| * mark_ -> mjw | 00:08 | |
| - mtm (QUIT: Ping timeout: 260 seconds) (~textual@47.202.75.129) | 01:03 | |
| + mtm (~textual@47.202.75.129) | 01:05 | |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-126-4-196.tukw.qwest.net) | 01:50 | |
| + colinsane (~colinunin@97-126-4-196.tukw.qwest.net) | 01:53 | |
| - colinsane (QUIT: Ping timeout: 244 seconds) (~colinunin@97-126-4-196.tukw.qwest.net) | 02:03 | |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@gnu.wildebeest.org) | 02:41 | |
| - cobra (QUIT: Ping timeout: 260 seconds) (~cobra@user/Cobra) | 02:49 | |
| - wickedshell (QUIT: Ping timeout: 246 seconds) (~wickedshe@2601:8c0:800:4baa:444a:1b9b:da9d:99ca) | 02:55 | |
| - paperManu (QUIT: Ping timeout: 252 seconds) (~paperManu@107.159.243.8) | 03:28 | |
| - op_4 (QUIT: Remote host closed the connection) (~tslil@user/op-4/x-9116473) | 04:05 | |
| + op_4 (~tslil@user/op-4/x-9116473) | 04:05 | |
| - aloo_shu (QUIT: Ping timeout: 252 seconds) (~aloo_shu@85.51.18.252) | 04:47 | |
| + aloo_shu (~aloo_shu@85.51.18.68) | 04:50 | |
| + wickedshell (~wickedshe@2601:8c0:800:4baa:3f1d:d321:6c17:f039) | 05:02 | |
| + gustav28 (~gustav@c-78-82-52-175.bbcust.telenor.se) | 10:02 | |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@47-186-65-73.dlls.tx.frontiernet.net) | 10:04 | |
| + jacobk (~quassel@47-186-65-73.dlls.tx.frontiernet.net) | 10:05 | |
| - aloo_shu (QUIT: Ping timeout: 244 seconds) (~aloo_shu@85.51.18.68) | 10:08 | |
| + aloo_shu (~aloo_shu@85.51.18.112) | 10:10 | |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-78-82-52-175.bbcust.telenor.se) | 12:09 | |
| minute | BoostisBetter: great to hear | 12:30 |
|---|---|---|
| + mjw (~mjw@gnu.wildebeest.org) | 12:31 | |
| amospalla | On my Pocket I have access to devices /dev/mmcblk2boot{0,1}, shouldn't these be hidden? | 13:00 |
| amospalla | I am playing with reform-emmc-debootstrap, recently, but I have also rebooted since then. | 13:01 |
| - mtm (QUIT: Ping timeout: 252 seconds) (~textual@47.202.75.129) | 13:03 | |
| + mtm (~textual@47.202.75.129) | 13:05 | |
| amospalla | josch: I'm trying to implement the FIXME for reform-emmc-bootstrap to allow running from nvme when boot is on sdcard | 13:15 |
| amospalla | I need to detect if boot is on emmc, but both sdcard and emmc are /dev/mmcblk devices. I guess I can not rely on /dev/mmcblk2boot{0,1} to be available because these should be hidden (IIRC). | 13:16 |
| josch | amospalla: why do you need to detect whether /boot is on sd-card? | 13:17 |
| josch | what you need to detect is whether your currently running system has its rootfs from SSD or not (and that's hard) | 13:18 |
| amospalla | I can only think on the emmc device size, because it has some MiB hidden, it should not class with the sdcard size, so a "there one and only one mmcblk with this size check" should work. | 13:18 |
| amospalla | sorry, eMMC, not sd-card. | 13:18 |
| josch | reform-emmc-bootstrap should not care where your current /bot is | 13:18 |
| josch | errr /boot | 13:19 |
| amospalla | I have the root detection already done, now I need to know if boot is on sd-card or emmc. | 13:19 |
| josch | because it sets your system up such that /boot will be on emmc -- irrespectivey where it was before | 13:19 |
| josch | why? | 13:19 |
| amospalla | I see. | 13:19 |
| josch | how did you do the root detection? | 13:19 |
| josch | i was not able to figure this out | 13:19 |
| amospalla | grep /proc mounts with a device | 13:19 |
| amospalla | josch: https://paste.debian.net/1344934/ | 13:20 |
| amospalla | line 62, using that check on other parts of the code can get wether the specified device is current live root | 13:21 |
| josch | amospalla: this does not work when your rootfs comes from /dev/mapper/reformvg-root, for example | 13:22 |
| amospalla | why? | 13:22 |
| josch | it also does not work is the rootfs is mounted with UUID= | 13:22 |
| josch | oh, typo, it does work with /dev/reformvg/root | 13:23 |
| josch | but instead of parsing /proc, maybe do: | 13:24 |
| josch | lsblk --noheadings --output=MOUNTPOINT /dev/reformvg/root | 13:24 |
| josch | another way is to use this: | 13:25 |
| josch | findmnt --fstab --noheadings --evaluate --mountpoint / --output SOURCE | 13:25 |
| josch | the latter has the advantage, that it also finds out where your / is mounted if your /etc/fstab uses a UUID= | 13:25 |
| amospalla | great, I'll use that | 13:27 |
| amospalla | ohhh, you already have set DEV_MMC, yay. | 13:28 |
| amospalla | should it abort when /boot is already on eMMC, but not abort when /boot is on sd-card? | 13:33 |
| amospalla | That is, when the tool is executed when system is running from nvme. If it is running from nvme and /boot is already on eMMC then it means that it booted correctly, so there should be no reason to overwrite it. | 13:36 |
| josch | in that case, why did the user run it? | 13:42 |
| josch | maybe make an interactive question informing the user of the situation and ask them whether they really want to proceed | 13:42 |
| amospalla | Right, I was just being defensive. | 13:42 |
| josch | remember, that even if /boot is mounted from emmc | 13:43 |
| josch | it's still possible that your kernel that is currently running came from a partition on your sd-card | 13:43 |
| amospalla | True, but in that situation the user has been messing and is already doing some "advanced things (tm)" | 13:43 |
| josch | yes, i imagine that one use-case of the script can be: user messed up and wants to get the system into a working state again | 13:44 |
| josch | (which is part of why i didn't want to name the tool "migrate"-something) | 13:44 |
| amospalla | That means comparing mount point and fstab | 13:44 |
| josch | (which is the primary functionality that minute wants to use this for) | 13:44 |
| amospalla | ok, I got a bit lost at this point, is it desired any change to current script? | 13:46 |
| amospalla | As it resides on git repository currently I mean. It has the following FIXME lines: | 13:49 |
| amospalla | # FIXME: support the scenario where we we are currently running from nvme | 13:49 |
| amospalla | # FIXME: in that case, fail if /boot is already on emmc | 13:49 |
| amospalla | I see, it is in the lines of "user wanta a working state again" | 13:52 |
| ch | with util-linux hat on: generally please use findmnt over blkid/lsblk; and always set the output options instead of doing some grep/awk parsing | 13:55 |
| amospalla | thank you, I didn't know that tool. | 13:56 |
| BoostisBetter | minute: just had the kb reset on its own again. Very good. Really think this was a good addition tot he kb. | 14:03 |
| + paperManu (~paperManu@107.159.243.8) | 14:05 | |
| + cobra (~cobra@user/Cobra) | 14:32 | |
| josch | ch: i actually have a util-linux question. I'm sometimes (for example in reform-emmc-bootstrap) deliberately using blkid over lsblk because its man page says that it reads directly from the device and since some tools write stuff to the device, i do want blkid for non-cached direct information, right? | 14:38 |
| josch | amospalla: yes "user wanta a working state again" is a good description. A bit more specific would be "user has rootfs on SSD and lost their /boot and wants to set it up on emmc"-tool | 14:40 |
| josch | amospalla: yes, i'd be looking forward to receiving a patch which implements the case where you run this from a system on SSD | 14:41 |
| ch | josch: kinda. but mostly you dont want blkid | 14:41 |
| josch | git grep lsblk | wc -l | 14:41 |
| josch | 15 | 14:41 |
| josch | git grep blkid | wc -l | 14:41 |
| josch | 4 | 14:41 |
| josch | git grep findmnt | wc -l | 14:42 |
| josch | 12 | 14:42 |
| josch | is there a reason why i should convert from lsblk to findmnt? lsblk seems to have all the options i want to use it in a script | 14:42 |
| ch | josch: so mostly blkid has illdefined behaviour, and might or might not give you what you want. new features go into lsblk/findmnt | 14:42 |
| josch | ch: what do you mean by "illdefined behaviour"? | 14:43 |
| josch | if you tell me that i should just blindly convert from blkid to lsblk i can do that | 14:43 |
| josch | but the man page, while it very directly pushes towards lsblk is still suggesting to me that there are valid uses of blkid | 14:44 |
| amospalla | josch: I'm only adding this: (1) allow running when current root is on nvme, (2) fail if eMMC has something mounted. | 14:45 |
| josch | amospalla: cool! :) | 14:45 |
| amospalla | And it does what it was doing before, whatever there is on eMMC, whatever it was booting from before, it just overwrites the eMMC to a working state. | 14:46 |
| josch | good | 14:46 |
| - cobra (QUIT: Quit: ZNC 1.8.2 - https://znc.in) (~cobra@user/Cobra) | 14:46 | |
| josch | minute: that's what you want, right? ^ | 14:46 |
| ch | josch: iirc blkid is ok, -if- you pass --probe and can/want to live with that | 14:50 |
| + cobra (~cobra@user/Cobra) | 14:50 | |
| ch | josch: if you want info that is consistent with udev etc, probably use lsblk | 14:50 |
| ch | josch: unfortunately the answer is "it depends" and what it depends on is not very clear | 14:50 |
| josch | i see, thank you! | 14:52 |
| - aloo_shu (QUIT: Ping timeout: 252 seconds) (~aloo_shu@85.51.18.112) | 14:57 | |
| + aloo_shu (~aloo_shu@85.51.18.40) | 14:59 | |
| - cobra (QUIT: Quit: ZNC 1.8.2 - https://znc.in) (~cobra@user/Cobra) | 15:00 | |
| josch | ch: while i have you here, how do i convert "blkid -s TYPE" to lsblk? For blkid, TYPE is "ext4", for example. But in lsblk, it is "part" which is not what i want. | 15:07 |
| ch | FSTYPE? | 15:08 |
| ch | depending on what you want (: | 15:09 |
| josch | ooooh | 15:10 |
| josch | it would help to have a list of the possible values somewhere | 15:10 |
| josch | yes, i want FSTYPE it seems | 15:10 |
| + cobra (~cobra@user/Cobra) | 15:10 | |
| ch | --list-columns gives you column names | 15:12 |
| ch | -O prints all rows with all cols | 15:12 |
| josch | those are a lot of columns | 15:13 |
| ch | yeah | 15:13 |
| josch | and i have no idea what half of them mean XD | 15:13 |
| ch | -O --json might be easier on the eye | 15:13 |
| josch | but i could've found FSTYPE via the column value, indeed | 15:13 |
| josch | I recently added jq as a Depends of reform-tools | 15:14 |
| josch | i didn't want to parse the parted output with shell anymore... | 15:14 |
| ch | :) | 15:14 |
| ch | use JSON; ? | 15:14 |
| josch | not sure whether a Depends on jq or libjson-perl is nicer. I have no strong opinion on that. | 15:16 |
| josch | It's working fine with jq now. | 15:16 |
| ch | well the script could stop being shell, mayb | 15:16 |
| josch | yes, but calling external programs in perl is also a pain and easier in shell :/ | 15:17 |
| josch | in any case, blkid is now out and lsblk is in: https://source.mnt.re/reform/reform-tools/-/merge_requests/110/diffs?commit_id=6a24903541d3b6d984c1edfd4d6c384509d438e8 | 15:20 |
| + josch89 (~josch@37.4.230.248) | 16:09 | |
| - josch89 (PART: !!unknown attribute: msg!!) (~josch@37.4.230.248) | 16:09 | |
| - aloo_shu (QUIT: Ping timeout: 252 seconds) (~aloo_shu@85.51.18.40) | 16:31 | |
| + aloo_shu (~aloo_shu@90.166.193.22) | 16:33 | |
| amospalla | josch: here is the patch https://paste.debian.net/1344950/ | 17:27 |
| josch | amospalla: can you send that as an email to josch@debian.org? Then I can leave comments inline | 17:28 |
| amospalla | sure | 17:28 |
| amospalla | josch: sent you the patch, but forgot to mention that I tested it on a Pocket Reform, both when booted from sd-card and nvme. | 18:10 |
| josch | thank you, i can test it on classic reform with imx8mq here | 18:14 |
| amospalla | doing this helped me understand boot things, now I think I could my root on f2fs or btrfs. | 18:15 |
| amospalla | I'll probably move back to emmc from nvme, my nvme takes a good portion of battery. | 18:16 |
| cwebber | ACTION checks forum | 18:49 |
| cwebber | Well definitely looks like I am not longer the only user who hit the blank screen thing anymore | 18:49 |
| josch | yup https://community.mnt.re/t/screen-turns-on-and-off-but-wont-boot/2938 | 18:51 |
| minute | cwebber: yeah. i'm really curious what exactly is the root cause, but i'll hopefully see that soon when your pocket arrives | 19:10 |
| cwebber | Yes hopefully! | 19:10 |
| cwebber | minute: I hope it's not this but if it turns out that the board is borked or something can I just buy one of the new boards and have you install it while it's there? | 19:11 |
| cwebber | The new boards seem interesting anyway | 19:12 |
| cwebber | But I shouldn't be preemptively pessimistic! | 19:12 |
| minute | cwebber: if something is borked we will replace under warranty | 19:42 |
| minute | cwebber: also, there isn't a new board for pocket... only for classic reform :3 | 19:42 |
| cwebber | Okie :) | 19:58 |
| cwebber | Thx 💜 | 19:58 |
| - ericsfraga (QUIT: Quit: ERC 5.6.1-git (IRC client for GNU Emacs 31.0.50)) (~user@2a00:23cc:b43d:4b00::3ce) | 20:09 | |
| mhoye | There's a new board for Classic Reform? | 21:26 |
| mhoye | ACTION clicks around madly | 21:26 |
| BoostisBetter | mhoye: I think the v3 board is the newest and I'm not sure it is out yet. | 21:26 |
| BoostisBetter | mhoye: I also know that the Reform Next boards should work with the Classic as well. | 21:27 |
| mhoye | I'll hold my breath | 21:27 |
| minute | BoostisBetter: that's not yet certain @ next, it would require a bunch of mechanical adapters that are not developed yet. | 21:29 |
| BoostisBetter | minue: roger! Really looking forward to the Next! | 21:39 |
| BoostisBetter | minute: when pushing the reset button on the kb controller, all you are doing is power cycling it right? | 21:41 |
| BoostisBetter | Nothing bad happens if you hold it down for a while longer or anything like that? | 21:41 |
| minute | BoostisBetter: not even power cycling, it's just a reset signal to the rp2040 cpu | 21:41 |
| BoostisBetter | minute: I had an issue where the kb firmware reset the kb, but the kb didn't seem connected as trackball and keys weren't doing anything | 21:42 |
| BoostisBetter | minute: I used the manual reset a few times and it came back up, but it took a couple of times (like 3) | 21:42 |
| minute | BoostisBetter: soon we should update your imx8mp uboot | 21:43 |
| minute | it has a new thing in it where it explicitly resets the usb hub at boot | 21:43 |
| minute | it could be that this reset not being done causes the instability | 21:43 |
| BoostisBetter | minute: you think that is the ticket? uboot is causing issues with the kb controller? | 21:43 |
| BoostisBetter | minute: ahh I see. That would make sense as well | 21:43 |
| minute | we discovered this a few days ago during re-installing imx8mp pocket motherboards, and each time booting them from emmc after install, the keyboard wouldn't work | 21:44 |
| minute | but the usb hub reset fixed that | 21:44 |
| BoostisBetter | minute: I am up to 6 days in uptime on the Pocket. so it seems like the kb and hub were always the issue and not the SoC | 21:44 |
| minute | nice @ uptime | 21:45 |
| BoostisBetter | minute: that is awesome to hear! Looking forward to the uboot update. Will it be announced? | 21:45 |
| BoostisBetter | minute: the reset functionality that is going into the uboot update, is it possible to expose this via the oled menu, such that when you reset the kb there, it is doing the same thing as uboot? | 21:50 |
| BoostisBetter | minute: please excuse my ignorane on the inner workings of uboot and the interface with the RP2040 | 21:51 |
| minute | BoostisBetter: the uboot update is already available but idk if reform-flash-uboot already knows about it cc josch | 21:51 |
| BoostisBetter | minute: gotcha, I asked him privately. | 21:51 |
| + savasten (~savasten@64.39.201.193) | 21:57 | |
| - aloo_shu (QUIT: Ping timeout: 244 seconds) (~aloo_shu@90.166.193.22) | 22:08 | |
| savasten | good day all, this can wait until Monday if people are busy relaxing this weekend. In an attempt to flash an updated system controller firmware meant to help with a black screen issue. i am stuck with an error from picotool stating No accessible RP2040/RP2350 devices in BOOTSEL mode were found. | 22:08 |
| savasten | command i am running is sudo picotool load build/sysctl.uf2 -f --vid 0x1209Â --pid 0x6d07 | 22:09 |
| BoostisBetter | savasten: why are you not using the update script instead of running picotool manually? | 22:10 |
| + aloo_shu (~aloo_shu@85.51.17.247) | 22:10 | |
| savasten | the script looks for a vendor id and product id that is no longer current after the previous update. | 22:15 |
| savasten | https://source.mnt.re/reform/pocket-reform/-/issues/4 | 22:15 |
| + jevans (~jevans@174-29-202-198.hlrn.qwest.net) | 22:17 | |
| savasten | └─[$] <> sudo ./update-sysctl-firmware.sh | 22:18 |
| savasten | picotool is already the newest version (2.0.0-3). | 22:18 |
| savasten | Summary: | 22:18 |
| savasten | Â Upgrading: 0, Installing: 0, Removing: 0, Not Upgrading: 0 | 22:18 |
| savasten | RP2040 System Controller device not found or more than one found, exiting. | 22:18 |
| - jevans (QUIT: Changing host) (~jevans@174-29-202-198.hlrn.qwest.net) | 22:21 | |
| + jevans (~jevans@user/jevans) | 22:21 | |
| minute | savasten: you can use fwupd with the cab file or you can use reform-mcu-tool | 22:25 |
| minute | (reform-mcu-tool does the initial reset instead of picotool now with our new usb vendor/product id) | 22:25 |
| minute | i'll try fwupd with the cab now on my system to double check if that works already... | 22:26 |
| minute | trying: fwupdmgr install sysctl.cab | 22:34 |
| minute | "no supported devices found" | 22:34 |
| minute | ah lol i forgot the usb jumpers on my pocket | 22:35 |
| minute | ah funny > Specified firmware is older than installed 'ga98e2bd < 20250107' | 22:37 |
| minute | fwupdmgr install --allow-older sysctl.cab | 22:38 |
| minute | ah > firmware signature missing or not trusted; set OnlyTrusted=false in /etc/fwupd/fwupd.conf ONLY if you are a firmware developer | 22:39 |
| minute | aha, now it does something | 22:40 |
| minute | > Successfully installed firmware | 22:40 |
| minute | savasten: ok so you can do: 1. sudo apt install fwupd 2. add the line OnlyTrusted=false to /etc/fwupd/fwupd.conf 3. fwupdmgr install --allow-older sysctl.cab | 22:41 |
| - aloo_shu (QUIT: Ping timeout: 252 seconds) (~aloo_shu@85.51.17.247) | 22:42 | |
| + aloo_shu (~aloo_shu@85.51.17.148) | 22:44 | |
| josch | minute: oh no... i totally missed that you released a new imx8mp u-boot git tag :( | 22:45 |
| minute | josch: aw, no worries | 22:46 |
| - savasten (QUIT: Ping timeout: 240 seconds) (~savasten@64.39.201.193) | 22:46 | |
| + savasten (~savasten@64.39.201.193) | 22:50 | |
| josch | hm... something is not right... | 22:52 |
| josch | minute: the build failed: https://source.mnt.re/reform/reform-imx8mp-uboot/-/jobs/7390 | 22:52 |
| josch | minute: and it strangely was started only an hour ago but the tag is from two days ago? | 22:52 |
| minute | josch: i reran it an hour ago because i saw that failure | 22:52 |
| minute | > EXTRAVERSION in u-boot/Makefile is different from that: | 22:53 |
| josch | oooh :) | 22:53 |
| minute | not sure what that was about | 22:53 |
| josch | one sec, i'll give you a MR with a more useful error message | 22:53 |
| josch | (and the fix for the error) | 22:54 |
| minute | oh nice | 22:54 |
| - savasten (QUIT: Ping timeout: 240 seconds) (~savasten@64.39.201.193) | 23:09 | |
| josch | minute: is this error message better describing what's going on? https://source.mnt.re/josch/reform-imx8mp-uboot/-/jobs/7394 | 23:11 |
| + savasten (~savasten@64.39.201.193) | 23:19 | |
| minute | josch: yes, much better, thanks! | 23:22 |
| josch | minute: hopefully even better: https://source.mnt.re/josch/reform-imx8mp-uboot/-/jobs/7399 | 23:26 |
| josch | minute: then you can merge this: https://source.mnt.re/reform/reform-imx8mp-uboot/-/merge_requests/5 | 23:29 |
| josch | it should be ready for you to put the tag 2025-01-12 to the top commit | 23:29 |
| josch | you have to do that manually as i cannot send you a tag via a MR | 23:29 |
| minute | josch: awesome, thank you | 23:29 |
| josch | sorry for the confusion | 23:29 |
| josch | minute: you set it to auto-merge (thank you) but it will have to fail with the latest commit | 23:30 |
| josch | because the latest commit expects there to be a tag 2025-01-12 but only you can add that to the reform/reform-imx8mp-uboot repository | 23:31 |
| minute | yeah i merged now immediately and will add a tag in the UI | 23:31 |
| josch | ah okay :) | 23:31 |
| + aloo_shu_ (~aloo_shu@85.51.17.85) | 23:32 | |
| - aloo_shu (QUIT: Ping timeout: 265 seconds) (~aloo_shu@85.51.17.148) | 23:32 | |
| * aloo_shu_ -> aloo_shu | 23:32 | |
| josch | yay it worked :) | 23:34 |
| josch | will be part of reform-tools 1.66: https://source.mnt.re/reform/reform-tools/-/merge_requests/110/diffs?commit_id=b5e5dfe5812f2dafce707983c8b0523a4037396f | 23:38 |
| - savasten (QUIT: Quit: Client closed) (~savasten@64.39.201.193) | 23:46 | |
| + savasten (~savasten@64.39.201.193) | 23:49 | |
| ch | > talking to the util-linux maintainer today motivated me to report this. | 23:56 |
| ch | what a way to start a bug mail! | 23:56 |
| ch | :) | 23:56 |
| josch | haha :D | 23:56 |
| josch | sorry XD | 23:56 |
| ch | (: | 23:56 |
| savasten | minute i followed the three steps. a few times even sudo once. after i run the command the system almost instantly powers off with just a flash of a new line under the command. then i need to cycle the power switch before it will power on. the firmware version does not seem the update. Current version:Â Â Â g2ab0d3a | 23:56 |
| savasten | should i be changing the jumpers on the main board? | 23:57 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!