tmlind | oh nice, good to hear pinephone has working pm for suspend now :) | 08:06 |
---|---|---|
* enyc meows | 09:30 | |
enyc | tmlind: useful =) alas not for n900 ?? | 09:30 |
Pali | freemangordon: Hi! Did you have a time to look at that u-boot rx51 usb code? | 14:17 |
scops | Hello :) i`m going to buy a pinephone for maemo leste since i really liked Diablo back in the day (i know maemo leste is more like maemo 5 than 4). so my question is; do i need to reflash and follow the steps on https://leste.maemo.org/PinePhone for every update or can i just "apt upgrade" or something like that? | 14:46 |
scops | and how useable for example in compare to Nokia n810/n800 is maemo leste in the current state of development? | 14:47 |
Wizzup | apt upgrade will work | 14:48 |
Wizzup | n810 didn't do calls, right? | 14:48 |
scops | yeah, i dont need a new phone... i'm looking for modern replacement for my old internet "tablet" ;) | 14:49 |
scops | so wwan functionality doesnt matter for me atm | 14:49 |
Wizzup | pinephone with keyboard should work pretty well | 14:52 |
Wizzup | there are some power management and 3d driver things to sort out, but it's not bad I think | 14:52 |
scops | does leste also include modern ports of old applications? for example gpe apps or the original internet tablet and maemo5 apps like the mailer, file management (....)? | 14:53 |
scops | i ask because imho modern gtk3 apps are somewhat slow on librem5 (i have one) and even slower on the pinephone | 14:54 |
scops | (hehe even my old n810 feels faster on something like ebook reading or file management than the librem5 with phosh xD) | 14:55 |
sicelo | :-) | 15:08 |
sicelo | As for old applications, some have been ported already, and others can be ported on request | 15:09 |
scops | how can i show the mouse on vbox? in the wiki there is a command switch for qemu... but on vbox? | 15:11 |
Wizzup | it should just be visible | 15:11 |
scops | it isnt... i'm testing the maemo-leste-1.0-amd64-20210530 build | 15:12 |
scops | it flashes for a sec if i scroll | 15:12 |
Wizzup | scops: what input device did you select in vbox | 15:20 |
Wizzup | https://leste.maemo.org/Virtual_Machine#VirtualBox | 15:20 |
Wizzup | If the mouse pointer is grabbed and no cursor shows, you can add: | 15:20 |
Wizzup | -usb -device usb-tablet | 15:20 |
Wizzup | to the command line. Alternatively if that dose not work you can instead add: | 15:20 |
Wizzup | -usb -device usb-mouse -machine vmport=off | 15:20 |
scops | thats what i said, that is a commandline switch for | 15:21 |
scops | qemu | 15:21 |
scops | not for virtualbox | 15:21 |
gordon1 | hi, got redirected here from #maemo, basically a question about n900 power management, why leste's current consumtion is 10x more than original freemantle? Do we know the reason? I got told that it's up to userspace cooperation and everything is ready for correct PM in kernel, is that so? And if that's so can I at least get proof-of-concept 8mA consumtion on idle with everything disable and no | 15:23 |
gordon1 | userspace software runing with recent kernel? | 15:23 |
buZz | 8mA? :O | 15:27 |
gordon1 | according to wiki idle/no sim/no wlan consumption is 7mA https://wiki.maemo.org/N900_Hardware_Power_Consumption | 15:28 |
buZz | did you see https://leste.maemo.org/Nokia_N900#Power_Management ? | 15:28 |
buZz | the most interesting there is > Currently it does not yet reliably reach any sleep modes. | 15:29 |
gordon1 | sure, but why? | 15:29 |
gordon1 | do we know the reason? | 15:29 |
buZz | well https://github.com/maemo-leste/bugtracker/milestone/22 is a start | 15:30 |
gordon1 | guys on #maemo told me that it is up to userspace, i.e. userspace stop CPU from going into those states? | 15:30 |
buZz | right, i'm not sure what you're asking, you want to run the kernel without any userspace? | 15:30 |
buZz | maybe you could replace /bin/init with a small script? :P | 15:31 |
buZz | not sure how you'd even readout power draw then, unless you're measuring externally | 15:32 |
gordon1 | yep, i've been there too, but only relevant issue there is power consumtion measurement tool, which just shows that cpu doesn't go below C1 for any significant amount of time, but i'm not sure if it answers the question "why?", maybe i just can't read the output of this tool | 15:33 |
gordon1 | yeah, that could be a start, good idea | 15:33 |
buZz | a linux system is many parts that can all be throwing rocks in the gears :) | 15:33 |
bencoh | you don't really need to "replace" /bin/init for that btw | 15:33 |
gordon1 | i think i can just read bq27200/current_now every N seconds, is that reliable source of current consumption? | 15:34 |
bencoh | you can just disable most services and see whether it actually starts entering idle states | 15:34 |
buZz | oh thats a nice route aswell yeah | 15:34 |
buZz | does framebuffer console work on the leste kernel? | 15:34 |
bencoh | gordon1: probably, and if not, there is some script that reads values from the bq27xx fuel gauge using i2cget | 15:34 |
gordon1 | bencoh: well, i thought more of doing some small initrd to make sure there is nothing running | 15:35 |
Wizzup | gordon1: we do know the reason, we don't hit OFF mode currently | 15:35 |
Wizzup | so something keeps kernel awake | 15:35 |
bencoh | gordon1: that would probably work as well | 15:35 |
Wizzup | could be powervr driver or something else | 15:35 |
Wizzup | it's solveable (and I got a later kernel to idle well without pvr loaded) | 15:35 |
buZz | does n900 use the same version of powervr as droid4 ? | 15:35 |
Wizzup | buZz: not currently, no | 15:35 |
buZz | could it? | 15:35 |
gordon1 | Wizzup: yey | 15:35 |
bencoh | (huh, I was about to ask whether it was pvr-related :( ) | 15:36 |
gordon1 | so, to summarize, we can get low power consumption without graphics? | 15:36 |
bencoh | Wizzup: does it reach the default arm idle state (based on wfi) at least? | 15:37 |
buZz | wizzup said he achieved that on a different kernel from the one inside leste now | 15:37 |
gordon1 | is there some "generic" fb stuff that i could use instead of powervr driver? | 15:37 |
gordon1 | buZz: well, i'm compiling my mainline 5.19.2 right now to give it a try, so don't think that's a problem | 15:38 |
Wizzup | gordon1: great, you can start trying it with minimal userspace, see if it hits off mode, then starting maemo | 15:38 |
Wizzup | keep in mind that leste requires powervr in the kernel, and we don't have that in recent lerkenl | 15:38 |
Wizzup | kernel* | 15:38 |
Wizzup | also 5.19? | 15:38 |
gordon1 | oops | 15:39 |
gordon1 | 5.12 | 15:39 |
Wizzup | I recently wrote a long message about pwr mgmt | 15:39 |
Wizzup | let me see if I can find it | 15:39 |
gordon1 | 5.12.9 | 15:39 |
lel | MerlijnWajer opened an issue: https://github.com/maemo-leste/bugtracker/issues/545 (N900: Try to hit OFF mode (low power consumption)) | 15:40 |
Wizzup | gordon1: ^ | 15:40 |
gordon1 | thanks, awesome! | 15:41 |
buZz | Wizzup: could you label that for the milestone? | 15:41 |
buZz | https://github.com/maemo-leste/bugtracker/milestone/22 <- that one | 15:41 |
buZz | well, not sure if its desirable, but at least keeps the info in 1 place? | 15:41 |
lel | MerlijnWajer milestoned an issue: https://github.com/maemo-leste/bugtracker/issues/545 (N900: Try to hit OFF mode (low power consumption)) | 15:42 |
buZz | \o | 15:42 |
gordon1 | how can i verify that it is in OFF mode? is it C7 on power-top? | 15:43 |
bencoh | I would check every /sys/devices/system/cpu/cpu0/cpuidle/state* | 15:44 |
gordon1 | >A shortcut to start on (2) and (3) is to load the latest kernel without powervr on the n900 and start working on OFF mode, without 3d acceleration | 15:45 |
gordon1 | (i'm ok with that) | 15:45 |
gordon1 | (i just need a terminal) | 15:45 |
gordon1 | bencoh: thanks | 15:47 |
gordon1 | thanks, i'll try to play with latest kernel and stuff like pmos/i3wm to check if it can do proper PM | 15:47 |
gordon1 | spoiler: with stock pmos/5.7.0 kernel it cannot | 15:48 |
buZz | quite sure i3wm has nothing to do with power management :P | 15:53 |
Wizzup | gordon1: you do need to enable off mode via sys or proc, and also try to boot to just init=/bin/sh | 15:59 |
gordon1 | ok | 16:00 |
* gordon1 first needs to bisect why 5.12 keeps panicing but 5.7 doesn't | 16:01 | |
sicelo | It's /sys, and recent kernels enable it by default | 16:01 |
sicelo | (Off mode, that is) | 16:01 |
sicelo | 5.12 was booting for me. Even early 5.13 | 16:02 |
Wizzup | gordon1: did you read my wall of text in the ticket I just made? https://github.com/maemo-leste/bugtracker/issues/545 | 16:04 |
gordon1 | i did, yes | 16:05 |
gordon1 | i took pmos config from 5.7, did olddefconfig and now getting panics | 16:05 |
gordon1 | can't read full stacktrace since i don't have UART | 16:05 |
gordon1 | if you have working 5.12 config i would really appreciate if you can share it | 16:06 |
sicelo | I'm on android phone currently, but will see if I have it handy | 16:06 |
sicelo | But did you try omap2plus_defconfig as is? | 16:07 |
gordon1 | well, i thought about that but superficial check showed that it lacks some rx51 options so i decided no to go with it | 16:08 |
sicelo | Options such as? | 16:08 |
sicelo | I'd definitely recommend starting with omap2plus, then enable what you need | 16:09 |
sicelo | I think 5.12 needed mmc aliases too | 16:17 |
gordon1 | i don't remember which one in particular, i just grepped by RX51 and found multiple =n | 16:21 |
buZz | maybe one was RX51_enable_hugepowerdraw | 16:26 |
buZz | :D | 16:26 |
sicelo | :-D | 16:27 |
Wizzup | I mean, we could work on OFF mode now, but then we'd have a n900 without 3d | 16:27 |
Wizzup | and if you haven't seen it - llvmpipe on the n900 is pure sadness | 16:27 |
Wizzup | expect 0.1fps | 16:27 |
gordon1 | do i need to do something in particular to enable off mode? | 16:33 |
Wizzup | gordon1: sicelo said that recently it was enabled by default, but the n900 wiki page on leste.maemo.org should contain a lot of info | 16:37 |
Wizzup | I don't recall currently, sorry | 16:38 |
Wizzup | gordon1: see: | 16:38 |
Wizzup | To enable the OMAP sleep debug LEDs (on the keyboard). Open keyboard, turn off the screen with the slider, and the LEDs will turn off if the phone ever reaches sleep modes. | 16:38 |
gordon1 | yeah, read that, but that's just indication.. | 16:39 |
gordon1 | but that might be helpful | 16:39 |
Wizzup | it is a solid indication since the leds never go off if you don't hit OFF mode | 16:39 |
Wizzup | unless you're in direct sunlight :-) | 16:39 |
Wizzup | brb | 16:40 |
gordon1 | well, booted into pmos recovery initrd with pretty much only /bin/sh running, according to cpuidle/states*/usage it spends most of the time in C2 still | 16:40 |
bencoh | you could check /proc/interrupts for more info | 16:41 |
sicelo | You enable it by echoing 1 to /sys/kernel/debug/pm_debug/enable_off_mode ... something like that | 16:41 |
gordon1 | i would if i find out how to type '|' =/ | 16:42 |
gordon1 | or '>' | 16:42 |
sicelo | Then you can use the LEDs for indication | 16:42 |
sicelo | You can load a console keymap or something | 16:42 |
gordon1 | nvm, it has tee | 16:42 |
bencoh | :] | 16:43 |
gordon1 | #1 is pm_wkup | 16:45 |
gordon1 | which sounds like no shit sherlock | 16:45 |
bencoh | wait, there is a "pm_wkup" interrupt? | 16:47 |
bencoh | sounds fun | 16:47 |
gordon1 | according to /proc/interrupts there is | 16:48 |
bencoh | doesn't show up in maemo5 kernel, so I guess it's ""new"" | 16:49 |
gordon1 | second one is gp_timer, third - 48070000.i2c | 16:49 |
gordon1 | those one >10k | 16:49 |
gordon1 | rest are under 1k | 16:49 |
gordon1 | *those ones | 16:49 |
gordon1 | that's 5.7.0 | 16:50 |
gordon1 | will try 5.12 with omap defconfig | 16:50 |
sicelo | You might find 5.12 is busier than 5.7 :-) | 16:52 |
gordon1 | 5.12.9 with omap2plus_defconfig results in black screen | 16:52 |
bencoh | alright, gp_timer comes from ./arch/arm/mach-omap2/timer.c | 16:53 |
bencoh | (looks like it's the clocksource on omap) | 16:53 |
bencoh | pm_wkup comes from ./arch/arm/mach-omap2/pm34xx.c and it has some funky comment | 16:55 |
gordon1 | i2c thing means something is polling some i2c device, doesn't it? | 16:57 |
bencoh | yeah | 16:57 |
bencoh | well, either actively polling, or reading some i2c registers after receiving some interrupt | 16:57 |
sicelo | gordon1: you need this patch, https://patchwork.kernel.org/project/linux-omap/patch/20210324132442.41907-1-tony@atomide.com/ | 16:58 |
sicelo | For 5.12.x black screen, that is | 16:59 |
gordon1 | sicelo: i don't use mmc right now, it just loads initrd | 16:59 |
gordon1 | but lemme try anyway | 16:59 |
gordon1 | sicelo: 5.12.9 already has that code | 17:00 |
sicelo | Ah. I checked against 5.12.7 which is the most recent I have on my laptop | 17:00 |
gordon1 | omap2plus_defconfig has no FB_OMAP2 enabled | 17:02 |
sicelo | Something else I found with N900 (at least the one I have) is that sometimes you just have to keep retrying boot ... maybe something doesn't reset properly once in a while, ¯\_(ツ)_/¯ | 17:03 |
sicelo | Uses drm fb | 17:03 |
gordon1 | 5.7.0 config uses fb_omap2 and works, i'll try to enable it | 17:03 |
gordon1 | didn't help | 17:07 |
sicelo | I was using drm for my tests | 17:08 |
gordon1 | yeah, omap2plus_defconfig even more unsuccessful than taking config from pmos, can't even get kernel output | 17:08 |
gordon1 | i probably need to solder UART | 17:09 |
sicelo | It needs patience :-) | 17:09 |
sicelo | UART would be best, of course | 17:09 |
gordon1 | oh, no patience involved, it returns back to off state after a while (i assume when panic timeout passes), so i can turn it on again | 17:10 |
Wizzup | gordon1: there's also the patch on the wiki to debug what is blocking deeper idle states | 17:10 |
sicelo | Meant, patience to find good config | 17:10 |
gordon1 | ah | 17:11 |
Wizzup | gordon1: iirc last time tmlind tried with 4.12 it just booted with omap2plus_defconfig | 17:11 |
Wizzup | 5.12* | 17:11 |
sicelo | Will upload config from 5.12-rc8 ... I might have had some unnecessary stuff, can't remember | 17:12 |
gordon1 | well, maybe there is something that is required compiled as module or i am doing something wrong preparing uImage | 17:12 |
sicelo | https://termbin.com/xl7g | 17:13 |
gordon1 | thanks! | 17:14 |
Wizzup | gordon1: what u-boot do you use | 17:14 |
gordon1 | mainline compiled from git | 17:14 |
Wizzup | gordon1: ok, so that has usbtty in there? | 17:15 |
gordon1 | the most fresh one, cloned it yesterday or the day before | 17:15 |
gordon1 | dunno, need to check | 17:15 |
bencoh | I thought they removed support for rx51? | 17:15 |
Wizzup | bencoh: no | 17:15 |
Wizzup | bencoh: just the USB part needs DM porting | 17:16 |
bencoh | oh, okay | 17:19 |
Wizzup | (soon, though) :) | 17:20 |
gordon1 | oh god i'm stupid | 17:22 |
gordon1 | forgot to add console=tty0 | 17:22 |
Wizzup | :) | 17:22 |
bencoh | :) | 17:22 |
sicelo | gordon1: once your kernel boots, you will likely also need https://patchwork.kernel.org/project/linux-omap/patch/20210417083839.46985-1-tony@atomide.com/ ... at least it was recommended to me, the last time I played with off mode (a month ago) | 17:22 |
Wizzup | it would be nice to have a way to do usbtty as console arg | 17:22 |
gordon1 | it still shuts down after few seconds for some reason after initrd is loaded | 17:22 |
sicelo | Haha, yes, it's in my config (console=tty0) | 17:22 |
Wizzup | gordon1: yes, watchdog needs kicking | 17:23 |
Wizzup | unless you disable it with flasher | 17:23 |
gordon1 | ah | 17:23 |
Wizzup | leste does all this for you with it's config | 17:23 |
gordon1 | how can i enable kicking? | 17:23 |
Wizzup | I don't know the exact config, we did the time investment in that before, maybe check our config | 17:23 |
Wizzup | but there is an option for kernel to kick watchdog until userspace takes over | 17:24 |
gordon1 | right, ok, i'll just grep for wdt or something | 17:24 |
Wizzup | https://github.com/maemo-leste/n9xx-linux/commit/91aeab75007ecec0de4854e52b5b0c65d7deb0c7 | 17:24 |
Wizzup | https://github.com/maemo-leste/n9xx-linux/commits/n900/v5.1-config | 17:24 |
Wizzup | actually this branch is the one you want https://github.com/maemo-leste/n9xx-linux/commits/n900/v5.1.21-config | 17:24 |
gordon1 | of course it's in module | 17:25 |
gordon1 | thanks, will turn that on now | 17:25 |
Wizzup | don't make it a module | 17:25 |
Wizzup | yeah | 17:25 |
sicelo | Seems you changed your mind about the config I shared - all that stuff is in there :-) | 17:26 |
sicelo | Anyway, it was still turning off before you see any output on screen? | 17:27 |
gordon1 | sicelo: yeah i thought i fixed mine with console=tty0, but it isn't the case so i'll try yours now | 17:27 |
gordon1 | because on mine keyboard doesn't work :( | 17:28 |
sicelo | um, that's weird | 17:28 |
sicelo | you might want to build UART tool, https://n900.elektranox.org/serial-adapter.html | 17:28 |
* sicelo still intends to | 17:29 | |
gordon1 | yeah, there has to be another way... | 17:29 |
gordon1 | too much PITA | 17:30 |
gordon1 | especially with 2.7V | 17:30 |
sicelo | i've generally been able to make it work without UART so far, but yes, it does annoy in the first stages, always :) | 17:30 |
Wizzup | gordon1: yes, usbtty | 17:30 |
gordon1 | yeah, exactly | 17:30 |
Wizzup | but it will prevent offmode from working | 17:30 |
gordon1 | well shit | 17:31 |
Wizzup | it's still useful to bring up other things | 17:31 |
gordon1 | it is | 17:31 |
gordon1 | potentially, what stops me (assuming i have kernel working more or less) to make g_serial gadget as console? | 17:31 |
bencoh | Wizzup: Ah, I guess that's what the comment in omap3_pm_init() is about (uart breaking OFF idle) | 17:32 |
sicelo | mmm, i thought that usbtty only works while in u-boot, and permanently dies once kernel takes over. | 17:32 |
sicelo | s/usbtty/u-boot usbtty/ | 17:32 |
Wizzup | sicelo: there should be usbtty support in kernel as gadget too | 17:32 |
Wizzup | if we keep it alive in the u-boot -> kernel transition, not sure | 17:33 |
gordon1 | sicelo: keyboard doesn't work in your config either =/ | 17:33 |
bencoh | Wizzup: I would expect the kernel driver to reset the USB conf, but dunno | 17:33 |
sicelo | does your keyboard work in u-boot? in fremantle? | 17:33 |
Wizzup | gordon1: does it work on leste | 17:33 |
Wizzup | bencoh: yeah | 17:33 |
gordon1 | it does | 17:33 |
Wizzup | bencoh: but it's still useful unless you need to debug a problem in between those states | 17:34 |
gordon1 | and it works on pmos 5.7.0 | 17:34 |
bencoh | I guess the usbtty driver spawns a ttyACM device? | 17:34 |
sicelo | very weird! never seen keyboard not working ever | 17:34 |
gordon1 | oh ofc | 17:34 |
bencoh | Wizzup: *nod* | 17:34 |
gordon1 | because it's compiled as module | 17:34 |
Wizzup | bencoh: u-boot or kernel? (also: don't know ;-)) | 17:35 |
bencoh | kernel can spawn an ACM device, but that's up to user to choose (or kernel builder) to choose; I was referring to uboot :) | 17:35 |
gordon1 | will g_serial/acm stop device from entering sleep? | 17:36 |
gordon1 | s/sleep/off/ | 17:36 |
gordon1 | btw is there some key combination/long press that makes hardware reset? | 17:37 |
* gordon1 tired of pulling/plugging battery in/out | 17:38 | |
gordon1 | finally it works now | 17:38 |
gordon1 | sicelo: thanks for config | 17:38 |
sicelo | yw | 17:38 |
buZz | afaik holding powerbutton for >10 secs? not sure | 17:39 |
sicelo | but yeah, there has to be something wrong somewhere with reset or some such - you may find it doesn't boot next time :P | 17:39 |
bencoh | gordon1: I wouldn't trust any power measurement with usb plugged in | 17:39 |
gordon1 | hey, i have 1 in state6/usage | 17:40 |
gordon1 | yey | 17:40 |
gordon1 | that's a start | 17:40 |
sicelo | in what? :) | 17:40 |
bencoh | cat state6/desc ? is that C6? | 17:40 |
gordon1 | /sys/devices/system/cpu0/cpuidle/state6/usage | 17:41 |
gordon1 | MPU OFF + CORE OFF | 17:41 |
bencoh | oh | 17:41 |
bencoh | neat | 17:41 |
gordon1 | i mean 1 is a bit on the low side compared to state1's 4010 | 17:42 |
gordon1 | but still | 17:42 |
gordon1 | and state4's usage is considerably higher | 17:42 |
bencoh | asuming you leave your device alone, it should go into state6 way more often | 17:43 |
bencoh | it's worth a try | 17:43 |
gordon1 | ok, lemme try | 17:44 |
gordon1 | can't quite turn off the screen from initrd | 17:45 |
gordon1 | is there some magic sysfs thing i can do for that? | 17:45 |
bencoh | I might be wrong, but I don't think it should really matter | 17:45 |
bencoh | unless the backlight is supposed to wake the cpu | 17:45 |
sicelo | it matters, iirc (and a lot!) :) | 17:45 |
bencoh | it matters for power consumption | 17:45 |
bencoh | but not for idle states | 17:45 |
bencoh | unless I missed something :) | 17:45 |
gordon1 | pm_wkup is still on top in /proc/interrupts, but gp_timer is gone (i guess that's "clockevent" now) | 17:46 |
sicelo | something like, echo 1 > /sys/devices/platform/omapdrm.0/graphics/fb0/blank ... although we're not sure if that really turns it off | 17:47 |
sicelo | s/turns it off/helps for this test/ | 17:47 |
bencoh | gordon1: yeah, apparently clockevent replaced gptimer | 17:48 |
bencoh | although it doesn't seem to be exactly the same thing, SoC-wise (I'm just reading commit messages) | 17:48 |
sicelo | at least to have OFF counts go up, screen must be off (at least that's what i was told here) | 17:49 |
gordon1 | yey it's 7 now | 17:50 |
bencoh | sicelo: oh, do you mean the fb driver does some active refresh? | 17:50 |
sicelo | gordon1: cat /sys/kernel/debug/pm_debug/count (some path like that) ... see if you can share what you get | 17:51 |
bencoh | (I mean, pushes frames to LCD even though there was no update) | 17:51 |
bencoh | (Oh, btw: fb2c599f056 / Fri Sep 11 18:12:09 2020 / ARM: omap3: enable off mode automatically) | 17:52 |
gordon1 | well, can't quite cop-paste it here | 17:52 |
gordon1 | is there something particular i need to be searching for? | 17:53 |
sicelo | yes this is why nowadays off mode is auto-enabled | 17:53 |
sicelo | maybe check the OFF value in core_pwrdm | 17:54 |
sicelo | and/or the others | 17:54 |
gordon1 | core_pwrdm: ... OFF:0 | 17:55 |
sicelo | haven't hit off mode yet :( | 17:55 |
sicelo | anyway, not a surprise ... you need to suspend all the ttys, etc. | 17:55 |
gordon1 | well, at least i have working kernel to play with | 17:58 |
sicelo | yes | 17:58 |
sicelo | :-) | 17:58 |
sicelo | i'll join you again soon as well - i was working on it, then got a busy new job ... things are slowly getting back to normal now | 17:59 |
sicelo | but i would suggest you do consider the UART thing - maybe where you live it's not going to be so difficult to get the components you need | 18:00 |
sicelo | gordon1: you'll need to do something similar to https://github.com/tmlind/droid4-pm/blob/master/scripts/openrc/droid4-pm#L88 (of course slightly different paths and names) | 18:07 |
sicelo | and be sure to blacklist the bluetooth driver (it doesn't really work anyway) | 18:08 |
gordon1 | i'm just going to turn if off in config | 18:11 |
gordon1 | hey kernel paniced again wtf | 18:11 |
gordon1 | i need to get this panic stacktrace somehow | 18:11 |
gordon1 | ah, it panics when i boot kernel with keyboard open | 18:29 |
gordon1 | i suddenly lost interest in this panic | 18:30 |
gordon1 | when i put UARTS to sleep it start shitting the console with ???? even after i removed console=ttyS2 from boot cmdline | 18:31 |
Wizzup | gordon1: is your back cover open? | 18:32 |
gordon1 | yes, it is, why? | 18:33 |
Wizzup | that (used to?) make sd card inaccessible | 18:33 |
gordon1 | interesting | 18:33 |
Wizzup | so if the panics are about not finding rootfs somehow | 18:33 |
Wizzup | I think it should be fixed now though | 18:33 |
gordon1 | it definitely is not rootfs panic since i have no rootfs | 18:33 |
Wizzup | ok | 18:33 |
gordon1 | i boot initrd only | 18:33 |
gordon1 | it has something about cpuidle in the stacktrace | 18:34 |
sicelo | maybe something recent ... keep checking - tmlind might also have ideas for you | 18:44 |
sicelo | i don't recall having any panics when last i tried | 18:44 |
gordon1 | nah, i'm in a process of bisecting that and it's somewhere in 5.9 kernel | 18:45 |
sicelo | ok | 18:45 |
gordon1 | i mean not sure if i want to continue this tedious process since i have workaround | 18:45 |
_uvos_ | Wizzup maybe also add hit off mode on omap4 to that bug | 18:58 |
_uvos_ | or make it about ret instead of off | 18:58 |
_uvos_ | gordon1 even if you dont continue please report the bug to the kernel ml with the best commits before/after you have | 19:01 |
Wizzup | _uvos_: omap4 hits ret, no? | 19:10 |
_uvos_ | Wizzup: yeah but omap4 can hit off kernel side. so either the bug is about off mode working nowhere on leste or ret not woring ok n900 | 19:18 |
_uvos_ | *omap4 cant hit off mode kernel side | 19:18 |
Wizzup | _uvos_: yeah not sure what the status of that is | 19:22 |
gordon1 | uvos: i'm not sure that it's going to be any help unless somebody would be able to reproduce this panic or i could extract the stacktrace | 21:43 |
gordon1 | i will probably continue the bisection | 21:44 |
diejuse | I made the second part of the video of Maemo Leste running on Android with the chroot technique. I hope you like it. https://www.youtube.com/watch?v=kipuT0VXzC4 | 22:29 |
bencoh | diejuse: that's pretty neat! | 22:35 |
bencoh | I didn't think we had that many mobile-friendly apps already available :) | 22:35 |
dreamer | the music under this video is somewhat disturbing .. | 22:39 |
dreamer | but indeed quite a lot of apps seem to work great | 22:40 |
dreamer | and that titan thing has bit of an odd resolution too | 22:40 |
diejuse | bencoh: Well, there are quite a few Debian applications. :) | 22:41 |
bencoh | yeah, but still | 22:42 |
bencoh | and it seems to work pretty smoothly too :) | 22:42 |
diejuse | dreamer: you are right but I just wanted a background music that did not give me copyright problems. It is randomly generated music. | 22:42 |
dreamer | hehe | 22:43 |
diejuse | dremear: Titan has a square screen like Blackberry Passport. I really love this screen because it is very readable. | 22:44 |
diejuse | bencoh: Everything works very fast. I honestly feel free as when I use my computer with Ubuntu. | 22:45 |
bencoh | :) | 22:46 |
bencoh | well, the 4x2GHz helps with that I guess | 22:46 |
bencoh | (I would probably give a try to the titan, if it didn't have a shared modem/cpu design) | 22:47 |
uvos | bencoh: you have one? | 22:47 |
bencoh | uvos: no, I don't, because of the shared modem/cpu issue | 22:57 |
bencoh | (like 99% of the modernish devices) | 22:57 |
Wizzup | bencoh: lots of those are from the repos | 22:59 |
Wizzup | bencoh: err, sorry, debian repos | 23:07 |
buZz | diejuse: fyi, you can use dr<tab> in most irc clients to auto-fill to 'dreamer: ' | 23:09 |
Wizzup | diejuse: nice video :-) | 23:10 |
Wizzup | diejuse: btw, it looks like the qt theme for dorian wasn't the correct one | 23:10 |
diejuse | buZz: I didn't know, thank you. I never used IRC until I searched for this channel. | 23:10 |
buZz | :) cool! | 23:12 |
buZz | irc is funny cause its optimized to txt chats since ~1988 | 23:12 |
uvos | "optimized" | 23:14 |
uvos | i think what the irc protcoll and extensions are concerned its more like "arcane" :P | 23:14 |
buZz | :D | 23:15 |
buZz | but this led to all kinds of optimizations in clients towards aiding expression | 23:15 |
diejuse | buZz: I really used it 15 years ago with MSN Messenger to try to flirt with women until facebook and social networks appeared. :D | 23:18 |
buZz | :D | 23:18 |
diejuse | Wizzup: Thank you! Where can I see what Dorian should look like? | 23:19 |
Wizzup | diejuse: well, the bar at the botton should match the gtk2 theme | 23:21 |
Wizzup | # env|grep QT | 23:21 |
Wizzup | QT_STYLE_OVERRIDE=maemo5 | 23:21 |
Wizzup | QT_QPA_PLATFORM=maemo | 23:21 |
diejuse | Wizzup: The icons in the bar below should look bigger and prettier, right? | 23:25 |
Wizzup | diejuse: yes | 23:32 |
diejuse | Wizzup: it worked. :) | 23:34 |
Wizzup | great | 23:34 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!