Wednesday, 2020-06-03

T42<edp_17> Good evening. Can someone tell me how I can let ril know the partition path where the radio lives, please? Thanks.00:32
T42<erfanoabdi> @KernelPanix [u make GSI SailfishOS ?], No I'm not planning to make any GSI for Sailfish06:32
rinigus@erfanoabdi, piggz: when grepping aosp sources for fpdata, I found `/data/system/users/[0-9]+/fpdata(/.*)` and `/data/vendor_de/[0-9]+/fpdata(/.*)` in system/sepolicy/private/file_contexts06:41
rinigusso, we could probably put uuserID into the path where it expects a number06:42
rinigusthat way it should comply with expected path and allow us to have multiuser supported out of the box06:42
rinigus@erfanoabdi: have you tried such path instead of using 006:43
T42<erfanoabdi> It's just on sepolicy06:44
T42<erfanoabdi> In frameworks i remember it was 0 only06:44
T42<erfanoabdi> But you can try it, this way isn't helping us in ubports:(06:44
rinigusspiiroin: looks like EV_MSC:MSC_GESTURE:4 works very well for wakeups when double taping fp sensor. thank you!07:22
rinigusis there a reason why mce has fake doubletap enabled by default? I am sure it wasn't set by me. what does it do?07:25
spiiroinrinigus: "fake doubletap" = double taps mce derives from regular touch events. in practice: "lpm" ui on screen + double tap -> wakeup to regular ui09:48
spiiroini.e. cases where touch firmware is not in double tap detection mode and ui is not supposed to handle touch events09:49
rinigusspiiroin: ok, so its idle if lpm is disabled, right?09:51
rinigusand not pulling touch events while device is expected to be just sleeping09:52
spiiroinrinigus: kind of but reverse causality or stte ;-) if kernel emits touch events, mce processes them. used for double tap in situations where UI has been told to ignore them. IIRC that is only: display==lpm || (display==off && suspend==disabled)10:01
spiiroinrinigus: but yes, that fake double tap setting can (and should) be enabled10:02
spiirointhat non-lpm part: "mcetool --set-suspend-policy=disabled" -> display is not truly powered off -> double tap works even if device fw does not support it @ expense of draining battery much faster10:06
rinigusOk, but we should have suspend enabled normally, so that path is mainly for developers10:07
rinigusspiiroin: thanks for insights, as always :)10:08
T42<edp_17> @kalube: Can you help me with the s-pen, please?12:08
T42<edp_17> I have added the evdevtablet into the droid-hal-device.conf, it looks like this:
T42<edp_17> Last time you told I should be able to make it work without messing around in the kernel as its drire is wacom. Can you let me know how, please? Thanks.12:09
kalube@edp_17 I'd recommend reading back through logs and rereading what I wrote, it shouldn't be too hard to hack in mouse emulation if you're sure it's the right approach. I'm afraid I don't really have time to help out right now12:17
T42<edp_17> @kalube: Thanks. Well, I started with that, reading back what you wrote. I'll try to work this out.12:18
T42<ankaos> How did you compile los-15.1? I don't see system_root.15:48
T42<ankaos> usr/libexec/droid-hybris/system/bin/logcat -b radio output16:17
T42<ankaos> where is there an error?16:29
riniguspiggz: how is it going? help needed or in the middle of the coding fun?17:15
T42<adampigg> rinigus: not had a change to really get stuck into it yet....i got asked to have a build for the volla phone ready for the weekend so had to do that17:22
T42<adampigg> if you want to tackle it, go for it ... i got as far as changing the map to a <uint32_t, QPair<uint32_t, QString> and adding a member for the current user id17:23
rinigus@adampigg: I can surely work on it. but I am not sure I need such a map... as I would have kept maps for different users separate.17:25
T42<adampigg> ah, ok, maybe17:25
riniguslet me try then. I'll make a new PR earlier and then you could give feedback early as well.17:26
rinigusthen maybe you could include it all into volla build as well17:26
rinigus@adampigg ^17:26
rinigus@adampigg: would you mind to review last night's PR though.17:28
rinigusthat fixed issues in our code with clear17:28
T42<adampigg> sure17:28
T42<adampigg> knock yourself out, im struggling with why this build doesnt boot!17:29
riniguspiggz: can I merge #21?17:38
T42<ankaos> ı see sim card. but not call not sms not net17:41
T42<adampigg> rinigus: merged17:50
rinigus@adampigg: thanks!17:51
T42<ankaos> I am having sim card network not found problem.18:45
piggzmal: rinigus: do you have a ld.config.28.txt that allows loading libselinux_stubs from /usr/libexec.... ?19:43
piggznevermind, found in the faq19:56
riniguspiggz: I presume you talk about ; don't think it enables anything special19:57
piggzrinigus: this line
piggzrinigus: finally, volla config is working with fingerprint out of the box21:14
piggz(using older version of fpd)21:14
zinstackHello, anybody here willing to check my journalctl log: Also, there's my kernel config, if that could be a problem: There's no display, only "vendor logo", if you can consider unlocked bootloader warning that. test_hwcomposer and surfaceflinger don't crash, but don't do anything either. Writing anything to /dev/fb0 will turn display off. mdss seems ok,21:29
zinstackstartup wizard seems to be going, but still, no picture. My platform is sdm632, believe it's caf/msm899621:29
T42<elros34> you did not disable selinux correctly:
T42<elros34> also enable CONFIG_FW_LOADER_USER_HELPER_FALLBACK21:37
T42<elros34> can't see any rc files loaded but maybe because it fail so early21:39
zinstack@elros34 Thanks for attention. Will try that now. Should rc files be registered in journal? The ones that disable services in particular don't seem to be having any footprint, but the still are considered, as without them journal is flooded with errors21:48
T42<elros34> android services must be started by droid-hal-init but first fix selinux21:51
T42<elros34> I think rc files are usually visible in journal21:52
zinstack@elros34 Mind telling me what they are? in case I omitted them completely21:53
T42<elros34> + some from /system21:54
zinstackAnd I'm on hybris-16.0, should those be in there too?21:55
T42<elros34> First fix selinux and flash new hybris-boot.img then you can worry about it. I am pretty sure I asked about it but get no answer so do you have /usr/sbin/ohmd? It fails for some reason21:56
zinstackohmd is present21:58
zinstack@elros34 Here's a new journal log: Looks pretty similar to what was before. Also, in spite of journal reporting ohmd failing, it hapilly runs (start too early, then restart on time?)22:06
T42<elros34> did you bother to use mer-kernel-check after making hybris-boot?22:07
zinstack@elros34 WAIT WHAT? I swear I ticked the bootparam. Must've been asleep while doing that22:10
zinstackHold on, selinux is not even in kernel. I will do it properly now, though, hold on22:12
T42<ankaos> how used to call?22:16
zinstack@elros34 So I did it properly, still dead22:45
T42<elros34> Do you have no selinux disabled at boot in logs?22:49
zinstack@elros34, it doesn't tell me it's loaded, nor unloaded22:54
zinstackI do, however, have selinux=0 in bootparams22:55
zinstackThe message should be right before Yama, I think, however, it's not. Weird22:56
T42<elros34> did you follow PR I linked? How about you show your current defconfig22:56
zinstack that one? - yes, still nothing. That's the new defconfig:
T42<ankaos> Currently, only calls are problematic in ril.22:59
T42<elros34> @zinstack so show fresh journal and zcat /proc/config,gz.  Also did you follow my advice to how to use gzip or do you still copy everything manually23:01
T42<elros34> do you really had CONFIG_DEFAULT_SECURITY="smack" by default? How about you start using github instead so I can see what changes you are making23:02
zinstack@elros34 Fresh journal log: Running config: I have resolved the problem with twrp, it was system_on_root thing, i do copy stuff manually still, I guess I've missed your message. I use github, it's just a little complicated for me to upstream today, I tested different kernel branch to see if it'd be different. It wasn't23:10
zinstackI do push occasionally, but since there were no progress today, I didn't.23:11
T42<elros34> zinstack you seems to use halium based defconfig with disabled CONFIG_DEFAULT_SECURITY_SELINUX  and some other changes. I suggest you to use pure lineage one and make only one selinux related change CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=0.23:21
zinstack@elros34 Actually I do. Today I thought of using kernel branch of halium as a base, since halium is based on hybris (what I think), I really have shot myself in a foot, heh. Ok, will try tomorrow, it's really late here23:25

Generated by 2.17.1 by Marius Gedminas - find it at!