| tmlind | Wizzup: best to test each operating point separately by disabling other c-states via sysfs | 06:43 |
|---|---|---|
| tmlind | Wizzup: usually the issue is is voltage related or emif configuration related | 06:43 |
| tmlind | uvos: interesting so what did you figure out on mainline sgx underperforming? | 06:44 |
| Wizzup | tmlind: ok, I will see if I can figure out where to change the c-states in sysfs if it's not just setting the cpu scaling frequency | 11:24 |
| Wizzup | arno11: I'll do the PR momentarily | 11:32 |
| arno11 | ok cool, thx | 11:32 |
| Wizzup | tmlind: uvos: so it looks to me like qminet creates /dev/qcqmi{0,1,2,3} on android and the dir we linked earlier on gitlab really is the only thing necessary for it | 15:41 |
| Wizzup | I wonder if qmicli can somehow just operate on these itnerfaces | 15:41 |
| Wizzup | well that and the four qmi{0,1,2,3} network interfaces I guess | 15:44 |
| Wizzup | here are some other mentions: https://forum.sierrawireless.com/t/network-interface-for-qcqmi/8909 | 15:48 |
| Wizzup | https://sigquit.wordpress.com/2014/06/11/qmiwwan-or-gobinet/ | 15:49 |
| Wizzup | this also mentions maybe what uvos was talking about | 15:49 |
| Wizzup | Note: Some Android devices also support QMI-capable chipsets through GobiNet (everything hidden in the kernel and the RIL). In this case, though, you may see that shared memory can also be used to talk to the QMI device, instead of a /dev/qcqmi port. | 15:49 |
| Wizzup | but I don't think this is the case here | 15:49 |
| Wizzup | (famous last words) | 15:49 |
| Wizzup | in general this seems like useful info btw | 15:50 |
| Wizzup | for me at least | 15:52 |
| arno11 | sicelo: Wizzup: if i pkill cm_pulse there is no negative impact, it restarts automatically and calls are working again...with a positive effect: (alerting) tone is working (again) now | 16:24 |
| arno11 | h-s-m, sometimes nokia-modem and now cmt_pulse. i wonder what's going on during boot | 16:27 |
| arno11 | ok issue solved for h-s-m | 16:46 |
| arno11 | renaming 15hildon-status-menu to 75hildon-status-menu (just after sphone in fact) makes h-s-m and availability BTN working correctly again | 16:49 |
| arno11 | freemangordon: ^^^ | 16:50 |
| arno11 | i'll try to start cmt_pulse a bit later now | 16:51 |
| arno11 | issue solved for cmt_pulse as well | 17:01 |
| arno11 | Wizzup: i need to make a small change to cmt_pulse PR | 17:03 |
| Wizzup | go for it | 17:05 |
| arno11 | could the very slow startup be related to 1 or several apps starting at the wrong time iyo ? | 17:05 |
| Wizzup | I don't know if the startup is very slow | 17:12 |
| Wizzup | does fremantle start much faster? | 17:12 |
| arno11 | yes a lot faster | 17:14 |
| arno11 | and an old leste img starts really faster too | 17:14 |
| arno11 | (PR updated btw) | 17:16 |
| arno11 | BTW for hildon-status-menu, is D4 affected or everything is ok with availibility BTN ? | 17:25 |
| Wizzup | it works for me | 17:27 |
| Wizzup | I don't know what problems you see specifically, but it works ok for me | 17:27 |
| arno11 | sorry for the lack of context: since december, availibility BTN in h-s-m was not working anymore on startup | 17:29 |
| arno11 | from a fresh install on n900 | 17:29 |
| arno11 | and starting it a bit later solved the issue | 17:29 |
| Wizzup | not sure what it would need then | 17:32 |
| arno11 | so (for n900) we need to modify h-s-m debian/rules or simply rename h-s-m xsession file in Xsession.post | 17:32 |
| arno11 | maybe we can modify debian/rules if there is no impact on other devices (dh_installxsession -u 'post 75' instead of dh_installxsession -u 'post 15') | 17:38 |
| Wizzup | ok, but you say this makes a difference i wonder why | 17:43 |
| Wizzup | like, what is making it bgreak | 17:43 |
| Wizzup | break* | 17:43 |
| arno11 | the fact that it starts before hildon-home, hildon-desktop and conversations imo | 17:47 |
| Wizzup | well conversations definitely has nothing to do with this | 17:49 |
| Wizzup | brb | 17:51 |
| Wizzup | correction, libqmi likely won't be able to talk to what I mentioned above, so it would indeed need to be integrated into qmi_wwan as non-usb backend | 18:17 |
| Wizzup | but that might be quite doable | 18:17 |
| freemangordon | arno11: I am against changing h-s-m start time unless we know why changing it make s a difference | 18:29 |
| freemangordon | also, there is no issue on d4 | 18:30 |
| freemangordon | or in VM in that regard | 18:30 |
| arno11 | ok | 18:30 |
| freemangordon | however... | 18:30 |
| arno11 | but remember i'm not a dev :P | 18:30 |
| freemangordon | I want to finish as much as I can with what I am doing currently (connui-cellular) | 18:30 |
| freemangordon | and then will see what's going on with n900/h-s-m | 18:31 |
| arno11 | ok | 18:31 |
| freemangordon | might take a week or so until I am on it | 18:31 |
| arno11 | ok (anyway i found a workaround for me and i'm happy with that) | 18:32 |
| arno11 | :) | 18:32 |
| freemangordon | right | 18:32 |
| freemangordon | Wizzup: sicelo: EDGE == UMTS == 2.5, right? | 18:32 |
| arno11 | no: edge == 2.5g | 18:32 |
| freemangordon | hmm | 18:32 |
| freemangordon | and what is UMTS? | 18:33 |
| arno11 | umts == 3G | 18:33 |
| freemangordon | ah | 18:33 |
| freemangordon | ok | 18:33 |
| arno11 | hsdpa == 3.5G | 18:33 |
| freemangordon | seems ofono does not report EDGE support | 18:33 |
| arno11 | but it works iirc | 18:33 |
| Wizzup | you need to check specific extra bits. like my older code does | 18:33 |
| freemangordon | Wizzup: care to point if you can, while I am searching? | 18:34 |
| Wizzup | also ofono has "edge" as technology | 18:34 |
| Wizzup | I am pretty sure | 18:34 |
| Wizzup | freemangordon: ofono_rat_change | 18:34 |
| freemangordon | but that's current RAT | 18:35 |
| freemangordon | not capabilities | 18:35 |
| Wizzup | what is gsm then, vs edge? | 18:35 |
| freemangordon | no idea :) | 18:35 |
| Wizzup | see my comment regarding NETWORK_MASK_EGPRS_SUPPORT | 18:35 |
| freemangordon | ok | 18:35 |
| Wizzup | at least thta might be relevant | 18:35 |
| freemangordon | that's what I need | 18:35 |
| Wizzup | note that btw we never show 3.5g in my current code | 18:36 |
| Wizzup | it always shows 3g for some reason (iiuc) | 18:36 |
| freemangordon | yes, this https://github.com/maemo-leste/connui-cellular/blob/ab3ddcb6f2c8971b21ade168a3f184b74f3f139a/lib/net.c#L233 | 18:37 |
| freemangordon | however, this needs modem to be connected to the net | 18:37 |
| freemangordon | but we need if modem supports edge at all | 18:38 |
| * freemangordon ckecks | 18:38 | |
| freemangordon | *checks | 18:38 |
| freemangordon | Features = sms net rat gprs ussd sim | 18:38 |
| freemangordon | Wizzup: ^^^ | 18:38 |
| freemangordon | this is what modem reports | 18:38 |
| freemangordon | oh, this is same as interfaces | 18:39 |
| Wizzup | I think you just use the current radio tech? | 18:40 |
| Wizzup | NetworkRegistration.Techology or whatever | 18:40 |
| freemangordon | yeah | 18:40 |
| freemangordon | seems I will have to | 18:40 |
| freemangordon | Wizzup: oh, see org.ofono.NetworkOperator | 18:41 |
| Wizzup | not all of these are present for all devices btw | 18:42 |
| Wizzup | this might well be, jfyi | 18:42 |
| freemangordon | this is where we have Technologies | 18:42 |
| freemangordon | yeah, right | 18:42 |
| freemangordon | ok, I think I have an idea how to continue | 18:43 |
| Wizzup | :) | 18:43 |
| Wizzup | important work | 18:43 |
| freemangordon | I already (almost) implemented multi-modem support | 18:43 |
| freemangordon | and get rid of libgofono | 18:44 |
| Wizzup | great | 18:44 |
| sicelo | Wizzup: I've seen 3.5G on my N900 with Leste :-) | 19:27 |
| Wizzup | sicelo: ok, maybe not 2.5G then | 19:37 |
| Wizzup | (that would match the codes comments) | 19:38 |
| arno11 | yes i can confirm too, 3.5G icon works | 19:39 |
| arno11 | not 2.5G | 19:39 |
| arno11 | but 2.5g network works for sure (even if it shows 2G icon) | 19:45 |
| sicelo | 21:09 < freemangordon> but we need if modem supports edge at all <<== org.ofono.RadioSettings.AvailableTechnologies. unfortunately it's optional, and most of our modems don't have it. there's no other way to know if a modem supports a particular tech | 20:26 |
| sicelo | actually scratch that ... we do have Available Techs | 20:30 |
| sicelo | it's just there's nothing for EDGE :-) | 20:31 |
| sicelo | N900 doesn't have it, mmm | 20:34 |
| sicelo | i think something's up with ofono/drivers/isimodem/radio-settings.c ... i can look at it someday | 20:48 |
| freemangordon | sicelo: yeah, I saw that ionterface | 21:03 |
| freemangordon | *interface | 21:03 |
| freemangordon | AvailableTechnologies = gsm umts lte | 21:03 |
| freemangordon | this is on the USB modem I use to develop in the VM | 21:04 |
| freemangordon | nothing about edge :) | 21:04 |
| sicelo | yes, nor 3.5G | 21:04 |
| freemangordon | mhm | 21:04 |
| sicelo | because you can't select 2.5 or 3.5 | 21:04 |
| freemangordon | mhm | 21:04 |
| sicelo | :-) | 21:05 |
| freemangordon | but again, this seems to be the network cell property | 21:05 |
| freemangordon | not modem one | 21:05 |
| freemangordon | maybe all modems are assumed to support it | 21:05 |
| freemangordon | and it depends on the tower etc | 21:05 |
| freemangordon | will see when I have more advanced code | 21:05 |
| sicelo | a cell/tower that supports 3.5 or 2.5 is also willing to use plain 2 or 3. so the negotiation for .5 is solely between modem & network. i never seen a phone where you can choose those modes | 21:10 |
| * sicelo looking forward to the upcoming code :-) | 21:10 | |
| sicelo | so isimodem driver doesn't have query_available_rats. damn | 21:22 |
| freemangordon | hmm, it should have | 21:29 |
| freemangordon | as on fremantle you can get network_supported_services | 21:32 |
| sicelo | i guess work on the ofono driver stalled before that could be done. we do have missing bits for the modem | 21:33 |
| freemangordon | get_radio_access_technology is the dbus call | 21:33 |
| sicelo | we're just lucky that a lot was already added | 21:33 |
| sicelo | fmg, yes but thay requires the driver to support it. | 21:33 |
| freemangordon | oh, sorry | 21:34 |
| freemangordon | it is Phone.Net.get_registration_status | 21:34 |
| freemangordon | anyway, I think I will manage to do that properly | 21:34 |
| sicelo | at least, even though AvailableTechnologies doesn't exist, you can still set "gsm" or "umts" on N900 as you need to. overall not a big deal | 21:43 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!