| T42 | Alfredo %lastname% was added by: Alfredo %lastname% | 13:03 |
|---|---|---|
| *** techgeekster_ is now known as techgeekster | 15:04 | |
| techgeekster | Hi everyone. 3-4 days ago I wrote about my port of zenfone3. @elros34 said that the vendror.keymaster3 service could be at fault. I restarted the whole build from scratch and disabled healthd and vendor.keymaster services and still I have no GUI. This is the journal https://paste.ubuntu.com/p/7Wy5dHPPMN/ | 16:13 |
| techgeekster | This is the logcat https://paste.ubuntu.com/p/XjzWxkZTyp/ The most weird error in logcat is this /vendor/bin/adsprpcd: vendor/qcom/proprietary/commonsys-intf/adsprpc/src/rpcmem_android.c:126:open /dev/ion failed! | 17:24 |
| T42 | <zinstack625> mdss_rotator seems to have wrong permissions | 17:27 |
| T42 | <zinstack625> And yeah, healthd seem necessary now that I think about it | 17:28 |
| T42 | <zinstack625> Not for gui, but for the long-term | 17:28 |
| T42 | <elros34> IIRC you already fixed /dev/ion and rotator issue. Didnt you forget about copying ueventd*.rc to out directory? | 17:29 |
| T42 | <zinstack625> mdss_rotator permissions are set in ueventd.qcom.rc on my device, and that probably should be patched somewhere automatically. Maybe it's set somewhere else on yours? | 17:37 |
| T42 | <zinstack625> Wait, so are system:system permissions for mdss_rotator even okay? | 17:43 |
| T42 | <elros34> no, but he already fixed it few days ago | 17:43 |
| techgeekster | I did copied the ueventd.rc to the out directory but the issue remains. I'll try again though. Maybe the fix is not implemented in my latest build | 17:43 |
| T42 | <zinstack625> @elros34 but that's what udev rule on river is... | 17:44 |
| T42 | <zinstack625> Not sure what's on the running system, but should be that? | 17:45 |
| T42 | <elros34> if it cause issue then you need to change it for example to system:graphics. Similar for msm_rotator on some devices | 17:48 |
| techgeekster | Hmm ok. I'll try now | 17:49 |
| T42 | <elros34> Based on journal it would be worth to check whether /vendor/firmware_mnt and /vendor/dsp symlinks points to same sane places. | 17:54 |
| techgeekster | @elros34 both firmware_mnt and dsp are empty on the device | 19:34 |
| T42 | <elros34> aren't they symlinks? Are they correct? | 19:35 |
| techgeekster | How can I check those? | 19:36 |
| T42 | <elros34> ls -al /vendor | 19:36 |
| techgeekster | lrwxrwxrwx 1 root root 14 Apr 4 2021 /vendor -> /system/vendor | 19:38 |
| T42 | <elros34> then /vendor/ or /system/vendor/ | 19:39 |
| techgeekster | drwxr-xr-x 2 root shell 4096 Jan 1 2009 | 19:41 |
| techgeekster | firmware_mnt drwxr-xr-x 2 root shell 4096 Jan 1 2009 dsp | 19:41 |
| techgeekster | Seem mounted but empty | 19:43 |
| T42 | <elros34> hm then I am not sure how they should looks like. You can flash lineage kernel and check in android how mountpoints and these dirs looks like. But before that could you remove audit=0 from kernel cmdline? People usually don't disable it | 19:45 |
| T42 | <elros34> I have checked your binder.c and maybe it would be worth to revert 1 or more comits. At least that which generate binder transaction error flood in your logs | 19:46 |
| techgeekster | Okk I'll try that and i'll post the results | 19:56 |
| T42 | iDubno was added by: iDubno | 23:14 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!