup doesn't work as expected with new upgraded version
2024-02-01 20:22:46 <clandmeter> mps: https://github.com/sophgo/linux-riscv/tree/sg2042-master
2024-02-01 20:23:44 <clandmeter> 6.8 with lots of sophgo patches
2024-02-01 20:24:19 <clandmeter> I want to give. this a try
2024-02-01 20:24:57 <clandmeter> im still struggling on how to generate the best config
2024-02-01 20:25:50 <mps> ikke: just tested, upgraded and flashed u-boot and upgraded kernel works fine on VF2
2024-02-01 20:26:04 <ikke> I have the issue now again that mtd0 does not exist :/
2024-02-01 20:26:08 <clandmeter> i guess this config is their minimum base config https://github.com/sophgo/linux-riscv/blob/sg2042-master/arch/riscv/configs/sophgo_mango_normal_defconfig
2024-02-01 20:26:28 <mps> flashcp -v /usr/share/u-boot/starfive_visionfive2/u-boot-spl.bin.normal.out /dev/mtd0
2024-02-01 20:26:40 <ikke> yeah, but no /dev/mtd0
2024-02-01 20:26:43 <mps> flashcp -v /usr/share/u-boot/starfive_visionfive2/u-boot.itb /dev/mtd3
2024-02-01 20:26:53 <clandmeter> so i could do a make sophgo_mango_normal_defconfig and build it
2024-02-01 20:26:54 <mps> huh, how?
2024-02-01 20:26:56 <ikke> I had that issue before
2024-02-01 20:27:09 <ikke> I rebooted before updating u-boot
2024-02-01 20:27:24 <mps> ikke: which kernel is booted
2024-02-01 20:27:42 <ikke> huh, 6.5.5
2024-02-01 20:28:12 <mps> hm, also it should have /dev/mtdX
2024-02-01 20:28:26 <ikke> yeah, but there is no /dev/mtd* at all
2024-02-01 20:28:28 <clandmeter> maybe it failed to load a module?
2024-02-01 20:28:42 <mps> cat /proc/mtd will show you status
2024-02-01 20:29:56 <mps> clandmeter: you got mango board?
2024-02-01 20:30:03 <clandmeter> does your dmesg show any erorr? or maybe try to load the module
2024-02-01 20:30:21 <clandmeter> the pioneer is the mango board
2024-02-01 20:30:39 <mps> ah really, didn't know
2024-02-01 20:31:09 <clandmeter> i have no idea why companies always use these code names and then come with some fancy marcom name
2024-02-01 20:31:42 <clandmeter> or maybe the pioneer uses the mango reference design 
2024-02-01 20:32:47 <mps> companies wants to confuse us as much as possible, then it is easier to sell all these shit^Wuseful things
2024-02-01 20:33:03 <clandmeter> i guess there is no minimum kernel config requirements for alpine? i mean there are, but not documented.
2024-02-01 20:33:34 <clandmeter> apart from the board specific requirements
2024-02-01 20:33:46 <mps> yes, defconfig is usually big
2024-02-01 20:34:50 <clandmeter> they also enable CONFIG_EXPERT
2024-02-01 20:35:09 <clandmeter> no idea why, but it blows up the kernel with a bunch of debug crap
2024-02-01 20:35:13 <mps> ikke: you can put all on mmc card and boot from it, skipping flash boot
2024-02-01 20:35:47 <ikke> I have it on an mmc car
2024-02-01 20:35:49 <ikke> card
2024-02-01 20:35:58 <mps> clandmeter: they want that you fix bugs. and because this enabled all debugs
2024-02-01 20:36:42 <ikke> I have an mmc and sdcard
2024-02-01 20:38:37 <mps> ikke: try from external mmc by settings boot switches to it
2024-02-01 20:38:51 <mps> this usually works
2024-02-01 20:38:54 <ikke> That's what I was booting of in the first place
2024-02-01 20:39:19 <ikke> /boot as far as I can tell contains the new kernel, but it still boots the old kernel
2024-02-01 20:39:20 <mps> and it doesn't work?
2024-02-01 20:39:42 <ikke> It boots, but still the old kernel
2024-02-01 20:39:46 <ikke> 6.5.5
2024-02-01 20:40:22 <mps> hm, did you had mounted boot partition mounted under /boot when you upgraded kernel
2024-02-01 20:43:29 <ikke> /dev/mmcblk1p3 on /boot type ext4 (rw,relatime)
2024-02-01 20:43:59 <mps> should be this one
2024-02-01 20:44:30 <ikke> https://tpaste.us/9MV0 
2024-02-01 20:46:12 <mps> could you scroll back on console to see what it actually boots
2024-02-01 20:46:35 <mps> u-boot says if it cannot boot selected kernel for some reason
2024-02-01 20:47:02 <mps> or, could you paste complete boot log
2024-02-01 20:47:47 <ikke> "Trying to boot from MMC2"
2024-02-01 20:49:29 <ikke> https://tpaste.us/nkOE 
2024-02-01 20:49:51 <ikke> *** Warning - bad CRC, using default environment
2024-02-01 20:50:39 <ikke> *** Warning - bad CRC, using default environment
2024-02-01 20:50:44 <ikke> Retrieving file: /dtbs-starfive/
2024-02-01 20:50:45 <ikke> ** File not found /dtbs-starfive/ **
2024-02-01 20:51:23 <mps> this is from vendor u-boot
2024-02-01 20:52:45 <mps> no, sorry I'm wrong
2024-02-01 20:54:44 <mps> for some reason I'm not sure you upgraded kernel properly though all looks ok
2024-02-01 20:55:53 <ikke> I'll try to load the kernel apk on it and install it again
2024-02-01 20:56:40 <mps> keep care about mount boot partition
2024-02-01 20:58:04 <mps> this line '** File not found /dtbs-starfive/ **' from boot log shouldn't be there
2024-02-01 21:01:09 <mps> do you have this 'fdtdir /dtbs-starfive/' in extlinux.conf for default boot option
2024-02-01 21:01:36 <mps> or only 'fdt' in front
2024-02-01 21:02:26 <ikke> from memory only fdt
2024-02-01 21:03:13 <mps> aha ok, check it when boots
2024-02-01 21:03:59 <mps> fdt needs full path of filename while fdtdir only directory
2024-02-01 21:05:32 <mps> will be afk for about 15 minutes
2024-02-01 21:19:31 <ikke> It still boots kernel 6.5.5, where does that kernel come from?
2024-02-01 21:24:34 <ikke> ok, so I removed the emmc and now it boots the correct kernel :/
2024-02-01 21:24:43 <ikke> so some how it still loaded the kernel from the emmc
2024-02-01 21:33:13 <mps> you should check boot_targets variable in u-boot
2024-02-01 21:33:38 <mps> it shows order of devices from which it will try boot
2024-02-01 21:35:38 <mps> current default in u-boot built in is 'boot_targets=mmc1 usb0 nvme0 mmc0 dhcp'
2024-02-01 21:39:26 <mps> or no, maybe I set this in my board flash and forgot what I did
2024-02-01 21:40:48 <mps> I wanted boot from mmc1 when it is plugged, then from usb and then nvme leaving emmc as last, dhcp I don't use
2024-02-01 21:43:15 <ikke> I just leave the emmc distached for now
2024-02-01 21:45:19 <mps> hm, according to source nvme should be first
2024-02-01 21:45:33 <ikke> I have nothing on the nvme
2024-02-01 21:45:40 <ikke> You see it trying though
2024-02-01 21:45:56 <ikke> ** No partition table - nvme 0 **
2024-02-01 21:45:58 <ikke> Couldn't find partition nvme 0:1
2024-02-01 21:47:32 <mps> that explains, next is usb and then emmc
2024-02-01 21:53:15 <mps> right `env default -a` sets this order 'boot_targets=nvme0 usb0 mmc0 mmc1 dhcp'. So I changed it and forgot
2024-02-01 21:55:47 <ikke> ah ok
2024-02-01 21:55:54 <ikke> well, now everything works again
2024-02-01 21:56:21 <ikke> I guess if I copy everything over from the sdcard to the emmc, it would work again as well
2024-02-01 21:57:02 <mps> probably, but check
2024-02-01 21:57:31 <mps> do you have /dev/mtd now
2024-02-01 21:57:48 <ikke> hmm, no
2024-02-01 21:58:10 <mps> and kernel is 6.7.3?
2024-02-01 21:58:11 <ikke> no /proc/mtd either
2024-02-01 21:58:18 <ikke> 6.7.3-0-starfive
2024-02-01 21:58:26 <mps> hm, strange
2024-02-01 21:58:34 <mps> modprobe mtd
2024-02-01 21:59:02 <ikke> no difference
2024-02-01 21:59:24 <mps> hm
2024-02-01 21:59:55 <ikke> the mtd module is loaded
2024-02-01 22:00:22 <mps> modprobe spi_nor
2024-02-01 22:00:42 <mps> and modprobe ofpart
2024-02-01 22:01:26 <ikke> mtd depends on spi_nor
2024-02-01 22:01:30 <ikke> so it was already loaded
2024-02-01 22:01:58 <ikke> loaded  both, no difference
2024-02-01 22:02:04 <mps> hm, strange again
2024-02-01 22:02:26 <ikke> oh, /proc/mtd exists
2024-02-01 22:02:37 <ikke> but output is empty (except for headers)
2024-02-01 22:05:43 <mps> try modprobe ofpart
2024-02-01 22:06:24 <ikke> I did that
2024-02-01 22:06:43 <ikke> https://tpaste.us/Xyml 
2024-02-01 22:13:46 <clandmeter> are you sure the dtb is loaded correctly?
2024-02-01 22:14:35 <ikke> let me try the V2-v1.3b one
2024-02-01 22:17:31 <ikke> yup, now it's there
2024-02-01 22:20:32 <ikke> mps: what should I provide to fdtdir?
2024-02-01 22:20:45 <ikke> I provided /dtbs-starfive/starfive/
2024-02-01 22:21:37 <mps> just 'fdtdir /dtbs-starfive/'
2024-02-01 22:22:12 <mps> new u-boot have subdirs named on SoC/vendor
2024-02-01 22:23:51 <ikke> Ok, now it works
2024-02-01 22:25:36 <mps> in new u-boot we now have all dtbs sorted in dirs named by vendors/SoC/class of boards/ and other well known names
2024-02-01 22:25:56 <ikke> I saw it
2024-02-01 22:26:02 <ikke> I've flashed uboot now
2024-02-01 22:26:19 <mps> \o/
2024-02-01 22:32:27 <clandmeter> it was missing dtb?
2024-02-01 22:33:11 <ikke> yeah
2024-02-01 22:36:25 <clandmeter> good to know
2024-02-02 14:56:43 <mps> ikke: (et all) I updated a little https://arvanta.net/alpine/alpine-on-visionfive/ and also uploaded new https://dev.alpinelinux.org/~mps/riscv64/visionfive-v2-mmc.img.xz image
2024-02-02 19:43:25 <mps> ikke: I built kernel for VF1 https://dev.alpinelinux.org/~mps/riscv64/linux-vf1-6.8_rc2-r0.apk
2024-02-02 19:44:10 <mps> named it as linux-vf1 to not be confused with linux-starfive which is nowadays for VF2
2024-02-02 20:55:31 <ikke> mps: cool, thanks. Will test it later 
2024-02-03 10:29:35 <mps> ikke: I updated https://arvanta.net/alpine/alpine-on-visionfive/ with description for VF1 and uploaded new image for it with 6.8-rc2 kernel
2024-02-06 14:13:21 <mps> kernel for VF2 is upgraded to 6.7.4, now jh7110_crypto doesn't crash on boot (doesn't crash at all)
2024-02-06 14:13:49 <mps> tested five time boot 'in a row' and it didn't crash
2024-02-06 14:16:10 <mps> and I uploaded new boot image https://dev.alpinelinux.org/~mps/riscv64/visionfive-v2-mmc.img.xz at which jh7110_crypto is not blacklisted
2024-02-06 14:16:47 <mps> also, upgraded cpufetch now supports riscv64
2024-02-08 19:38:51 <koshershekels> https://appealforassistance.notion.site/Appeal-for-Assistance-abcca31346e944b38e09ebabb4208152
2024-02-09 03:25:20 <iggy> I'd be tempted to just ban *.compute.amazonaws.com, but that might be a _little_ heavy handed
2024-02-09 05:10:50 <valerius> it will be like the old days when we banned *.ipt.aol.com
2024-02-09 17:36:41 <mps> heh, a little self boasting https://forum.rvspace.org/t/rvspace-pioneers-2023/4107
2024-02-14 14:57:35 <ncopa> mps: congrats!
2024-02-14 15:09:59 <ikke> Nice!
2024-02-14 15:14:51 <mps> thanks both
2024-02-21 19:56:43 <mps> ncopa: you told earlier that you would like to use visionfive V2 as router and I told that it have some delays in interactive use of remote works
2024-02-21 19:57:46 <mps> now I tested with latest kernel which have some updates to network drivers and found that interactive use is now quite acceptable
2024-02-21 19:57:59 <ikke> interesting
2024-02-21 19:58:10 <mps> I made router with it and works fine
2024-02-21 20:00:00 <mps> not fast networking as cudy wireless device with mediatek ethernet bridge/switch but anyway decided to use it for now
2024-02-21 20:00:58 <mps> mostly because can't install alpine on cudy (only openwrt)
2024-02-21 20:01:28 <mps> on VF2 alpine works very nice and fast
2024-02-21 20:02:33 <mps> ikke: also I use 3.3V pin for powering fan and now is not so noisy but keep temp acceptably normal
2024-02-21 20:03:48 <mps> dovecot on VF2 with nvme is a lot faster than on previous arm32 server with ssd disk
2024-02-21 20:05:54 <mps> will see if can add wifi access point with usb dongle
2024-02-22 10:44:21 <ncopa> what kind of routing speed can I expect? a pass-through data stream
2024-02-22 11:03:39 <mps> idk for sure. I have about over 300Mb/s in ookla speedtest (I have to use pppoe and mss clamping which lower throughput) but lot more with iperf in local net
2024-02-22 11:04:16 <mps> best option is to test it locally and see
2024-02-22 21:59:55 <ikke> mps: is that the kernel with the improved network performance?
2024-02-22 22:00:53 <mps> ikke: I think so though I don't have data from previous versions to compare
2024-02-22 22:01:55 <mps> but some improvements are added for ethernet drivers are added in latest two kernels
2024-02-22 22:02:22 <mps> two major version kernels*
2024-02-22 22:02:43 <ikke> Mine is on 6.7.3
2024-02-22 22:03:01 <ikke> So I could use that as a baseline
2024-02-22 22:03:12 <mps> I think it have these improvements
2024-02-22 22:03:16 <ikke> ah ok
2024-02-22 22:03:30 <mps> I just pushed 6.7.5
2024-02-22 22:03:35 <ikke> yes, I noticed that
2024-02-22 22:04:29 <mps> I'm now satisfied with it as router, and mail server
2024-02-22 22:04:56 <ikke> I don't use it for anything right now
2024-02-22 22:04:57 <mps> not perfect but good
2024-02-22 22:05:21 <ikke> but I have it as a bridge to the sv1 board
2024-02-22 22:05:35 <ikke> that works fine
2024-02-22 22:05:46 <mps> nice to hear
2024-02-22 22:06:01 <ikke> haven't done any network testing though
2024-02-29 16:48:59 <clandmeter> https://labs.scaleway.com/en/em-rv1/
2024-02-29 16:49:14 <clandmeter> OS alpine :)
2024-02-29 16:57:44 <mps> what kernel it runs for alpine?
2024-02-29 16:58:22 <clandmeter> Dunno
2024-02-29 17:39:04 <ncopa> i wonder if we could ask scaleway for one of those machines
2024-02-29 17:59:58 <mps> good, thead SoC is enabled in linux-edge
2024-02-29 18:04:58 <mps> maybe is time to build u-boot for th1520 SoC
2024-02-29 18:11:18 <mps> hm, it is for lichee-pi
2024-02-29 18:11:39 <mps> but anyway we can build it
2024-02-29 21:47:29 <ncopa> Linux alpine-riscv 5.10.113+ #1 SMP PREEMPT Thu Feb 29 09:44:59 UTC 2024 riscv64 Linux
2024-02-29 21:48:35 <ncopa> [    0.000000] Linux version 5.10.113+ (root@runner-uzktd58qn-project-6120-concurrent-6) (riscv64-linux-gnu-gcc (Ubuntu 11.4.0-1ubuntu1~2
2024-02-29 21:48:35 <ncopa> 2.04) 11.4.0, GNU ld (GNU Binutils for Ubuntu) 2.38) #1 SMP PREEMPT Thu Feb 29 09:44:59 UTC 2024
2024-02-29 21:48:42 <ncopa> the kernel is built with ubuntu
2024-02-29 21:55:52 <mps> I see, and it is 5.10 series
2024-02-29 21:55:53 <ikke> booooh :P
2024-02-29 21:56:21 <mps> so, alpine is not alpine there
2024-02-29 21:56:36 <ncopa> well, it is pretty much alpine userspace
2024-02-29 21:57:01 <ncopa> they have done very little with the userspace. pretty much a vanilla alpine
2024-02-29 21:57:02 <mps> ncopa: do you know what board they use
2024-02-29 21:57:12 <ncopa> im trying to figure that out
2024-02-29 21:57:47 <ncopa> cool it has bluetooth...
2024-02-29 21:57:52 <mps> maybe `cat /proc/device-tree/model` could help
2024-02-29 21:58:17 <ncopa> thank you sir!
2024-02-29 21:58:28 <ncopa> Scaleway EM-RV1-C4M16S128-A
2024-02-29 21:58:44 <mps> never heard for it
2024-02-29 21:58:44 <ncopa> looks like they built their own boards
2024-02-29 21:59:09 <mps> seems you are right
2024-02-29 21:59:25 <ncopa> or they fed the dev tree with their data
2024-02-29 22:00:25 <mps> to me it looks very similar to lichee-pi
2024-02-29 22:00:27 <ncopa> they use 1.3G
2024-02-29 22:00:48 <ncopa> but they also include the kernel sources and build, which is what takes all the space
2024-02-29 22:01:36 <mps> you mean alpine rootFS?
2024-02-29 22:01:43 <ncopa> yes
2024-02-29 22:02:01 <ncopa> VERSION_ID=3.20.0_alpha20231219
2024-02-29 22:02:01 <ncopa> PRETTY_NAME="Alpine Linux edge"
2024-02-29 22:02:13 <ncopa> makes sense. we dont have stable risc-v releases yet
2024-02-29 22:02:27 <ncopa> i found this article https://bret.dk/scaleway-risc-v-testing/
2024-02-29 22:02:45 <ncopa> which also mentions the similarities with lichee-pi
2024-02-29 22:05:19 <mps> hm maybe somewhat modified lichee-pi
2024-02-29 22:06:04 <mps> so I think sophgo is better solution for us to concentrate on it
2024-02-29 22:11:01 <ncopa> clandmeter how is it going with the milk-v boxes?
2024-02-29 22:11:17 <ikke> They are running, but he said it was not that stable
2024-02-29 22:11:27 <ikke> Frequent IO stalls
2024-02-29 22:11:38 <ncopa> what kernel?
2024-02-29 22:11:52 <ikke> 6.1.72-5-sophgo