Thursday, 2019-12-12

r0kk3rzyou would have been exceptionally lucky if everything worked in the first build ;)00:09
T42<edp_17> :) Yes, I wanted to be something like that.00:09
T42<edp_17> But probably my knowledge is not deep enough to achieve that.00:10
T42<edp_17> Not in any way. :)00:10
r0kk3rzit almost never happens00:10
r0kk3rzeven for experienced porters00:10
T42<edp_17> For the hammerhead I made 16 builds while it started to be usable. Then another 3 with changes that needed to reflash the image. Plus two because the MTP didn't work and got help here.00:12
T42<elros34> your defconfig have disabled CONFIG_MEMCG but mer-kernel-check doesn't complain so  pastebin your out/*/.config or zcat /proc/config.gz00:13
T42<edp_17> Certainly.00:13
T42<elros34> do  you run mer-kernel-check against your defconfig or .config file?00:14
T42<edp_17> seemingly against .config00:15
T42<edp_17> That was in the hadk.00:15
T42<elros34> so I don't get it00:15
T42<elros34> anyway you need this enabled:
T42<edp_17> my .config is:
T42<edp_17> Thank you. In the meantime I created and flashed another image. (removed the android.mount and android_external.mount + included all changes except this last one into the kernel.)00:23
T42<edp_17> Still, stuck on the samsung logo and telnet doesn't work.00:24
T42<edp_17> still don't have /system mounted00:25
T42<edp_17> can I mount it manually?00:26
r0kk3rzyou have telnet?00:27
T42<elros34> how about you provide fresh journaltl instead creating new images over again? In latest your systemd stuck so sure you can mount system but it will not help00:27
T42<edp_17> I agree.00:30
T42<edp_17> The fresh 'journalctl --no-tail -f':
T42<edp_17> Sorry @r0kk3rz, I just see your question. By default I do not have telnet. I need to manually assign a mac then the IP address to get that.00:37
r0kk3rzok, thats not so bad00:38
T42<elros34> udev still waits for something, and you didn't remove every files related to your custom mount units. Alsonally show your /lib/systemd/system/system.mount00:38
T42<edp_17> system.mount:
T42<edp_17> I did remove all related stuff. At least under the droid-configs/sparse folder.00:45
r0kk3rzand you rebuilt the config package?00:45
T42<elros34> system.mount looks good but one of the unit in /lib/systemd/system/*.mount must be wrong00:46
T42<elros34> android_external.mount: Where= setting doesn't match unit name. Refusing00:46
T42<edp_17> Yes I did rebuilt the configs with : 'rpm/dhd/helpers/ --configs'00:46
T42<edp_17> Right after I compiled the kernel.00:47
T42<edp_17> Hmm. In /lib/systemd/system I still have the android.mount and android_external.mount. How, if I deleted them from droid-configs/sparse?00:49
r0kk3rzcheck if the file is still on the device00:49
T42<edp_17> If I delete these two files from the device and reboot, can that help?00:49
T42<elros34> ... I told to do this on device and grab new journal hour ago00:49
T42<edp_17> I know. But on my device at that time there was no sailfish installed. I was building the image with all you recommended (including removing these files) then I flashed this onto the device.00:51
T42<edp_17> Now I deleted these two files from the device.00:52
T42<elros34> if you are not sure whether your changes are included you can always unpack droid-hal or droid-config*rpms from droid-local-repo/$DEVICE/00:52
r0kk3rzsounds like you missed something in your build process00:53
T42<edp_17> Seemingly I did.  :(00:53
T42<edp_17> Apologies I didn't want to waste your time.00:54
T42<edp_17> Either mine.00:54
r0kk3rzi wonder if your kernel changes made it in either00:56
T42<edp_17> Good point.00:56
T42<edp_17> journalctl:
T42<elros34> r0kk3rz: I dont get his kernel changes, here memcg is disabled but mer-kernel-check in droid-hal-$DEVICE.log doesn't complain abou it00:57
r0kk3rzyou did repackage the hal right? with build_packages.sh00:59
T42<edp_17> The memcfg change is not in this build. Elros told me after the build was baked.01:00
r0kk3rzyou can flash the boot.img with the kernel changes01:01
T42<edp_17> These are the steps I have done:
T42<edp_17> @r0kk3rz: Where can I find the boot.img?01:05
r0kk3rzout/ from the kernel build process01:06
T42<elros34> out/target/product/$DEVICE/hybris-boot.img NOT boot.img01:06
r0kk3rzthats the shortcut for doing kernel changes01:06
T42<edp_17> Thanks. I was looking in out/target/product/herolte/obj/KERNEL_OBJ/arch...01:07
r0kk3rzfind is your friend01:08
T42<elros34> you should consider this to get rid of noise from your logs
T42<edp_17> Thank you.01:10
T42<edp_17> I have added the memcg changes and applied this last patch to remove the debug messages. Now I am compiling the kernel.01:18
T42<edp_17> Something is wrong because it is compiling the kernel for more than 10 minutes.01:30
T42<edp_17> Huh, it has completed and flashed on the device.01:36
T42<edp_17> mer-kernel-check:
r0kk3rzget the new journal01:40
r0kk3rzyou should also try and fix those warnings01:41
T42<edp_17> I got similar in hammerhead build so I may come back to them later, once the image works ok.01:43
T42<edp_17> fresh journal:
r0kk3rzsystem mounted?01:54
T42<edp_17> No, unfortunately. :(02:00
r0kk3rzit doesnt look like its even trying02:00
T42<elros34> it's still same, you should have "Mounting Droid mount for ". Have you checked all mount units in /lib/systemd/system/*.mount? Maybe just backup and remove them (leave only system.mount) to confirm that this is the issue02:02
T42<edp_17> I have a few ones:
T42<edp_17> I have spotted something else. In my edp17 branch here:
T42<edp_17> the dhd is pointing to an unkown place.02:18
T42<edp_17> On my pc the dhd folder is refreshed from mer-hybris.02:18
T42<edp_17> How can I push this into  dhd?02:18
T42<elros34> cd to dhd and make sure it points to mer-hybris: git remote -v then git checkout to latest commit02:24
T42<edp_17> The reason I need this , this is used in my local_manifest. And because the dhd is pointing to nowhere, it is always pulling down the old helpers file from ISDriim repo.02:24
T42<edp_17> the git remote -v returns with DRiim's repo.02:25
T42<edp_17> however I refreshed it from mer-hybris.02:27
T42<elros34> just to make sure: you used latest dhd from mer--hybris when building packages not 2 yesrs old?02:28
T42<edp_17> Now I am not sure anymore. I think, better if I delete and re-do the content in rpm folder.02:29
T42<elros34> go to dhd and run git log02:29
T42<edp_17> But if you click on the 'dhd @ 06faa02' link in my repo, that takes you a 404 page. So, when I try to use this in my local_manifest, it won't work.02:32
T42<elros34> this git log looks good except latest commit. I have no idea what do you need local_manifest at this stage, you already synced all your repos. I didn't run repo sync in my HABUILD since ages and I wouldn't even try to not loose my local changes02:35
T42<edp_17> Well, this should be ok then.02:36
T42<elros34> this commit is not ok 06faa020039fa50208b9b7e4a729b2521d6a0b4c02:36
T42<edp_17> Why?02:37
T42<edp_17> That was when I refreshed dhd from mer-hybris.02:38
T42<edp_17> I think I got it. When I refreshed dhd, I committed my work when I was IN dhd folder, not from rpm02:38
T42<edp_17> So, probably I need to revert that commit and refresh dhd while I am in rpm, not in dhd.02:40
T42<elros34> git reset —hard is your friend. Anyway I need to sleep a little, leave a feedback whether removing those mount units (except system) helps cause I have no other ideas why systemd stuck02:42
T42<edp_17> Of course. Thank you for your support so far. Sleep well, good night. Probably I'll go too sleep soon.02:42
T42<edp_17> I have removed all .mount except system.mount, rebooted and checked the mounts. Still there is no /system anywhere.02:50
T42<edp_17> To be continued...02:51
T42<birdzhang> @edp_17 where is your kernel source? point me the github url02:53
T42<edp_17> Hi @birdzhang !02:53
T42<birdzhang> on my device(vince), i have to edit boot dtsi02:53
T42<edp_17> I used this kernel (cm-14.1):
T42<edp_17> I am soon fainting out so, probably need some sleep. :(02:57
T42<edp_17> Good night everyone.02:57
T42<birdzhang> Good night02:58
T42<adampigg> quick silica question.... i have a 'base' Page{} qml file, with a flickable, pulldownmenu and an item07:41
T42<adampigg> the is exposed as a default property so 'implementation' pages can add a flickable to it07:42
T42<adampigg> i want to do something similar with the pulldown menu, but when I expose that and add items to it, they all appear ontop of each other07:43
T42<adampigg> i exposed it as menu._contentColumn07:43
T42<adampigg> and add menuitems as an array07:43
T42<adampigg> eg, menu: [......]07:43
T42<adampigg> onlything i can think to try is adding using javascript, setting the parent to the menu contentcolumn, but im not sure how to set the onClicked handler doing it that way07:45
T42<adampigg> q1q1.>07:46
Mister_Magister@adampigg i believe qt questions go to #sailfishos channel :P07:55
T42<adampigg> yeah yeah, but i dont have that on telegram08:04
r0kk3rzyou wanna what with the who?08:19
T42<adampigg> r0kk3rz, populate a pulleymenu from a 'subclass' of a Page08:44
T42<adampigg> by exposing the menu control as a property08:44
T42CYB3R0ID694 was added by: CYB3R0ID69411:42
T42<CYB3R0ID694> Anyone tried to port Sailfish on OnePlus 6?11:45
malsomeone was already porting it11:46
T42<CYB3R0ID694> If anyone building & need tester, ping me.. I would love to be tester12:12
Mister_Magistermal: so what do we do with udev rules12:14
malfix them?12:15
Mister_Magisterbut how12:15
Mister_Magisteridk how12:15
Mister_Magistermal: c…can you help me with that or no chance12:30
malI can't help right now, try to search online how to make absolute links in udev rules12:31
* Mister_Magister i can't do it rn either12:31
deathmistmal: any news about libhybris submodule update? I've been using my fork @ for over a week and I've seen no issues so far myself, this at least provided fixed audio out of the box on 16.014:56
Mister_Magistermal: i don't think you can make absolute symlinks15:03
malpretty sure you can but might need some magic15:05
maldeathmist: we have been thinking if we should include some things there before submodule update15:05
*** ahjolinna_ is now known as ahjolinna19:22
rinigusvknecht & mal: when comparing xperia 10 and tama guides, looks like we do miss when building systemimage vendorimage20:02
rinigusright now cannot say whether that patch is touching anything involved in systemimage or vendorimage20:02
rinigusfor hybris-hal it was applied20:03
vknechtrinigus, funny you speak of it, just experiencing "dbus-daemon[2648]: Failed to set up security class mapping (selinux_set_mapping():Invalid argument)."20:11
vknechtguess should disable again in /etc/selinux/config20:12
rinigusvknecht: I will try to see where do those files end up on devices meanwhile20:13
malrinigus: seems quite important20:31
rinigusmal: maybe, checking whether its used on android or hybris side. have to apply them and see whether they will influence the build20:32
vknechtdisabling in /etc/selinux/config cleared that dbus fail20:33
rinigusmal: if it is on android side then the logic of keeping android as-is will not apply for selinux enabled builds. or we will have to patch it while installing20:33
malrinigus: well missing dbus will cause everything to fail20:43
rinigusmal: agreed! I have to rebuild full image, so it will take a while. right now I do disable selinux in /etc config.20:45
malI'm porting fp2 to new android base, again20:47
rinigusmal: on xperia 10, do you have dbus in /system/etc/selinux? `grep -r dbus` . I suspect it ends up there20:50
rinigusfp2 - which base?20:51
malandroid 7, fairphone finally released the official sources for their android 7 base20:51
piggzrinigus: on your port, have you had firewall issues with the usb connection?20:51
malI was also considering lineage 16 or 1720:51
malrinigus: there are two lines with dbus in that folder20:53
riniguspiggz: don't think so, but I may have tested for developer only.20:53
rinigusmal: thanks! ok, so that's what I am missing. and I suspect vknecht as well. will rebuild systemimage and see if I can get selinux up and running in permissive mode on xz2 with these files changed20:55
rinigusmal: and I presume some hits are in /vendor/etc/selinux as well20:59
malI think could be in /vendor/etc/selinux/precompiled_sepolicy21:00
riniguswhen checking with drmservice (assuming its similar), I've got a hit in /vendor/etc/selinux/plat_pub_versioned.cil:21:02
riniguswell, will know tomorrow :)21:03
T42<edp_17> Good evening.23:43
T42<edp_17> I had an idea with my S7 build. As it still stuck on the samsung logo and seemingly there is no /system is mounted. So, I downloaded DRiim's image from xda and installed it on my S7. (That is an older image Then I wanted to copy over the /lib/systemd/system/system.mount into my build to see it boots up with that. But before23:47
T42I did this, compared the system.mount in both images and realized they are the same.23:47
r0kk3rzyeah theres nothing wrong with the file23:48
T42<edp_17> So it seems my image doesn't boot NOT because of the system.mount in my image.23:48
T42<edp_17> Then I thought it mus be the kernel in my image.23:48
T42<edp_17> So, I flashed my kernel over when DRiim's image was installed on my S7 and rebooted. It booted up correctly, so it is not the kernel either.23:49
r0kk3rzthe kernel didnt seem that unhappy23:49
T42<edp_17> I am lost now.23:49
T42<edp_17> Hi @r0kk3rz23:49
T42<edp_17> What I spotted is Driim's image doesn't get the max address + ip address either. To access that, I also needed to do the manual trick that Elros told my on the other day.23:50
T42<edp_17> Well, anyway, this is where I am now and I am out of idea what to change to get my image boot up.23:51
T42<edp_17> I have a fully build image and all the rpm packages. Is there any way to install them all on Driim's image?23:52
r0kk3rzyou dont want to do that23:53
T42<edp_17> I know this is very messy, but don't have any other idea.23:53
T42<edp_17> I thought if I flash my kernel on his image and replace all packages, then I am technically using the one that I built.23:54
r0kk3rzflailing around madly wont help23:54
r0kk3rzsystemd has a bunch of diagnostic tools we can use to figure out wtf is going on23:54
T42<edp_17> I see, then It was a desperate but bad idea.23:55
r0kk3rzbecause for some reason its not mounting system, which is obviously the problem23:55
T42<edp_17> yes23:55
T42<edp_17> The problem is that I am not at all familiar with this Linux version (Sailfish) and I never really used any diagnostic tools on other version either.23:56
r0kk3rzfor eg. systemd-analyze plot > /home/nemo/startup.svg23:56
T42<edp_17> systemd is a black whole to me.23:56
r0kk3rzthis will give you a picture of the boot order23:56
T42<BusterBg_18> Hi all, what do I need to debug to get lipstick to start?23:57
r0kk3rzyes, this is porting, you're going to need to learn some things23:57
T42<edp_17> systemd-analyze command not found23:57
T42<edp_17> Yes I am willing and happy to learn!23:57
T42<BusterBg_18> Trying once again to port hybris-16 and as always not getting a damn GUI23:57
T42<edp_17> I meant systemd is a black hole rather than whole. :)23:59
r0kk3rz@BusterBg_18: check things are booting properly first, droid-hal-init finishes properly, other things start up23:59

Generated by 2.17.1 by Marius Gedminas - find it at!