Friday, 2021-08-06

T42<dgmarie2> does anybody know what might be causing this error00:25
T42<dgmarie2> it happens when running rpm/dhd/helpers/ --droid-hal00:26
piggzbit of a pain, maximum artifact size in 1gb ... and my sfos 4.1 image is coming in at just over that!!!07:30
riniguspiggz: build locally? github allows to attach larger files to the releases - see
piggzrinigus: you know i hate local-anything!08:26
piggzi think first ill remove some of the dev packages from the image08:26
riniguspiggz: not the big fan either, but it turned out to be less pain to run mic locally. at least for me08:27
T42<NotKit> it can also run on GitHub Actions in theory08:27
piggzeasy fix will be if removing gdb, strace etc will bring me in just close enough!08:39
T42<elros34> @dgmarie2 try to replace 'kernel_release=$(cat out/target/product/%{device}/*/*/include/config/kernel.release)' with 'kernel_release=$(cat out/target/product/%{device}/*/KERNEL_OBJ/include/config/kernel.release)' in
T42<edp_17> @elros34 : I am not sure what is disabled in the init*.rc file but I put them all here:
T42<edp_17> Regarding the hwcomposer, I am not using any specific one. At leats there is nothing in my hybris/mw. I only use your qt5-qpa-hwcomposer-plugin.10:14
T42<edp_17> I only can see the power related stuff in udevadm monitor -p:
T42<elros34> @edp_17 about init*rc file check git log in your device/vendor/* and you will know whether you disabled any service which might be hdmi/usb/slimport related. I am talking about android part of hwcomposer (hwcomposer.*.so from /system/lib/), this one trigger hotplug function in qpa-hwcomposer-plugin. Sources should be somewhere in hardware/*/*/libhwcomposer or something like that11:15
T42<elros34> what about that sysfs value I asked/11:16
T42<elros34> something is missing on android part, there is nothing slimport related in udev. Have you tried on cm12 whether it works?11:22
mighty17[m]@herrie1982! Hi remember me (the guy with Samsung Galaxy Tab 2 :P) finally got halium working just wanted to thank you11:39
T42<edp_17> @elros34 : Nothing is changed in device/vendor/*:
T42<edp_17> about hwcomposer, this is what I have on device:
T42<elros34> so, then in your source you must have corresponding source code for it.  If you run make hwcompose.msm897 you should see which source it use12:42
T42<elros34> in your $android_root*12:42
T42<edp_17> Yes, I have a source for that (plus for some other versions too):
T42<Herrie1982> OK great! 3.4 kernel? (re @SailfishFreenodeIRCBridgeBot: <mighty17[m]>@herrie...)12:46
T42<edp_17> @elros34 : make hwcompose.msm897 in HABUILD:
T42<edp_17> @elros34 : make hwcompose.msm897 in host (mer/android/droid):
T42<elros34> did you run regular commands in habuild before? Obviously I made a typo and you copy/paste it12:48
T42<elros34> time to clean up my keyboard..12:49
T42<edp_17> Yeah, I copy pasted. 😊12:52
T42<edp_17> @elros34 : I got it. 😃 hwcomposer.msm8974 is built:
T42<edp_17> On your question, I couldn't try TV out port on the CM12.1 rom itself, because it doesn't boot. It stays on the boot logo.12:59
T42<edp_17> @elros34 : The value of /sys/module/slimport/parameters/enable_irq is Y13:02
T42<edp_17> There is nothing else in /sys/module/slimport/parameters/13:02
mighty17[m]<T42> "<Herrie1982> OK great! 3.4..." <- Yeah kernel 3.4 :D13:06
T42<elros34> @edp_17 why did you choose not bootable room for sfos base?  Anyway try to write 0 then 1 to enable_irq and check dmesg, logcat whether it generates any logs. Maybe something needs to be configured on usb so your converter is properly detected13:15
T42<elros34> btw what kind of converter do you use to connect phone to tv13:17
T42<dgmarie2> To build gst-droid I need to use an older version, but it automatically gets updated when I run -gg14:42
mighty17[m]Yup that's broken afaik14:44
T42<elros34> @dgmarie2 --help will help you avoid auto update14:47
T42<elros34> tip  --offline switch14:48
T42<dgmarie2> Yeah I remember now, thank you14:48
T42<edp_17> @elros34 : Why did I chose a not bootable rom? This was the first port I've done with absolute no knowledge. I just re-used what somebody else used on an existing port (I think The Real John Galt). Then, I stick with it.14:49
T42<edp_17> @elros34 : I wrote Y and N into /sys/module/slimport/parameters/enable_irq, but nothing changed in either logcat not in dmesg -w.14:58
T42<edp_17> The converter (cable) I use is the same that I am using on the same device under Android to connect to the TV. (That android is 6.0.1)14:59
T42<edp_17> I guess need to switch to a base that boots and has working HDMI output. Right?15:00
T42<elros34> you mean new port? it's not worth15:01
T42<edp_17> @Herrie1982 : I've downloaded both but unfortunately, Los16 doesn't install as a secondary rom with multirom.15:01
T42<edp_17> @elros34 : Yes, that's how I meant.15:01
T42<elros34> there is very little chance to have it working in usable state even if you manage to get it working at all15:02
T42<edp_17> Then, how can we find out what's wrong? I built that Should I do anything with that?15:02
T42<edp_17> I see.15:02
T42<elros34> read source code of hwcomposer and see how it triggers hotplug, probably waits for some events from kernel either via udev or sysfs then figure out why this is not generated15:03
T42<edp_17> I'll try. Thanks.15:04
T42<elros34> not sure if that is correct branch you use but according to kernel should generate devices/virtual/switch/hdmi. But that is for pure hdmi not sure how slimport is handled15:09
T42<elros34> maybe you have something in /sys/devices/virtual/switch/15:10
T42<edp_17> I have this in my device tree:
T42<edp_17> Yes there is a hdmi in /sys/devices/virtual/switch/15:13
T42<edp_17> And there is state which has 0:
T42<edp_17> And it remains 0 even when I attach the cable.15:14
T42<elros34> IIRC this state changed to 1 when I connect hdmi on my device15:15
T42<edp_17> This what I think it should do.15:15
T42<edp_17> The /sys/devices/virtual/switch/hdmi/power/runtime_status value is unsupported15:16
T42<elros34> try again with enable_irq but use 0/1 instead Y
T42<elros34> probably that will not change anything but at least you can build that driver with some additional prints too see whether it detect anything15:32
T42<elros34> according to logs at least hdmi_audio should change when you connect cable:
T42<edp_17> I used 0/1 but it wrote N/Y into it anyway.15:56
T42<elros34> yeah probably that is expected15:57
T42<dgmarie2> i have 'Provides: ofono-configs' and 'Obsoletes: ofono-configs-mer' in my droid config.16:10
T42<dgmarie2> the previous porter for my device had these in their config and it helped with an error in the build but it seems like it's causing a problem now16:11
T42<dgmarie2> I saw that someone had this issue before but I'm not sure how they fixed it. Anybody got any ideas?16:31
T42<dgmarie2> this seems useful: 'Solution 2: install droid-config-davinci-1-202108061550.aarch64 (with vendor change)' how do I choose this solution?16:32
T42<elros34> install droid-config manually in target via sb2 -t <target> -m sdk-install -R zypper in16:55
T42<dgmarie2> thank you17:05
T42<edp_17> @elros34 : I'll check the audio when I connect the cable. Thanks.18:32
T42<edp_17> @elros34 : how can I add those additional prints and how can I build that driver?18:34
T42<edp_17> @elros34 : For me the /sys/devices/virtual/switch/hdmi_audio/state is also remains 0 even if I attach the cable. Something is clearly missyng to trigger the hdmi stuff.19:38
T42<edp_17> Will check what you recommended: "read source code of hwcomposer and see how it triggers hotplug, probably waits for some events from kernel either via udev or sysfs then figure out why this is not generated"19:39
T42<elros34> yeah maybe you could flash android kernel, even without booting gui you should be able to adb shell and check whether these sysfs state files are changing.19:41
T42<edp_17> Good idea.19:41
T42<elros34> the line I send you previously in hwcomposer source is place where hotplug is called but obviously if kernel do not generate any  information about hdmi then it doesn't work19:42
T42<edp_17> Thanks. I'll try.19:58
T42<edp_17> @elros34 : Do you have idea why the aarch64 build has a bootloop on the treltexx? It bootlops so quickly that I cannot get into the device for logs.20:41
T42<edp_17> That device also has mhl 3 tv out. :-)20:42
T42<elros34> no, but it shouldn't be very hard to enable persistent logging and getting journal20:43
T42<edp_17> How to do it, if cannot get into the device?20:43
T42<edp_17> It doesn't even appear on my linux host when I attach with a cable to it.20:44
T42<edp_17> Nothing comes up in dmesg, so cannot manually bring the network up.20:45
T42<elros34> remove init.log in recovery and check whether is recreated if yes then init_enter_debug2 should stop bootloop20:47
T42<elros34> silly question, is that device even aarch64 capable?20:55
T42<elros34> looks like no21:03
T42<edp_17> It is because it's the exynos version.21:28
T42<edp_17> I'll try what you said about init.log and init_enter_debug2. 🤞21:28
T42<elros34> yeah forget about it, your kernel and device is 32bit21:34
T42<edp_17> I've seen in the kernel somewhere that was armv7hl but the exynos version is advertised as 64bit model, so I thought could build aarch64 for it.21:37
T42<edp_17> Perhaps I was wrong, wasn't I? :(21:38
T42<edp_17> Now, I have found this "64-Bit Exynos 5433 Confirmed, Will Only Run In 32-Bit Mode On Galaxy Note 4"21:39
T42<elros34> ouh21:39
T42<edp_17> Plus " can anyone please confirm if Samsung has enabled 64 bit mode on the Note 4 ?" ... "No, 64bit is not enabled. And it will never be." ... "The Note 4 was never meant to be a 64-bit model. The fact that the Exynos processor they used in some of the variants is 64-bit capable doesn't mean much."21:44
T42<edp_17> So it seems I built that aarchm64 image for basically nothing. 😭21:44
T42<edp_17> I thought could get WayDroid running on it.21:44
T42<edp_17> Anyway, I still can get the HDMI (mhl) and/or fingerprint scanner and camera working. 😊21:45
T42<edp_17> I'll go back to create armv7hl builds for the note 4.21:46
T42<edp_17> Thanks for your help anyway!21:46
T42<elros34> good look:)21:46
T42<elros34> luck*21:46
T42<edp_17> Both apply. Lol. :)21:49

Generated by 2.17.1 by Marius Gedminas - find it at!