| dsc_ | Wizzup: thx @ n900 | 04:30 |
|---|---|---|
| dsc_ | I have 10/10 faith in PowerVR SGX530 GPU running my application without a problem (lie) | 04:31 |
| dreamer | which application is that? :) | 04:36 |
| dsc_ | dreamer: https://twitter.com/maemoleste/status/1465697460043464707 | 04:38 |
| dreamer | ah, neat :) | 04:38 |
| dsc_ | (ga naar bed :D) | 04:38 |
| dreamer | (zelluf!) | 04:38 |
| dsc_ | net wakker :P | 04:38 |
| dreamer | oh, mogge | 04:38 |
| dsc_ | moggel | 04:39 |
| * dreamer wel zo crashen ja | 04:39 | |
| dsc_ | :P | 04:39 |
| Wizzup | uvos: iiirc tmlind mentioned the bridge | 12:07 |
| Wizzup | not sure what wasn't working though :) | 12:08 |
| Wizzup | freemangordon: where does osso-abook store contacts? in eds right, and that stores it where? | 12:50 |
| mighty17[m] | <uvos> "clk rate mismatch: 170666667 !=..." <- It's 3x, so I should try 170666667/2? | 14:17 |
| Wizzup | wondering if we should ask the nemomobile folks for help with audio | 15:24 |
| Wizzup | they seem active and might be more willing to help somehow - https://blog.mlich.cz/2021/12/nemomobile-in-december-2021/ | 15:24 |
| uvos | Wizzup: yes i know, i was hopping this was the problem, but seemingly omapdrm refuses to even try and activate the dsi signal with the bridge as the endpoint | 15:29 |
| uvos | dont know why atm | 15:29 |
| uvos | mighty17[m]: sure but really you should try and find out where 170666667 comes from | 15:30 |
| uvos | mighty17[m]: its proubly something its caluclating from your pannel timeings or something like that | 15:30 |
| uvos | just trace it backwards | 15:30 |
| mighty17[m] | So you're saying that I'll have to patch things in kernel possibly? | 15:31 |
| uvos | maybe not, could be something in your dts if you have the timeings there - or your pannel driver if you have one (you should) | 15:32 |
| uvos | point is idk where it gets the rate from - i think it calculates it from the timeings, so to figure out where it comes from you have to read the code in the backtrace | 15:33 |
| uvos | (thats whats its for after all) | 15:33 |
| mighty17[m] | uvos: It's simple panel, maybe I need to upstream the original one? | 15:34 |
| uvos | no simple pannel should work in theory (but a pannel driver is perferd afaik) but idk how often im supposed to say it, read the code to the functions in the backtrace to figure out where the rate comes from.... | 15:36 |
| uvos | or run the pannel at the "correct" rate | 15:37 |
| uvos | where it dosent work | 15:37 |
| uvos | and compear the reigster state with android | 15:37 |
| uvos | to figure out whats wrong | 15:37 |
| mighty17[m] | Yea yes, I get the part about backtrace, it's just that I'm being lazy at times for it :( | 15:37 |
| uvos | then you can go backwards from there | 15:37 |
| uvos | yeah i mean theres no free lunch im afraid | 15:37 |
| mighty17[m] | uvos: Rwmem right? | 15:37 |
| uvos | yeah | 15:37 |
| uvos | rwmem can ready any register if you know the address to it | 15:38 |
| mighty17[m] | i dunno the address in android tho | 15:39 |
| uvos | the TRM tells you | 15:39 |
| uvos | and ofc the registers addresses are the same on any os | 15:39 |
| mighty17[m] | https://github.com/MightyM17/linux_openpvrsgx/blob/espresso-dev/drivers/gpu/drm/omapdrm/dss/dss.c#L654 just a warning tho? :o | 15:40 |
| uvos | you can download the omap4's trm easy on google | 15:40 |
| uvos | yeah but your pannel dosent work at the correct rate @60Hz | 15:40 |
| uvos | so you should understand the warning no? | 15:40 |
| mighty17[m] | uvos: https://www.ti.com/pdfs/wtbu/OMAP4430_ES2.0_Public_TRM_vJ.pdf yeah i need it for pinmuxing | 15:40 |
| mighty17[m] | uvos: yeah, its at 30Hz if i use /2 value, but anyways i'll try to understand backtrace | 15:41 |
| uvos | mighty17[m]: right @trm i would dump the entire range starting at 0x48040000 | 15:43 |
| uvos | and compear | 15:43 |
| uvos | it will be vastly the same most likely | 15:44 |
| uvos | letting you find differences hopfully easily | 15:44 |
| mighty17[m] | neat! thats something i can do currently | 15:44 |
| uvos | and then for those registers that are different read what they do and see i fany makes sese | 15:44 |
| mighty17[m] | uvos: omap4-l4 would also give that right? | 15:44 |
| uvos | but i would read the backtrace code first | 15:45 |
| uvos | since thats less work | 15:45 |
| mighty17[m] | uvos: thats what im doin :D | 15:45 |
| uvos | idk what omap4-l4 is | 15:45 |
| mighty17[m] | the dtsi | 15:45 |
| uvos | ok | 15:46 |
| mighty17[m] | rwmem will compile in android? | 15:48 |
| uvos | mighty17[m]: sure do what i explain here for cpacprw https://github.com/IMbackK/cpcaprw | 15:49 |
| mighty17[m] | thanks! | 15:50 |
| mighty17[m] | <uvos> "mighty17: right @trm i would..." <- `-ash: ./rwmem: Permission denied` | 15:55 |
| mighty17[m] | even as sude | 15:56 |
| mighty17[m] | sudo* | 15:56 |
| uvos | on mainline? | 15:57 |
| mighty17[m] | yeah | 15:57 |
| uvos | theres a kernel config option that resetrcs access to /dev/mem to certain addresses iirc | 15:57 |
| mighty17[m] | hm will check | 15:58 |
| mighty17[m] | uvos: regarding android, does it need busybox and stuff to run? or just TWRP will be fine? | 16:17 |
| mighty17[m] | also `make_standalone_toolchain.py --arch arm --force --api 21 --stl gnustl --install-dir=$(SOME_PATH)` api matters here? as in will i have to use api21 rom on device as well | 16:27 |
| mighty17[m] | `export CROSS_COMPILE=$(SOME_PATH)/binary` binary doesnt exist in `$(SOME_PATH)` | 16:33 |
| Wizzup | so thumb2 builds are not possible with omap2plus_defconfig since armv6/omap2 doesn't do thumb, let's disable it for our builds | 16:38 |
| uvos | we allreday disable omap2 | 16:39 |
| Wizzup | maybe, not in the droid4-linux that I have atm | 16:39 |
| Wizzup | sre seems to be MIA mostly btw, he was usually so quick to respond :) | 16:40 |
| uvos | mighty17[m]: the $(SOME_PATH) isent to be taken literally ofc | 16:40 |
| uvos | he as allways been MIA for as long as i have been around at least | 16:41 |
| uvos | wrt response speed | 16:41 |
| uvos | Wizzup: https://github.com/maemo-leste/droid4-linux/blob/maemo/beowulf-experimental/debian/patches/0004-remove-omap2-from-defconfig.patch | 16:42 |
| Wizzup | that is not enough | 16:42 |
| Wizzup | it also needs to remove ...MULTI_v6=y | 16:42 |
| uvos | we should do that then | 16:43 |
| Wizzup | yes, I am | 16:43 |
| uvos | idk if we want thumb kernel tho | 16:43 |
| uvos | dosent omap3 have semi broken thumb? | 16:43 |
| Wizzup | well it's the last resort for me to see if I can get the n900 to boot/work reliably :) | 16:43 |
| uvos | is that worked around in kernel? | 16:43 |
| Wizzup | everything else I tried failed | 16:43 |
| Wizzup | yes, the errata is worked around | 16:43 |
| uvos | well thumb should not help | 16:44 |
| uvos | if it dose its just maksing something | 16:44 |
| Wizzup | if it is masking something, then that is progress | 16:44 |
| Wizzup | our stable 5.1 n900 kernel is thumb2 | 16:44 |
| Wizzup | tmlind: with thumb2 kernel the problem is still there on the n900, so I'm pretty sure it's a regression | 17:04 |
| Wizzup | maybe I should build plain 5.15 and see if it also occurs - maybe our droid4 patches cause problems somehow? | 17:05 |
| uvos | sure yeah its possible | 17:05 |
| mighty17[m] | <uvos> "mighty17: the $(SOME_PATH) isent..." <- i know, but that build dir doesnt have binary | 17:09 |
| uvos | did i make a slight typo maybe | 17:10 |
| uvos | ? | 17:10 |
| uvos | could be binarys | 17:10 |
| uvos | or bin | 17:10 |
| mighty17[m] | bin | 17:10 |
| uvos | regarding the api version | 17:12 |
| uvos | no you just need an andoroid with api >21 | 17:12 |
| mighty17[m] | coolio! | 18:04 |
| mighty17[m] | <uvos> "or bin" <- `make: error while loading shared libraries: libffi.so.8: cannot open shared object file: No such file or directory` | 18:05 |
| Niklas[m] | Hello,is there also a x86 build of Maemo Leste?I'd like to try it on a small tablet which has not enough memory for Arch+Gnome. | 19:04 |
| buZz | well, kinda | 19:05 |
| buZz | there's maemo-leste-1.0-amd64-20211031.qcow2.xz | 19:05 |
| buZz | for running in qemu, or a .vdi for doing the same in vmware/virtualbox , iirc | 19:05 |
| buZz | not completely sure how far that would get you | 19:05 |
| buZz | https://phoenix.maemo.org/view/Images/job/leste-image-virtual/ | 19:05 |
| uvos | thats one option | 19:06 |
| buZz | Niklas[m]: but i'd love to hear how far you can get :D | 19:06 |
| uvos | the option i would reccomend is install devuan normaly, add our repos to your sources and install hildon-meta | 19:06 |
| buZz | ah, yes good point | 19:06 |
| Niklas[m] | And if I convert them to a ISO,can I boot them on real hardware?Does it support UEFI (this stupid thing has no legacy boot -.-) | 19:06 |
| buZz | but are there x86 packages in repo then? | 19:06 |
| uvos | buZz: yes | 19:06 |
| buZz | you cant convert a .vdi to .iso | 19:07 |
| uvos | amd64 | 19:07 |
| buZz | uvos: sweet | 19:07 |
| buZz | Niklas[m]: so grab a devuan iso , install distro with uefi or whatever you want | 19:07 |
| uvos | we dont build x86-32 tho | 19:07 |
| uvos | so depends on how old his tablet is | 19:07 |
| buZz | then add repo and install hildon-meta | 19:07 |
| buZz | 32bit only would be >15yo, i guess? | 19:07 |
| Niklas[m] | The CPU supports 64bit,but the UEFI is 32bit-only | 19:08 |
| uvos | some atom processors where pretty recent | 19:08 |
| buZz | Niklas[m]: ah one of -those- | 19:08 |
| buZz | lol | 19:08 |
| uvos | your uefi dosent support bios emulation? | 19:08 |
| Niklas[m] | Nope | 19:08 |
| buZz | those are quite a hell to work with :P | 19:08 |
| uvos | great | 19:08 |
| uvos | :P | 19:08 |
| buZz | i know fedora actively supports this 32bit-uefi-64bit-rootfs setup | 19:09 |
| buZz | or used to at least, i think fedora is over | 19:09 |
| Niklas[m] | I mean,it has like a dozen times more settings than the BIOS of my laptop,but legacy boot is not among them :/ | 19:09 |
| Niklas[m] | I have Arch Linux installed currently.It supports 32bit UEFI with 64bit OS just fine,but since some updates ago I haven't been able to use a GUI because desktop environments getting more and more bloated and 1GB of RAM (not upgradeable) is too few for them | 19:10 |
| Niklas[m] | And it never ran smoothly.I hope Maemo can do better,as it still supports the Nokia N900 for example which has even less | 19:11 |
| buZz | Niklas[m]: why not just install fluxbox , openbox , blackbox , windowmaker , whatever | 19:12 |
| uvos | i mean if you get away from kde/gnome arch has no issues with 1gb on lxqt or something like i3, sway | 19:12 |
| buZz | any of the <10MB ram using window managers | 19:12 |
| buZz | twm , dwm | 19:12 |
| Niklas[m] | Yeah,but this thing has no hardware keyboard.GNOME is the only one which works quite well with touchscreen-only and a 7 inch screen | 19:13 |
| uvos | the benifit i gues is that we have a touch frendly ui | 19:13 |
| uvos | oh 7 inches | 19:13 |
| uvos | that bearly a tablet at all :P | 19:13 |
| buZz | my biggest tablet is a 8" omap4330 | 19:14 |
| buZz | oh, omap4460 perhaps? | 19:15 |
| buZz | seems postmarket supports it nowadays :P https://wiki.postmarketos.org/wiki/Archos_80_G9_(archos-a80) | 19:15 |
| buZz | ah, no it doesnt | 19:16 |
| Niklas[m] | Which Devuan release should I download now to add the Maemo repos later? | 19:17 |
| buZz | i -think- maemo leste is still on beowulf? | 19:18 |
| buZz | > Maemo Leste is based on Devuan Beowulf (Debian Buster) and all the "supported" devices ship with recent Linux (mainline-based) kernels. | 19:18 |
| buZz | yes | 19:18 |
| Niklas[m] | Thanks,I'll try that | 19:19 |
| freemangordon | Niklas[m]: beware that hildon-desktop requires 3D acceleration | 19:59 |
| Niklas[m] | If GNOME worked before,that should be supported,right? | 20:00 |
| freemangordon | umm, not sure | 20:01 |
| freemangordon | does gnome require 3D? | 20:01 |
| Niklas[m] | I don't know,I normally don't use GNOME if I don't have to :D | 20:01 |
| freemangordon | well, I use it all the time, however, I never been interested if it requires 3d or not. But I doubt it does | 20:02 |
| freemangordon | anyway, what is the GPU in the thingie you want leste on? | 20:02 |
| Niklas[m] | It's some dual-core Intel Atom cpu with integrated graphics | 20:03 |
| Niklas[m] | Intel HD Graphics probably? | 20:03 |
| freemangordon | oh, you should have drivers then | 20:04 |
| Niklas[m] | That's good but I didn't have luck with booting another OS so far,will try again in a few days. | 20:14 |
| Wizzup | freemangordon: old gnome didn't, new gnome most likely does | 20:59 |
| Wizzup | i.e. gnome 14.04 can't be compared to modern gnome | 20:59 |
| uvos | gnome has required 3d accell since gnome3 with gnome-shell | 21:13 |
| Wizzup | right | 21:13 |
| uvos | gnome3 is older than 2014 i think | 21:13 |
| uvos | so even fmgs gnome needs accel | 21:14 |
| uvos | unless he uses gnome-legacy? or classic? idk | 21:14 |
| uvos | there was some fallback mode for a while i think | 21:14 |
| freemangordon | gnome-session-flashback | 21:14 |
| freemangordon | :) | 21:14 |
| Niklas[m] | I used the normal Gnome 3.Maybe it was even updated to Gnome 40,I don't remember. | 21:14 |
| freemangordon | yeah, given the GOU is intel you shouldn;t have issues | 21:15 |
| freemangordon | *GPU | 21:15 |
| uvos | Niklas[m]: yes you did the falback thing was dropped a long time ago | 21:15 |
| uvos | not sure when exactly but 3+ years at least | 21:15 |
| Niklas[m] | Graphics acceleration really shouldn't be a problem and never was.Booting a 64bit OS on 32bit UEFI is,however. | 21:16 |
| uvos | yeah | 21:16 |
| uvos | fun | 21:16 |
| Niklas[m] | Maybe I can start the Devuan USB stick using the GRUB which is already installed on internal storage. | 21:16 |
| uvos | yeah sure once you have grub its easy | 21:27 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!