libera/#maemo-leste/ Wednesday, 2025-01-08

Wizzupfreemangordon: hm, maybe daedalus src is just not properly up to date with debian bookworm branch00:45
gnarfacei think there usually isn't much more than an hour or two delay unless something has gone wrong, but stuff sometimes does...00:56
freemangordonWizzup: so, if you cloned xorg from daedalus then obviously daedalus use different xorg from bookwork07:07
gnarfaceoh, 3 of the xorg packages are forked actually07:08
gnarfacexserver-xorg-core, xserver-xorg-dev, and xserver-xorg-legacy all appear to have the "devuan" substring ni the version07:09
gnarface*in07:09
freemangordondo you know where devuan git repo for those lives?07:10
freemangordonyeah, 2:21.1.7-3+deb12u8devuan107:11
freemangordonhttps://git.devuan.org/devuan/xorg-server.git07:12
gnarfacethat's the right one, afaik07:13
freemangordonhmm, https://git.devuan.org/devuan/xorg-server/commits/branch/suites/daedalus-security/config/udev.c07:13
freemangordonthere reverts *are* there07:13
freemangordonthis is in security though07:13
freemangordonso, the question remains - what did Wizzup clone?07:14
gnarfacedebian source package maybe?07:15
gnarfacenot sure if the changes would show up there07:15
gnarfacethey might, i just don't know07:15
gnarfacedidn't really follow the whole backlog either, sorry07:16
freemangordonno, that's not a source package, we have a history in our git repo07:16
freemangordongnarface: here, in our repo https://github.com/maemo-leste-upstream-forks/xorg-server/commits/maemo/daedalus/config/udev.c07:16
freemangordonwe miss 3 "Revert "linux: ...." commits" that both bookwork and daedalus seem to have07:17
freemangordonsee here for example https://git.devuan.org/devuan/xorg-server/commits/tag/devuan/2%2521.1.7-3+deb12u7devuan1/config/udev.c07:17
freemangordonWizzup: latest TI SGX blobs are in the repo, the only thing remaining is to unravel the xorg mistery08:54
freemangordonalso, uvos' 'charge mode' UI does not work08:55
Wizzupfreemangordon: ok, but please provide some details on x10:43
Wizzuplike, do you know for a fact that this revert is in our repo?10:44
Wizzupbecause I don't see it on maemo/daedalus branch10:44
WizzupI type 'git log config/udev.c' and it does not show10:44
freemangordonWizzup: sorry, can't parse10:49
freemangordonthe issue is that we miss those reverts in our repo10:49
freemangordonwhat more details do you want me to provide?10:50
Wizzupok, you -want- the reverts10:50
freemangordonyes10:50
freemangordonwell, I 'want' our code to be in par with the one in daedalus10:50
Wizzupright10:50
freemangordonI have no idea if there are more differences10:50
Wizzup$ git rebase debian-bookworm10:51
WizzupCurrent branch maemo/daedalus is up to date.10:51
Wizzup:D10:51
freemangordonbut that's not correct10:51
Wizzupso git is lying10:51
freemangordonno, wait10:51
freemangordonyou are using wroing repo, to start with10:51
freemangordon*wrong10:51
Wizzuphttps://salsa.debian.org/xorg-team/xserver/xorg-server this is wrong?10:51
freemangordon https://git.devuan.org/devuan/xorg-server10:51
freemangordonno, see ^^^10:52
Wizzupdevuan has their own x changes?10:52
freemangordonexactly ;)10:52
Wizzupwe never used that before10:52
WizzupI always used X10:52
Wizzuper10:52
Wizzupdebian10:52
Wizzupbut ok10:52
Wizzupand than what, daedalus/security?10:52
Wizzupthen*10:53
freemangordon2:21.1.7-3+deb12u7devuan110:53
freemangordonsecurity and main repos have same version, according to apt-cache10:53
freemangordonoh, wait10:53
freemangordonno, in security is 2:21.1.7-3+deb12u8devuan110:54
freemangordonso yes, security10:54
Wizzupye10:54
WizzupI just rebased on this10:54
Wizzupso what is our suffix in the vresion10:54
Wizzup2:21.1.7-3+deb12u8devuan1 -> 2:21.1.7-4+7m4 or so?10:54
WizzupI normally just make it 21.1.7.0 or so10:54
freemangordon2:21.1.7-3+deb12u8devuan1+7m4 should be it10:55
freemangordonor10:55
WizzupI don't know if this will work in our CI10:55
freemangordon2:21.1.7-3+deb12u8devuan1leste110:55
Wizzupcan we even provide our own +?10:55
WizzupI think CI might just cut it off, I don't know10:56
freemangordonfor tag yes10:56
Wizzupat long as devuan can -never- override our x10:56
Wizzup:)10:56
Wizzupas long as*10:56
freemangordonwell, if you want to be sure, you have to increase the epoch as well10:56
WizzupI should have just done 21.1.7.0 :P10:56
WizzupI should have just done 21.1.7.3.0* :P10:57
Wizzupno, wait10:57
Wizzupjust waking up here10:57
Wizzupso I can do this: 2:21.1.7.0-1 or 3:21.1.7-110:57
Wizzupyou can pick :)10:58
freemangordonWizzup: we should use apt preferences to avoid devuan/debian upgrading our packages10:58
freemangordonwhat will happen if newer release appears in backports?10:58
freemangordonhmm, wait10:58
freemangordonupgrades ignore backports10:58
freemangordon3:21.1.7-1 is the proper one, however, we must somehow not that this is based on daedalus version10:59
freemangordon*note10:59
freemangordonthat's why 3:21.1.7-3+deb12u8devuan111:00
freemangordonand ci will add +7m411:00
freemangordonor 3:21.1.7-3+deb12u8devuan1leste111:00
freemangordonthis will become 21.1.7-3+deb12u8devuan1leste1+7m411:01
freemangordonand tag will be 21.1.7-3, UUIC11:01
freemangordonor, see what we do for mesa11:02
freemangordon22.3.6+sgx111:02
freemangordonthis is the tag11:02
freemangordonso, we can use 2:21.1.7+leste1-3+deb12u8devuan111:03
freemangordonWizzup: this ^^^11:03
freemangordontag will be 21.1.7+leste111:04
Wizzupok11:25
WizzupI think that's crazy but ok :D11:25
Wizzuplet's just hope CI will accept it11:26
Wizzupwhat else do we need to make a new d4 image?11:26
Wizzupit's running now (xorg-server)11:28
Wizzupstill don't see the reverts :D11:29
Wizzuphttps://git.devuan.org/devuan/xorg-server/commits/branch/suites/daedalus-security/config/udev.c11:29
Wizzupoh they are there11:30
Wizzupbut not in git log? wtf?11:30
WizzupI'm just lost11:30
WizzupI did 'git checkout -t devuan/suites/daedalus-security'11:31
Wizzupand the commits are not there for config/udev.c in 'git log'11:31
Wizzupgit branch --contains af4c84ce8855e84c0ad89b929bc972e884f0b8e311:33
Wizzupsays it's there11:33
Wizzupbut github doesn't show them11:34
freemangordonlet me clone it11:38
freemangordonlocal clone history shows tham11:43
freemangordon*them11:43
freemangordongithub not11:43
freemangordonthat's crazy :)11:44
freemangordonhmm, also, the code is not there it seems11:46
freemangordonbut lets see what will be build11:47
Wizzuplol, yeah I guess :D11:47
freemangordonhmm, lemme check another file11:48
Wizzupgit lo | grep af4c84ce8855e84c0ad89b929bc972e884f0b8e3 does show it11:50
Wizzupgit log11:50
freemangordoncommits are missing from here as well11:50
freemangordonhttps://github.com/maemo-leste-upstream-forks/xorg-server/commits/maemo/daedalus/hw/xfree86/os-support/linux/lnx_platform.c11:50
freemangordonreverts11:50
freemangordonbut, I see them in the local clone11:50
WizzupI think github is lying11:50
freemangordongithub went crazy11:50
freemangordonyeah11:50
Wizzupwe'll switch to something else eventually :)11:51
Wizzup-eventually-11:51
Wizzupoh, xorg-server source failed11:51
Wizzuplooks like some newline in gbp.conf11:51
Wizzupmisisng line11:51
freemangordonyeah, missing the section header11:53
freemangordon  p, li { white-space: pre-wrap; }  [DEFAULT]11:54
Wizzupadded it already11:57
Wizzupit's building now11:57
freemangordonoh, the current version in the repo is higner :(12:02
freemangordon*higher12:02
freemangordonWizzup: will have to remove xserver from the repos first12:03
Wizzupok,np12:12
Wizzupdone12:12
Wizzupgoing into repos now12:19
Wizzupanything else before I make a new img?12:19
Wizzuphuh12:20
Wizzup+ reprepro -b /srv/repository/leste -v --waitforlock 1000 --ignore=wrongdistribution processincoming daedalus xorg-server_21.1.7+leste1-3+deb12u8devuan1+4m7_amd64.changes12:20
WizzupError: trying to put version '2:21.1.7+leste1-3+deb12u8devuan1+4m7' of 'xnest-dbgsym' in 'daedalus|main|amd64',12:20
Wizzupwhile there already is the stricly newer '2:21.1.7.1-1+4m7' in there.12:20
Wizzupok, I see12:20
freemangordonWizzup: yeah, the new xorg pulls libseat13:02
freemangordonbut, still segfaults with the old omap driver13:02
freemangordonso, I will push a fix for it, no need to investigate what got broken13:03
mkfxhi13:08
freemangordonhi!13:09
Wizzupah, so you have a fix?13:15
freemangordonmhm13:16
freemangordonpushing13:16
Wizzupgreat13:16
freemangordonWizzup: we can use apt preferences.d to make sure our packages will not be overwritten13:18
freemangordonWizzup: ready13:20
freemangordonwe shall try fullscreen qt application with new blobs, to see if it still crashes the GPU13:21
freemangordonWizzup: hmm, wait, it didn't survive the reboot13:32
freemangordonlemme see what happens13:32
Wizzupok13:47
Wizzupmaybe your xorg server wasn't replaced13:47
Wizzupif you had the newer version13:47
freemangordonit was14:20
freemangordonwith the old version it was working14:20
freemangordonnot it does not work as user14:20
freemangordonas root it is ok14:20
freemangordonlemme poweroff to gather the logs14:21
Wizzupok14:30
freemangordon[    51.974] (EE) OMAP(0): ERROR: Cannot set the DRM interface version.14:47
freemangordonthis comes from   OMAPOpenDRMMaster()14:48
freemangordonhttps://github.com/maemo-leste/xf86-video-omap/blob/master/src/omap_driver.c#L15914:50
freemangordonno idea what's going on14:50
Wizzupand if you install your version back, does it work?15:15
freemangordonWizzup: my version of what? DDX driver?17:12
freemangordonno, it didn't, but I didn't check the error as I assumed it was segfaulting17:13
freemangordonhmm...17:13
freemangordonlemme check something17:13
freemangordonpermissions are ok17:14
freemangordonWizzup: hmm, elogind does not run17:30
inkyi've got my bionic.18:44
inkybattery was zero, it would not charge. it first showed some progress, but then i saw it stopped draw energy and the battery image disappeared from the screen.18:45
inkyi charge it in separate charger now, hope to have leste running tonight.18:45
sicelowhat's a simple, lightweight drawing program that I could run on N900? something that would allow me to work at pixel level?19:07
* sicelo still trying to work on maeotp icon19:09
Wizzupback19:54
Wizzupfreemangordon: as in, the service, or what?19:54
freemangordonyes, it is disabled as a service, but that shoudl nto be a problem19:55
freemangordonanyway, before pulling daedalus xorg, the same driver was working fine19:56
freemangordonit was working with the bookworm one19:56
Wizzuplol19:56
WizzupI think you mean you had daedalus org19:56
Wizzupxorg19:57
freemangordonyes, as I told you, it was booting stright to h-d19:57
Wizzupbut not maemo/daedalus xorg19:57
freemangordonno, the one we had in our repos and git was working properly with v 0.7.1 of ddx driver19:57
freemangordonafter that you rebased on devuan xorg19:58
freemangordonand now it works as root, but not as user19:58
Wizzuphmmm20:00
Wizzupthat does sound like elogind issues20:00
freemangordonyes20:00
Wizzupis elogind installed?20:01
Wizzupoh... I see20:01
freemangordonit is20:01
Wizzupdevuan changed debian/control to use 'libseat'20:01
freemangordonhmm?20:01
freemangordonyes20:01
Wizzupwhatever that is, as opposed to elogind20:01
freemangordondo we need seatd?20:01
freemangordonor something?20:02
Wizzupor we just use elogind20:02
WizzupI don't know what libseat is20:02
freemangordonmaybe we shall ask on #devuan20:02
Wizzupdepends on how much more you want to change things around :)20:03
Wizzupdevuan added debia/patches/* for libseat20:04
freemangordonyes, but why?20:04
Wizzupbecause they don't like elogind I guess20:04
Wizzupdon't know, up for whatever you prefer20:05
freemangordonwhat patches do you mean?20:05
freemangordonI see nothing in debian/patches20:05
freemangordonah, wait20:06
freemangordonopenat(AT_FDCWD, "/dev/dri/card1", O_RDWR|O_LARGEFILE|O_CLOEXEC) = 1020:42
freemangordonioctl(10, DRM_IOCTL_SET_VERSION, 0xbeb0be84) = -1 EACCES (Permission denied)20:42
freemangordonWizzup: ^^^20:42
Wizzupwhat are the perms on those?20:43
freemangordonfine20:43
Wizzupexlusive access problems?20:44
freemangordonno, seems like some DRM security20:45
freemangordonhmm https://elixir.bootlin.com/linux/v6.12.6/source/drivers/gpu/drm/drm_ioctl.c#L58520:46
Wizzupok, just say the word and I am happy to go back to elogind :D20:47
freemangordonafter reading https://github.com/kennylevinsen/seatd?tab=readme-ov-file I think it is better to spend some tme on it20:48
Wizzupok20:48
Wizzupdoes debian use libseat or is that a devuan only thing?20:48
freemangordonseems devuan only20:48
freemangordondebian is systemd :)20:49
Wizzupthen why did it work with elogind before?20:54
Wizzuphttps://packages.debian.org/bookworm/elogind20:55
freemangordonno idea20:55
Wizzupalso in trixie and sid20:55
Wizzupwhy don't we stick with elogind? it works on debian and devuan20:55
freemangordondevuan xorg uses libseat20:55
freemangordonI asked for help on both #devuan and #kennylevinsen20:56
freemangordonlets see20:56
Wizzupok, but then we can use debian xorg and have it work on devuan still?20:56
freemangordonsure20:56
freemangordonbut, if we do things like that, why not use upstream Xorg?20:56
freemangordonI mean - we can wait for a day, no? we have a working combo we can use if we can't solve the issue20:57
freemangordonBTW, I am not *that* sure issue is with libseat20:58
freemangordonmaybe elogind provides some more power to the user20:58
freemangordonlemme check something20:58
freemangordon[ 5556.003234] omapdrm omapdrm.0: [drm:drm_ioctl [drm]] comm="seatd" pid=4275, dev=0xe201, auth=1, DRM_IOCTL_SET_MASTER21:31
Wizzupfreemangordon: we don't us upstream xorg because then we don't have m-i-s21:44
Wizzupand yse, we can wait for a day21:44
freemangordonm-i-s?21:44
Wizzupfreemangordon: hah so seatd is occupying it?21:44
Wizzupmaemo-input-sounds21:44
Wizzupand also vibrations :)21:44
freemangordonwhy?21:44
Wizzupbecause their xrecord is broken21:45
freemangordonah21:45
Wizzupand I submitted a fix, and it was ignored21:45
freemangordonyeah, seatd makes itself drm master21:45
freemangordonbut it should drop it21:45
Wizzupif you use seatd, but we used to use elogind and session stuff around it21:49
WizzupI am not sure if whatever we use now is seatd aware21:50
arno11sicelo: btw with ofono2, 'set tech pref any' gives highest prio to gsm. weird, it was the opposit with ofono1 (higher prio to umts)22:29
siceloarno11, mmm sounds weird indeed. I'll also check from my end.23:13
arno11sicelo: ok23:19
siceloarno11, mine seems to be fine23:58
siceloplus it's on "any" by default23:58

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!