freemangordon | uvos: yes, the questions "how mce tells x11 to disable input devices" is still relevant | 07:14 |
---|---|---|
freemangordon | *question | 07:15 |
freemangordon | scratch that, found it | 07:17 |
freemangordon | heh | 07:34 |
freemangordon | mce: x11-ctrl: unable to open display | 07:34 |
freemangordon | root does not have access to X display | 07:49 |
freemangordon | Wizzup: any objections adding "xhost +local:" to startup scripts? | 07:49 |
rafael2k | I here cheering for you. | 07:49 |
rafael2k | xhost + | 07:50 |
rafael2k | : ) | 07:50 |
freemangordon | yeah | 07:50 |
freemangordon | rafael2k: could you upgrade your device to chimaera-devel and confirm xorg runs as user 'user'? | 08:33 |
rafael2k | yes | 09:00 |
freemangordon | rafael2k: 'yes, I could' or 'yes, it runs as user'? :) | 09:08 |
freemangordon | uvos: it turned out mce cannot open the display :) | 09:08 |
freemangordon | so no input devices got disabled | 09:08 |
rafael2k | freemangordon, sorry, I'll just apt-get upgrade with chimaera-devel repo and let you know, just a sec | 09:09 |
freemangordon | sure, no hurry, I just wasn't sure what that 'yes' was supposed to mean | 09:09 |
rafael2k | The following NEW packages will be installed: | 09:15 |
rafael2k | autologin xinit | 09:15 |
rafael2k | insserv: Script zzzremove-first-boot-flag has overlapping Default-Start and Default-Stop runlevels (S 1 2 3 4 5) and (0 1 2 3 4 5 6). This should be fixed. | 09:17 |
rafael2k | got this too ^ | 09:17 |
rafael2k | Installing new version of config file /etc/X11/Xsession-post ... | 09:18 |
rafael2k | Installing new version of config file /etc/X11/Xsession.d/00settings ... | 09:18 |
rafael2k | Installing new version of config file /etc/init.d/xorg ... | 09:18 |
rafael2k | It will take a bit as the kernel is also being upgraded (6.1.13 -> 6.1.15) | 09:19 |
rafael2k | rebooted - it is up | 09:24 |
rafael2k | autologin, startx and xinit as "user" | 09:26 |
rafael2k | Xorg as "root" | 09:26 |
rafael2k | Xsession-post as "user" too | 09:28 |
freemangordon | ok, seems we'll have to tell it to run as "user" | 09:29 |
freemangordon | rafael2k: does loginctl lists x11 session | 09:29 |
freemangordon | it should be cN | 09:29 |
rafael2k | oginctl | 09:30 |
rafael2k | SESSION UID USER SEAT TTY | 09:30 |
rafael2k | 1 1000 user pts/0 | 09:30 |
rafael2k | c2 1000 user seat0 | 09:30 |
rafael2k | 2 sessions listed. | 09:30 |
freemangordon | cool | 09:44 |
rafael2k | lemme know any test needed | 09:45 |
freemangordon | sure, thanks! | 09:46 |
rafael2k | : )) | 09:46 |
freemangordon | uvos__: do we car that much if startx runs Xorg as root? or, shall we divert /etc/X11/Xwrapper.config and add needs_root_rights=no? | 09:48 |
freemangordon | *care | 09:49 |
sicelo | why would omap4 (samsung tab 2) break on thumb2 kernel? freemangordon any ideas? | 11:02 |
sicelo | is that perhaps why leste kernel isn't thumb2? | 11:02 |
freemangordon | sicelo: what is the contaxt? | 11:47 |
freemangordon | *context | 11:47 |
Wizzup | freemangordon: what does that do @ xhost? | 11:49 |
Wizzup | freemangordon: ah, that was the problem? | 11:50 |
freemangordon | mhm | 11:50 |
freemangordon | mce was unable to open the display so it was not disabling input devices | 11:50 |
freemangordon | with upgraded maemo-system-services we should be back to normal | 11:51 |
Wizzup | lol | 11:51 |
Wizzup | good find | 11:52 |
freemangordon | at least here I am seeing something like 40mA on d4 | 11:52 |
freemangordon | with gprs on | 11:52 |
freemangordon | Wizzup: maybe upgrade and confirm idle draw on your device is normal | 11:54 |
Wizzup | I will | 11:54 |
freemangordon | well, better upgrade :) | 11:54 |
Wizzup | I did :) | 11:56 |
sicelo | freemangordon: working with mighty17[m], i was trying to use a common kernel config for his samsung espresso and my n900. it works, as long as THUMB2_KERNEL is not set. when set, the espresso doesn't boot correctly. unfortunately, he hasn't got uart, so for now we can't tell what breaks things with THUMB2 | 12:03 |
uvos | maybe build a thumb2 kernel for d4 | 12:10 |
uvos | possibly its broken for omap4 in general | 12:11 |
uvos | and d4 is easy to debug | 12:11 |
uvos | the leste kernel is not not useing thumb2 beacause of any know issue | 12:12 |
uvos | its not thumb2 because thats upstream default, and i dident change it | 12:12 |
Wizzup | I think we have the same kernel for d4 and n900 now | 12:13 |
uvos | yes | 12:13 |
Wizzup | aren't we enabling thumb kernel? or just userspace? | 12:13 |
uvos | just userspace | 12:13 |
uvos | we have never used thumb2 kernel on d4 afaik, and since we absorbed the n900 into the d4 kernel it has not been thumb2 either | 12:14 |
uvos | freemangordon: we build our own xorg anyhow, so changeing the config file seams no issue. we dont care THAT mutch, but it is a real security thing. | 12:23 |
freemangordon | uvos: if we are going to change Xorg, I would say we better fix it properly | 12:35 |
freemangordon | this: https://github.com/freedesktop/xorg-xserver/blob/master/hw/xfree86/xorg-wrapper.c#L235 is buggy | 12:35 |
freemangordon | because it unconditionally checks for modesetting capability on dri devices | 12:36 |
freemangordon | IIUC, there no way pvr driver to have modesetting caps :) | 12:37 |
freemangordon | I wonder if it wouldn;t be easier to just remove pvr node from /dev/dri | 12:46 |
freemangordon | BTW, we have the same issue with lima | 12:46 |
freemangordon | so it is not d4 only | 12:47 |
freemangordon | umm, pvr only | 12:47 |
sicelo | i would test thumb2 with my droid 4, but i'm going to be physically away from it for around 2-3 months | 12:48 |
bencoh | sicelo: may I ask which espresso you were referring to? because samsung chips have been shipped with arm64 kernels for quite a while | 13:02 |
bencoh | (and/or which exynos SoC if you know) | 13:03 |
sicelo | samsung tab 2 | 13:03 |
sicelo | it has omap4430 :-) | 13:03 |
bencoh | ah, I see | 13:03 |
freemangordon | Wizzup: cannot build xorg on device or in VM, requires libsystemd-dev :( | 13:51 |
Wizzup | heh | 14:15 |
Wizzup | I can help in 30 mins or so | 14:15 |
freemangordon | Wizzup: what about https://pastebin.mozilla.org/NHJousDv ? | 14:53 |
freemangordon | or anybody who has a clue in that regard :) | 14:58 |
Wizzup | freemangordon: and you can't install libsystemd-dev? | 15:19 |
rafael2k | xorg should not need systemd, this is strange, as it runs on OSes without systemd | 15:34 |
Wizzup | if you want to build it with logind support you need libsystemd-dev | 15:35 |
Wizzup | iiuc | 15:35 |
Wizzup | and on debian they didn't make it so that you can use libsystemd-dev with elogind installed, again iiuc | 15:35 |
Wizzup | freemangordon: you can just tag it and run on CI | 15:35 |
rafael2k | what a shit | 15:35 |
rafael2k | :( | 15:35 |
arno11 | freemangordon: new dist-upgrade works on n900 | 15:54 |
arno11 | power cons back to normal | 15:54 |
arno11 | idle 47 | 15:54 |
arno11 | i've tested lot of stuff on n900. most things works fine | 15:59 |
arno11 | with chimaera | 16:00 |
sicelo | and modem? | 16:01 |
arno11 | modem is the main problem | 16:01 |
sicelo | still broken? i would like to see the dmesg, etc. as requested yesterday | 16:02 |
arno11 | ok | 16:02 |
arno11 | [ 0.486175] kprobes: kprobe jump-optimization is enabled. All kprobes are optimized if possible. | 16:03 |
arno11 | [ 16.314483] ehci-omap 48064800.ehci: new USB bus registered, assigned bus number 1 | 16:03 |
arno11 | [ 16.410766] ohci-platform 48064400.ohci: new USB bus registered, assigned bus number 2 | 16:03 |
sicelo | i've got working modem in vanilla 6.3-rc2, i.e. without powervr, but it shouldn't affect modem in any way | 16:03 |
sicelo | arno11: modem is not on USB ... | 16:04 |
arno11 | sorry wrong command | 16:04 |
arno11 | in fact no mistake | 16:05 |
arno11 | grep modem and usb result | 16:05 |
sicelo | :-) | 16:06 |
sicelo | s/vanilla// | 16:06 |
arno11 | mdbus2 gives no modem as well | 16:08 |
arno11 | only bluetooth lol | 16:09 |
sicelo | zgrep NOKIA_MODEM /proc/config.gz | 16:10 |
arno11 | CONFIG_NOKIA_MODEM=m | 16:10 |
sicelo | ok, at least that's a good start | 16:11 |
arno11 | yes indeed | 16:11 |
sicelo | zgrep OMAP_SSI /proc/config.gz | 16:11 |
arno11 | CONFIG_OMAP_SSI=m | 16:12 |
sicelo | btw, does the modem on your N900 work, e.g. in Fremantle? | 16:12 |
arno11 | yes no probs | 16:12 |
sicelo | dmesg | grep 'ssi\|modem' | 16:12 |
arno11 | usb results lol | 16:13 |
sicelo | i don't understand ... | 16:14 |
arno11 | see previous copy paste | 16:14 |
arno11 | ehci usb bla bla | 16:14 |
arno11 | but no modem | 16:14 |
sicelo | i'm not sure this comes from that dmesg + grep command | 16:14 |
arno11 | i confirm | 16:15 |
sicelo | oh, haha, i see. it's the "ssi" | 16:15 |
sicelo | ssi exists in possible :p | 16:15 |
sicelo | lsmod | grep ssi | 16:16 |
arno11 | no result | 16:17 |
sicelo | lsmod | grep 'nokia\|hsi' | 16:17 |
sicelo | weird | 16:17 |
sicelo | so the module doesn't load, but it exists | 16:18 |
arno11 | yep | 16:18 |
arno11 | nokia hsi no result | 16:18 |
arno11 | if no sim card is detected ? | 16:19 |
sicelo | modem is independent of SIM. should still probe without it | 16:19 |
arno11 | ok | 16:19 |
sicelo | post your dmesg somewhere | 16:19 |
arno11 | ? | 16:21 |
arno11 | only dmesg ssi\|modem give results but not related to modem | 16:23 |
buZz | are we merging https://marc.info/?l=linux-omap&m=167819256924890&w=2 soon? :D | 16:27 |
buZz | curious how the charging will get improved on cpcap | 16:27 |
buZz | guess i could just do my own kernel again, but meh :P | 16:28 |
sicelo | buZz: if you do build kernel, please humor me and enable THUMB2_KERNEL :-) | 16:52 |
buZz | :D | 16:52 |
sicelo | I'm dying to know if it works on d4 or not | 16:52 |
freemangordon | Wizzup: I installed libsystemd-dev, in the cost of removing couple of packages | 17:47 |
freemangordon | so, any thoughts about https://pastebin.mozilla.org/NHJousDv ? | 17:48 |
uvos | i mean its not technicly true | 17:53 |
uvos | you could have a device with a render node that only supports ums | 17:53 |
uvos | that would be wierd, and dosent exist | 17:53 |
uvos | but its not impossib;e | 17:54 |
uvos | honestly the should just drop ums support form xorg allready... | 17:54 |
buZz | is xorg even still being developed? | 17:56 |
uvos | sure | 17:57 |
uvos | its commit history is still quite healthy | 17:57 |
uvos | even if you ignore xwayland stuff | 17:57 |
uvos | and its still the dominat display server | 17:58 |
uvos | at least like 1 year ago it was sill like 90% | 17:59 |
Wizzup | freemangordon: no real thoughts, if it works, let's do it | 19:32 |
freemangordon | uvos: while I agree, the current check is even worse | 21:28 |
freemangordon | BTW, if you have better idea how to fix that, please share | 21:32 |
freemangordon | oh, ok, why is xserver-xorg-legacy installed in the first place? | 21:59 |
freemangordon | Wizzup: we shall remove that from the image | 21:59 |
freemangordon | yep, removing that package from d4 makes Xorg run as user, without any patches | 22:04 |
sicelo | :-) | 22:08 |
sicelo | btw, there are chimaera images around already? or need dist-upgrade? | 22:09 |
uvos | freemangordon: any idea why https://github.com/maemo-leste/sphone/blob/6ca8fcc19d28e6f89163132fa5a8258a995680c6/src/modules/contacts-ui-abook.c#L95 never finds any contacts? | 22:16 |
uvos | the exact same query finds the contact if used with e_book_client_get_contacts_sync | 22:16 |
freemangordon | uvos: maybe enable abook debug messages | 22:18 |
freemangordon | see https://github.com/maemo-leste/osso-abook/blob/master/lib/test-contact-chooser.c#L32 | 22:19 |
freemangordon | sicelo: yes, afaik | 22:21 |
freemangordon | (there are images) | 22:21 |
uvos | OSSO_ABOOK_DEBUG=all sphone -v -v dosent appear to print anything | 22:21 |
uvos | is there something else | 22:21 |
uvos | is osso_abook_debug_init required to get any output? | 22:21 |
freemangordon | G_MESSAGES_DEBUG=all | 22:24 |
freemangordon | also, seems osso_abook_aggregator_new() is async | 22:25 |
uvos | ah | 22:25 |
uvos | ok | 22:25 |
uvos | doc dosent tell me this | 22:25 |
uvos | http://maemo.org/api_refs/5.0/beta/libosso-abook/OssoABookAggregator.html#osso-abook-aggregator-new | 22:26 |
freemangordon | but, why don't you use the functions I propsed? | 22:26 |
freemangordon | *proposed | 22:26 |
uvos | because i want to use the same query | 22:26 |
freemangordon | ok | 22:26 |
uvos | the idea being that later sphone can genrate ebook quers for eatch backend | 22:27 |
freemangordon | ok | 22:27 |
uvos | and then both the eds and the abook module can use those | 22:27 |
uvos | not sure how to tell when osso_abook_aggregator_new is ready | 22:27 |
freemangordon | see https://github.com/maemo-leste/osso-abook/blob/master/lib/eds.c#L128 | 22:28 |
freemangordon | I think you should connect to "ebook-status" signal | 22:29 |
uvos | i presume havein a aggregator open at all times is bad | 22:33 |
uvos | (ie uses power because its updateing presence or sutch) | 22:33 |
uvos | othwerise it could just have a roster at all times and check OssoABookAggregatorState when it wants to query | 22:34 |
uvos | https://github.com/maemo-leste/modest/blob/dc9929645ad2880b6a91848ce6dc514cacfb5c43/src/hildon2/modest-address-book.c#L93 | 22:35 |
uvos | modest seams to solve this problem with a dirty hack | 22:36 |
Wizzup | freemangordon: cool, shall we conflict it on in hildon-base ? | 22:42 |
Wizzup | or do we install it explicitly? | 22:43 |
uvos | it was needed for n900 | 22:43 |
uvos | pre ddk1.17 | 22:43 |
uvos | so presumably we install it explicitly | 22:43 |
Wizzup | I can't find it in leste-config, hildon-base or image-builder | 22:46 |
freemangordon | who install it then? | 22:56 |
freemangordon | *installs | 22:56 |
uvos | hmm osso_abook_aggregator_get_state never returns anything except OSSO_ABOOK_AGGREGATOR_PENDING (no mater how long it has been) | 22:57 |
uvos | must be doing something else wrong too | 22:57 |
Wizzup | freemangordon: maybe if you don't install elogind it gets installed automatically, but not removed when you install elogind | 23:00 |
Wizzup | just a guess | 23:01 |
Wizzup | we can conflict with it in any case I think? | 23:01 |
Wizzup | well, if you remove it on the d4 now, and dist-upgrade doesn't pull it back in... we're fine to conflict with it | 23:01 |
Wizzup | maybe good to also have it tested on the PP | 23:03 |
Wizzup | just tested, it doesn't get pulled back in | 23:04 |
arno11 | got the modem working on n900: module is not loaded on startup | 23:12 |
arno11 | otherwise it works | 23:12 |
arno11 | but the phone menu settings is not working | 23:13 |
uvos | i mean no shit | 23:13 |
uvos | its blacklisted https://github.com/maemo-leste/leste-config/blob/master/leste-config-n900/etc/modprobe.d/blacklist-omap-ssi.conf.leste | 23:13 |
Wizzup | uvos: wait, we still blacklist the modem? | 23:18 |
Wizzup | I thought the issue was resolved | 23:18 |
uvos | yes, aperantly | 23:18 |
uvos | sure maybe someone solved the underlying issue | 23:19 |
uvos | but no one updated leste-config | 23:19 |
arno11 | i can confirm: modem blacklisted | 23:19 |
Wizzup | uvos: yes this is the case I believe | 23:21 |
Wizzup | sicelo: ^^ shall I update the blacklist? | 23:21 |
sicelo | of course. blacklist should be removed | 23:21 |
Wizzup | arno11: btw I'm just getting my n900 set up, so I can try to get you a trace tomorrow for the shutdown issues you're seeing | 23:22 |
Wizzup | sicelo: just checking :) | 23:22 |
arno11 | i sent a message no more issue with shutdown | 23:23 |
arno11 | in fact most things works great now | 23:25 |
arno11 | excepting this bloody phone menu settings | 23:25 |
Wizzup | the setting applet? | 23:25 |
arno11 | yes | 23:25 |
Wizzup | yeah... that's not finished... | 23:25 |
arno11 | ok | 23:26 |
Wizzup | it requires a non significant amount of work, but it's not -impossible- | 23:27 |
Wizzup | basically the phone settings uses connui-cellular which wraps C functions to ofono calls | 23:27 |
Wizzup | and the C functions made certain assumptions about when to block that don't make it super pleasant to wrap | 23:28 |
Wizzup | I did a bunch, but some things aren't even implemented at all / just stubbed | 23:28 |
arno11 | ok | 23:28 |
Wizzup | enough for the signal and operator name and basic stuff to work though | 23:28 |
Wizzup | pin/puk too | 23:29 |
arno11 | yeah | 23:29 |
Wizzup | we can look at it some time, see what else needs to be tackled. | 23:29 |
Wizzup | I guess for the n900 we still want call audio | 23:30 |
Wizzup | pavel offered to help with some of it at fosdem when I talked with him | 23:30 |
sicelo | oh, he did? | 23:31 |
Wizzup | yes, I thought I wrote that here | 23:31 |
sicelo | i thought he all but quit n900. nice to know :-) | 23:32 |
Wizzup | he prefers the droid, but he offered to help | 23:32 |
arno11 | that's nice | 23:32 |
Wizzup | he said he can get it to work, but we have to fix the latency, but he thought it was maybe some pa issue | 23:32 |
Wizzup | maybe even just the default pa buffer latency | 23:33 |
Wizzup | (I think) | 23:33 |
arno11 | or resampling | 23:33 |
arno11 | with pa | 23:33 |
Wizzup | if we can get it to work with alsa, even with poor quality, that'd be a big step towards actually getting it to work | 23:34 |
sicelo | i did get it working once myself. just need to get back to it | 23:35 |
Wizzup | maybe soon is a good time :D | 23:36 |
Wizzup | <-- says he who constantly bites of more than he can chew to the point that he can't even bite anymore :p | 23:36 |
sicelo | :-D | 23:37 |
uvos | https://github.com/maemo-leste/libhildonmime/pull/4 | 23:45 |
uvos | freemangordon: ^^ | 23:45 |
uvos | allows one to open urls in modest | 23:45 |
Wizzup | uvos: great :) | 23:47 |
Wizzup | arno11: can you share your module blacklist for pm? | 23:49 |
arno11 | pm? | 23:52 |
sicelo | power management, i.e. how you get 47mA standby | 23:54 |
arno11 | ok i m tired lo | 23:54 |
arno11 | lol | 23:54 |
arno11 | sec | 23:54 |
arno11 | omap3_isp | 23:54 |
arno11 | ehci_omap | 23:55 |
arno11 | omap_hdq | 23:55 |
arno11 | hci_nokia | 23:55 |
arno11 | hci_uart | 23:55 |
arno11 | sometimes less than 47 | 23:56 |
Wizzup | great, this will probably also help with debugging RET | 23:57 |
Wizzup | I wonder if we just add these to leste-config blacklist for now | 23:57 |
uvos | this is with ret right? 47mA seams like it would need to hit ret | 23:57 |
Wizzup | he said without | 23:57 |
uvos | oh ok | 23:57 |
Wizzup | I'm going to check | 23:58 |
arno11 | we can try if we are sure they will not interact with other stuff | 23:58 |
Wizzup | uvos: does maemo-6.1.y contain the n900 modem fix? | 23:59 |
Wizzup | I just pulled it, but still searching | 23:59 |
uvos | Wizzup: idk whats the modem fix? | 23:59 |
uvos | its only 15mA more than d4 gets with ret, so pretty good | 23:59 |
arno11 | cool | 23:59 |
Wizzup | I see my 'force pio for now', but iirc sicelo tracked it down | 23:59 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!