libera/#maemo-leste/ Tuesday, 2021-11-16

Wizzupfreemangordon: what about these:  -DLINUX -DSGX540 -DSUPPORT_DMABUF00:13
Wizzupbencoh: getting this btw: man: error while loading shared libraries: libmandb-2.8.5.so: cannot open shared object file: No such file or directory00:16
Wizzupmaybe some divert or lib missing00:16
WizzupLD_LIBRARY_PATH="$LD_LIBRARY_PATH:/amd64/usr/lib/man-db" man00:19
Wizzupthis works00:19
Wizzup(this is a problem because dh_installman errors)00:20
bencohI don't think you should need to do that, but I remember having some issue with mandb00:20
bencohdh_installman doesn't seem to fail here though00:20
bencohwhich package did you fail to build?00:20
Wizzupxf86-video-omap00:20
bencohfrom git?00:21
Wizzupfrom our git yeah00:22
Wizzupbencoh: does 'man' work from the shell?00:22
Wizzupfreemangordon: pushed to pvr-exa-pkg00:22
Wizzup(on xf86-video-omap)00:23
bencohWizzup: looks like it does00:23
Wizzupweird00:23
bencohI think it didn't work at some point, but it does now, even in the container I created yesterday00:23
bencohso I guess I fixed that somehow (?)00:23
Wizzupfreemangordon: so what needs fixing still is that the build needs pvr2d.so and libsrv_um.so, which are part of device-specific packages, so I don't know what to ship in -dev00:24
bencohdpkg-buildpackage: info: binary-only upload (no source included)00:24
bencohreal    1m26.293s00:24
Wizzupweird00:24
bencohusing the source from apt-get source00:24
WizzupI mean yeah it should work if man works :)00:24
Wizzupbut my 'man' command errors out00:24
Wizzupwhich debian also invokes00:24
Wizzupwith the LD_LIBRARY_PATH change it also builds for me00:24
bencohI think you're missing a divert because of some package that wasn't installed at the time you ran the divert lines00:24
Wizzuphm00:25
bencoh(back to the xorg build-dep thing)00:25
Wizzuphehe00:25
Wizzupcan I just re-run the divert lines or will that not work?00:25
bencohI wonder what would happen if you re-run it for stuff that was already diverted00:26
Wizzupright, that's what I mean00:27
bencohyou could try running the ones related to man/man-db00:28
bencoh(or you could re-create a container)00:28
Wizzup# /wrapper/amd64-divert.sh /usr/bin/mandb00:28
WizzupAmdifying /usr/bin/mandb ...00:28
WizzupLeaving 'local diversion of /usr/bin/mandb to /usr/bin/mandb.armhf'00:28
Wizzup/usr/bin/ln: failed to create symbolic link '/usr/bin/mandb': File exists00:28
Wizzupyeah I don't want to re-create it atm00:28
Wizzupah00:29
Wizzupit was one of the last ones, now it works00:29
Wizzup/wrapper/amd64-divert.sh /usr/lib/man-db/libman-2.8.5.so00:29
Wizzup/wrapper/amd64-divert.sh /usr/lib/man-db/libmandb-2.8.5.so00:29
bencohnice :)00:29
bencohoh btw, neat trick to get a decent working area: lxc.mount.entry = /home/bencoh mnt/bencoh none bind,optional,create=dir00:29
bencohthen create a user in container with the same uid00:29
Wizzupprobably not a bad idea yeah00:29
bencohthat way you can keep working from host as usual00:30
bencohand just build in lxc00:30
Wizzupyeah I'll do that if end up using this more often00:30
Wizzupit was great for the mesa stuff for sure00:30
bencoh:)00:30
Wizzupalso for the other pkgs00:30
bencohit could be even faster for mesa assuming I find a way to optimize meson without messing the platform/arch detection process00:31
bencohI wonder if that's do-able00:31
WizzupI don't think meson is really the slow part tbh00:31
Wizzupbut buggy it is for sure :)00:31
bencohactually it is00:31
Wizzupbencoh: my 16 cores were pegged all the time with gcc's00:31
bencohthe full build took 19mn here (i7 laptop / 4c / 8th), and I could see meson (python) take a few cores for a significant amount of time00:32
Wizzuphm, yeah, it was way faster here, but the initial setup took a bit of time (maybe 15s?)00:33
bencohapparently for mesa it runs stuff even between compilations00:33
bencohah, 15s isn't much then00:33
Wizzupdownside of not using automake I guess @ in between compilations00:33
bencohyeah00:34
bencohautotools-based builds are freaking fast since I diverted dash/m4/libtool and friends00:34
Wizzupuvos: this interesting to you? https://www.olx.bg/d/ad/motorola-droid-2-a955-nov-ne-polzvan-v-kutiyata-si-CID632-ID8aA3y.html?isPreviewActive=0&sliderIndex=300:52
uvosWizzup: a955, same as a855 except with 2x the ram. Also sim locked unless its the global variant (hard to tell) then its unlocked outside the us like xt894.00:56
uvos25euros is also about the going rate in germany of a953/2/0 that are not simlocked00:56
uvosso not usefull no00:56
uvosalso locked bootloader btw unlike a855 but like a85x00:58
uvosa955 = droid 2 a95x = milestone 200:58
uvosa855 = droid 1 a85x = milestone 100:58
Wizzupit's the global one it seems01:02
Wizzupon the photo01:02
freemangordonWizzup: hmm, maybe I shall dload()07:29
freemangordonand not link07:29
freemangordonyeah, that will solve the issue07:29
freemangordonor, build using d4 headers/libs07:30
Wizzupfreemangordon: just let me know10:08
tmlindd4 refused to start.. turns out i dropped it and the battery flex cable got torn. works fine after swapping the battery13:23
tmlindno pstore logs naturally :)13:24
sicelommm, the battery is loose then?13:39
Wizzuptmlind: hehe13:47
tmlindi had replaced the battery so not glued down with the tape, probably a piece of two sided tape under the battery would have saved it13:54
tmlindit fell corner first, i was mostly worried i had cracked the lcd13:55
tmlindconcrete floor with a thin carpet on top13:55
Wizzuplet's hope we can find a pallet and then we won't have to worry about dropping it, hehe13:55
tmlindyup, at least no need to pay 300 units for service..13:56
Wizzupfreemangordon: anything I can do now to help wrt packaging?13:56
uvosnew sphone is on the way, with old storage backend remove + new conversations view replaceing the old sms-history window.14:13
uvosjust a preview, there are some known issues, for one gtk_text_view_scroll_to_mark dosent work on the subclass hildon_text_view for some reason so the hildon variant dosent scroll down when a message arrives14:15
uvosalso sphone wont ring if the conversations view is showing the contact that the new sms came from, even if the window is not focused (will fix, just fyi)14:16
d4irchi14:33
dsc_:)14:33
Wizzupthis doesn't do the scrolling performance justice, but it's with ffmpeg capturing the desktop: https://wizzup.org/d4-output.webm14:40
Wizzupit's actually really smooth in the conversation view14:40
dsc_and the application starts quickly :)14:44
Wizzupwarm start at least is like 2s max, probably faster14:44
freemangordonn900?15:16
Wizzupd4, I didn't try it on the n900 yet15:16
Wizzupin any case it will be *running* in the background so the startup time is irrelevant mostly15:16
freemangordonah, right15:16
Wizzupbut qml scrolling was I think 3d accelerated15:16
freemangordonWizzup: re packaging - what else remains?15:17
Wizzupit was much smoother than say (currently) scrolling in the wireless list15:17
uvosyes it is15:17
uvos(so it widgets btw)15:17
freemangordonit should be, yes15:17
Wizzupfreemangordon: I think we need to package 5.15 kernel and build mesa in the repos with my patches15:17
freemangordonok15:17
freemangordonplease do :)15:17
Wizzupfor mesa, I think parazyd has problems with it needed libllvm from backports15:17
freemangordonwhy?15:17
Wizzupso we need to either import that libllvm or try to build it with grandma ancient version of libllvm15:17
freemangordonwhat is the issue with backports?15:18
WizzupI think that we can't just automatically get it from backports, and most people don'r have backports on15:18
Wizzupso we would have to download the debs from backports and add it to our own15:18
Wizzup+ repos15:18
freemangordonah, I see15:18
freemangordonwell, not big issue, right?15:18
WizzupI think so, but parazyd manages this so I tend to just do what he says is right :)15:19
Wizzupalso building the ddx in the repo I think15:19
Wizzupdid you see my branch of it?15:19
freemangordonddx is not ready15:19
freemangordonno15:19
freemangordonI mean - we need headers package first, which komes from kernel build15:19
Wizzuphttps://github.com/maemo-leste/xf86-video-omap/tree/pvr-exa-pkg15:19
freemangordon*comes15:19
Wizzupfreemangordon: no, it's in sgx-ddk-um-dev15:20
Wizzupyou can build it now with that package15:20
Wizzupsee pvr-exa-pkg branch15:20
freemangordonyes, I see it, but sgx-ddk-um-dev should come from kernel15:20
Wizzupmaybe I should not have put the headers there and we need to get it from the kernel somehow15:20
Wizzupyikes :(15:20
freemangordonyes15:20
* Wizzup doesn't like linux kernel packaging15:20
freemangordonwell, kernel shall be packaged anyways15:21
freemangordonone more package out of it should not be a big issue15:21
freemangordonso, please have kernel packaged first15:21
freemangordonwe are still discussing with Tomi how to have that TILER patch upstreamed15:22
Wizzupis there a branch that we can build now?15:22
freemangordonsure, tmlind's15:23
Wizzupthat has all your patches?15:23
freemangordonI think so15:23
freemangordonhmm, wait15:23
freemangordonplease for whatever seem appropriate, I'll push the latest there15:24
freemangordon*fork wherever15:24
WizzupI don't follow15:24
freemangordonthere are new version of the patches+patches I will sent later on that are not on tmlind's repo15:24
freemangordondo we have tmlind's repo cloned on leste?15:25
Wizzupyes but not 5.15 yet I think15:25
Wizzuphttps://github.com/maemo-leste/droid4-linux/tree/droid4-pending-pvr-omapdrm-v5.11 this is latest we have15:25
freemangordonok, I'll push 5.15 there15:26
Wizzupparazyd: can you look at building the kernel for experimental with tmlind's 5.15 stuff?15:26
Wizzupok15:26
freemangordonhow to name the branch? keep in mind I plan to force-push at least once :)15:32
Wizzuppvrsgx-wip-5.1515:33
freemangordonok15:33
Wizzupwe can always make copies of it15:33
freemangordonok15:33
uvoswe prbubly still want to keep pending-pvr seperate from pending-15:35
uvosie pvr contains only pvr + its chainges and pending contains changes that should land in mainline eventually15:36
uvosie droid4-pending-pvr-omapdrm-v5.15 = mainline v5.15 + pvr + hacks needed to make it work.15:37
Wizzupfor packaging it for our -experimental repo I don't think it matters15:37
uvosie droid4-pending-omapdrm-v5.15 = mainline v5.15 + stuff we want to upstream15:37
dsc_Excuse my spam16:07
dsc_but I need to test16:07
dsc_a few lines16:07
dsc_it will be over soon16:07
dsc_ok, test completed16:07
dreameralmost a haiku .. not quite16:07
dsc_dreamer: :P16:09
dsc_https://plak.infrapuin.nl/selif/8kbuwf3u.png16:09
dreamersuch maemo, very wow16:10
calebtheythem[m]damn, im kinda excited to get to play with this ui on a real device17:18
calebtheythem[m]i tried to boot maemo on the OnePlus 6 some time ago, but Xorg / mesa wasn't very happy about freedreno there lol17:19
parazydWizzup: Yeah, tonight, if it's ready.17:22
Wizzupcalebtheythem[m]: let me poke my buddy and see if he can ship the phone tomorrow or on thursday17:23
Wizzuphm, what didn't work wrt freedreno?17:23
calebtheythem[m]Wizzup: aha thanks, im in no rush17:24
calebtheythem[m]iirc it boots and managed to sort of display the wallpaper, but everything looked very broken17:25
Wizzuphm, interesting17:25
calebtheythem[m]that was way back on 5.9 though, sdm845 kernel support has come a long way since then (and i think you guys have newer aarch64 builds now?)17:25
Wizzupmaybe we need to see if we can debug this a bit more17:25
Wizzuphow did you try to boot maemo last time, was that on pmos?17:25
calebtheythem[m]if you have recent mesa it should basically just work17:26
calebtheythem[m]i grabbed an aarch64 rootfs from you (i think the raspberry pi 4 one), installed my kernel modules and firmware and booted it with a custom ramdisk17:27
uvosxorg did work with no issue on xt1602 for me some time ago17:27
uvoson pmos (dident try hildon )17:27
calebtheythem[m]i suspect things would be different now, I'll try and have a go some time this week17:27
Wizzuplet me know, I could try to help setting up a/the image17:36
mighty17[m]tmlind: i have https://github.com/Unlegacy-Android/android_kernel_ti_omap4/blob/3.4/common/arch/arm/plat-omap/include/plat/hardware.h as include in my sound driver, is it needed now in mainline?18:57
tmlindmighty17[m]: no that's old legacy stuff, drivers should not include it21:04
Wizzupparazyd: sorry, I missed what your reply was to, the kernel?21:09
parazydWizzup: Yes. If you have it ready, just give me links.21:11
WizzupI think it is ready21:11
Wizzupfreemangordon: oh, you didn't push?21:12
Wizzupfreemangordon: can you push?21:12
parazydWizzup: Also commit the mesa build fixes if you haven't already21:12
freemangordonnot now, I want to finish the patch first21:12
freemangordonparazyd: I dont; have any21:12
freemangordonbesides adding pvr to be build in debian/rules21:12
parazydI wrote to Wizzup, he had some21:12
freemangordonah, sorry :)21:13
Wizzupfreemangordon: so what kernel do you want us to build then?21:13
freemangordonWizzup: maybe wait till tomorrow21:13
freemangordonlets not hurry it, I need couple of hours to finish the patch21:13
freemangordonsorry about that, but see the last mail I sent to Tomi21:13
Wizzupso do we just to 5.15 pending then just to get the packaginb stuff going?21:14
freemangordonyes21:14
freemangordonmakes sense21:14
Wizzupparazyd: then this https://github.com/tmlind/linux/tree/droid4-pending-v5.1521:14
freemangordonuse tmlind's 5.1521:14
parazydack21:14
freemangordononce I have the patch ready, I will push them on top21:14
Wizzupyeah if you can do 5.15 now, then I can add patches later parazyd21:15
parazydWizzup: Should we have a new omap-linux for this or just use the same droid4 repo?21:20
Wizzuplet's use the droid repo for now and push it to experimental21:20
parazydWhat about all these patches? https://github.com/maemo-leste/droid4-linux/tree/maemo/beowulf-devel/debian/patches21:24
uvoswe need:21:25
uvosuart-with-charger.patch21:25
uvosfbdev.patch21:25
uvos0001-ASoC-simple-card-utils-add-support-for-componants-pr.patch21:26
uvos0002-ARM-dts-motorola-mapphone-add-interrupt-for-headphon.patch21:26
uvosand the defconfig changes21:26
uvosand im not sure about bus-ti-sysc-fix-flakey-idling-of-uarts.patch21:27
uvospobubly thats in 5.15 allready21:27
uvosmostly only 0001-power-cpcap-battery-add-a-module-option.patch and very likely bus-ti-sysc-fix-flakey-idling-of-uarts.patch can be droped21:29
uvosand 0002-allow-busy-ret.patch21:30
parazydok21:33
uvosmight want to also drop anbox-support.patch21:35
uvosnot sure what is the point of this on omap421:35
Wizzuplet's keep it for now21:38
Wizzupwe have it packaged anyway, and with maybe we cna make it work now21:38
uvossure building the modules isent a big deal21:40
* freemangordon is afk22:00
parazydWizzup: It's ready22:34
parazydhttps://github.com/maemo-leste/droid4-linux/tree/maemo-5.1522:34
parazydhttps://github.com/maemo-leste/droid4-linux/tree/maemo/beowulf-experimental22:34
parazydHad to manually rebase a few patches, so apologies for any potential build errors in advance :p22:35
Wizzupparazyd: check, ty, which pathes?22:36
parazydhttps://github.com/maemo-leste/droid4-linux/commit/75ee1aefb2856414dc99e3d14b090fc5010f5d1522:37
parazydMainly 0001-ASoC-simple-card-utils-add-support-for-componants-pr.patch22:38
Wizzupuvos: can you check if that's ok - I think it's yours?22:38
Wizzupparazyd: great :)22:38
parazydI tagged the release so I think you can just build it22:39
parazydWhenever22:39
Wizzupshall we just try on jenkins?22:41
parazyddoit.jpg22:41
parazyd:D22:41
WizzupI started it I think22:50
parazydok, it'll take a while :)22:52
ashleyit's just exciting and heartwarming to see Maemo development happening -- over a decade (!) after Fremantle was released :) thank you everyone for your hard work!22:58
parazyd<322:59
Wizzupashley: we're all very motivated to use it as our daily driver :D23:03
ashley\o/ I'd love to use it as my daily driver as well, but alas, the N900 is 3G-only and 4G speeds are sadly a much-needed feature for me since these days I browse the Internet more than I make calls etc. but I've still considered picking up a N900 or couple for software development purposes...why do the 4G-enabled QWERTY smartphones need to be so darn expensive?!23:04
Wizzupyeah LTE would be nice to have23:06
parazydShit I should do the fxtec bringup sometime23:07
Wizzupparazyd: yes23:07
Wizzuplol23:07
Wizzupother stuff is more important atm though23:07
ashleyindeed; better to have one (or a couple) devices as functional as possible before starting to work or continuing existing work on porting it to other devices23:10
WizzupI really hope we can find a lot more droid4 devices23:18
dreamerthey're shutting down 3G here in a couple months :(23:26
dreamerI thought N900 was "2.5G" only?23:26
Wizzupdreamer: KPN is, t-mobile is not23:28
Wizzupbut yeah they will keep it for 1 or 2 more years max I guess23:28
ashleydroid4...motorola asanti? aka "that one cool 4G QWERTY phone that America got but we here in Europe didn't, but it's not like it'd work with European 4G bands"?23:35
Wizzupyes, I think it's maserati23:35
Wizzup(not asanti?)23:36
Wizzupbut you're otherwise correct23:36
Wizzuphttps://leste.maemo.org/Motorola_Droid_4 this one23:36
MartijnBraam[m]asanti is the photon q23:36
ashleyaah; nice!23:38
Wizzuphttps://wizzup.org/conversations-wip.mp423:41
Wizzupparazyd: ^23:42
WizzupI'll make a better video later, but I promised dsc :D23:43
parazydLooks lovely!23:49
Wizzupyeah it's smooth23:51
Wizzupthere are also themes btw but the camera was too annoying to hold to show more23:52
uvosWizzup: very nice23:52
uvosparazyd: patch looks fine23:52
parazydYeah it'll build now23:54
parazydI merged in this branch https://github.com/openpvrsgx-devgroup/linux_openpvrsgx/tree/letux-pvrsrvkm-5.15-rc123:54
parazydInto droid4-pending-v5.1523:54
parazyd(Hope that's the right one)23:54
uvosparazyd: i would/ i do merge https://github.com/tmlind/linux_openpvrsgx/tree/droid4-pending-pvr-omapdrm-v5.15 instead23:57
Wizzupoh yeah23:57
parazydugh23:58
parazydok23:58
* parazyd adds it into todo.md23:59

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