2026-04-08.log

+ f_ (16abab341f@postmarketOS/funderscore)00:02
- aloo_shu (QUIT: Remote host closed the connection) (~aloo_shu@90.166.193.71)00:06
joschminute: commit 11a2a193c6c764184a0f1303569345f7f954c4e6 is the culprit. I replaced the proprietary x86_64 aml_encrypt_g12b executable with gxlimg in November last year and this seems to work fine for classic Reform u-boot but not for pocket reform u-boot.00:16
+ aloo_shu (~aloo_shu@90.166.193.71)00:20
+ stephano (~stephano@71.238.14.13)01:08
minutejosch: awesome, thanks for figuring this out!01:15
minutecc f_01:15
joschrepk (gxlimg author) is in #linux-amlogic but i want to write up a proper issue report once i have a bit more of an idea of what is actually going on01:18
stephanojosch: looking into my build now. i thought i was calling build.sh, but if that's true, then I think that means I used gxlimg.01:41
- mjw (QUIT: Ping timeout: 272 seconds) (~mjw@gnu.wildebeest.org)02:06
+ pomel0 (~pomel0@user/pomel0)02:49
- pomel0 (QUIT: Remote host closed the connection) (~pomel0@user/pomel0)02:53
- wickedshell (QUIT: Ping timeout: 248 seconds) (~wickedshe@2601:8c0:c7c:3572:88d2:d101:12d9:4dc1)03:02
- paperManu (QUIT: Ping timeout: 255 seconds) (~paperManu@173.206.232.56)03:38
- paperManu_ (QUIT: Ping timeout: 248 seconds) (~paperManu@173.206.232.56)03:39
- stephano (QUIT: Quit: Textual IRC Client: www.textualapp.com) (~stephano@71.238.14.13)04:02
+ wickedshell (~wickedshe@2601:8c0:c7c:3572:67f9:1fab:6fc2:beb0)04:38
- ephase (QUIT: Read error: Connection reset by peer) (~ephase@82.66.198.11)05:02
+ ephase (~ephase@82.66.198.11)05:24
- ephase (QUIT: Read error: Connection reset by peer) (~ephase@82.66.198.11)05:57
+ oliverD (~Thunderbi@user/oliverd)06:13
+ ephase (~ephase@82.66.198.11)06:21
- oliverD (QUIT: Ping timeout: 255 seconds) (~Thunderbi@user/oliverd)06:23
+ oliverD (~Thunderbi@user/oliverd)06:26
- oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd)06:58
+ pomel0 (~pomel0@user/pomel0)07:25
+ oliverD (~Thunderbi@user/oliverd)07:53
- shebang (QUIT: Ping timeout: 248 seconds) (~shebang@136.51.81.5)08:02
- oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd)08:07
+ oliverD (~Thunderbi@user/oliverd)08:13
- oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd)08:15
- pomel0 (QUIT: Ping timeout: 248 seconds) (~pomel0@user/pomel0)08:45
- marty (QUIT: Ping timeout: 246 seconds) (~marty@146.70.185.34)08:54
+ marty (~marty@static-23-234-103-77.cust.tzulo.com)08:57
- rodion_goritskov (QUIT: Remote host closed the connection) (~rodion_go@mail.goritskov.com)09:09
+ rodion_goritskov (~rodion_go@mail.goritskov.com)09:13
- rodion_goritskov (QUIT: Remote host closed the connection) (~rodion_go@mail.goritskov.com)09:46
+ rodion_goritskov (~rodion_go@mail.goritskov.com)09:47
+ oliverD (~Thunderbi@user/oliverd)10:17
- oliverD (QUIT: Ping timeout: 248 seconds) (~Thunderbi@user/oliverd)10:57
- potash (QUIT: Quit: The Lounge - https://thelounge.chat) (~potash@user/foghorn)11:18
+ potash (~potash@user/foghorn)11:20
+ siviq (~siviq@user/siviq)11:22
- siviq (QUIT: Client Quit) (~siviq@user/siviq)11:23
+ mjw (~mjw@gnu.wildebeest.org)11:36
+ paperManu (~paperManu@204.244.186.187)12:25
+ paperManu_ (~paperManu@204.244.186.187)12:29
+ oliverD (~Thunderbi@user/oliverd)13:05
* mjw -> Guest249813:10
* Guest8326 -> mjw13:10
- Guest2498 (QUIT: Ping timeout: 248 seconds) (~mjw@gnu.wildebeest.org)13:11
- oliverD (QUIT: Ping timeout: 245 seconds) (~Thunderbi@user/oliverd)13:13
- jah (QUIT: Ping timeout: 244 seconds) (~jahkosha@user/jahkosha)13:25
- iank (QUIT: Quit: ZNC 1.8.2+deb2+deb11u1 - https://znc.in) (~iank@fsf/staff/iank)13:25
+ jahkosha (~jahkosha@user/jahkosha)13:25
* jahkosha -> jah13:26
+ iank (~iank@fsf/staff/iank)13:30
- paperManu (QUIT: Ping timeout: 246 seconds) (~paperManu@204.244.186.187)14:06
- paperManu_ (QUIT: Ping timeout: 246 seconds) (~paperManu@204.244.186.187)14:11
lidstahmmh, so, no display when booting openbsd with an hdmi screen plugged in (logical, but didn't cost anything to test)14:12
lidstahfun fact, however: now the LPC OLED displays only a white menu bar (circle-b/circle-s and so on works, or blindly selecting an invisible menu item works too)14:14
lidstahlpc rst button didn't correct this behavior either (and now, ofc, battery reading needs recalibrating if I'm not mistaken)14:15
f_what'd I miss?14:19
f_oops14:21
f_josch, minute: I'll try booting these on my bpi-cm4 then14:23
f_are there any full uart logs of broken v.s. working somewhere?14:27
f_quite weird though that the broken pocket reform image works on classic reform14:35
f_also, BL2 has no support for loading from NVMe14:39
f_oh wow it's getting so much weirder14:42
f_https://ircb.dersco.re/uploads/funderscore/f8188a91-f_-paste.txt it's failing to do anything with the fip header at all, before even dram init14:43
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)14:45
f_ah my SD card is borked sorry14:51
f_I can confirm that https://source.mnt.re/reform/reform-a311d-uboot/-/jobs/19489/artifacts/file/meson-g12b-bananapi-cm4-mnt-pocket-reform-flash.gxlimg.bin doesn't boot while https://source.mnt.re/reform/reform-a311d-uboot/-/jobs/19489/artifacts/file/meson-g12b-bananapi-cm4-mnt-pocket-reform-flash.bin is booting fine14:53
+ antti- (~antti@user/antti)14:56
- antti (QUIT: Ping timeout: 244 seconds) (~antti@user/antti)14:57
* antti- -> antti14:57
+ bkeys (~Thunderbi@134.22.115.162)14:59
f_there's an @AML header present in the aml_encrypt_g12b image but not in the gxlimg image at offset 0x86780... and a bunch of BL31-related things that are different I think15:01
+ bkeys1 (~Thunderbi@134.22.115.162)15:02
- bkeys (QUIT: Ping timeout: 265 seconds) (~Thunderbi@134.22.115.162)15:03
* bkeys1 -> bkeys15:03
joschf_: you are reading https://source.mnt.re/reform/reform-a311d-uboot/-/issues/1 ?15:04
joschf_: i put all artifacts (including the temporary ones) here: https://source.mnt.re/reform/reform-a311d-uboot/-/jobs/19495/artifacts/browse15:04
joschf_: the only difference between the two u-boots should be the defconfig and the dts. I cannot explain how either of these two could trigger a broken bootloader.15:05
joschf_: "quite weird though that the broken pocket reform image works on classic reform" -- it does not. Broken pocket reform image has a reset loop everywhere (also on classic reform). But classic reform u-boot can boot pocket reform (if boot.scr sets the correct device tree for linux because obviously $fdtfile will be wrong)15:07
f_ahhh misread15:10
f_ok thanks15:11
joschf_: i suspected that maybe the problem exists because the build.sh script builds the pocket refor mimage after the classic reform image and maybe some artifacts from the first build wrongly make it into the second. But the problem is the same even if i build the pocket reform image first.15:13
joschand the fact that swapping out gxlimg for aml_encrypt_g12b fixes it suggests that it's not a problem with the build script15:13
joschalso very weird that stephano has a working build which uses gxlimg when built on their own hardware and not MNT gitlab CI15:14
- bkeys (QUIT: Remote host closed the connection) (~Thunderbi@134.22.115.162)15:21
f_I suspect "BL33 CHK" to be a 'generic BL3X' error15:37
f_most of the diffs are in BL31 not BL33 (afaict)15:37
- paperManu (QUIT: Ping timeout: 248 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)15:45
+ mark_ (~mjw@gnu.wildebeest.org)15:49
+ paperManu (~paperManu@bras-base-mtrlpq0776w-grc-30-70-24-176-176.dsl.bell.ca)16:09
joschf_: do you think you see anything obvious? Otherwise I'd file an issue with glximg on github16:25
f_I'll compare the artifacts and see if something changes16:30
- mark_ (QUIT: Ping timeout: 245 seconds) (~mjw@gnu.wildebeest.org)16:30
f_because you say gxlimg boots on classic reform... there shouldn't be any changes between classic and pocket (apart from u-boot DTS)16:31
joschyes, exactly16:35
- chrcav (QUIT: Ping timeout: 265 seconds) (~chrcav@user/chrcav)16:35
joschyou should make the same observation on your bpi-cm416:37
+ chrcav (~chrcav@user/chrcav)16:41
f_I did16:42
+ vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50)17:33
- marty (QUIT: Ping timeout: 255 seconds) (~marty@static-23-234-103-77.cust.tzulo.com)17:37
+ marty (~marty@143.244.47.76)17:39
- xktr (QUIT: Ping timeout: 245 seconds) (~xktr@user/xktr)17:42
+ xktr (~xktr@user/xktr)17:43
- paperManu (QUIT: Ping timeout: 248 seconds) (~paperManu@bras-base-mtrlpq0776w-grc-30-70-24-176-176.dsl.bell.ca)18:00
+ bkeys (~Thunderbi@8.sub-97-181-39.myvzw.com)18:04
+ siviq (~siviq@user/siviq)18:14
- siviq (QUIT: Client Quit) (~siviq@user/siviq)18:15
+ pomel0 (~pomel0@user/pomel0)18:47
+ zmc (~zmc@user/zmc)18:52
joschthat's awesome, thank you!19:04
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)19:07
- bkeys (QUIT: Read error: Connection reset by peer) (~Thunderbi@8.sub-97-181-39.myvzw.com)19:07
- paperManu (QUIT: Ping timeout: 268 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)19:12
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)19:14
+ mark_ (~mjw@gnu.wildebeest.org)19:17
+ shebang (~shebang@136.51.81.5)19:17
+ hlep (~hlep@user/hlep)19:46
- mjw (QUIT: Killed (osmium.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d)19:54
* mark_ -> mjw19:54
+ mark_ (~mjw@2001:1c06:2486:a800:a09a:fc1c:5a8:e74d)20:01
- gachikuku_ (QUIT: Remote host closed the connection) (~gachikuku@kamil.gr)20:02
+ gachikuku_ (~gachikuku@kamil.gr)20:02
- shebang (QUIT: Quit: Konversation terminated!) (~shebang@136.51.81.5)20:09
- hlep (QUIT: Quit: nyaa~) (~hlep@user/hlep)21:01
+ voltaire28_ (~jlafon@28.162.2.93.rev.sfr.net)21:12
- paperManu (QUIT: Ping timeout: 248 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)22:56
+ paperManu (~paperManu@204.244.186.187)23:11
+ paperManu_ (~paperManu@204.244.186.187)23:13
joschminute, f_: updating to the latest gxlimg version from git fixes the a311d u-boot problem23:18
f_josch: huh23:23
f_sorry I got sidetracked :(23:23
joschf_: quite the opposite! I hope you didn't spend too much of your time on this23:24
f_I was still kinda curious23:24
joschhttps://source.mnt.re/reform/reform-a311d-uboot/-/jobs/19496/artifacts/browse23:24
joschthese are built with gxlimg and both work for me23:25
joschf_: I'm curious as well but I'm afraid I'll probably not find the time to figure out *why* this works now23:25
joschi've left in the temporary artifacts in this build in case anybody [tm] is curious enough23:25
joschbut i think it's enough for me to have found a solution23:26
f_It might have been this https://github.com/repk/gxlimg/commit/f157d7b147c47b6d2172f278e2c62d59136c7ffd23:26
f_looking at diffs I notice the @AML hdr wasn't actually missing, it was offset23:26
+ oliverD (~Thunderbi@user/oliverd)23:28
- oliverD (QUIT: Client Quit) (~Thunderbi@user/oliverd)23:30
- voltaire28_ (QUIT: Quit: WeeChat 4.8.1) (~jlafon@28.162.2.93.rev.sfr.net)23:30
+ oliverD (~Thunderbi@user/oliverd)23:30
joschminute: once either you or f_ or stephano have verified independent of me that this works I can merge this change and you can tag a new release23:31
joschoh no wait, apparently i have the necessary privileges to tag myself -- in that case i'm just waiting for another person to verify this. I'll ping stephano in the gitlab issue23:31
f_the pocket reform binary work for me23:32
joschperfect, thank you!23:33
f_you're welcome!23:35
- oliverD (QUIT: Quit: oliverD) (~Thunderbi@user/oliverd)23:47

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!