*** phlixi is now known as Guest6207 | 13:17 | |
*** phlixi_ is now known as phlixi | 13:17 | |
*** phlixi is now known as Guest6219 | 15:14 | |
*** phlixi_ is now known as phlixi | 15:14 | |
T42 | <edp_17> @elros34 , @b100dian : Here is strace -f -s 256 /vendor/bin/hw/android.hardware.biometrics.fingerprint@2.1-fpcservice : https://paste.ubuntu.com/p/jt7hRk5Rfz/ | 20:22 |
---|---|---|
mal | @edp_17 wondering if this is bad openat(AT_FDCWD, "/dev/qseecom", O_RDWR) = -1 EACCES (Permission denied) | 20:24 |
mal | can you check permissions of that | 20:24 |
T42 | <edp_17> mal: Here it is: https://paste.ubuntu.com/p/JxWcBNRJR4/ | 20:25 |
T42 | <elros34> check rc file and start service as same user you find there | 20:26 |
mal | or change to 666 for testing | 20:26 |
T42 | <edp_17> strace after chmod 666 : https://paste.ubuntu.com/p/K4B3m3cztQ/ | 20:28 |
T42 | <edp_17> No permission denied. | 20:28 |
T42 | <edp_17> Not on /dev/qseecom at least | 20:28 |
T42 | <edp_17> openat(AT_FDCWD, "/dev/uinput", O_WRONLY|O_NONBLOCK) = -1 EACCES (Permission denied) | 20:29 |
T42 | <edp_17> /dev/uinput: https://paste.ubuntu.com/p/TMSdsDpGVH/ | 20:30 |
T42 | <elros34> another which requires system user, fight with issues created by itself or start process properly | 20:30 |
T42 | <elros34> not this time | 20:30 |
T42 | <elros34> do you see some output from kernel in dmesg? | 20:31 |
T42 | <elros34> can you show whole rc which which normally start this? | 20:32 |
T42 | <edp_17> dmesg: https://paste.ubuntu.com/p/sV7GGCZGqq/ | 20:32 |
T42 | <edp_17> I don't know which rc starts this. Let me find. | 20:33 |
T42 | <b100dian> I have that .rc file with user system and group system input uhid. So running strace --user=system should help | 20:33 |
T42 | <b100dian> /vendor/etc/init/android.hardware.biometrics.fingerprint@2.1-service.rc on my L18.1 device (sorry, L17.1 is not handy atm) | 20:34 |
T42 | <b100dian> usually `find / -name \*.rc -exec grep fingerprint {} \; -print` (pretty sure grep can do this directly but that's in my rom) | 20:35 |
T42 | <b100dian> and yes, dmesg -w in parallel would be very wise, maybe access denied is "cannot find some other /dev needed" | 20:35 |
T42 | <elros34> not enough because uinput only accesible for uhid user:group (re @b100dian: I have that .rc file...) | 20:39 |
T42 | <b100dian> I assumed with the uid strace would aet the corresporing gid(s), is there another parameter for that? | 20:40 |
T42 | <b100dian> *set | 20:41 |
T42 | <edp_17> Content of /vendor/etc/init/android.hardware.biometrics.fingerprint@2.1-service.rc: https://paste.ubuntu.com/p/MrYBCqCB6b/ | 20:41 |
T42 | <edp_17> I'll run that find too. | 20:42 |
T42 | <edp_17> The result of find (I ran it as root): https://paste.ubuntu.com/p/Jv2WXnfMV4/ | 20:42 |
T42 | <edp_17> strace with —user=system: https://paste.ubuntu.com/p/vHnNN2pYjH/ | 20:45 |
T42 | <edp_17> and dmesg -w while strace was running: https://paste.ubuntu.com/p/xsn74hkJ2W/ | 20:45 |
T42 | <edp_17> Something is clearly missing: "Could not find 'android.hardware.biometrics.fingerprint@2.1::IBiometricsFingerprint/default' for ctl.interface_start" | 20:49 |
T42 | <b100dian> something like `runuser -u system -g system -G uhid -G input strace `... ? | 20:58 |
T42 | <b100dian> maybe add that 9015 gid too | 20:58 |
T42 | <b100dian> also, run `dmesg -w` beforehand and only paste what's happening when you run the service | 20:58 |
T42 | <edp_17> I've just did a chmod 666 /dev/uinput | 20:58 |
T42 | <edp_17> strace: https://paste.ubuntu.com/p/JBczn57W23/ | 21:01 |
T42 | <edp_17> dmesg -w: https://paste.ubuntu.com/p/Yv6G8Sx47z/ | 21:01 |
T42 | <elros34> sometimes it's better to add strace to rc file innstead fighting with right use, groups andright timing, see all your dmesg is filled with restarting service | 21:02 |
T42 | <edp_17> Do you meant to add strace -f -s 256 —user=system into here: https://paste.ubuntu.com/p/MrYBCqCB6b/ | 21:04 |
T42 | <edp_17> And yes, the dmesg is full of restart of that service. | 21:05 |
T42 | <b100dian> > fchownat(AT_FDCWD, "/data/vendor/.fps/tac_ver", 1000, 9015, 0) = -1 EPERM (Operation not permitted) | 21:06 |
T42 | <b100dian> same with ic_ver | 21:06 |
T42 | <b100dian> do you have such files and with what permission(s)? | 21:06 |
T42 | <b100dian> *don't change them yet :D | 21:07 |
T42 | <edp_17> Permission of /data/vendor/.fps/tac_ver : https://paste.ubuntu.com/p/Yv6G8Sx47z/ | 21:07 |
T42 | <b100dian> Sounds like that 9015 gid .But wrong pastebin above?^^ | 21:08 |
T42 | <edp_17> /data/vendor/.fps/ic_ver is the same | 21:08 |
T42 | <edp_17> Sorry, I haven't submitted the pastebin. Here they are: https://paste.ubuntu.com/p/r4gB3VgwJk/ | 21:08 |
T42 | <elros34> no, more like strace -s 256 -ff -o /data/somewritabledir/fpc or something like that (re @edp_17: Do you meant to add ...) | 21:11 |
T42 | <b100dian> I wonder what that 9015 is that is not in your /etc/group | 21:17 |
T42 | <edp_17> The 9015 is not there: https://paste.ubuntu.com/p/PZBPdp6T3h/ | 21:18 |
T42 | <elros34> check in ueventd.rc | 21:22 |
T42 | <elros34> or maybe it was different file, if not then in device tree | 21:22 |
T42 | <edp_17> I've added strace into the rc file like: https://paste.ubuntu.com/p/PZBPdp6T3h/ | 21:23 |
T42 | <edp_17> Reboot. But there is nothing in /data/edp17 | 21:23 |
T42 | <elros34> did you make it writable for everybody? | 21:23 |
T42 | <edp_17> Yes: drwxrwxrwx 2 root root 3488 Oct 13 22:13 edp17 | 21:24 |
T42 | <elros34> this past url is wrong | 21:24 |
T42 | <edp_17> Damn, again. SOrry. https://paste.ubuntu.com/p/7G7w3fR6N5/ | 21:25 |
T42 | <elros34> add full path to strace binary | 21:25 |
T42 | <edp_17> Okay. | 21:25 |
T42 | <edp_17> Content of /ueventd.rc: https://paste.ubuntu.com/p/y6jJvCsm8n/ | 21:26 |
T42 | <elros34> search in source for this fingerp 9015 | 21:27 |
T42 | <elros34> or blindly create that group and see what will happen | 21:27 |
T42 | <elros34> https://github.com/RealJohnGalt/MotoG6-Ali-Firmware-Dump/blob/master/vendor/etc/group | 21:29 |
T42 | <edp_17> Okay, now I have fpc file(s) It seems a new appears time to time: https://paste.ubuntu.com/p/8ZmhwFK4y9/ | 21:30 |
T42 | <edp_17> Here is one of them: https://paste.ubuntu.com/p/Vw9tPRCCmY/ | 21:31 |
T42 | <edp_17> Hmm, that 9015 is missing on my device /vendor/etc/group : https://paste.ubuntu.com/p/w3PTckgxqh/ | 21:33 |
T42 | <edp_17> I have: vendor_fingerp::5015: | 21:34 |
T42 | <edp_17> and don't have fingerp::5015: | 21:34 |
T42 | <b100dian> worth a try though, it is definitely used/useful | 21:41 |
T42 | <b100dian> Elros: nice find | 21:42 |
T42 | <edp_17> So, I should add those two lines, right? | 21:42 |
T42 | <elros34> that is not a problem anymore if straced properly (see there is not EPERM anymore) but maybe for something else | 21:43 |
T42 | <edp_17> What I spotted was. In Vlad's binder-list there was a com.fingerprints.extension. I've found /vendor/lib64/com.fingerprints.extension\@1.0.so that I symlinked to /usr/libexec/droid-hybris/system/lib64/ but fingerprint service is still crashing. | 21:45 |
T42 | <b100dian> according to the proprietary files you have that but I don't think you need to link it, why wouldn't it be found there (ref: https://github.com/edp17/android_device_motorola_ocean/blob/lineage-17.1/proprietary-files.txt#L287 ) | 21:46 |
T42 | <edp_17> Aha, thanks. I'll remove that symlink then. | 21:47 |
T42 | <edp_17> writev(3, [{iov_base="\0\344\24\t<\fg\242{\313-", iov_len=11}, {iov_base="\6", iov_len=1}, {iov_base="fpc_tac\0", iov_len=8}, {iov_base="sensor_command, Failed to send command: 0 to TA, status code: -9\0", iov_len=65}], 4) = 85 | 21:52 |
T42 | <b100dian> dmesg time again? | 21:57 |
T42 | <edp_17> In the meantime I set chmod 666 on these three: /data/vendor/.fps/tac_ver , /data/vendor/.fps/ta_ver and /data/vendor/.fps/ic_ver | 22:05 |
T42 | <edp_17> Didn't help | 22:05 |
T42 | <edp_17> dmesg: https://paste.ubuntu.com/p/WKXpxYTsHn/ | 22:07 |
T42 | <edp_17> Also added those two 9015 lines into /vendor/etc/group : https://paste.ubuntu.com/p/DWHW9rfxQy/ | 22:09 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!