jk_0 | https://pastebin.com/MZUYRxP8 | 00:45 |
---|---|---|
jk_0 | I managed to get ps xua for you | 00:46 |
jk_0 | Wizzup: https://pastebin.com/JbdkakwY | 01:00 |
jk_0 | and that is pstree -l :) | 01:00 |
freemangordon | yesss, 2 kernel patches later and h-d works in landscape :) | 08:49 |
uvos | freemangordon: what kind of patches? | 08:50 |
freemangordon | in omapdrm driver (to correctly set GEM buffers size) and pvr driver (to assume BOs sg memory size to be page size aligned) | 08:51 |
uvos | ok | 08:51 |
freemangordon | will post them later on so you'll see | 08:51 |
uvos | great | 08:53 |
buZz | aligment to get TILER working? | 08:59 |
freemangordon | alignment to get SGX render on TILER BOs | 09:00 |
buZz | nice :) | 09:04 |
buZz | will droid4's hw ever allow dual outputs? screen + hdmi? | 09:04 |
buZz | non-mirrored | 09:05 |
uvos | yes | 09:05 |
buZz | omg :) exciting | 09:05 |
uvos | works fine atm (well not in hildon but thats h-ds fault) | 09:05 |
uvos | this is btw also how motorola used it | 09:05 |
buZz | apt install pd gem gl4es + that working makes droid4 almost a viable device for video performances | 09:05 |
uvos | android on the phone display + x11 on the hdmi display | 09:05 |
buZz | hmhm, but only on their initial firmware i think? i should really fix my lapdock soon, keyboard is all mushy | 09:06 |
uvos | yes only on 2.3 | 09:06 |
uvos | later its just android | 09:06 |
buZz | lol > Lot of 5.Motorola Lapdock 100 Droid Razr Cracked screen, NP. Parts or repair | 09:08 |
buZz | 35 euros ex everything | 09:08 |
buZz | :P | 09:08 |
parazyd | uvos: osso-icons | 09:09 |
parazyd | (Unsure if you got your reply) | 09:09 |
uvos | parazyd: no thanks | 09:17 |
freemangordon | tmlind: is there a way to set /dev/dri cards order? | 09:35 |
freemangordon | or, this shall be done through udev? | 09:35 |
Wizzup | parazyd: I thought could also ship their own icons? | 09:38 |
parazyd | Wizzup: Yeah any program can, as long as it's installed under the same (hicolor I think) directory | 09:41 |
Wizzup | freemangordon: parazyd: I'll continue on the packaging today, hope to get it done in between some work | 09:45 |
parazyd | ok | 09:45 |
parazyd | lmk how I can help | 09:45 |
Wizzup | mesa, or another pkg from here: https://github.com/maemo-leste/bugtracker/issues/524#issuecomment-968103934 | 09:46 |
Wizzup | I am doing the sgx ddk stuff and headers from xf86-video-omap | 09:46 |
Wizzup | and it should all go into -experimental | 09:46 |
Wizzup | there's also a 5.15 kernel build for -experimental based on latest droid4 pending + maybe the patches fmg mentioned | 09:47 |
Wizzup | maybe the least painful thing to do is to try with latest mesa release without pvr for -devel, and then we can try to add the pvr stuff on top of that | 09:48 |
Wizzup | I'm setting up a cross container with mostly native performance using bencoh's method for stuff as well | 09:48 |
parazyd | Alright, will check after lunch today | 09:49 |
Wizzup | just poke me then, I might have been looking at mesa already | 09:49 |
Wizzup | bencoh: fakeroot-sysv is not installed after the debootstrap for amd64 it seems | 09:59 |
Wizzup | chrooting to the amd64 one and installing it | 10:00 |
Wizzup | bencoh: iirc I also had to install devuan-keyring on the armhf container | 10:04 |
uvos | is HildonTextView thumb scrollable? | 10:27 |
bencoh | Wizzup: if you followed the steps it should be | 10:46 |
bencoh | Wizzup: I tried yesterday | 10:47 |
Wizzup | it wasn't installed, but I installed it manually | 10:47 |
Wizzup | pretty sure I followed all the steps this time | 10:47 |
Wizzup | assuming you're taking fakeroot-sysv | 10:47 |
bencoh | but I updated it with a patched fakeroot-sysv script to allow working with both armhf and amd64; I dunno if the version you have is patched | 10:47 |
Wizzup | talking* | 10:47 |
bencoh | Wizzup: including the xorg build-dep? | 10:47 |
Wizzup | bencoh: you're right, that's the one I skipped... :) | 10:48 |
bencoh | :) | 10:48 |
Wizzup | in any case it works now | 10:48 |
bencoh | yay | 10:48 |
Wizzup | I haven't tried to build anything yet, but everything else seems ok | 10:48 |
Wizzup | (I need to add leste sources and such) | 10:48 |
Wizzup | btw the xorg dep is after the fakeroot-sysv so I can't matter for that | 10:48 |
Wizzup | maybe it's a debootstrap on gentoo vs debian thing | 10:49 |
bencoh | ? | 10:49 |
bencoh | ah | 10:49 |
Wizzup | right, with regards to fakeroot not being installed | 10:49 |
bencoh | I just realized the build-deb thing is in the armhf container, not in the chroot | 10:49 |
Wizzup | yeah, but the sed on the fakeroot binary is also before that step | 10:50 |
Wizzup | the only things I had to change wrt the notes was (1) suspend lxc create and install qemu before it needs it (2) install fakeroot in amd64 container/chroot (3) install fakeroot in armhf | 10:51 |
bencoh | Wizzup: the build-dep happens in armhf, so basically it isn't related in any way, I was mistaken :) | 10:52 |
bencoh | so we should probably add it to the list of stuff to install | 10:52 |
bencoh | (for the amd64 chroot) | 10:53 |
Wizzup | probably yeah | 10:54 |
Wizzup | but then the order needs to change still | 10:55 |
Wizzup | as in the sed needs to then happen after the install | 10:55 |
bencoh | Wizzup: the sed is amd64-related, the build-dep is armhf-related | 10:55 |
Wizzup | yes | 10:56 |
Wizzup | but if you meant to add fakeroot to this: | 10:56 |
Wizzup | chroot /amd64 | 10:57 |
Wizzup | apt-get install bash-completion bash-static build-essential debhelper file flex bison gcc-8-arm-linux-gnueabihf g++-arm-linux-gnueabihf locales man-db quilt perl ninja-build meson xsltproc | 10:57 |
bencoh | ah, I see what you mean now :) | 10:57 |
Wizzup | then you need to still run the sed to create fakeroot-sysv-patched after that | 10:57 |
bencoh | those should go to debootstrap as well | 10:57 |
bencoh | I haven't changed it yet | 10:57 |
bencoh | (debootstrap can take a list of packages to install) | 10:57 |
Wizzup | ah check | 10:57 |
Wizzup | brb | 10:58 |
bencoh | tbh I'm feeling lazy about it, but I should make that thing a proper script | 10:58 |
Wizzup | then we could put it on the wiki, yeah | 11:04 |
bencoh | we can probably ship it as a tarball meanwhile (that's what scratchbox used to do) | 11:06 |
Wizzup | parazyd: at least I think having an -experimental repo ready would be useful | 11:36 |
parazyd | Wizzup: It should be working already | 11:52 |
Wizzup | ok | 11:52 |
Wizzup | so there's nothing in there atm that we need to purge first? | 11:52 |
parazyd | Lemme check | 11:52 |
parazyd | There's only a 5.10 droid4 kernel | 11:53 |
parazyd | Should I remove it? | 11:53 |
Wizzup | yeah although ir probably doesn't matter | 11:57 |
parazyd | Sure, if you build a new one it'll be replaced. | 11:59 |
parazyd | (And the -devel version is newer anyway) | 12:00 |
parazyd | Wizzup: ok if I nuke the current mesa repo? We add no patches, it's just newer than beowulf. | 12:12 |
parazyd | I'm gonna try to make it more easier to maintain | 12:12 |
dreamer | parazyd: <3 gl4es package. works great ootb so far :) | 12:13 |
parazyd | oh nice | 12:13 |
dreamer | (buZz and I are testing with pd+Gem) | 12:13 |
parazyd | lmk if I should pull new stuff from upstream | 12:14 |
dreamer | on my pandora I have the upstream version, but the rest of that OS is so incredibly outdated :) | 12:14 |
dreamer | :( * | 12:14 |
* dreamer can't wait to one day figure out how to boot devuan on the pandora .. | 12:14 | |
Wizzup | parazyd: define current | 12:14 |
Wizzup | parazyd: we need to fork mesa for pvr stuff | 12:15 |
Wizzup | so removing the repo doesn't make sense to me | 12:15 |
parazyd | Wizzup: https://github.com/maemo-leste-upstream-forks/mesa | 12:15 |
Wizzup | yes, we need it | 12:15 |
parazyd | I just want to start from scratch | 12:15 |
Wizzup | maybe save the branch somewhere | 12:15 |
Wizzup | but yeah sure feel free to hack on it | 12:15 |
parazyd | *nod* | 12:15 |
Wizzup | I think bencoh's container thing should make this much more painless to test btw | 12:16 |
Wizzup | much less painful* | 12:16 |
Wizzup | I volunteer to test a build later today if you want | 12:21 |
parazyd | Started working on it now | 12:23 |
Wizzup | *nod* | 12:23 |
Wizzup | cool :) | 12:23 |
Wizzup | ideally we could build 21.2.5 | 12:23 |
Wizzup | and add the pvr patches on top of that | 12:23 |
parazyd | Yep | 12:24 |
Wizzup | since that would help lima devices (presumably) | 12:24 |
parazyd | That's my plan | 12:24 |
Wizzup | cool | 12:24 |
parazyd | Need to see if I have to update glvnd and libdrm first | 12:24 |
Wizzup | right | 12:24 |
uvos | heh so i made a thread view for sphone | 12:44 |
uvos | and it crashes xorg with 100% precision when rotated with it shown | 12:44 |
uvos | on d4 | 12:44 |
uvos | like him dose soemtimes | 12:44 |
Wizzup | well that's good news in the sense that we have a reproducible bug | 12:44 |
uvos | i was looking for a repoducable case for this | 12:45 |
uvos | yeah | 12:45 |
uvos | but i think fmg is just gona save me at this point | 12:45 |
Wizzup | btw, parazyd just got the reboot on upgrade again and it looks like mce cannot get dbus when it restarts and then the device reboots after 10 tries | 12:45 |
uvos | with ddk1.17/new -video-omap | 12:45 |
Wizzup | we're trying to see if the dbus addrs in /tmp/ get clobbered somehow | 12:45 |
Wizzup | what is threaded view for sms even though? | 12:46 |
Wizzup | afaik there is no way to 'respond to a thread' ? | 12:46 |
uvos | sure there is | 12:46 |
uvos | Wizzup: like android | 12:46 |
Wizzup | I don't understand | 12:46 |
uvos | well every contact has one thread, you can have multiple threads like email foc | 12:47 |
uvos | *ofc | 12:47 |
Wizzup | so what's a thread about it? | 12:47 |
Wizzup | or do you mean what maemo does as well | 12:47 |
uvos | im not describing it correctly | 12:47 |
uvos | its like android or telegram or irc or whatever works | 12:47 |
uvos | there are rooms | 12:47 |
uvos | with recipiants | 12:47 |
parazyd | You mean "groups" ? | 12:49 |
uvos | yeah execpt a group is just one perso | 12:50 |
uvos | since its sms | 12:50 |
parazyd | Right | 12:50 |
uvos | (only available backend atm anyways - its backend selectable) | 12:50 |
bencoh | maemo has no real "thread" support (at least in Conversation), but it has "conversations" | 12:50 |
uvos | yeah sphone is same now | 12:51 |
uvos | i just dont know the terminology | 12:51 |
bencoh | then it's a good start I'd say :) | 12:51 |
dsc_ | did someone say conversations | 12:52 |
Wizzup | heh | 12:52 |
dsc_ | ;) | 12:52 |
bencoh | :) | 12:52 |
lel | parazyd deleted a repository: https://github.com/maemo-leste-upstream-forks/libglvnd | 12:53 |
lel | parazyd created a repository: https://github.com/maemo-leste-upstream-forks/libglvnd | 12:53 |
lel | parazyd edited a repository: https://github.com/maemo-leste-upstream-forks/libglvnd | 12:54 |
lel | parazyd deleted a repository: https://github.com/maemo-leste-upstream-forks/libdrm | 12:58 |
lel | parazyd created a repository: https://github.com/maemo-leste-upstream-forks/libdrm | 12:58 |
lel | parazyd edited a repository: https://github.com/maemo-leste-upstream-forks/libdrm | 12:59 |
uvos | hehe the moment a bot is reporting on your mistakes in real time | 13:00 |
parazyd | I'm just resetting the repos to a clean slate. | 13:01 |
parazyd | Actually the latest unstable builds just fine on beowulf | 13:01 |
parazyd | Now onto mesa | 13:02 |
uvos | oh ok | 13:02 |
freemangordon | with PVR EXA disabled, d4 boots to hildon-desktop :) | 13:07 |
tmlind | nice, yeah i think the dri nodes are created by udev | 13:10 |
freemangordon | I think we shall create some udev rules (or whatever) so card0 to be omapdrm, not pvr | 13:16 |
uvos | ENV{ID_PATH}=="?*", KERNEL=="card*", SYMLINK+="dri/by-path/$env{ID_PATH}-card" | 13:17 |
uvos | overide it with something that links to dri/by-path/$(whateveromapdrm ist) | 13:18 |
freemangordon | uvos: please, could you do that | 13:18 |
uvos | sure | 13:18 |
freemangordon | we need that for d4 and Nokia devices only | 13:19 |
freemangordon | well, not really | 13:19 |
uvos | do we want the sgx node at all? | 13:19 |
freemangordon | we need it for devices with SGX GPU :) | 13:19 |
uvos | or is it non operatable | 13:19 |
freemangordon | uvos: no idea | 13:19 |
uvos | ok | 13:19 |
freemangordon | I can try without it later on | 13:20 |
uvos | we need the sgx node for kmscube/plain kms/drm applicaitons or? | 13:20 |
freemangordon | and, it seems we'll have to rename pvr_dri to omapdrm_dri | 13:20 |
freemangordon | don;t know | 13:20 |
freemangordon | and cannot test ATM | 13:20 |
uvos | ok | 13:20 |
uvos | ill do it later | 13:21 |
freemangordon | thanks! | 13:22 |
dsc_ | Re: my droid4, I'm now on an USB wall charger, when I connect the d4 it turns on, into kexecboot, but quickly turns off again | 13:33 |
dsc_ | I am getting the feeling it is not actually charging | 13:33 |
dsc_ | *or* kexecboot automatically picks an item from the list (maemo) and turns black because my OS is broken (?) | 13:34 |
Wizzup | I'd just leave it on the wall for ~30 mins | 13:34 |
dsc_ | yeah, it wont boot into stock android too | 13:35 |
uvos | dsc_: pick stock android | 13:35 |
uvos | ok | 13:35 |
uvos | its to empty | 13:35 |
dsc_ | ok | 13:35 |
parazyd | Can we somehow protect against this happening? | 13:36 |
uvos | no | 13:36 |
uvos | well we do | 13:36 |
uvos | with mce | 13:36 |
parazyd | How does Android do it? | 13:36 |
uvos | but if it happes becasue the device is off for a long time and self0discarges we cant protect | 13:36 |
uvos | so the probem is: | 13:36 |
uvos | cpcap charges by itself untill some voltage is accived (unless the voltage is below 2.5 or so volts then it wont charge beause its unsafe) | 13:37 |
uvos | then cpcap boots the device | 13:37 |
uvos | android immidatly enters its charge mode | 13:37 |
uvos | but with kexecboot, kexecboot loads first | 13:37 |
uvos | the battery is then below androids emergency threshold by the time it starts | 13:38 |
parazyd | aha | 13:38 |
uvos | so its too late and it powers off again | 13:38 |
dsc_ | some sort of visual indicator it is charging at least would be nice somehow | 13:39 |
parazyd | The green LED should be on | 13:39 |
uvos | the green led is cpcap's hw signaling charge | 13:39 |
dsc_ | it is not | 13:39 |
parazyd | fwiw I sometimes charge it with a DC-only USB cable (no data pins) and it helps | 13:40 |
parazyd | Not sure why | 13:40 |
parazyd | I think it works for buZz too | 13:40 |
uvos | if the id pin is not pulled low | 13:40 |
uvos | cpcap dosent boot ever | 13:40 |
uvos | so it just charges while off | 13:40 |
dsc_ | I might have an USB cable that is not capable of delivering power (?) | 13:41 |
dsc_ | :D | 13:41 |
uvos | no | 13:41 |
uvos | you might have an usb calble thats capapble of delivering data | 13:41 |
dsc_ | right | 13:41 |
uvos | anyhow if this happens i just remove the battery and give it a bit of juice | 13:41 |
uvos | i gues you dont have anything to do that | 13:41 |
dsc_ | correct | 13:42 |
uvos | dsc_: ok try what Wizzup said, maybe the charge led is just off because the android kernel wrote the reg to turn this feature off before it powered down | 13:43 |
ruleh | try plugging it into a pc usb port, seems to work for me | 13:43 |
uvos | otherwise your out of luck i think | 13:43 |
tmlind | ruleh: hi again! | 13:43 |
dsc_ | kk | 13:43 |
ruleh | HI :) | 13:43 |
tmlind | the cpcap charger stuff should sort of behave nowadays | 13:44 |
uvos | tmlind: this is about cpcaps trickel charge behavior before boot | 13:44 |
ruleh | the low battery charging seems to be highly charger dependent for some reason | 13:44 |
tmlind | uvos: right, sorry that was for ruleh to summarize the status since few years ago.. | 13:45 |
uvos | oh ok :) | 13:45 |
ruleh | there is also this situation where the device gets power over usb to run things like fastboot but doesn't actually charge the battery. | 13:47 |
uvos | ruleh: sure yeah for a855 there was the "factory" cable that powerd the device over usb and booted fastboot only | 13:49 |
uvos | not sure if its implemented on xt894 | 13:49 |
uvos | looks like this was with pullup on the id pin | 13:50 |
tmlind | hmm yeah fastboot cable with id pin pulled up will always power up the device using usb bus, even if no battery. not sure booted stock os charges battery though | 13:51 |
uvos | ok on a855 this forces fastboot mode, no way to boot stock os. so thats not how xt894 works, interesting | 13:52 |
ruleh | would an usb otg cable do the same? | 13:53 |
uvos | no | 13:53 |
uvos | otg pulls the pin down | 13:53 |
parazyd | uvos, Wizzup: btw I just got 7.5G of Nov 15 12:53:53 localhost maemo-launcher[5733]: error accepting connections (Operation not supported) | 13:54 |
parazyd | In /var/log/syslog | 13:54 |
parazyd | It keeps filling up | 13:54 |
uvos | neat | 13:54 |
parazyd | lol | 13:54 |
* parazyd reboots | 13:55 | |
bencoh | best way to test the sdcard ever | 13:55 |
parazyd | It was in the VM, but regardless | 13:56 |
Wizzup | parazyd: maybe restarting maemo-launcher is problematic? | 13:56 |
uvos | my /var/log/syslog looks fine | 13:57 |
uvos | (on d4) | 13:57 |
parazyd | Perhaps | 13:57 |
parazyd | I'm trying to build mesa so I don't want to restart services atm | 13:57 |
uvos | parazyd: this isent wasent on the same boot as where you upraged maemo-launcher? | 13:57 |
parazyd | No, I already got ke-recv or something reboot my VM when I did a dist-upgrade | 13:58 |
bencoh | is dist-upgrade supposed to work nowadays? | 14:01 |
uvos | yes but dsme/ke-recv broke it some time ago | 14:01 |
Wizzup | bencoh: upgrade is mostly dist-upgrade for our purposes, unless you meant actually moving from one release to another | 14:11 |
buZz | parazyd: yep, blocking data pins on a USB cable makes charging (from zero) a -lot- easier | 14:26 |
dreamer | blocking? | 14:39 |
bencoh | shorting to ground iirc | 14:40 |
bencoh | (that's what chargers should do) | 14:40 |
bencoh | (not sure if it's really shorting it to ground or having a pull-down resistor though) | 14:40 |
freemangordon | uvos: do you want me to cc you in TILER/omapdrm mails I send to Tomi? | 15:06 |
freemangordon | Wizzup: same question to you | 15:06 |
freemangordon | though, linux-omap is in CC so you should already have them | 15:07 |
uvos | freemangordon: its unlikely that i can add anything, but i am happy to follow along. | 15:07 |
freemangordon | yeah, it is not about helping, just FYI | 15:07 |
uvos | and im on linux-omap only in with the spamy email | 15:07 |
uvos | so yes | 15:07 |
freemangordon | ok, will cc | 15:08 |
Wizzup | freemangordon: yes please | 15:08 |
freemangordon | ok | 15:08 |
freemangordon | sicelo: hmm, who is that guy? should I know him? | 15:11 |
freemangordon | I mean - I follow the discussion on the ML, but not sure shall I do anything about | 15:12 |
buZz | you mean pavel? | 15:17 |
buZz | oh different ML i guess | 15:18 |
buZz | bencoh: no, just not connecting the datapins | 15:18 |
buZz | so, left floating | 15:18 |
bencoh | shorting D+/D- together? | 15:18 |
freemangordon | buZz: I mean Lucas Fryzek | 15:18 |
buZz | i dont know this person :) | 15:19 |
buZz | bencoh: i have a 'usb condom' that only allows GND and +5V to connect through | 15:19 |
buZz | the datapins are unconnected | 15:19 |
bencoh | ah, that | 15:19 |
buZz | i think i even pulled em out of the usb plugs? | 15:19 |
Wizzup | freemangordon: | 15:21 |
Wizzup | glmark2 Score: 92 | 15:21 |
Wizzup | neat | 15:21 |
freemangordon | yeah :) | 15:24 |
freemangordon | Xorg segfaults from time to time though, but I'll investigate | 15:24 |
Wizzup | *nod* | 15:24 |
freemangordon | this is with PVR EXA | 15:25 |
freemangordon | without it it is stable | 15:25 |
Wizzup | check | 15:25 |
freemangordon | so, we have usable 5.15/1.17 :) | 15:25 |
freemangordon | I'll have to resend some patches though | 15:25 |
uvos | great :) | 15:26 |
uvos | hows perf with no exa? | 15:26 |
freemangordon | lemme check | 15:26 |
uvos | i gues fineish? | 15:26 |
freemangordon | yes | 15:26 |
lel | parazyd created a repository: https://github.com/maemo-leste-upstream-forks/directx-headers | 15:26 |
uvos | cant be worse than current ddk1.9 :) | 15:26 |
freemangordon | the same like 1.9 | 15:26 |
uvos | ok | 15:26 |
lel | parazyd edited a repository: https://github.com/maemo-leste-upstream-forks/directx-headers | 15:27 |
* freemangordon runs glmark without exa | 15:27 | |
parazyd | (mesa dep, in case you wanna make jokes :P ) | 15:27 |
uvos | was about to | 15:27 |
uvos | fmg is makeing great progress, but i think dx12 is out of the question on pvr | 15:27 |
bencoh | :] | 15:28 |
freemangordon | well, why not, while I am ont it :p | 15:28 |
freemangordon | parazyd: do we really need to compile directx suport? | 15:28 |
bencoh | (I wouldn't be surprised if the poulsbo board received windows/directx support) | 15:28 |
freemangordon | I guess we cah disable that from the options | 15:28 |
bencoh | (intel / gma / pvr) | 15:28 |
parazyd | It's just easier than making a mess out of the repo. I want to have a unified mesa for all architectures. | 15:29 |
parazyd | And it just compiled in my VM | 15:29 |
parazyd | So next I'll rebase PVR on top. | 15:29 |
uvos | yeah i mean there is no reason not to have dx support on gallium drivers when leste is run on amd64 | 15:30 |
freemangordon | parazyd: ok, I understand that, but still, on no arch we will ever support there will be need fo directx, no? | 15:30 |
uvos | freemangordon: gallium supports dx on linux so yeah you can use it on leste on a pc | 15:30 |
uvos | and wine dose do so | 15:30 |
parazyd | You can, but the reason I'm building this is there's no directx flag in the compile options. | 15:30 |
freemangordon | so, you mean windows how provides dx support to linux guest? | 15:30 |
freemangordon | s/how/host | 15:31 |
freemangordon | parazyd: ok | 15:31 |
uvos | mesa just implements dx natively you could use it in regular linux applications to (no one dose this ofc) | 15:31 |
freemangordon | what? omg. well, ok | 15:32 |
freemangordon | I think it is time to wipe it all out and start from scratch :D | 15:32 |
freemangordon | next letter is 'm', so 'minux'? | 15:32 |
freemangordon | uvos: glmark2 Score: 44 | 15:33 |
uvos | freemangordon: slightly better than ddk1.9 | 15:33 |
parazyd | freemangordon: https://149366088.v2.pressablecdn.com/wp-content/uploads/2018/04/microsoft-loves-linux-1.jpg | 15:33 |
parazyd | :p | 15:33 |
uvos | freemangordon: but really ddk1.9 colapses with h-d running | 15:33 |
uvos | glmark score is like 20 iirc | 15:33 |
freemangordon | well, that's noprmal | 15:34 |
freemangordon | *normal | 15:34 |
freemangordon | you have SW copy and SW compositing | 15:34 |
uvos | yeah | 15:34 |
freemangordon | that'll be fixed with EXA | 15:34 |
* freemangordon needs some rest | 15:35 | |
freemangordon | ttyl | 15:35 |
uvos | bye | 15:35 |
parazyd | Wizzup: ok, everything's here I think: https://github.com/maemo-leste-upstream-forks/mesa/tree/maemo/beowulf-experimental | 16:28 |
parazyd | Wizzup: Can you try building it for arm? | 16:29 |
parazyd | (I don't have a proper lxc setup right now) | 16:29 |
Wizzup | ok | 16:30 |
Wizzup | parazyd: going to verbose about this | 16:35 |
Wizzup | I run apt-get build-dep . | 16:36 |
Wizzup | and get: | 16:36 |
Wizzup | The following packages have unmet dependencies: builddeps:. : Depends: meson (>= 0.52) but 0.49.2-1 is to be installed | 16:36 |
Wizzup | E: Unable to correct problems, you have held broken packages. | 16:36 |
Wizzup | so first thing to do is install it from backports? | 16:36 |
Wizzup | doing so now | 16:36 |
parazyd | Yep | 16:37 |
parazyd | Wizzup: oh and you have the rest of the deps in beowulf-experimental, so enable it | 16:38 |
parazyd | Latest libdrm and libglvnd | 16:38 |
Wizzup | yeah I did enable it | 16:38 |
Wizzup | I get this: | 16:39 |
Wizzup | ../meson.build:305:2: ERROR: Problem encountered: tegra driver requires nouveau driver | 16:39 |
freemangordon | tmlind: is there any change to have my patches *after* 5.15.2 merge | 16:39 |
freemangordon | *chance | 16:39 |
freemangordon | I have to send new versions, but I am not sure how to do that | 16:40 |
freemangordon | I tried to git rebase -i, but it fails | 16:40 |
Wizzup | what fails exactly? | 16:42 |
Wizzup | parazyd: any idea what to do next? | 16:43 |
Wizzup | trying something | 16:44 |
freemangordon | the rebase itself | 16:44 |
freemangordon | CONFLICT (content): Merge conflict in drivers/spi/spi-tegra20-slink.c | 16:44 |
Wizzup | well, what's the conflict? | 16:46 |
freemangordon | ok, I guess I can revert aand apply on top | 16:46 |
Wizzup | right | 16:46 |
Wizzup | parazyd: I added nouveau next to the pvr section in meson.build and it's going now, let's see | 16:47 |
Wizzup | ah, nope, didn't help | 16:47 |
Wizzup | meson /o\ | 16:49 |
tmlind | freemangordon: hmm well your patches appear after 5.12.2 if you do git log 5.12.2.. but not sure what you're trying to do | 16:50 |
tmlind | i thought you wanted the stable 5.12 merged in? | 16:50 |
tmlind | next week it's 5.12.3 and again they patches get buried :) | 16:50 |
freemangordon | tmlind: yes, they appear after, but are actually before it seems | 16:51 |
freemangordon | I want to squash a change I made into one of those patches | 16:51 |
freemangordon | by using git rebase -i, but it wants to everything between 5.15-rc and 5.15.2 | 16:51 |
freemangordon | and fails | 16:52 |
parazyd | Wizzup: Hold on, something came up here, sorry | 16:52 |
Wizzup | k, np, I'll do some other stuff meanwhile | 16:52 |
freemangordon | tmlind: top commit is (HEAD -> droid4-pending-pvr-omapdrm-v5.15, tmlind/droid4-pending-pvr-omapdrm-v5.15) Merge tag 'v5.15.2' into droid4-pending-pvr-omapdrm-v5.15 | 16:52 |
freemangordon | so it seems you have applied my patches and then merged 5.15.2 on top | 16:53 |
freemangordon | not really an issue, I'll just revert and reapply | 16:53 |
tmlind | ok | 16:56 |
Wizzup | parazyd: https://pastebin.com/raw/9pDk5H9K | 17:09 |
Wizzup | looks like something lacks a _GNU_SOURCE | 17:10 |
vectis | quit | 17:14 |
Wizzup | https://gitlab.freedesktop.org/mesa/mesa/-/issues/5067 | 17:14 |
Wizzup | parazyd: we will need at least one patch to make it build on armhf, but not a big deal | 17:17 |
Wizzup | I also have some changes | 17:17 |
d4irc | Hello Wizzup! | 17:23 |
dsc_ | :o | 17:23 |
Wizzup | yo | 17:24 |
parazyd | Wizzup: Feel free to commit | 17:24 |
Wizzup | freemangordon: uvos: is the pvr mesa driver a DRI driver? | 17:25 |
Wizzup | d4irc: <script>alert('hi');</script> | 17:28 |
Wizzup | parazyd: freemangordon: https://dpaste.com/BAYQH8D4W | 17:29 |
Wizzup | I guess that should check for omap perhaps? | 17:29 |
uvos | Wizzup: dri has several meanings, but i think whatever your really asking is to be awnserd with yes. | 17:34 |
Wizzup | It seemed to be a weird meson check | 17:34 |
Wizzup | I just commented the check for now | 17:34 |
bencoh | Wizzup: if you're talking about mesa, fmg's version built without me patching anything | 17:35 |
bencoh | (with meson 0.56) | 17:35 |
Wizzup | did you use dpkg-buildpackage -b -uc ? | 17:35 |
Wizzup | I doubt it somehow | 17:36 |
bencoh | I used -b | 17:36 |
Wizzup | well debian/rules did not seen to contain DRI_DRIVERS += pvr | 17:36 |
Wizzup | so it wasn't being built | 17:36 |
Wizzup | in any case this is latest mesa with pvr on top | 17:36 |
bencoh | ah, I added that of course | 17:36 |
bencoh | oh, latest mesa, nevermind then | 17:36 |
Wizzup | and they messed u _libdrm_checks in various ways | 17:36 |
Wizzup | including out of bounds access (lol) | 17:36 |
Wizzup | I'll circle back with parazyd about that | 17:36 |
Wizzup | uvos: did you fix this when rebasing on latest mesa? https://dpaste.com/F683TS8DC | 17:39 |
Wizzup | parazyd: anything you touched? ^^ | 17:41 |
dsc_ | Wizzup: contrary to popular belief I was not born yesterday, as such I give all my Text{} qml components the "textFormat: Text.PlainText" property :P | 17:42 |
uvos | Wizzup: no i applied the chromeos patches directly and those simply applied fine. | 17:43 |
uvos | Wizzup: i gues this is a difference in fmgs re pvr | 17:43 |
Wizzup | those values don't exist in the gl_config struct | 17:43 |
Wizzup | so maybe they applied fine but it does not build | 17:44 |
uvos | well im using it | 17:44 |
uvos | so it must have build :P | 17:44 |
uvos | or rather i used to use it | 17:44 |
Wizzup | what version? | 17:44 |
Wizzup | 21.2.5? | 17:44 |
uvos | uh | 17:44 |
Wizzup | seems quite unlikely since it doesn't even build freedreno without patches | 17:44 |
uvos | idk really i just cloned mesa-git some time ago (2-3 months maybe | 17:45 |
uvos | ) | 17:45 |
uvos | ill check in a sec | 17:45 |
Wizzup | in any case I can't continue as this requires in-depth mesa knowledge, but we're getting there I guess | 17:45 |
Wizzup | well I could maybe comment them if this is not used elsewhere | 17:46 |
Wizzup | I'll try again later, maybe the patches were not applied well | 17:49 |
Wizzup | dsc_: hehe | 17:51 |
d4irc | Wizzup: :> | 17:52 |
dsc_ | https://plak.infrapuin.nl/f7ymqupb.mp4 | 17:54 |
bencoh | yay | 17:56 |
uvos | freemangordon: btw swaping card0/1 is not as trivial as i thought | 17:56 |
uvos | so /dev/dri/by-path/* are links and /dev/dri/card* are the real files, in oposit to what i anticipated | 17:57 |
uvos | you can make udev move the card* files | 17:57 |
uvos | but its not very nice | 17:57 |
uvos | (to do that) | 17:57 |
bencoh | dsc_: does it work on top of telepathy? | 17:57 |
dsc_ | not right now but it will | 17:58 |
dsc_ | i just hooked it up for IRC to get some mock data | 17:58 |
bencoh | ah, alright | 17:58 |
uvos | dsc_: very nice :) | 17:58 |
dsc_ | :> | 17:58 |
uvos | mine dosent look as nice :P | 17:58 |
bencoh | yeah, it really looks great | 17:59 |
dsc_ | thanks | 18:00 |
uvos | apperantly im really a rabbit. | 18:00 |
dsc_ | xD | 18:00 |
uvos | Wizzup: 21.2.0-rc | 18:04 |
Wizzup | these break it | 18:07 |
Wizzup | https://github.com/maemo-leste-upstream-forks/mesa/commit/4daef7ffe373f067249e052d96dcac145984b6e8#diff-72cebfc89f85afe1b55728bc8e8332f6ab63a4b45c5bc024145f98c92d3faaf8 | 18:07 |
Wizzup | https://github.com/maemo-leste-upstream-forks/mesa/commit/c212283ef8e7b877dac17294bc447495665ad519#diff-72cebfc89f85afe1b55728bc8e8332f6ab63a4b45c5bc024145f98c92d3faaf8 | 18:07 |
Wizzup | so I guess we could just remove them from pvr as well | 18:07 |
uvos | maybe yeah | 18:09 |
Wizzup | in general https://github.com/maemo-leste-upstream-forks/mesa/commits/master/src/mesa/drivers/dri/common/utils.c | 18:09 |
uvos | so dsc_ | 18:10 |
uvos | can i take a peek at the code? | 18:10 |
Wizzup | bencoh: this setup is neat btw | 18:20 |
bencoh | Wizzup: :) | 18:25 |
Wizzup | I fixed the mesa pvr compilation problems and 21.2.5 built (no idea if it works) | 18:27 |
freemangordon | I can test later on | 18:55 |
freemangordon | Wizzup: what is the repo url? | 19:04 |
Wizzup | freemangordon: sorry, do you mean a repo for the deb pkgs? | 19:08 |
freemangordon | yes | 19:09 |
Wizzup | or the patch? | 19:09 |
Wizzup | ok | 19:09 |
Wizzup | I will upload | 19:09 |
Wizzup | it's not in a repo atm | 19:09 |
freemangordon | ah | 19:09 |
Wizzup | you will need libdrm and glvnd stuff from beowulf-experimental though I think | 19:09 |
freemangordon | Wizzup: well, I asked for url of beowulf-experimental :) | 19:10 |
Wizzup | sec | 19:10 |
Wizzup | deb https://maedevu.maemo.org/leste beowulf-experimental main contrib non-free | 19:10 |
Wizzup | freemangordon: https://wizzup.org/debs.tar.bz2 | 19:11 |
Wizzup | that is https://github.com/maemo-leste-upstream-forks/mesa/tree/maemo/beowulf-experimental + https://dpaste.com/EURCTX4QL.txt + https://dpaste.com/8EVC2YKRR.txt (patches wip) | 19:13 |
freemangordon | hildon-connectivity-wlan : Depends: libicd-network-wpasupplicant-dbus-n900 but it is not installable or | 19:23 |
freemangordon | libicd-network-wpasupplicant-dbus-common but it is not installable | 19:23 |
freemangordon | weird | 19:23 |
Wizzup | I think others remarked on this too, I haven't hit it yet | 19:24 |
Wizzup | brb 10-15 mins | 19:24 |
freemangordon | hmm, I upgraded and now TS does not work :( | 19:38 |
freemangordon | uvos: xinput does not list TS, any idea? | 19:41 |
Wizzup | could it be the udev rule for ts buttons? | 19:44 |
freemangordon | yes, but I can;t remember what was that | 19:44 |
freemangordon | ah, found it | 19:45 |
freemangordon | 85-input-devices.rules.leste? | 19:45 |
freemangordon | I shall remove that,right? | 19:46 |
freemangordon | yeah, that was it | 19:47 |
Wizzup | z/win 28 | 19:52 |
Wizzup | oops | 19:52 |
devrtz[m] | Hey there, I'm trying to organize a devroom for mobile linux related topics at fosdem and just wanted to ask if there would be some interest here in helping out or even giving a talk should my submission be accepted | 19:52 |
Wizzup | fosdem is remote this year, right? | 19:53 |
devrtz[m] | yup | 19:55 |
Wizzup | what kind of help do you need? | 19:57 |
devrtz[m] | Should it be accepted sorting through submitted talks. During the event: Coordinating with speakers, making sure everything goes smooth, that sort of thing. I'm actually not a 100% sure of all the things that need doing as I've never organized a devroom before. | 20:02 |
devrtz[m] | I do have some help of people who did though :) | 20:02 |
freemangordon | Wizzup: libgl1-mesa-dri:armhf depends on libllvm8 (>= 1:8~svn298832-1~); however: | 20:02 |
freemangordon | mesa-vdpau-drivers:armhf depends on libllvm8 (>= 1:8~svn298832-1~); | 20:03 |
freemangordon | ah, buster-backports | 20:04 |
Wizzup | right | 20:04 |
Wizzup | devrtz[m]: hm... I'd probably be more likely to talk about something than do a lot of emails since I already feel a bit overworked as-is, but perhaps others would like to do that? | 20:05 |
Wizzup | I know some folks here were also in-person fosdem volunteers in the past | 20:05 |
devrtz[m] | Yeah I actually saw the talk at the main track in '20 :) | 20:07 |
devrtz[m] | Ok, cool, should the devroom get accepted I will post the CfP here! | 20:08 |
Wizzup | I'm sure it'll get accepted, mobile linux is bigger than ever I think | 20:08 |
uvos | wait is our 5.15 leste kernel for expiramental w/o ts-buttons now? | 20:08 |
Wizzup | that probably depends entirely on fmg's config and what he used exactly, we don't have a -experimental kernel yet afaik | 20:09 |
uvos | oh ok | 20:09 |
freemangordon | Wizzup: mesa seems to be fine | 20:11 |
uvos | freemangordon: did you see this "[17:56] <uvos> freemangordon: btw swaping card0/1 is not as trivial as i thought" and the msgs below it? | 20:12 |
uvos | how bad do we need this? what i would have to do with udev is a bit of a hack | 20:12 |
freemangordon | yes, but don;t know how to comment on that :) | 20:12 |
uvos | ok | 20:12 |
freemangordon | don;t know | 20:12 |
Wizzup | freemangordon: that's great | 20:12 |
freemangordon | uvos: I am not sure how to fix that | 20:13 |
freemangordon | lemme try something | 20:13 |
ruleh | loading omapdrm before pvr should make sure that the numbering is "correct", no? | 20:13 |
uvos | so whats using card0? | 20:13 |
freemangordon | I guess, but we can;t guarantee that afaik | 20:13 |
uvos | video-omap? | 20:13 |
uvos | mesa? | 20:13 |
freemangordon | mesa | 20:13 |
ruleh | modprobe.d can enforce some sort of ordering I think | 20:13 |
uvos | yeah that would work too but i dont think its so great. | 20:14 |
freemangordon | lemme see if symlinking pvr_dri to omapdrm_dri will help | 20:14 |
uvos | might be better to ajust whatever uses the device to check what its binding to | 20:14 |
Wizzup | freemangordon: so as far as you are concerned we can build this mesa in -experimental ? | 20:15 |
freemangordon | yes | 20:15 |
Wizzup | ok | 20:15 |
freemangordon | well, I think "export MESA_LOADER_DRIVER_OVERRIDE=pvr" in /etc/profile/d is the least hacky solution | 20:16 |
uvos | freemangordon: do you know where the logic is in mesa what plugin+rener node to use | 20:17 |
freemangordon | no | 20:17 |
uvos | ok | 20:17 |
Wizzup | it's should be a decent solution for now yeah | 20:23 |
freemangordon | so please, whoever maintains those, add it for d4 and n900 | 20:24 |
uvos | parazyd: ^^^ | 20:25 |
Wizzup | leste-config yeah? | 20:30 |
Wizzup | for experimental? | 20:30 |
freemangordon | dunno | 20:30 |
freemangordon | but yeah, for experimental | 20:31 |
uvos | yeah leste config | 20:31 |
uvos | export MESA_LOADER_DRIVER_OVERRIDE=pvr should not do anything bad on current devel i think | 20:31 |
uvos | but yeah experimaental is more safe | 20:31 |
freemangordon | this mesa gives glmark 85 in portrait and 96 in landscape | 20:34 |
freemangordon | that's good I guess | 20:34 |
freemangordon | hildon scrolling: *** FPS: 83 *** | 20:40 |
uvos | might want to cap hildon at 60 | 20:46 |
uvos | even in the absence of vsync | 20:46 |
uvos | no need to torture the battery | 20:47 |
freemangordon | yeah | 20:47 |
Wizzup | freemangordon: was there anything you wanted the sgx-ddk-um to depend on? | 22:30 |
freemangordon | I don;t think so | 22:32 |
Wizzup | ok | 22:32 |
Wizzup | freemangordon: what about /usr/include/sgx/{hwdefs,include4} ? | 22:41 |
Wizzup | unless you really want to use the full kernel path? | 22:41 |
Wizzup | freemangordon: and for the .pc file, you want the -I CFLAGS, and what libs do you want to link against? | 22:45 |
freemangordon | Wizzup: -lsrv_um -lpvr2d | 23:13 |
Wizzup | ok | 23:20 |
Wizzup | freemangordon: I guess we want the ti443x and ti343x to conflict | 23:21 |
Wizzup | # pkg-config --cflags sgx-ddk-um | 23:29 |
Wizzup | -I/usr/include/pvrsgx -I/usr/include/pvrsgx/include4 -I/usr/include/pvrsgx/hwdefs | 23:29 |
Wizzup | # pkg-config --libs sgx-ddk-um | 23:29 |
Wizzup | -lsrv_um -lpvr2d | 23:29 |
Wizzup | freemangordon: can I push to xf86-video-omap for the new -dev pkg and pkgconfig stuff? | 23:51 |
Wizzup | I guess we also need some meta pkg for the sgx-ddk-um libs that pulls in the device libs | 23:52 |
Wizzup | we could also make them go in separate components I guess | 23:52 |
Wizzup | maybe 'Provides: sgx-ddk-um-libs' ? | 23:53 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!