n900000 | hi | 06:16 |
---|---|---|
n900000 | am I doing something wrong or is wifi broken on the latest n900 image? | 06:16 |
n900000 | the select connection popup is empty | 06:16 |
stano | hi n900000 | 06:23 |
stano | idk | 06:23 |
stano | does it need calibration? | 06:23 |
n900000 | calibration? | 06:28 |
n900000 | I'm dd'ing the 6/27 img right now, lets see if it works on that one | 06:29 |
stano | i don't know if n900 needs calibration. droid4 does | 06:31 |
n900000 | yeah nothing about calibration on the n900 page | 06:32 |
n900000 | I guess most people using leste moved on to the droid 4? | 06:32 |
stano | quite a few still use n900 | 06:32 |
stano | did you flash uboot with 0xFFFF? | 06:33 |
stano | r install it with fremantle? | 06:33 |
stano | maybe you can point me to getting n900 going | 06:33 |
n900000 | flashed with 0xFFFF | 06:33 |
n900000 | my freemantle won't install uboot for some reason | 06:34 |
stano | how long do you wait before sliding the unit open? | 06:34 |
stano | the leste image on microsd has to be inserted *before* doing 0xFFFF? | 06:35 |
stano | here i just made a fun ring tone http://0x0.st/-O0u.flac | 06:39 |
n900000 | you have to boot with the keyboard open | 06:39 |
n900000 | and yes put the sd card in first | 06:40 |
stano | k | 06:40 |
n900000 | Ok wifi works on the 6/27 image | 06:43 |
n900000 | I think the latest one is broken then | 06:43 |
n900000 | no idea how to file that as a bug though | 06:43 |
stano | https://github.com/maemo-leste/bugtracker/issues | 06:44 |
n900000 | ah I don't have a github account | 06:44 |
stano | have you been using leste on n900 for a while? | 06:48 |
stano | or are you new to it | 06:48 |
n900000 | completely new | 06:50 |
n900000 | I just got my n900 today lol | 06:50 |
n900000 | I have a droid 4 on the way too | 06:50 |
stano | congrats! :) | 06:50 |
n900000 | and waiting on that pinephone keyboard | 06:50 |
n900000 | thanks | 06:50 |
stano | do you know if flashing the n900 can be done with it connected to a usb hub? | 06:51 |
n900000 | I was planning to use PMOS on it but I wasn't happy with the performance | 06:51 |
n900000 | I've been connecting it straight but it should probably work | 06:51 |
stano | pmos on n900 or droid4? | 06:55 |
n900000 | n900 | 06:55 |
stano | didn't know it was ported to n900 wow | 06:56 |
n900000 | I'm gonna stick with leste for both most likely | 06:56 |
n900000 | even mpd was stuttering on PMOS | 06:56 |
stano | would you perhaps test some emulators on n900 | 06:56 |
n900000 | games? sure | 06:57 |
stano | ok that helps motivate me | 06:57 |
n900000 | back in the freemantle days I remember seeing people running snes and PS1 emulators on it | 06:57 |
n900000 | when the phone was $800 lol | 06:57 |
stano | drpocketsnes is made for 16 bpp screens | 06:57 |
stano | pcsx is closer to ready | 06:58 |
n900000 | nice | 06:58 |
n900000 | There's a thread on the maemo forums with one of the leste devs running pcsxr on his droid 4 | 06:59 |
stano | i wouldn't say he's a developer. tinkerer maybe. | 07:00 |
stano | developers actually write software | 07:01 |
n900000 | I dont care lmao | 07:01 |
sicelo | n900000: yes n900 wifi broken on leste, it seems | 07:09 |
sicelo | might be as simple as wpa_supplicant version, etc. | 07:10 |
n900000 | it worked on the 2nd to latest image | 07:11 |
n900000 | 6/27 | 07:11 |
n900000 | is there a way to only have audio go through headphones? | 07:38 |
n900000 | when I plug headphones in the audio plays on both the n900 speakers and the headphones | 07:38 |
stano | that's certainly what should happen n900000 | 07:59 |
n900000 | is there a way to route it to just the headphones? I installed pulsemixer and none of the output settings change the default behavior | 08:02 |
stano | i seem to recall without the pulseaudio stuff, i could use alsamixer to turn up and down headphones vs speakers, on the droid 4 | 08:04 |
stano | maybe nobody has gotten around to setting up pulseaudio correctly for n900 yet | 08:05 |
n900000 | yup ended up doing it through alsamixer | 08:12 |
n900000 | just muting the speaker function channel | 08:12 |
n900000 | weird how the volume rocker doesn't adjust the volume too | 08:13 |
stano | considering it's a few guys making this stuff for a handful of non-paying customers, it works amazingly well :P | 08:16 |
n900000 | I completely agree | 08:17 |
stano | pulseaudio is more or less the only way to handle dynamic audio routing for things like phone calls | 08:21 |
stano | if that's not on your requirements it can be disabled | 08:22 |
freemangordon | n900000: https://github.com/maemo-leste/bugtracker/issues/489#issuecomment-879793412 | 08:32 |
stano | wow /sys/class/power_supply/battery/voltage_now is giving me 3229000 and no beeps or shutdown | 11:48 |
stano | this is with the EB41 on droid4. maybe firmware is still confused from experimental battery. | 11:48 |
stano | darnit i missed the shutdown, now i have to recharge to full, then discharge to full, then recharge to full again to calibrate battery monitor | 12:15 |
Wizzup | morning | 12:17 |
stano | morning Wizzup | 12:17 |
Wizzup | I will look at the n900 wifi stuff today, I just need to set up my power supply and such | 12:20 |
stano | what hardware are most leste-users running right now and what are you hoping them to be running it on in near future Wizzup ? | 12:37 |
Wizzup | We don't really do analytics | 12:47 |
Wizzup | I'd say the n900 and droid4 are common, pinephone also somewhat | 12:47 |
Wizzup | I think those are also a good first bet in the near future. | 12:47 |
stano | looks like linux might be possible on Tegra X1 tablets | 14:09 |
stano | anyone using leste on a tablet yet? | 14:10 |
Wizzup | I had/have it working on the pinetab | 14:17 |
stano | well the google pixel c would be another potential candidate, as it uses the 'X1' cpu which has linux support (same cpu as jetson nano) but i don't see anyone with a working display driver for that tablet screen | 14:57 |
stano | they can be had around 110 euro | 14:57 |
stano | deliver around 500 nits brightness with a very powerful gpu, which also has linux drivers | 14:57 |
bencoh | regarding pulseaudio / calls / automatic switching ... what kind of automagic does pulseaudio do anyway? | 15:00 |
Wizzup | I think pulseaudio can act on plug events for headphone and such | 15:01 |
Wizzup | it can also cork streams if certain types activate | 15:01 |
Wizzup | e.g. just having mumble running can break your browser audio | 15:01 |
Wizzup | (there's a lot of info on this online) | 15:01 |
bencoh | wait, what? | 15:01 |
Wizzup | https://github.com/mumble-voip/mumble/issues/1069 | 15:01 |
bencoh | (the plug events is possible in-kernel with alsa, and doesn't sound impossible to replace in userspace) | 15:02 |
Wizzup | yep | 15:02 |
bencoh | Wizzup: ah, the class thing, yeah | 15:02 |
stano | https://github.com/pixelc-linux | 15:02 |
bencoh | I have to admit that is handy | 15:02 |
Wizzup | need to for a bit, bbl :) | 15:03 |
Wizzup | (sorry) | 15:03 |
bencoh | :) | 15:03 |
mighty17[m] | hello, is there omap gptimer support in mainline, asking as gp timer creates a pwm signal for the backlight of my device | 18:31 |
tmlind | mighty17[m]: yes grep for ti,omap-dmtimer-pwm | 18:38 |
mighty17[m] | yeah just found out its called dmtimer and not gptimer | 18:38 |
mighty17[m] | https://github.com/torvalds/linux/blob/master/Documentation/devicetree/bindings/pwm/pwm-omap-dmtimer.txt nice | 18:38 |
mighty17[m] | tmlind: any ideas how i can get a correct value for the pwm? | 18:40 |
mighty17[m] | https://pastebin.ubuntu.com/p/x7bXX6PNVg/ | 18:40 |
mighty17[m] | like here https://github.com/torvalds/linux/blob/5bfc75d92efd494db37f5c4c173d3639d4772966/arch/arm/boot/dts/am3517-evm.dts#L147 how do u know its 5000000 | 18:44 |
tmlind | mighty17[m]: in the pwm timer case, the timer is just feeding the programmed clock rate to the connected device, maybe that's considered sufficient for a backlight | 19:40 |
mighty17[m] | where do u get programmed clock rate | 19:52 |
uvos | check the rate on android? | 20:04 |
uvos | either by reading in the vendor kernel or by dumping the corrisponding registers | 20:05 |
Wizzup | I see this on n900: | 20:22 |
Wizzup | nl80211: Scan SSID | 20:22 |
Wizzup | nl80211: Scan extra IEs - hexdump(len=10): 7f 08 00 00 00 00 01 00 00 40 | 20:22 |
Wizzup | nl80211: Scan trigger failed: ret=-22 (Invalid argument) | 20:22 |
Wizzup | wlan0: State: SCANNING -> DISCONNECTED | 20:22 |
mighty17[m] | uvos: https://github.com/LineageOS/android_kernel_ti_omap4/blob/cm-14.1/arch/arm/mach-omap2/board-espresso-display.c#L131 | 20:22 |
mighty17[m] | 32kHz to ns doesnt seem to work | 20:23 |
Wizzup | the internet said it might be related to WNM being enabled in debian, but I am not sure if it was disabled previously or if it is only enabled now | 20:25 |
Wizzup | https://bugs.archlinux.org/task/61119 could it be this? | 20:28 |
uvos | why would that appear on n900 only tho | 20:30 |
uvos | it uses the same driver as d4 no? | 20:30 |
uvos | wl12x | 20:30 |
Wizzup | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=833507 this suggests it might be the 'mac randomisation' | 20:33 |
Wizzup | (which NM does, but we also do that, don't we?) | 20:33 |
sicelo | No, n900 uses not wl12x | 20:34 |
Wizzup | https://lists.infradead.org/pipermail/hostap/2017-December/038142.html | 20:34 |
Wizzup | wl1251_spi 16384 0 | 20:34 |
Wizzup | wl1251 69632 1 wl1251_spi | 20:34 |
uvos | ahit wl1251 | 20:34 |
uvos | *ah its | 20:34 |
uvos | i thought n900 has wl1271 | 20:35 |
uvos | right totally different driver then | 20:35 |
uvos | carry on | 20:35 |
Wizzup | the above link seems likely | 20:35 |
Wizzup | to be the cause/fix | 20:35 |
Wizzup | parazyd: do you know the old version of wpa_supplicant, before we got the new one? | 20:35 |
parazyd | What about it? | 20:36 |
parazyd | It's upstream from Debian | 20:36 |
Wizzup | yes, and it looks like the bug might be either in the config they (now) use, but I need to know what versions we used and what they now have | 20:36 |
parazyd | https://packages.debian.org/source/buster/wpa | 20:37 |
Wizzup | also I wonder if we can undo changing the mac addr of the n900 | 20:37 |
Pali | in 2.6.28 kernel both wl1251 and wl1271 are handled by wl12xx kernel driver... later in kernel ňsupport for wl1251 was forked/moved into wl1251 driver and wl1271 and new chips stays in wl12xx driver | 20:37 |
Wizzup | parazyd: ok, where did we get the newer version from? | 20:38 |
Wizzup | ok, newer is 2.2.9 | 20:38 |
parazyd | Newer is 2.9 | 20:38 |
parazyd | I think unstable | 20:38 |
Wizzup | bullseye I think yeah | 20:39 |
parazyd | https://salsa.debian.org/debian/wpa | 20:39 |
Wizzup | well I don't know WTH to do | 20:39 |
parazyd | All is here so you could see git log | 20:39 |
Wizzup | I guess I can try to build custom kernel and see what changed | 20:39 |
Wizzup | but I wonder if 2.2.9 is built with WNM enabled whereas 2.2.7 had it disasbled | 20:39 |
Wizzup | disabled* | 20:39 |
uvos | any one know where wpa's bugtracker is | 20:40 |
uvos | seams not to have one | 20:40 |
Wizzup | no, looks like WNM is still disabled | 20:40 |
uvos | maybe one shoult dry wpa git first to see if its fixed, i find it hard to belive that you cant ask the kernel if WNM params should be added | 20:41 |
Wizzup | mac addr randomisation is the other potential reason https://groups.google.com/g/linux.debian.bugs.dist/c/ZOWNZzUyr48/m/lALgGUSdCgAJ?pli=1 | 20:41 |
Wizzup | uvos: this is CONFIG_WNM in wpa_supplicant, not in kernel | 20:41 |
parazyd | We currently have rules to make a persistent mac | 20:42 |
Wizzup | both 2.7 and 2.9 have it disabled it seems | 20:42 |
Wizzup | parazyd: yes, so we change it from the 'default' on, whatever it could be | 20:42 |
parazyd | The default is random | 20:42 |
uvos | Wizzup: yeah i know but having to do what https://lists.infradead.org/pipermail/hostap/2017-December/038142.html suggests seems wrong | 20:42 |
uvos | the kernel should have some whay to tell you what the interface supports... | 20:42 |
Wizzup | parazyd: that's uncommon I would say amongst most wifi devices | 20:43 |
Wizzup | parazyd: where can I disable the randomisation / custom setting of mac addr? | 20:43 |
parazyd | That's how the n900 kernel driver worls | 20:43 |
Wizzup | yes, please tell me how I can disable the userspace call | 20:43 |
parazyd | Because IIRC the mac is saved in cal? | 20:44 |
parazyd | It's somewhere in /etc | 20:44 |
parazyd | Not at home rn, sorry | 20:44 |
Wizzup | looks like it's 00mac | 20:44 |
parazyd | Ah yes | 20:44 |
Wizzup | lol it sources a file that doesn't exist on my n900 | 20:44 |
Wizzup | /etc/network/n900-staticmac | 20:45 |
Wizzup | how does that even work | 20:45 |
parazyd | It should be created after the first time | 20:45 |
Wizzup | well I've used this for 1+ year | 20:45 |
Wizzup | so that can't be the issue | 20:45 |
sicelo | Btw, when I was running sid on n900, I didn't have issues with wifi. Can't remember what version it was though | 20:45 |
parazyd | So first it creates a random mac | 20:45 |
parazyd | And then saves it | 20:45 |
parazyd | Subsequently it should load the created file | 20:45 |
Wizzup | yes, and the file does not exist | 20:46 |
Wizzup | so something is quite wrong | 20:46 |
Wizzup | I'll reboot and see if I can figure it out | 20:46 |
parazyd | It should run on pre-up iirc? | 20:46 |
parazyd | So no need for reboot | 20:46 |
uvos | well previoulsy the permanent setting was broken but wifi still worked see https://github.com/maemo-leste/bugtracker/issues/475 | 20:47 |
Wizzup | well, the interface went online just an hour ago when I did the update | 20:47 |
uvos | (permanent mac setting that is) | 20:47 |
Wizzup | uvos: yes, but it is possible that the new version broke there somehow | 20:47 |
uvos | sure | 20:47 |
Wizzup | this is so annoying | 20:48 |
Wizzup | wl1251 has no params, great | 20:50 |
uvos | this enableing the use of hidden wifi thing is really expensive eh? | 20:50 |
uvos | :P | 20:50 |
Wizzup | yup | 20:51 |
Wizzup | I'm mostly just annoyed not at the bug but that we shipped something that broke everyone's n900 wifi and there seems to be no simple fix | 20:51 |
sicelo | Just to be sure ... the recent images still have rfkill in kernel? | 20:52 |
parazyd | note-to-self: Provide also a generic runlevel without hildon for debugging | 20:52 |
Wizzup | sicelo: I did apt update | 20:53 |
Wizzup | didn't reboot | 20:53 |
Wizzup | restarted icd2 and wpa_spplicant | 20:53 |
Wizzup | and wifi broke | 20:53 |
Wizzup | so this is not related to kernel | 20:53 |
Wizzup | (changes) | 20:53 |
sicelo | Ok | 20:53 |
Wizzup | uvos: btw https://hostap.epitest.fi/bugz/ | 20:53 |
Wizzup | that's what google gave me at least | 20:54 |
Wizzup | well, I suppose all there's left to do is mail their ml | 20:56 |
Wizzup | hm, it looks like there should be some way to get more debug https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/net/wireless/ti/wl1251/wl1251.h?id=f63b4c971f5fb1310f145785c3b2b77651ef129e#n69 | 20:58 |
Wizzup | ah nope https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/net/wireless/ti/wl1251/wl1251.h?id=f63b4c971f5fb1310f145785c3b2b77651ef129e#n53 | 20:59 |
Wizzup | so a kernel recompile/build is in order I suppose | 21:00 |
* Wizzup bbl | 21:00 | |
Wizzup | parazyd: we can just get the old wpa_supplicant pinned for n900 or something stupid? | 21:10 |
Wizzup | I'd then upgrade the icd plugin to work with it | 21:10 |
sicelo | What's the version right now? | 21:11 |
Wizzup | https://packages.debian.org/source/bullseye/wpa | 21:11 |
parazyd | uh | 21:12 |
parazyd | That's possible | 21:12 |
parazyd | But we can't do a downgrade | 21:12 |
Wizzup | so the only option left ahead of us/me/someone is to bisect wpa supplicant on the n900 | 21:14 |
Wizzup | I think | 21:14 |
sicelo | Mmm, looks like bullseye has same version as sid according to that page. Worked (with linux 5.12) | 21:14 |
Wizzup | sicelo: how is wpa_supplicant started for you? | 21:15 |
sicelo | systemd | 21:15 |
sicelo | Mmm, no, connman | 21:15 |
parazyd | So what happens on the n900 generally? | 21:16 |
sicelo | Or both :-) | 21:16 |
Wizzup | sicelo: I mean in 'ps xua' | 21:16 |
parazyd | If you manually start wpasup | 21:16 |
Wizzup | parazyd: are you asking me? | 21:16 |
parazyd | Yes | 21:16 |
Wizzup | what happens is simple: | 21:16 |
Wizzup | > scan | 21:16 |
Wizzup | OK | 21:16 |
Wizzup | <3>CTRL-EVENT-SCAN-FAILED ret=-22 | 21:16 |
Wizzup | 100% of the time | 21:16 |
Wizzup | -22 = EINVAL | 21:16 |
sicelo | One more thing to check ... did we perhaps lose -wext in wpa-supplicant cmdline? | 21:16 |
parazyd | ugh | 21:16 |
Wizzup | so the kernel rejects the scan request because the scan requets contains invalid scan data | 21:16 |
Wizzup | sicelo: we have both wext and the other | 21:17 |
Wizzup | root 3023 0.0 2.1 7244 5496 ? S 20:54 0:00 /sbin/wpa_supplicant -s -P /run/wpa_supplicant.wlan0.pid -i wlan0 -u -D nl80211,wext -c /etc/wpa_supplicant/wpa_supplicant.conf | 21:17 |
sicelo | Try putting wext first | 21:17 |
sicelo | Ahead of nl, that i | 21:18 |
Wizzup | ioctl[SIOCSIWSCAN]: Inappropriate ioctl for device | 21:18 |
sicelo | *is | 21:18 |
Wizzup | <3>CTRL-EVENT-SCAN-FAILED ret=-1 | 21:18 |
Wizzup | sicelo: are you sure we used wext? | 21:20 |
sicelo | yes we must use wext | 21:20 |
sicelo | what starts wpa_supplicant for us (Leste)? | 21:21 |
Wizzup | sicelo: I don't see wext working | 21:23 |
Wizzup | sicelo: via dbus I think | 21:23 |
sicelo | i'll test my old debian sid install in a moment | 21:25 |
sicelo | and if luck permits, i can try out a Leste image too | 21:26 |
Wizzup | usbnet should work | 21:28 |
sicelo | the images in question are the ones in phoenix.* ? | 21:32 |
sicelo | ah, i see they're the same | 21:35 |
Wizzup | sicelo: yes, the images are also broken | 21:38 |
Wizzup | and any update to older images will also break | 21:39 |
sicelo | sure. i'm downloading newest | 21:39 |
Wizzup | I checkedps xua before I upgraded, and wext was not first, btw | 21:40 |
Wizzup | so I am pretty sure we were *not* using wext before | 21:40 |
Wizzup | sicelo: let me know, I'm going to do something else for a bit | 21:44 |
Wizzup | but I think we'll have to either bisect wpa_supplicant (and build our own with a fix), or rebuild kernel with the wl1251 code changed to debug what it thinks is 'invalid' | 21:44 |
uvos | did you strace the ioctl that returns EINVAL? | 21:46 |
Wizzup | seems to be this: | 21:49 |
Wizzup | 3130 sendmsg(5, {msg_name={sa_family=AF_NETLINK, nl_pid=0, nl_groups=00000000}, msg_namelen=12, msg_iov=[{iov_base={{len=52, type=0x17 /* NLMSG_??? */, flags=NLM_F_REQUEST|NLM_F_ACK, seq=1626378496, pid=3720350778}, "\x21\x00\x00\x00\x08\x00\x03\x00\x03\x00\x00\x00\x08\x00\x2d\x00\x04\x00\x01\x00\x0e\x00\x2a\x00\x7f\x08\x00\x00\x00\x00\x01\x00"...}, iov_len=52}], msg_iovlen=1, msg_controllen=0, | 21:49 |
Wizzup | msg_flags=0}, 0) = 52 | 21:49 |
Wizzup | 3130 recvmsg(5, {msg_name={sa_family=AF_NETLINK, nl_pid=0, nl_groups=00000000}, msg_namelen=12, msg_iov=[{iov_base={{len=36, type=NLMSG_ERROR, flags=NLM_F_CAPPED, seq=1626378496, pid=3720350778}, {error=-EINVAL, msg={len=52, type=0x17 /* NLMSG_??? */, flags=NLM_F_REQUEST|NLM_F_ACK, seq=1626378496, pid=37 | 21:49 |
Wizzup | maybe we're somehow telling it to scan a specific ssid? seems weird to me though | 21:52 |
Wizzup | nl80211: Add NL80211_SCAN_FLAG_FLUSH | 21:54 |
Wizzup | maybe this is the problem somehow? | 21:54 |
n900000 | hello | 22:05 |
n900000 | having a small issue getting kexecboot running on my new droid 4 | 22:05 |
n900000 | I followed the wiki to update android, and to install kexecboot | 22:05 |
sicelo | root@devuan-n900:~# grep WEXT /boot/config-5.1.21 | 22:06 |
sicelo | # CONFIG_CFG80211_WEXT is not set | 22:06 |
n900000 | but on boot it just goes straight to fastboot with "boot failure" at the top | 22:06 |
n900000 | happens with or without a leste flashed sd card inserted | 22:06 |
uvos | mbm determined that you violated mbm signature varification | 22:07 |
uvos | how did you upgrade? | 22:07 |
n900000 | I just followed the "Upgrade Android" section of the wiki | 22:07 |
Wizzup | hmm.... | 22:08 |
n900000 | could get past the initial setup out of the box because I don't have a SIM card sadly | 22:08 |
n900000 | couldn't* | 22:08 |
uvos | hmm | 22:08 |
n900000 | fastboot also says that the device is locked with status code:0 | 22:08 |
n900000 | is that relevant? | 22:08 |
uvos | no | 22:08 |
uvos | did the new android boot before you installed kexecboot? | 22:09 |
n900000 | yeah I got the little android logo with the loading bar | 22:09 |
uvos | i dont know remeber what the bootlogo for moto android looks like | 22:09 |
uvos | i think you might be describing the recovery | 22:10 |
uvos | did it boot to android fully | 22:10 |
n900000 | maybe yeah | 22:10 |
uvos | what do you mean maybe | 22:10 |
n900000 | I could try reflashing the update and let you know | 22:10 |
uvos | no | 22:10 |
n900000 | maybe that was the recovery | 22:10 |
uvos | wait | 22:10 |
uvos | so first lets erase utags again | 22:12 |
uvos | fastboot erase utags | 22:12 |
n900000 | ok | 22:12 |
uvos | i assume you left allow-mbmloader-flashing-mbm.bin installed | 22:12 |
n900000 | yeah I had it installed | 22:12 |
n900000 | ok utags erased | 22:13 |
Wizzup | here's something funny. | 22:13 |
uvos | ok so grab VRZ_XT894_9.8.2O-72_VZW-18-8_CFC.xml.zip | 22:13 |
Wizzup | I reverted back to the older wpa and libicd plugin | 22:13 |
Wizzup | and if I type 'scan' on wpa_cli, it fails the same way | 22:13 |
Wizzup | but the dbus scan call does not | 22:13 |
uvos | heh | 22:13 |
uvos | fun | 22:13 |
Wizzup | of course, neither mdbus2 or dbus-send can actually send a{sv} args | 22:13 |
n900000 | yup re-downloaded it | 22:14 |
uvos | ok | 22:14 |
uvos | unpack it | 22:14 |
Wizzup | (because why would likfe ever be easy...) | 22:14 |
uvos | and then fastoot flash boot boot.img | 22:14 |
uvos | fastboot flash system system.img | 22:14 |
uvos | fastboot flash recovery recovery.img | 22:14 |
uvos | also device_tree device_tree.bin (might not be the right name sec) | 22:15 |
uvos | then fastboot erase data | 22:15 |
uvos | and fastboot erase cache | 22:15 |
uvos | and fastboot reboot | 22:15 |
Wizzup | actually mdbus2 can: mdbus2 -s fi.epitest.hostap.WPASupplicant /fi/w1/wpa_supplicant1/Interfaces/1 fi.w1.wpa_supplicant1.Interface.Scan "{'Type': 'active'}" | 22:16 |
n900000 | Hmm when I try to flash boot.img | 22:17 |
n900000 | I get | 22:17 |
n900000 | fastboot: error: cannot get boot partition size | 22:17 |
uvos | might need flash:raw | 22:17 |
uvos | altho i dont think d4 needs that | 22:17 |
uvos | are you sure its a d4? :P | 22:17 |
n900000 | lol it absolutely is | 22:17 |
uvos | hm ok wierd | 22:17 |
uvos | anyhow flash:raw | 22:18 |
n900000 | flash:raw works | 22:18 |
uvos | hmm something is wrong here | 22:19 |
Wizzup | uvos: sicelo: parazyd: yeah so via dbus scanning in wpa_supplicant works | 22:20 |
uvos | its fastboot flash devtree device_tree.bin | 22:20 |
uvos | btw | 22:20 |
Wizzup | also on newer wpa_supplicant | 22:20 |
Wizzup | and wpa_cli's 'scan' not working is true for both versions | 22:20 |
Wizzup | the only reason I upgraded to newer wpa is because the dbus interface does funky stuff with hidden APs | 22:21 |
uvos | n900000: could you ppaste your mbm version | 22:21 |
sicelo | Wizzup: ah, great then | 22:22 |
n900000 | everything is working up until fastboot erase data | 22:22 |
n900000 | Erasing 'data' FAILED (remote: 'no such partition') | 22:22 |
n900000 | fastboot: error: Command failed | 22:22 |
n900000 | sure no prob, how do I get it? | 22:23 |
Wizzup | sicelo: problem is not solved | 22:23 |
Wizzup | but some more info was gathered | 22:23 |
uvos | sry its userdata | 22:23 |
uvos | n900000: second line in fastboot mode | 22:23 |
n900000 | ah ok | 22:23 |
uvos | eg 0a.77 | 22:23 |
uvos | also fastboot oem fb_mode_clear | 22:24 |
n900000 | got it, I just rebooted and it's giving me the red motorola logo | 22:24 |
uvos | great | 22:24 |
uvos | so mbm is happy | 22:24 |
n900000 | and it booted into android | 22:24 |
uvos | ok | 22:24 |
uvos | great | 22:24 |
uvos | now flash utags again | 22:25 |
uvos | (i assume you flashed kexecboot to bspw allready | 22:25 |
uvos | ) | 22:25 |
n900000 | kexecboot is already flashed unless the above overwrote it, yup | 22:25 |
uvos | no | 22:25 |
uvos | ok reboot into fastboot again | 22:25 |
uvos | and flash utags again | 22:25 |
n900000 | should I let android boot up fully before going back into fastboot? It's at that red spinning logo | 22:26 |
uvos | should not be stictly nesscary but you might break android if you just shut it down during inital boot | 22:27 |
n900000 | nvm the android setup booted up fine | 22:27 |
uvos | ok | 22:27 |
n900000 | ok mbm is | 22:27 |
n900000 | 0A.77 | 22:27 |
uvos | ok thats fine | 22:27 |
uvos | now flash utags | 22:27 |
n900000 | done | 22:28 |
n900000 | reboot? | 22:28 |
uvos | yes | 22:28 |
n900000 | eyyyyy | 22:28 |
n900000 | it worked | 22:28 |
uvos | great | 22:28 |
n900000 | ok let me put in my sd card | 22:29 |
n900000 | this pin thing is very annoying btw | 22:29 |
uvos | yes | 22:29 |
n900000 | perfect, booting in leste now | 22:31 |
n900000 | any idea what went wrong? I more or less copypasted the wiki instructions | 22:32 |
n900000 | also thanks a ton, I love the leste community | 22:32 |
uvos | proububly just a single bit that got flipped in your download or inital flash | 22:32 |
uvos | so some partition failed signature varfication | 22:33 |
n900000 | gotcha just bad luck then | 22:33 |
n900000 | I have 2 more droid 4's we'll see how those go when I flash them later | 22:33 |
n900000 | Oh wow the capacitive buttons work | 22:35 |
uvos | idk why everyone is so supprized by this :P | 22:35 |
stano | what's the actual mapping of those buttons? | 22:36 |
uvos | ? | 22:36 |
uvos | the keycodes? | 22:36 |
n900000 | pretty much exactly what you'd think | 22:36 |
stano | to keyboard, or some hildon function? | 22:36 |
n900000 | home is multitasking back is back etc | 22:36 |
stano | if the back is 'Esc' it solves one of the big SDL problems | 22:38 |
n900000 | also going from PMOS to leste surprises you with how much polish leste has. The hardware acceleration definitely helps too | 22:38 |
uvos | no back is f8 | 22:38 |
stano | k | 22:38 |
uvos | i just wrote a driver that makes those keys a keyboard | 22:38 |
uvos | with f6,7,8 and xf86search | 22:38 |
uvos | they are xgrabkey'd by hildon | 22:39 |
Wizzup | ok, I have something that works on the n900, but now to see if it still works with hidden APs | 22:39 |
freemangordon | uvos: f7/f8 are reserved | 22:40 |
freemangordon | for volume keys | 22:40 |
freemangordon | so please, do not reassign that to anything else | 22:40 |
uvos | nope | 22:40 |
uvos | we are going xf86volup down | 22:40 |
stano | the back closes sdl apps but f8 doesn't close the same app on desktop. <confused> | 22:40 |
uvos | im sorry | 22:40 |
uvos | anyhow yes i want ot change the d4 keys anyhow | 22:41 |
freemangordon | uvos: where do you do that? | 22:41 |
uvos | i think they should be home menu etc | 22:41 |
uvos | freemangordon: change the keys? | 22:41 |
uvos | in dts | 22:41 |
stano | i am so happy it closes the apps | 22:41 |
stano | thanks uvos | 22:41 |
freemangordon | using f7/f8 for other functions but volume | 22:41 |
uvos | having the android keys be f* was silly in retrospect | 22:41 |
stano | all kinds of things use f7/f8 | 22:41 |
sicelo | freemangordon: i'd let it go :) | 22:41 |
freemangordon | volume/zoom | 22:41 |
n900000 | Wizzup I was having wifi problems with the latest leste image last night and now I'm having the same thing on the droid 4 | 22:41 |
freemangordon | stano: no, OS uses those all over the p[lace | 22:42 |
stano | xf86volup / down seems right to me | 22:42 |
n900000 | n900 was just giving me a black pane when trying to scan networks | 22:42 |
freemangordon | sicelo: we can;t because it is not volume applet only | 22:42 |
uvos | freemangordon: yeah your never going to have the volume keys be f7/f8 on all devices ever | 22:42 |
uvos | freemangordon: anyhow lets not get into this | 22:42 |
uvos | yes i intend to reassing the android keys | 22:42 |
freemangordon | uvos: lets not go into this now, but we'll have to soon or later | 22:43 |
stano | i don't see the argument. volup/down buttons shouldn't be mapped to f7/f8 | 22:43 |
uvos | stano: the problem is not should | 22:43 |
freemangordon | stano: did you use maemo before? | 22:43 |
Wizzup | n900000: the n900 wifi problem is very specific | 22:43 |
uvos | the problem is that its hardcoded for lots of platforms to be xf86volup and xf86voldown | 22:43 |
uvos | like x86 etc | 22:43 |
stano | seems good to me to have them xf86volup/voldown | 22:43 |
uvos | so unless we want to pach lots of input dirvers | 22:43 |
uvos | its just not going to happen | 22:44 |
Wizzup | n900000: what are you seeing with the d4? | 22:44 |
freemangordon | ok, lets have a separate discussion on that in 2 weeks (I am MIA starting Sunday) | 22:44 |
freemangordon | but the so-called "volume keys" on android have much more functions on maemo | 22:45 |
uvos | they have several funcitons on android too | 22:45 |
uvos | and this isent about android | 22:45 |
Wizzup | we had this discussion before I think | 22:45 |
n900000 | Wizzup I don't know if its a problem exactly, but it took a few mins after setting up the connection to actually connect. | 22:45 |
freemangordon | uvos: yes, we discussed a bit | 22:45 |
uvos | Wizzup: yeah | 22:45 |
n900000 | and thanks for working on the N900 issue, I reverted back to the 6/27 image for the time being | 22:45 |
freemangordon | and I still miss the answer to the question - what will happen if you press volume keys on the BT headset? | 22:46 |
Wizzup | freemangordon: can you test hidden ap stuff tonight or tomorrow for me? | 22:46 |
sicelo | Wizzup: what did you find (n900 wifi) | 22:46 |
freemangordon | actually parazyd tested on android and the rezult was - nothing happened | 22:46 |
Wizzup | I forgot why I switched to scanning using wpa control interface rather than dbus | 22:46 |
freemangordon | Wizzup: sorry, no, too late here already | 22:46 |
uvos | freemangordon: sure thats a stiky situation, we could have the volume applet only react to the dbus singals | 22:47 |
parazyd | Yeah | 22:47 |
uvos | and then since we want the dbus signals to be sent by mce | 22:47 |
uvos | we can have mce filter out bt devices | 22:47 |
parazyd | I can do further tests if needed | 22:47 |
freemangordon | uvos: but really, lets not discuss now | 22:47 |
Wizzup | parazyd: hidden ap ones? | 22:47 |
uvos | freemangordon: sure | 22:47 |
freemangordon | volume ones :) | 22:47 |
parazyd | But on my.bt headaset nothing happened | 22:47 |
parazyd | Wizzup: bt headphones | 22:47 |
Wizzup | ok | 22:48 |
Wizzup | well I'm close to ready to just stop working on hidden ap and at least restore normal func | 22:48 |
freemangordon | why is that? | 22:48 |
Wizzup | the amount to which this broke everything is just insane | 22:48 |
freemangordon | what is "this"? | 22:48 |
Wizzup | working on the hidden ap stuff | 22:48 |
freemangordon | wpa_supplicant? | 22:48 |
Wizzup | that's one of it yeah | 22:48 |
n900000 | sorry dumb question, on the n900 is there a key combo for Tab? Or am I forced to use the onscreen one in osso-xterm? | 22:49 |
Wizzup | sicelo: plain and simplem the wpa control interface 'scan' never works on the n900 | 22:49 |
uvos | we should fix that really | 22:49 |
Wizzup | sicelo: it does something different from the dbus scan interface | 22:49 |
uvos | Wizzup: did you mail the list? | 22:49 |
freemangordon | Wizzup: did you try wext driver? | 22:49 |
Wizzup | freemangordon: yes, that does not work at all | 22:49 |
Wizzup | uvos: wpa_supplicant? no | 22:49 |
freemangordon | hmm, it used to | 22:49 |
uvos | yeah we should ask for advice i think | 22:50 |
Wizzup | we don't even have wext enabled | 22:50 |
Wizzup | (in kernel) | 22:50 |
freemangordon | anyway, I am too sleepy now to produce anything sane | 22:50 |
freemangordon | so... good night guys | 22:50 |
uvos | gn8 | 22:50 |
n900000 | also sxiv works great on leste, even full screening. keyboard mapping and be remapped to work on the hardware keyboards too. I like it better than mihphoto | 22:50 |
n900000 | What do you guys use for web browsing? | 22:51 |
uvos | firefox-esr | 22:51 |
uvos | works really well on d4 | 22:51 |
uvos | (not on n900 ofc) | 22:51 |
Wizzup | parazyd: shall we enable WEXT in kernel? | 22:51 |
n900000 | how about the poor n900 with it's 256mb of ram lol | 22:51 |
n900000 | the freemantle browser is actually really fast on simple webpages | 22:52 |
uvos | your kinda sol on n900 relly surf at least starts but its too mutch for it | 22:52 |
parazyd | Wi | 22:53 |
uvos | maybe you would have luck with link2 or dillo or something | 22:53 |
parazyd | Wizzup: I lf it's needed, sure | 22:53 |
parazyd | wft keyboard, sorry | 22:53 |
sicelo | Wizzup: i guess if you have nl80211 working with dbus, and that's what Leste uses, then no need for WEXT | 22:53 |
n900000 | yeah I expected as much, all I really need is newsboat and it works fine though | 22:53 |
Wizzup | sicelo: until everything works (e.g. hidden aps), we can't say it works | 22:53 |
Wizzup | wpa_supplicant acts differently over dbus than it does over control interface | 22:53 |
n900000 | sorry more dumb questions, is video playing hardware accelerated on any of these? | 22:54 |
uvos | no | 22:54 |
sicelo | okay. wext is definitely far more reliable - actually 5.14 has wext enabled for the n900 | 22:54 |
uvos | hardware decoding works nowhere (pp maybe?) | 22:54 |
uvos | hardware presentation sorta works | 22:54 |
uvos | on d4 i have gotten it to work only in firefox | 22:54 |
n900000 | pinephone has that new gstreamer patch | 22:54 |
n900000 | hopefully we can have an mpv build that accelerated on the pp | 22:55 |
n900000 | For music I'm using MOC on the n900 | 22:55 |
uvos | anyhow hardware prestation performance is broken by the pvr setup being broken (x clients (no the comp manager) are copying frames on cpu) | 22:55 |
n900000 | with alsamixer to mute the external speaker | 22:55 |
n900000 | ahh ok, sounds like a ton of work to even attempt doing | 22:56 |
Wizzup | sicelo: I wonder why we have it disabled then | 22:56 |
Wizzup | parazyd: can you add WEXT for a -devel kernel for n900? | 22:56 |
Wizzup | maybe look at the defconfig sicelo mentions | 22:56 |
sicelo | yes i saw. that's why i think if it worked with nl80211 + dbus, no need to change. nl80211 is the better option long term, of course | 22:57 |
Wizzup | yes, but hidden ap does not work due to the scan dbus scans work in wpa_supplicant | 22:58 |
Wizzup | due to the way dbus scans work* | 22:59 |
Wizzup | I don't know what to do here really | 22:59 |
uvos | i mean if it works for sicelo on sid with wpa_cli then we just need to figure out what the diff is between that and us no? | 23:00 |
Wizzup | could be wext | 23:00 |
uvos | sounds like its wext | 23:00 |
uvos | yeah | 23:00 |
Wizzup | yeah, but also wpa_supplicant being such a mess in some things | 23:00 |
Wizzup | parazyd: yes, I think WEXT is needed | 23:01 |
uvos | we could also check what changes if anny are in the wl1251 driver between 5.6 or what we have on n900 atm and 5.13 or whatever sicelo used with sid | 23:04 |
uvos | sicelo: ^^^ | 23:05 |
Wizzup | my memory is so hazy, I don't even recall why we wanted the newer wpa_supplicant | 23:09 |
Wizzup | good we have irc logs | 23:09 |
bencoh | because new is better, and red is faster! | 23:13 |
sicelo | :-) | 23:13 |
uvos | heh | 23:13 |
uvos | i think Wizzup was encountering a bug iirc | 23:14 |
Wizzup | yeah but I don't recall what | 23:51 |
Wizzup | sicelo: with wext, can you check if you can connect to hidden wlan? | 23:51 |
sicelo | i can try, but tomorrow. have to turn myself in now. didn't have success booting debian - i really don't know why recent kernels are hit-and-miss with booting. the same kernel can boot one moment, and not boot the next :-/ | 23:55 |
Wizzup | ok | 23:58 |
Wizzup | I'm building libicd-network-wpasupplicant 0.1.9 now | 23:58 |
Wizzup | let's hope it doesn't break wifi for folks in devel, but fixes it for n900 folks | 23:58 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!