Saturday, 2021-10-16

T42<adampigg> Theodan,.mal, rinigus, needed for pro1 and pinephone07:42
rinigus@adampigg: I would expect pro1 be the same as tama (or very close). now pinephone could be a bit different. maybe you should check on that?08:16
T42<adampigg> Sure08:20
ThaodanCurrently it's broken devices with internal mmc since some shenanigans that udisks2 does it device if an mmc is removable or not.11:26
ThaodanUdisks2 needs to be updated/patched.11:26
rinigusThaodan: but can't we help udisk2 via its config files? so, the port would include some config to help it out?11:35
Thaodanrinigus: udisk just needs to be updated12:03
Thaodanthe kernels handling of what's removable or not has some tuning in udisk2 and newer versions are better in this12:04
piggzrinigus: ping15:27
T42<Conrad> hey, i am trying to understand the porting and wanted to work on the Z3 compact tablet, which is heavily outdated. however, i even failed with the basics to reproduce the available images. I posted to problem talk.maemo.com15:29
T42<Conrad> is there anything obvious I miss?15:29
riniguspiggz: pong15:48
piggzrinigus: its fine, sorted :)15:49
riniguspiggz: great!15:49
T42<elros34> Conrad Try in platform sdk: But if you want to work on port than better read whole HADK pdf:
T42<Conrad> ok, thanks. i will have look15:53
Thaodanb100dian: You don't need to make your custom_data handling so complicated. I changed yours to allign with the one found in that works fine. The issues lies elsewhere I think.16:46
T42<b100dian> Ok with that commit, but you dont't have the two others: moving to heap and fix for dbus request id regression
*** mal_ is now known as mal17:25
T42<Conrad> So, this sdk which i can install via that tutorial is the basis for the docker Images like the one from coderus?17:26
Thaodanb100dian: the last two can be squashed and the typecasting you do there can be skipped.17:42
T42<b100dian> Yes, I'll make it more readable once we get it properly working with ngf-qt18:17
ThaodanThe erasing is the issue18:17
Thaodanthe effect id is wrong then18:17
Thaodaneffect id has to match one of the predefined effects during erasew18:18
T42<b100dian> It happens when the effect is in flight?18:18
T42<b100dian> I have some local changes for that, but far from perfect. Ill push them tonight for testing18:19
T42<b100dian> I still wasnt able to.use the custom data [1] output parameter from my driver at least18:20
T42<b100dian> To set a correct timeout18:20
ThaodanThe effect has to be erase after play18:22
ThaodanSee here:18:23
T42<b100dian> That sounds like a good idea18:23
ThaodanSo the erase during that is done if the effect is cached always has to be done18:23
kquote03[m]Hello. Now I'm building the RPMs, but i seem to get a weird error when running `rpm/dhd/helpers/ --droid-hal`
kquote03[m]sdk-assistant lists my target successfully, so idk why it is complaining about that, but more importantly I have no idea why the RPMS folder is empty.18:35
kquote03[m]And another question, is a black screen and an immediate reboot an expected result of flashing hallium-boot on the boot partition of an otherwise standard LOS installation (sorry, i couldn't help my self)18:35
Thaodan   /home/kenan1099/hadk/droid-hal-a5y17lte.log  check that18:36
ThaodanWell you should flash hybris boot18:37
T42<b100dian> Thaodan: I've pushed my working copy to a branch named temp. It adds effects to a queue from where they are erased before playing the next one but I don't see the "reports back %d ms" ever displaying anything else than 018:37
kquote03[m]I apologize for my lack of reading comprehension18:40
kquote03[m]now it halts becuase it can't find ~/hadk/installroot/default.prop18:42
kquote03[m]i tried touching ~/hadk/installroot/default.prop, but no dice18:43
kquote03[m]* no dice (anyways seems like a bad idea)18:43
Thaodan post logs there is no help in telling us what it says18:44
kquote03[m]    /home/kenan1099/hadk/droid-hal-a5y17lte.log18:48
kquote03[m]<Thaodan> " post logs there is no help in..." <- i feel that im a victim of not being able to straighten a dog's tail. apologies for repeatedly not listening to this advice18:50
ThaodanYou  have to fix the device config18:53
kquote03[m]you mean the .spec or the local_manifest?18:53
kquote03[m]ah the kernel config18:54
Thaodanthe log says it call18:54
kquote03[m]Thanks for the tip18:54
Thaodanb100dian: tried that branch, it works good except that the vibrations are a little to long18:55
ThaodanIt was better before18:55
ThaodanYou can declare custom directly as __s16 no need to explicitly cast from int to __s1618:57
Thaodan@b100dian: Which userspace vibrator software is used for your device? I don't think CUSTOM_DATA_LEN higher then 3 is needed.18:59
T42<b100dian> That I don't know. I need to study Vibrator.cpp from your links. Of course it can be 3 as is in both our drivers.19:02
ThaodanVibra erase effect: Invalid argument I hit that sometimes when it tries to erase the effect for a second time.19:02
Thaodanhere my vibrator cpp19:03
T42<b100dian> Yes, I'll look there. What is your current status w.r.t the ngfd PR + ngf-qt? MIne, with the 'temp' branch above, is that weak effect on keyboard vibrates 9 out of 10 keystrokes approx. and service is lost in about 1day - so there is some leak still there.19:08
ThaodanMaybe the way you allocate them mem? before no malloc was used.19:09
T42<b100dian> no, I mean effect leak. but yeah, one idea is to have the custom-data int parsed and stored separatedly and the custo_data be re-inited for each upload instead of once at the beginning for all19:12
T42<elros34> kquote03[m] comment out /default.prop from About reboot, I think it shouldn't happen unless you did not make all needed changes19:12
Thaodanthat's why need log:
Thaodansometimes I get one erase after another19:16
Thaodanthan Vibra erase effect: Invalid argument happens19:16
ThaodanI think for android the custom data is also set each time Play is called19:17
Thaodanafter rebooting the intensity of each vibration is fine19:19
Thaodanah they do setup ff_effect each time they play.19:20
ThaodanIs there maybe a race condition in erasing effects?19:21
T42<b100dian> because I cannot get this to work yet, yes maybe there's a timing race
Thaodanwhat does it do there? From what I read t19:26
Thaodanthe effect was played in line 11019:26
Thaodancan you skip the caching of the effect since it is already stored in the global table after the ini is parsed?19:28
T42<b100dian> it reads back the milliseconds+useconds the effect will take to play. The data[1] and [2] are reporting that back right after upload19:32
T42<Rogeriu> hi20:20
T42<Rogeriu> port for redmi note 8?20:20
T42<XAP2P> (re @Rogeriu: port for redmi note ...)20:24
T42<XAP2P> oh wow20:25
T42<XAP2P> 4.2 (re @XAP2P:
T42<XAP2P> @birdzhang good job, thx :D (re @XAP2P: 4.2)20:25
Thaodan@b100dian: that's specific to the FF_CUSTOM?20:27
T42<b100dian> yes20:27
ThaodanDo we need that or no?20:28
T42<b100dian> since the custom effect is not presenting any duration, it is a good idea to use what is returned as a timer so that effects don't overlap20:33
ThaodanHm that's a good idea20:33
Thaodanwithout a waveform its simpler in hardware but more complicated in software I guess20:34
T42<b100dian> I think I found my error if you take the last commit of, I though it was ms + usecs, but it is seconds + mseconds actually. I don't seem to get missed vibrations now on the keyboard20:34
T42<b100dian> yeah, my driver also reads those timings from firmware, so they are not well know beforehand20:35
T42<b100dian> Thaodan: many things can be refactored as you've alreay shown - the temp branch is only if you want to test this w.i.p. If all goes well after in a couple of days I'll re-write the PR and commits properly20:36
ThaodanTried your latest commit, it works as good as before just some times effects are erased a second time20:45
T42<b100dian> I'll try to think why that happens21:14
kquote03[m]hey, I've run into an issue and the log file is very large (verbose.log), may i upload it gzipped on here or using any file hosting service?21:15
T42<elros34> full output from console is usually enough21:36
T42<Rogeriu> thanks (re @XAP2P:
ThaodanRinigus: Do you have hdr working with android 10?23:26

Generated by 2.17.1 by Marius Gedminas - find it at!