| freemangordon | tmlind: uvos: why is d4 modem using AT interface and not qmi? | 08:20 |
|---|---|---|
| tmlind | freemangordon: i could never get the notifications to trigger on qmi, they seem to be forced to uart ts27.010 only | 08:38 |
| tmlind | sure if you figure out a way to make things work with qmi that would simplify things :) | 08:38 |
| tmlind | seems to be a modem firmware specific configuration | 08:39 |
| freemangordon | I see | 08:53 |
| tmlind | who knows, maybe the modem configuration could be updated with some commands | 08:54 |
| freemangordon | well, either ways we don;t have qmi calls working (in upstream) | 08:55 |
| freemangordon | so I will spend some time on it | 08:55 |
| tmlind | ok cool. so for now we've been mostly using qmi commands, and just kick the modem on uart pdu notifications so the qmimodem sees the event | 08:56 |
| tmlind | let me find the commit for you | 08:56 |
| freemangordon | tmlind: in the meanwhile https://git.kernel.org/pub/scm/network/ofono/ofono.git/commit/?id=8af8136f61742be963b67b88f18e85c12509e0f9, https://lore.kernel.org/all/1707981467-2955-2-git-send-email-ivo.g.dimitrov.75@gmail.com/ and https://lore.kernel.org/all/1707981467-2955-3-git-send-email-ivo.g.dimitrov.75@gmail.com/ :) | 08:57 |
| freemangordon | perhaps I should have cc-ed you | 08:57 |
| freemangordon | next time | 08:58 |
| freemangordon | I guess that should work on d4 as well | 08:59 |
| tmlind | ok great | 09:01 |
| tmlind | so for kicking the qmimodem on uart notifications, see mot_qmi_trigger_events() | 09:01 |
| tmlind | ideally we'd find a way to re-route notifications back to the qmimodem by using some at or qmi commands.. | 09:02 |
| tmlind | and then plain qmimodem should just work | 09:02 |
| freemangordon | and it lacks voice calls suppoer :D | 09:02 |
| freemangordon | but, yeah, that's the plan | 09:02 |
| freemangordon | will take a while though, I want to finish what I was doing initially before jumping to qmimodem (cpl phone applet) | 09:03 |
| tmlind | and i think right now if a qmimodem data connection gets disconnected, qmimodem driver won't notice that and modem consumes extra 80 mW of power until the data connection is shut down | 09:03 |
| freemangordon | and then will jump to d4 driver | 09:04 |
| tmlind | so at least disconnect notification using mot_qmi_trigger_events() is currently missing and causing poor power consumption when moving around and getting disconnected by the cell network | 09:04 |
| tmlind | ok | 09:04 |
| freemangordon | oh | 09:04 |
| freemangordon | ok, lets save that somewhere | 09:04 |
| tmlind | yeah we get some 0 ending uart notfication, but should call mot_qmi_trigger_events() | 09:04 |
| tmlind | so uart knows data got disconnected, qmimodem driver does not | 09:05 |
| tmlind | probably easy to reproduce with some tin foil over the phone :) | 09:05 |
| freemangordon | already did that when I was playing with signal strength iirc :) | 09:06 |
| tmlind | hehe ok :) | 09:06 |
| freemangordon | I think it is some 'enable event' command that's missing | 09:07 |
| tmlind | ok | 09:07 |
| tmlind | it was wizzup who accidentally discovered the mot_qmi_trigger_events() style kick will notfiy the qmimodem btw :) | 09:07 |
| tmlind | hmm so i guess one option would be to call mot_qmi_trigger_events() every three seconds when there's an active voice or data connection | 09:08 |
| tmlind | or every 10 seconds, then qmimodem should behave | 09:09 |
| tmlind | it would not help with incoming sms or voice calls though | 09:09 |
| freemangordon | my bigger concern is what to do with qmimodem voice patches that float around inet | 09:10 |
| tmlind | possibly we could also have the kernel serdev driver notify kernel qmi on any events.. | 09:11 |
| tmlind | weird that the qmimodem voice patches are not merged | 09:11 |
| freemangordon | mhm | 09:11 |
| freemangordon | maybe they were never sent | 09:11 |
| tmlind | yeah maybe, wasn't there some mailing list change over the past few years or something? | 09:12 |
| freemangordon | I see them only on pmos gitlab repo | 09:12 |
| freemangordon | there was | 09:12 |
| tmlind | in any case ofono development stalled pretty much over 10 years ago already it seems :( | 09:12 |
| freemangordon | but is very active lately | 09:12 |
| tmlind | ok good to hear | 09:12 |
| tmlind | nothing stopping you from sending the voice patches though | 09:13 |
| tmlind | if the pmos folks don't use ofono, they probably lost interest | 09:13 |
| freemangordon | yeah , see https://git.kernel.org/pub/scm/network/ofono/ofono.git/log/ | 09:13 |
| freemangordon | right, that's what sicelo said as well | 09:13 |
| freemangordon | what stops me is that I am not sure I'll be able to test them | 09:14 |
| tmlind | make a voice call? | 09:14 |
| freemangordon | from the VM ;) | 09:14 |
| freemangordon | but yeah, will try to | 09:14 |
| tmlind | d4 or n900? | 09:14 |
| freemangordon | x86_64 VM | 09:14 |
| freemangordon | that's what I mostly use for development | 09:14 |
| tmlind | heh you could connect over usb to d4 modem from your x86_64 vm :) | 09:15 |
| freemangordon | I have EC20 that came with Anakon | 09:15 |
| freemangordon | *Anakin | 09:15 |
| tmlind | ok | 09:15 |
| freemangordon | tmlind: umm, how? like, is this some gadget? | 09:15 |
| freemangordon | if that's possible, that'd be great | 09:16 |
| tmlind | no.. the d4 firmware bootloader whatever cell testing mode remuxes the qmimodem to the micro-usb | 09:16 |
| tmlind | you can do the same from linux too, the code is missing in the phy-cpcap-usb though | 09:16 |
| freemangordon | ah, more development :) | 09:16 |
| tmlind | let me find the define for it | 09:16 |
| freemangordon | I'll pass and just use EC20, all this tiny gaps are vastly distracting me | 09:17 |
| freemangordon | *these | 09:17 |
| freemangordon | unless it is very easy | 09:17 |
| tmlind | yeh np, fyi i think the gpio mux option in phy-cpcap-usb.c is CPCAP_MDM_RX_TX to route the modem to the micro-usb | 09:18 |
| freemangordon | ok, thanks | 09:18 |
| tmlind | that kind of hints that there must be some way to make the d4 modem behave as qmimodem only | 09:18 |
| freemangordon | when it comes to d4 I'll do the development on the device, it is fast enough | 09:19 |
| freemangordon | right | 09:19 |
| tmlind | maybe the bootloader in that case sends some command to the modem so the notifications are routed to qmimodem | 09:19 |
| freemangordon | andoid is using AT notifications as well, right? | 09:19 |
| tmlind | hmm or actually, maybe it's some command over uart when android boots | 09:20 |
| freemangordon | most-probably | 09:20 |
| tmlind | so booting to the bp only mode to android, and adb logcat radio and look at the commands when pc uses a data connection might reveal something | 09:21 |
| freemangordon | tmlind: pretend that I have no idea about android :) - what is bp mode? | 09:21 |
| freemangordon | maybe uvos or Wizzup can help with this | 09:22 |
| tmlind | if you boot to the firmware bootloader menu with vol up and down pressed, theres' an option for bp only mode (base band only i think) | 09:22 |
| freemangordon | ah | 09:22 |
| tmlind | so if you boot android to bp only mode, then connect d4 to a pc, it shows up as a qmimodem and ttyUSB serial ports | 09:23 |
| freemangordon | that sounds great | 09:24 |
| freemangordon | lemme try it | 09:24 |
| tmlind | hmm but likely adb logcat radio won't work in that case as the micro-usb is routed to the modem.. | 09:24 |
| tmlind | mainline kernel should have the modem usb id's, assuming they don't change for the bp mode | 09:25 |
| freemangordon | https://pastebin.com/eC1m6RjF :D | 09:26 |
| freemangordon | ofono recognized that as some AT modem | 09:27 |
| tmlind | hehe cool :) | 09:29 |
| tmlind | try a qmi voice call and see what happens? | 09:29 |
| freemangordon | I can't, in VM this is upstream ofono I use to send patches | 09:30 |
| freemangordon | will have to revert to stock first | 09:30 |
| tmlind | ok well just add the qmimodem voice patches but not the maemo uart hacks | 09:30 |
| freemangordon | but not now, have to run | 09:30 |
| tmlind | ok me too, ttyl | 09:30 |
| freemangordon | anyway, thanks, will keep you informed | 09:31 |
| tmlind | ok thanks | 09:31 |
| Wizzup | freemangordon: we use the pinephone, doesn't that use qmi modem for voice calls | 11:54 |
| Wizzup | s/use/support/ | 11:54 |
| freemangordon | yes, PP uses qmimodem, for voice calls, but this is not upstreamed | 12:03 |
| Wizzup | isn't that what you're looking for, or? | 12:06 |
| Wizzup | maybe I got confused | 12:06 |
| freemangordon | it is, but it is not in a shape to be upstreamed | 12:07 |
| Wizzup | btw, the modem pass through would be kinda of useful on the d4 too, much like the n900 can pass through it's modem over usb | 12:07 |
| Wizzup | freemangordon: right | 12:07 |
| freemangordon | ok, | 12:08 |
| freemangordon | but I am not going to implement that now :) | 12:08 |
| Wizzup | sure | 12:08 |
| freemangordon | wait, what? https://git.kernel.org/pub/scm/network/ofono/ofono.git/commit/?id=07e00b7215cd98aa033f2ebc4ff169662d95d34c | 12:15 |
| Wizzup | pavel worked it on a while ago over just usb iirc | 12:17 |
| freemangordon | it still uses AT cpommands though | 12:17 |
| freemangordon | *commands | 12:17 |
| freemangordon | I will try switching to qmi | 12:18 |
| Wizzup | hm? | 12:19 |
| Wizzup | I think just using qmi over usb is a dead end for d4 | 12:19 |
| Wizzup | the power usage is completely unacceptable | 12:19 |
| freemangordon | ok | 12:20 |
| freemangordon | anyway, at least we have something upstreamed | 12:20 |
| freemangordon | could you repeat again what is the requirement for moving away from g_at_chat? | 12:22 |
| freemangordon | Wizzup: hmm: | 12:22 |
| freemangordon | ofono_devinfo_create(modem, 0, "qmimodem", data->device); | 12:22 |
| freemangordon | this is what I see in our ofono | 12:23 |
| freemangordon | in motmdm_pre_sim() that is | 12:23 |
| freemangordon | seems I lack the details on what "the power usage is completely unacceptable" is supposed to mean | 12:24 |
| Wizzup | yes, our modem is a mix of qmi over usb and at commands | 12:24 |
| Wizzup | I don't remember the details, but tmlind would hopefully | 12:24 |
| Wizzup | 'our modem' -> 'our ofo\no modem' | 12:25 |
| uvos__ | Ä btw, the modem pass through would be kinda of useful on the d4 too, much like the n900 can pass through it's modem over usbÄ | 12:41 |
| uvos__ | d4 can do this in hardware | 12:41 |
| uvos__ | i dont quite remember the details but its one of the options you can choose in fastboot | 12:41 |
| uvos__ | "seems I lack the details on what "the power usage is completely unacceptable" is supposed to mean" if you use qmi only you need to keep the usb if awake at all time so the modem can send you updates when a call or notification comes in | 12:42 |
| uvos__ | this has unacceptable power consumption implications | 12:43 |
| uvos__ | however this dosent mean you need to use the at interface and the gpio irqs for calls as sutch | 12:43 |
| uvos__ | but you do need to use them to catch the modem telling you theres an event underway so you can wake the qmi interface | 12:44 |
| tmlind | uvos__: the usb interface will autoidle though when no traffic so it may not be that bad | 13:10 |
| tmlind | here's a broken mz616 32 on ebay btw if anybody wants to buy it for parts, seems to ship only to germany https://www.ebay.com/itm/204346488943 | 13:10 |
| tmlind | so presumably it has a micro-sd | 13:11 |
| tmlind | seems the price is kind of expensive but maybe the seller would take an offer of something lower | 13:12 |
| diejuse | Is there a way to open a window in Maemo with scrollbars? Many applications display content that overflows the screen and cannot be accessed. It also happens with apps that overflow content when switching from landscape to portrait mode making them unusable. For example miphoto is not usable in portrait mode. | 13:23 |
| diejuse | *mihphoto | 13:23 |
| uvos__ | nope | 13:52 |
| uvos__ | you could make xorg run in some higher resolution and rescale to the display using randrs transformation matrix | 13:53 |
| uvos__ | but this would be a very extream solution | 13:53 |
| uvos__ | and switching would require a reboot | 13:53 |
| uvos__ | tmlind: 40euro for a broken decade old tablet, insane | 13:54 |
| buZz | xrandr itself can also do scaling | 14:05 |
| buZz | never really got it to work well, though | 14:05 |
| buZz | ah, thats what you said :D randrs | 14:05 |
| tmlind | uvos__: maybe make an offer for 10€? :) | 14:21 |
| buZz | Wizzup: found a car-mount? :) | 14:22 |
| uvos__ | buZz: randr is even more correct as randr is the extension in xorg and xrandr is just a cli program the can to be used to speak the client side of the randr protocoll :P | 14:24 |
| uvos__ | \me pedantic | 14:25 |
| buZz | yeah :D | 14:25 |
| * uvos__ pedantic | 14:25 | |
| buZz | thanks <3 | 14:25 |
| * buZz likes pedantics | 14:25 | |
| buZz | uvos__: i want a carmount for droid4 , to mount a droid4 -inside- my 3D printer | 14:25 |
| buZz | and run 'klipperscreen' on it, and show interface of printer with touch controls | 14:26 |
| buZz | :) | 14:26 |
| buZz | https://github.com/KlipperScreen/KlipperScreen | 14:26 |
| buZz | maybe even with USBethernet instead of wifi | 14:26 |
| uvos__ | neat | 14:27 |
| uvos__ | maybe a mz609 would be even better suited to this task | 14:27 |
| buZz | i think it'll be ideal :) | 14:27 |
| buZz | i dont know, but for droid4 Wizzup has some 'carmounts' | 14:28 |
| buZz | whichever device, i just want to mount it solidly to the frame | 14:28 |
| buZz | and i have 2x droid4 here, so why not that :) | 14:28 |
| Wizzup | you might also be able to 3d print something | 14:39 |
| Wizzup | I will have to search for a while, doing some renovation and these aren't accessible | 14:40 |
| buZz | Wizzup: there's no real rush | 14:47 |
| buZz | just trying to find a project -_- , some IRL emotional crap happened over last months, and i'm still recovering | 14:47 |
| buZz | need to find something that grabs my attention away well enough ;) | 14:47 |
| buZz | not sure if this would be it though | 14:48 |
| dsc_ | buZz: what 3d printer do you have? | 14:51 |
| buZz | for this project? https://nurdspace.nl/CoreAnet | 14:52 |
| buZz | i also have a Malyan M100 | 14:52 |
| buZz | and at the hackerspace we have ...8? 3d printers, but just 2x Creality Ender3 are hooked up | 14:52 |
| buZz | https://i.imgur.com/HIOi1Fo.jpeg | 14:53 |
| dsc_ | nice | 14:54 |
| tmlind | fyi rebasing the pending mapphone patches on v6.8-rc4 here | 14:55 |
| buZz | yeah its pretty | 14:55 |
| buZz | current printarea is ~20x30x20cm , but i could increase that in all directions still | 14:55 |
| buZz | and superquiet and fast | 14:55 |
| tmlind | for a 3d printer use of mz609 with the touchscreen the touchscreen driver remains problematic | 14:56 |
| tmlind | uvos__: or did you figure out some recent touchscreen branch for mz609? | 14:56 |
| buZz | -most- i want the screen for is just displaying temps etc, but controlling printer from it would be nice yeah | 14:59 |
| buZz | i can do it all remotely too, over http | 14:59 |
| tmlind | yeah ok | 15:00 |
| uvos__ | tmlind: oh no i forgot about that | 15:50 |
| uvos__ | miz61x then | 15:50 |
| uvos__ | or allwiner tablet | 15:50 |
| tmlind | yeah | 16:04 |
| tmlind | pushed out v6.8-rc4 based branches | 16:04 |
| tmlind | https://github.com/tmlind/linux/commits/mapphone-pending-v6.8/ | 16:04 |
| tmlind | https://github.com/tmlind/linux_openpvrsgx/commits/mapphone-pending-pvr-omapdrm-v6.8/ | 16:04 |
| tmlind | looks like the console on the lcd issue is still there and needs to be bisected | 16:05 |
| tmlind | very lightly tested as always :) | 16:06 |
| Wizzup | tmlind: for mz60x, is it just the ts that does not work, or also the lcd? | 16:21 |
| tmlind | Wizzup: mz617 (and mz615 and mz616) are pretty close to working with mainline, different ts controller | 16:51 |
| tmlind | had an oops in the v6.8-rc4 based branch at gsd_receive_buf() | 16:51 |
| tmlind | at least got the pstore trace, i disabled the pstore compression option so it's more usable now | 16:51 |
| tmlind | gotta go now | 16:52 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
