freemangordon | Wizzup: repo keys? | 08:32 |
---|---|---|
freemangordon | guys, do you have clue about Pavel's comment here https://git.kernel.org/pub/scm/network/ofono/ofono.git/tree/plugins/droid.c?id=07e00b7215cd98aa033f2ebc4ff169662d95d34c#n197 ? | 08:33 |
tmlind | freemangordon: pavel had patches to make use of the buggy /dev/ttyUSBn at modem interface, i think over ttyUSB4 | 08:50 |
freemangordon | tmlind: ok, but I lack the whole picture, like, why he preferred AT over QMI? | 09:00 |
freemangordon | anyway, I sent an email to him (you're in CC) | 09:00 |
freemangordon | tmlind: also, what is gsmmux? | 09:01 |
freemangordon | and why do we need it? | 09:01 |
tmlind | freemangordon: gsmmux is the multiplexed serial port interface /dev/gsmtty* | 09:02 |
freemangordon | but we have that, no? | 09:02 |
freemangordon | this is what we use in leste ofono, iiuc | 09:02 |
tmlind | freemangordon: yes we have that and that's handling the ofono notifications and kicking the qmimodem | 09:02 |
freemangordon | so, Pavel's comment is no longer valid then | 09:03 |
tmlind | well you can also use the buggy /dev/ttyUSB0 too.. | 09:03 |
freemangordon | ok, but what is buggy about it? | 09:03 |
tmlind | it eventually hangs the modem | 09:03 |
tmlind | and is a very limited interface | 09:04 |
freemangordon | and we don't know why/when (the hang)? | 09:04 |
tmlind | some commands hang the modem right away, some command are buggy and hang it at some point, don't recall the details | 09:04 |
freemangordon | ok, I see | 09:04 |
freemangordon | so, upstream driver is useless then, no? | 09:05 |
tmlind | afaik /dev/gsmtty* is the only way to access voice call audio and gnss features on the modem unless some of these can be done over the qmimodem somehow | 09:05 |
tmlind | well pavel might still use the upstream driver, not sure | 09:05 |
freemangordon | I think they shall work through qmi | 09:05 |
tmlind | and the interface is threre | 09:05 |
freemangordon | well, booting it in BP mode exports QMI interface | 09:06 |
tmlind | well if you figure out a way to get modem notifications work with just qmi only, that would help a lot of course | 09:06 |
tmlind | send an sms to the bp mode modem and see if x86 sees anything happening on the qmimodem? | 09:07 |
freemangordon | I can try, but before doing so I would like to have a clear picture on what to do with upstream driver | 09:07 |
tmlind | which driver, the ttyUSB4? just leave as it is :) | 09:07 |
freemangordon | I can't as it works on usb id | 09:07 |
freemangordon | and takes over | 09:07 |
freemangordon | also, qmicli does not allow me to listen for notifications | 09:08 |
tmlind | hmm but we have already patches for that? | 09:08 |
freemangordon | tmlind: sure, btu I want to upstream everything | 09:08 |
freemangordon | ofono devs are really active lately and very cooperative | 09:09 |
tmlind | yeah well adding a motorolamodem may be needed to get it all working and have the ttyUSB stuff too | 09:09 |
freemangordon | I don;t want to miss that opportunity | 09:09 |
freemangordon | also, they are even willing to discuss/fix dbus API etc | 09:09 |
tmlind | yeah sure nice | 09:09 |
freemangordon | to me it makes absolutely no sense to continue using 1.x given that 2.4 is out and 2.5 or 3.0 is on its way | 09:10 |
tmlind | agreed | 09:10 |
freemangordon | and I will spend time on make things happen, given that I have the right idea on what has to be done | 09:11 |
tmlind | heh yeah great | 09:11 |
freemangordon | ok, lets wait for Pavel's reply | 09:12 |
tmlind | because the /dev/ttyUSB* are there, we should not drop the modem handling for it but instead add motorolamodem that supports qmi and other interfaces | 09:14 |
tmlind | so three interfaces total like pavel's comments listed | 09:14 |
freemangordon | ok, but droid.c uses atmodem | 09:14 |
freemangordon | which is buggy | 09:14 |
tmlind | yup | 09:15 |
freemangordon | so this shall be converted to use qmimodem | 09:15 |
freemangordon | IIUC | 09:15 |
freemangordon | but, I don;t want to send patches for that over Pavel's head | 09:15 |
tmlind | hmm well you can't convert a ttyUSB interface implemented in the modem's firmware to use qmimodem :) | 09:16 |
freemangordon | esp given that most probably I lack details | 09:16 |
freemangordon | umm, sorry, can't parse | 09:16 |
tmlind | well the modem firmware provides the /dev/ttyUSB* interfaces to linux, they are separate from the qmimodem interface | 09:17 |
freemangordon | what is this https://github.com/maemo-leste-upstream-forks/ofono/blob/maemo/chimaera-devel/plugins/motmdm.c#L612 then? | 09:17 |
freemangordon | btw QMI is over /dev/cdc-wdm0 AFAIK | 09:20 |
tmlind | that's the motorola modem part that uses the /dev/gsmtty* interfaces | 09:20 |
freemangordon | ok, but this is QMI, no? | 09:20 |
tmlind | no not qmi, but kicks the qmimodem for notifications | 09:20 |
tmlind | so /dev/cdc-wdm0 is qmi yeah, /dev/gsmtty* is multiplexed uart, and /dev/ttyUSB* are the buggy at modem interfaces | 09:21 |
freemangordon | umm, I got lost :) motorollamodem.c kicks modem to send notifications through QMI interface, no? | 09:21 |
tmlind | yeah based on the notifications coming from /dev/gsmtty* interfaces | 09:22 |
freemangordon | exactly | 09:22 |
freemangordon | but, ignoring that we don;t know how to tell QMI to send notifications without modme being kicked, we use qmimodem to communicate with it | 09:22 |
tmlind | likely there's some modem nvram setting that routes the notifications to either /dev/gsmtty* or to /dev/cdc-wdm0 | 09:22 |
tmlind | yeah we should be able to all the communication over qmi in theory | 09:23 |
freemangordon | yes, but besides notifications, it is all qmimodem, not gsmmuz | 09:23 |
freemangordon | right | 09:23 |
tmlind | yeah | 09:23 |
freemangordon | but dis is totally different to what we have in upstream | 09:23 |
freemangordon | *this | 09:23 |
freemangordon | where it is atmodem all over the place | 09:24 |
freemangordon | no gsmmux or QMI | 09:24 |
tmlind | right, upstream only has the support for /dev/ttyUSB* interface | 09:24 |
freemangordon | which seems to be useless for any real life usage as it hangs the modem :) | 09:24 |
freemangordon | so, my plan is to get rid of that and push our driver upstream, but I want to hear from Pavel first | 09:25 |
tmlind | more or less yeah, but the interface still exists, so should be possible to support that too | 09:25 |
freemangordon | umm, I am not sure that's possible | 09:25 |
freemangordon | as that means we will have to modem interfaces in ofono for a single modem | 09:26 |
tmlind | the motorola modem currently configures both qmi and /dev/gsmtty* interfaces | 09:26 |
tmlind | so should be able to also configure ttyUSB interfaces | 09:26 |
freemangordon | but why if it hangs the modem? | 09:26 |
tmlind | only some features hang the modem :) | 09:27 |
freemangordon | heh | 09:27 |
tmlind | gotta go here, bbl | 09:27 |
freemangordon | bbl | 09:27 |
Wizzup | freemangordon: hm? | 10:54 |
Wizzup | freemangordon: I don't think so, must be some mafw thing | 10:54 |
Wizzup | freemangordon: I'll take a look in any case | 11:08 |
inky | > no more crash or seg fault :) | 11:20 |
inky | please package it. | 11:20 |
freemangordon | tmlind: did anyone try AT+QURCCFG ? | 16:23 |
freemangordon | hmm, does not seem to work | 17:00 |
diejuse | I have finally found the culprit of the problem of the portrait mode breaking when pressing the space bar key with other virtual keyboards. | 19:47 |
diejuse | If I don't start hildon-input-method everything works fine. For some reason it creates a conflict in the space bar key pressed by other virtual keyboards. | 19:48 |
sicelo | freemangordon: since ofono code is still fresh in your mind ... maybe you can help me find where the name of the modem is 'created' ... the "2" in /n900_2 is annoying sometimes :p | 20:23 |
sicelo | not that it matters too much though | 20:24 |
inky | wow! | 20:45 |
freemangordon | sicelo: ofono_modem_create() | 20:58 |
freemangordon | looks like the modem is created twice | 20:59 |
freemangordon | or even 3 times | 20:59 |
freemangordon | as afaik it starts from 0 | 20:59 |
sicelo | that sounds like a problem :-/ | 21:02 |
sicelo | looks like it's because the modem has many sysfs entries: hsi-char, cmtspeech, n900-modem, phonet0 | 21:27 |
sicelo | anyway, don't fix it if it ain't broken :p | 21:27 |
arno11 | Wizzup: i got keyboard (sdl input) working with drnoksnes. fullscreen works too. Now i have few troubles with sound and gui is broken. but games work fine :) i'll be very busy during the 2 next weeks but will try to upload stuff asap if you want to try it | 22:31 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!