Elektra has joined #maemo-leste
Elektra is now known as Guest88
ashley has quit [Ping timeout: 255 seconds]
<Wizzup> heh, our mesa is newer than bullseye mesa
akossh has quit [Quit: Leaving.]
<Wizzup> freemangordon: random question, but should tinymail depend on telepathy-glib to have telepathy support enabled?
ashley has joined #maemo-leste
ashley is now known as Guest9393
Guest88 has quit [Ping timeout: 252 seconds]
<Wizzup> For Chimeara, I hope that the only things that will require significant work at this point are: qt pkgs, eds telepathy, mafw-shared
<Wizzup> the remaining py2 pkgs are probably doable, and a bunch depends on osso-abook-dev which we don't have now because of eds telepathy
<Wizzup> tinymail + gtkhtml3 + modest might need some work too, but a lot is pretty much there
<Wizzup> uvos: please look if we need to rebase iio-sensor-proxy
Guest9393 is now known as ashley
ashley has joined #maemo-leste
ashley has quit [Changing host]
Elektra has joined #maemo-leste
Elektra is now known as Guest6274
ashley has quit [Ping timeout: 256 seconds]
thunderysteak has quit [Ping timeout: 265 seconds]
xmn has quit [Ping timeout: 260 seconds]
joerg has quit [Ping timeout: 265 seconds]
joerg has joined #maemo-leste
sicelo_ has joined #maemo-leste
peterM_ has joined #maemo-leste
dsc__ has joined #maemo-leste
dsc_ has quit [*.net *.split]
sicelo has quit [*.net *.split]
linmob has quit [*.net *.split]
danielinux has quit [*.net *.split]
sicelo_ is now known as sicelo
ceene has joined #maemo-leste
Twig has joined #maemo-leste
<freemangordon> Wizzup: seems I am missing something - how is tinymail related to telepathy?
danielinux has joined #maemo-leste
mardy has joined #maemo-leste
alex1216 has joined #maemo-leste
Twig has quit [Remote host closed the connection]
Guest6274 is now known as ashley
ashley has quit [Changing host]
ashley has joined #maemo-leste
pere has quit [Ping timeout: 260 seconds]
akossh has joined #maemo-leste
pagurus has joined #maemo-leste
pere has joined #maemo-leste
<Wizzup> freemangordon: it is not, it's just not building atm
uvos has joined #maemo-leste
<uvos> havent looked at the pr
<uvos> but if that works like the amd/ati ddx its just adding double buffering to x'es own rendering
<uvos> which hildon allready dose too
<uvos> or should anyways
<Wizzup> the mr has a descr
<Wizzup> "This adds support for TearFree page flips to eliminate tearing without the use of a compositor."
<uvos> right yes that sounds exactly like the amd ddx
<uvos> wrt iio-sensor-proxy
<uvos> i only build this for us beacuse, while it is in debian, it was not in devuan
<uvos> beacuse old iio-sensor-proxy versions hard depended on systemd
<uvos> this is no longer the case
<uvos> and i presume devuan now has iio-sensor-proxy
<uvos> is there a package search webpage for devuan somewhere? i ususaly just use the debian one
<uvos> so you can just drop this package
<uvos> huh wait, the package dosent contain a init script
<uvos> just a systemd unit
<Wizzup> uvos: you added various patches for groups and stuff too
<uvos> what the?
<uvos> Wizzup: those are not nessecary
<Wizzup> uvos: well that's expected, debian doesn't support anything else
<uvos> i just enabled some features we dont use
<uvos> Wizzup: but this is devuan
<Wizzup> doesn't it just show all pkgs?
<uvos> its from http://deb.devuan.org/merged chimaera/main amd64
<uvos> so i gues no?
<uvos> this seams like an oversight
<Wizzup> devuan merged debian packages into its own
<Wizzup> merges*
<Wizzup> I wrote the code that does it way back
<uvos> okay, well thats.. not great
<Wizzup> I wouldn't assume that it doesn't just come from debian
<Wizzup> well it's a fork, not an overlay on top of debian
<uvos> so who to contact in devuan to fix this?
<Wizzup> to be clear I helped rewriting the code and make it fast, that's all
<Wizzup> uvos: it's a debian pkg, not a devuan one, so you need to convince debian to support something other than systemd, I guess?
<Wizzup> I don't think there's a devuan maintainer for it
<Wizzup> but you could ask in the channel I guess
<uvos> whats the channel name?
<uvos> #devuan
<uvos> ?
<Wizzup> I think so
<uvos> Wizzup: will you also be archiveing the repos from -forks you remove?
<uvos> its pretty confusing otherwise
<Wizzup> uvos: yes, I did plan to take some ation
<Wizzup> ation*
<Wizzup> action*
<Wizzup> see the bottom of the first comment https://github.com/maemo-leste/bugtracker/issues/644
<uvos> lots of work i know, but really we should just switch to systemd, it would be alot better for us since, while im not a compleat fan of the kitchen sink approche, we really use the features it provides (with our own hacky tools)
<uvos> and example of this is dsme lifeguard and reboot behavior, systemd can do both of these
<uvos> Wizzup: ok great!
<Wizzup> not going to happen with this release
<uvos> right
<Wizzup> we could consider -also- supporting unit files, but the way the whole thing went down still annoys me
<uvos> yes politics where horrible, as is the absorbtion of other stuff like udev into the systemd orbit, but systemd the init system itself is not bad
<Wizzup> well, in any case, we'll keep dsme around for now, but I agree it's a bit of a mess, the shutdown in particular
<uvos> the shutdown isent the main problem imo
<Wizzup> while we reboot/shutdown, dsme tries to restart services that stop working because they depend on something else that was stopped
<freemangordon> Wizzup: if we shutdown through dsme, this will not happen
<freemangordon> or at least meakle dsme know that we do shutdown
<freemangordon> *make
brabo has quit [Quit: leaving]
<Wizzup> freemangordon: ok
brabo has joined #maemo-leste
<Wizzup> freemangordon: np if not, but do you think you could look at eds tp today? I have an itch to finish this chimaera stuff this week and this is a blocker for anything that wants to use abook
xmn has joined #maemo-leste
<uvos> Wizzup: btw i dont think we need sdl2 anymore either
<uvos> i think kmsdrm is enabled upstream
<uvos> and the other patches wehere about directfb we dont need anymore with ddk1.7
<uvos> *17
<Wizzup> yeah, I marked it was probably skip
<Wizzup> it as*
<uvos> same story with sdl1
<uvos> hmm besides --enable-video-x11-xrandr=yes
<uvos> need to look at the flags upstream
<Wizzup> well we still need to build our own mesa I am sure (glesv1)
<uvos> not sure how this is related, but yes
<Wizzup> well, debian is disabling gles1 in various places
<Wizzup> so there's a good chance they also disabled it from libsdl1
<freemangordon> Wizzup: will try, but no promise
<Wizzup> freemangordon: ok, ty
<freemangordon> I am a bit of disappointed with kernel maintainers and need a bit of rest
<Wizzup> freemangordon: it'll be hard to test I am sure, but if it just compiles first that's probably enough for now
<uvos> Wizzup: sdl1 dosent appear to have any such flag
<freemangordon> sure, it is just that I want to be in a mood for it :)
<Wizzup> ok
<uvos> Wizzup: btw i have another xt1602
<Wizzup> oh, cool
<uvos> soon anyways
<Wizzup> I'm still waiting for the xt1602 I ordered to arrive
<uvos> i bought my gf a new phone and will collect hers :P
<Wizzup> nice :)
<Wizzup> I'm still lookng for another to send to Tony, since this'll be my first, so if you see one on ebay
<uvos> i can send it someone if i get a xt1604
<Wizzup> not sure of the exact model
<uvos> seams extreamly expensive
<uvos> i bought mine for like 20 euros xD
<Wizzup> oh
<Wizzup> I found another..
<Wizzup> lol, they only do pickup and it's on the border with belgium and germany (3+ car drive)
<Wizzup> 3h+
<uvos> sounds.. also expensive :P
<Wizzup> this one is 25 eur, dropped them a msg
<Wizzup> we'll see
<Wizzup> the irony is that the website tells me I am a "world changer" by buying second hand, and that the CO2 saved is the amount of a trip from amsterdam to brussels, which is maybe only twice the distance of picking it up
<uvos> xD
elastic_dog has quit [Ping timeout: 264 seconds]
<uvos> there allways some on ebay-kleinanzeigen btw
<uvos> usualy no modell number ofc
<uvos> but it bought 2 blind and both ended up being xt1602 so i presume this is the most sold version in germany at least
<uvos> yeah
<Wizzup> it ought to be either xt1602 or xt1604
<Wizzup> well is this kleinanzeigen not to klein to ship to .nl? or maybe we can ship to you?
<uvos> its just an informal classifieds page
<uvos> you contact the sellers directly and negotiate whatever
elastic_dog has joined #maemo-leste
<Wizzup> right
<uvos> probubly better if done in german
<uvos> what do you want ill ask
<Wizzup> not sure, I think we'd just want a xt1602 and xt1604 for tony
<Wizzup> (or at least one of them)
<Wizzup> uvos: hm there seem to be more in .nl than I thought
<Wizzup> oh, maybe these aren't the play versions?
<uvos> careful yeah
<uvos> motorola g4 is a totaly different phone
<Wizzup> confusing :D
<uvos> lenovo are idiots with nameing
<freemangordon> it seems we have some pvr issue as well in 6.1
<freemangordon> dmesg is full of ERROR: PVRSRVMapFullDmaBuf failed: PVRSRV_ERROR_OUT_OF_MEMORY - Unable to allocate required memory fd 25
<Wizzup> we also disabled all compaction with leste-config
<Wizzup> it's probably not related, just wanted to mention ...
<freemangordon> I doubt it is
<Wizzup> what are you running when you see this?
<freemangordon> I run gtkperf a couple of times
<freemangordon> and now it happens even if I scroll h-d
<Wizzup> smells like a leak
<freemangordon> so, to me it looks pvr driver run out of MMU space
<freemangordon> yes, but I see nothing wrong in dri or gem buffers in /sys/kernel/debug/...
<freemangordon> smells like a driver bug
<freemangordon> Xorg has 39 fds
<Wizzup> ok
<freemangordon> 303 dma_buf objects ~13MB in total
<Wizzup> if it perhaps some config option we're missing? I very vague recall needing something to prevent fragmentation
<freemangordon> no, this is SGX MMU, I don;t think kernel options have anything to do with it
<freemangordon> 402 gem objects ~14MB
<freemangordon> tiler_map looks perfect
<freemangordon> everything points to a leak in the driver
<uvos> fun
<freemangordon> yeah
<freemangordon> there was major rewrite of dma_fence code
<freemangordon> I guess related
<freemangordon> once I finish playing with cos estimator I will see if this is steadily reproducuble
<freemangordon> *SOC
<freemangordon> The fun part is that if you don't look in dmesg or ignore a little stutter, you will not understand that there is any issue :D
<Wizzup> I haven't seen it in dmesg yet
<Wizzup> (using it for a few days, ftr)
<Wizzup> but sounds like gtkperf is a good way to reproduce
<freemangordon> omap driver falls back to SW render and works
<Wizzup> hildon-application-manager also needs porting to newer apk-pkg interfaces
<Wizzup> made a start, but taking a break from that now
<uvos> for some reason pvr is also a bit slower in 6.1 maybe related, maybe not
<uvos> but openlara seams to underperform
<freemangordon> I guess it is related
<uvos> i havent seen the dmesg however
<uvos> just greped
<freemangordon> will try to repro
<freemangordon> not now, have to attend mtg
<freemangordon> ttyl
norayr has left #maemo-leste [Disconnected: Received SIGTERM]
norayr has joined #maemo-leste
pere has quit [Ping timeout: 256 seconds]
<Wizzup> tmlind: btw I have a xt912 here so can try the kexecboot stuff this weekend I hope :)
pere has joined #maemo-leste
thunderysteak has joined #maemo-leste
akossh has quit [Read error: Connection reset by peer]
<uvos> tmlind: so i gues we are not going to be using the fact that the bootloader on xt912 is unlockable
<uvos> we could use the mainline kernel for kexecboot i gues, altho presumably this would expose problems in our cpcap drivers expecting the register state to be initalized by the android kernel
akossh has joined #maemo-leste
<tmlind> i don't think at least my xt910 is unlockable, also flashing xt912 modem fails
<uvos> yes xt910 (and all verizon devices for that matter) is not
<tmlind> lcd is some mipi dsi amoled.. needs some days of tinkering, no output quite yet
<uvos> mhm
<tmlind> but i think the lcd is the same on xt912 too
<tmlind> so after that xt912 should behave pretty much like d4
<tmlind> i guess only european xt910 are unlockable? mine seems to be ca rogers/fido
<uvos> according to xda its complicated
<uvos> on release there was a dev editon that was unlockable
<tmlind> ok, fastboot oem unlock disabled on mine
<uvos> at some point the just starting shipping the unlockable bootloader on all devices (in europe at least)
<tmlind> ok
<uvos> *they
<uvos> im not sure if this means you can flash the later/euopean bootloader
<uvos> probubly not
<tmlind> probably can flas xt912 modem on it then..
<tmlind> for the unlocked variants
<uvos> in generaly this would be atypical for android devices
<uvos> unlocked bootloader ususaly just means the android partitons
<uvos> are not signed
<tmlind> yeah modem could be still efused
dsc__ is now known as dsc_
<tmlind> Wizzup: yeah if you get a chance see log for the info to dump, no rush
<buZz> did ppl already notice the mce update changed the auto-backlight-adjustments?
<uvos> it should not have
<uvos> change how exactly?
<buZz> it does get brighter on a bright light, but never goes back to dimmer
<uvos> that is correct and has been so for a long time
<buZz> oh eh weird, it did for me
<buZz> should i be self-setting it lower every time?
<uvos> no
<uvos> so it resets on display off
<uvos> so while the display is on it will only increase the brightness
<uvos> never lower it
<buZz> ah indeed, weird that it did lower it for me then
<uvos> this is to prevent it changing the brightness to often
norayr has left #maemo-leste [Error from remote client]
<uvos> there is a config option
<uvos> if you want the old behavior
<uvos> in mce.ini
<buZz> ah ty
<uvos> its been this way since at least 1 year
<buZz> hehe weird, i must have had a old mce.ini
<uvos> as the top of the file suggest you shal not change mce.ini at all
norayr has joined #maemo-leste
<buZz> yeah well :P
norayr has left #maemo-leste [#maemo-leste]
norayr has joined #maemo-leste
<Wizzup> uvos: ok so the xt1602 I found for 25 euro is not an option, for some reason the person outright does not want to ship it
<buZz> boo
alex1216 has quit [Quit: WeeChat 2.3]
<freemangordon> hmm, on allwinner tablet battery driver reports 44% for 3646mV
<freemangordon> that can't be true
<bencoh> why not?
<bencoh> 3.7V should be ~50%
<freemangordon> hmm
<bencoh> it might be slightly off (5~10%) but it's a decent approx, especially if it's voltage-based only
<freemangordon> on d4 it looks different
<freemangordon> maybe because it is HV battery
<freemangordon> so it has different discharge curve
<bencoh> charging at 4.35V vs 4.2V you mean? yeah, it would impact the nominal voltage slightly
<freemangordon> mhm
SuperMarioSF has joined #maemo-leste
<freemangordon> because on 'calibrated' d4 battery, voltage for 44 % is ~4V
<bencoh> uh
<bencoh> 4V/44% sounds plain wrong
<bencoh> either your battery is dead, or ... I dunno
<freemangordon> could be
<freemangordon> this is new battery
<bencoh> it might be """new"""
<freemangordon> that lasts almost forever
<bencoh> oh
<freemangordon> calibrated @ 1620mAh
<bencoh> I'd love a link to buy such a battery then :D
<bencoh> I'm not happy with mine
<freemangordon> @ 4.2V, nominal capacity 2250mAh@4.35
<freemangordon> this is the same battery uvos implanted
<freemangordon> I did the same and I think buZz too
<bencoh> is it a standard battery? where does it come from?
<freemangordon> no, you need soldering iron :)
<bencoh> ah
<bencoh> can you still close the phone?
<freemangordon> lemme try to find it
<freemangordon> yes
<bencoh> neat
<SuperMarioSF> woah
<freemangordon> it fits perfectly
<SuperMarioSF> hack into a new battery
<freemangordon> you just have to replace the electronics
<bencoh> sounds perfect
<SuperMarioSF> sounds explosive
<bencoh> SuperMarioSF: :D
<freemangordon> no, why?
<freemangordon> it is just a PCB
<SuperMarioSF> yeah
<SuperMarioSF> PCB on a battery
<freemangordon> this
<bencoh> I dunno how cpcap deals with battery level, maybe then assume the battery has a lower capacity than that and does "electron counting" as well
<bencoh> oh, polarcell <3
<freemangordon> it does coloumb counting
<bencoh> I'll have to order one I guess, thanks
<bencoh> well then if it has a total capacity hardcoded somewhere then it might result in invalid reporting
<SuperMarioSF> well I don't want to deal with flex cable...
<freemangordon> removing original electronics is a bit tricky, as it is glued by some black polimer
<bencoh> such as 40%/4V
<freemangordon> bencoh: no, cpcap does coloumb counting
<freemangordon> that can;t be hardcoded
<SuperMarioSF> btw it is likely harrd coded into the PCB on battery
<freemangordon> no
<freemangordon> pcb on the battery is nvram with some parameters for android
<freemangordon> one of them is design capacity/voltage
<bencoh> freemangordon: total capacity might be written somewhere
<freemangordon> it is
<freemangordon> but that does not make any difference for upstream cpcap-battery driver
<bencoh> and is it still 1600~1700, or did you write 2250 there?
<freemangordon> it is 2250@4.35
<freemangordon> but we charge it to 4.2 only
<bencoh> well, I dunno what it does then, something's wrong
Juest has quit [Ping timeout: 260 seconds]
<freemangordon> no idea
<bencoh> yeah, I complained about charging 4.35V which resulted in swollen battery at some point
<SuperMarioSF> btw second XT883 arrived. condition is great
<freemangordon> was that you now tmlind doesn;t want my patches ? :)
<bencoh> woopsie :D no idea, I reported about that years ago, I think other people complained about something later on
<freemangordon> bencoh: in case you didn't follow:
<freemangordon> I am trying to implement some voltage based SOC estimation algo
<freemangordon> calculating Ri in real-time
<freemangordon> and using some formula found over the inet
<bencoh> voltage-only is tricky when under load
<freemangordon> that's why Ri
<bencoh> still tricky, but interesting :)
<bencoh> do we have a sense resistor (I guess yes) and do we have access to the raw readings?
<freemangordon> so, on d4 it seems to work fine - calculates Ri ~= .130mOhm, which is about right
<freemangordon> uvos confirmed yesterday that voltage readings are the same as what is measured with DMM on battery terminals
<freemangordon> so, with some massaging (high time constants for Ri and estimated SOC filtering), this works pretty much ok on d4
<freemangordon> like, usually difference with calibrated value is 1-3%, rarely 5
<freemangordon> so I decided to try on my allwinner tablet
<freemangordon> not a good idea - calculated Ri is ~0.09 that sounds sane for 3x1300 mAh batteries in parallel
<freemangordon> but, estimation is totally wrong
<freemangordon> bencoh: so in short - every 30 seconds I am taking 16 samples every second and calculate Ri=(Umax-Umin)(Imin-Imax)
<freemangordon> oops, it is Ri=(Umax-Umin)/(Imin-Imax)
<freemangordon> and then use that Ri to calculate the OCV of the cell
<freemangordon> oh, I thin kI know what's wrong
<freemangordon> I changed one of the constants
<bencoh> wait, all that in software?
<freemangordon> yes
<bencoh> that all sounds terrible battery wise :D
<freemangordon> not really
<freemangordon> also, this is just a POC
<bencoh> :)
<freemangordon> in RL it will take samples every 2 minutes or something
<bencoh> it's pretty nice as a POC tbf yeah :)
<freemangordon> actually I don;t see this affecting battery life on my d4
<bencoh> hmm
<bencoh> it might mean that something else runs in the background and keeps us awake (actually we know there is)
<bencoh> (s/there is/something does/)
<freemangordon> either ways device is woken >15 times/s
<bencoh> ah
<bencoh> okay
<freemangordon> yeah, restoring the constant made it behave correctly
<bencoh> nice :)
<freemangordon> Calculated r = 0.071186, Ri = 0.052247 for Imax=-956mA@3602mV, Imin=-1251@3581mV
<freemangordon> calculated (current 25 25, filtered 28) measured 29
<freemangordon> if you are curious I can provide the code
<bencoh> I'm more interested in the method, what's your input? battery voltage, and battery current?
<freemangordon> yes
<freemangordon> that's all
<bencoh> and you just get min/max values out of 16 samples?
<freemangordon> yes
<bencoh> interesting
<freemangordon> well, if you draw the most simplified equivalent scheme...
<freemangordon> just E and Ri
<bencoh> E?
<freemangordon> OCV
<bencoh> ah
<freemangordon> E like, ummm... don;t know it in English, sec
<freemangordon> cannot find it, but it is voltage as E, not voltage as U if you know what I mean :)
<bencoh> I think I do yeah
<bencoh> like how we represent the voltage source in ciruits, with U = E - Ri
<bencoh> R being the internal resistance
<freemangordon> exactly
<freemangordon> but I don't know the english term for E
<bencoh> it might really be OCV
<freemangordon> yeah, maybe
SuperMarioSF has quit [Quit: Client closed]
<freemangordon> If I translate from Bulgarian, it should be "Electromotive voltage"
<freemangordon> seems to be EMF in english
<bencoh> indeed yeah
<freemangordon> "EMF is the maximum potential difference between two points of the battery when no current flows from the source in the case of an open circuit"
<bencoh> I learned a similar word in french :)
<freemangordon> have to go to the store, bbiab
SuperMarioSF has joined #maemo-leste
<SuperMarioSF> Wizzup, images of second phone were sent to you via email.
<SuperMarioSF> this is a good condition one, no gunk, no sticky surface.
<SuperMarioSF> with almost complete accessories (one less battery, missing a unimportant paperwork)
<SuperMarioSF> even came with its quick start screen protection film 🤣
Twig has joined #maemo-leste
<uvos> freemangordon: you have to scale the soc estimation by voltage_max_design
<uvos> the nominal voltage of a hv lipo is about 3.8
<freemangordon> yeah
<freemangordon> also, I think I shall dig my master degree thesis and remember how inverse regression models were calculated ;)
<bencoh> ^^
<uvos> ok yeah thatl help
<freemangordon> yeah, I was able to control a robotic arm about 2 meters long with 2 degrees of freedom to draw a square (1m side or something) with 1m/s and less than 1% error (IIRC). Simulation in matlab, but still :)
<freemangordon> by using a simple matrix math and inverse regression
<uvos> neat neat
<freemangordon> mhm. So i think something similar shall be useful to simulate a battery
ceene has quit [Ping timeout: 268 seconds]
<freemangordon> OMG!!! glamor cares about GLES2?
SuperMarioSF has quit [Quit: Client closed]
Twig has quit [Remote host closed the connection]
<Wizzup> freemangordon: heh
<uvos> just in time for debian folks to decide to disable gles2 in mesa :P
<Wizzup> lol
<Wizzup> @SuperMarioSF: thakns for the email!
mardy has quit [Quit: WeeChat 3.5]
norayr has left #maemo-leste [Disconnected: Received SIGTERM]
norayr has joined #maemo-leste
norayr has left #maemo-leste [Disconnected: Received SIGTERM]
norayr has joined #maemo-leste
Juest has joined #maemo-leste
norayr has left #maemo-leste [Disconnected: Received SIGTERM]
norayr has joined #maemo-leste
norayr has left #maemo-leste [Disconnected: Received SIGTERM]
norayr has joined #maemo-leste
norayr has left #maemo-leste [Disconnected: Received SIGTERM]
norayr has joined #maemo-leste
uvos has quit [Ping timeout: 260 seconds]