00:58.02 | *** join/#openmoko-cdevel jluis (~jluis@111.Red-79-151-229.dynamicIP.rima-tde.net) |
00:59.02 | *** join/#openmoko-cdevel wolfspraul (~wolfsprau@mimi.q-ag.de) |
01:44.17 | *** join/#openmoko-cdevel floyd2 (~floyd@p508AA2C9.dip.t-dialin.net) |
02:10.17 | *** join/#openmoko-cdevel ferluszt (~peter@p5DD28B3F.dip0.t-ipconnect.de) |
02:49.07 | *** join/#openmoko-cdevel NuttyBunny (~cnegrete@187.152.208.138) |
04:32.16 | *** join/#openmoko-cdevel DocScrutinizer (~halley@openmoko/engineers/joerg) |
05:46.31 | *** join/#openmoko-cdevel wolfspraul (~wolfsprau@mimi.q-ag.de) |
06:48.22 | *** join/#openmoko-cdevel ThibG (~ThibG@81-64-13-28.rev.numericable.fr) |
06:59.04 | *** join/#openmoko-cdevel GarthPS (~quassel@qrc29-1-82-245-206-103.fbx.proxad.net) |
07:07.18 | GarthPS | moin |
07:19.17 | Lopi | moin |
07:19.24 | *** join/#openmoko-cdevel radekp (~radek@63.120.broadband10.iol.cz) |
07:22.19 | *** join/#openmoko-cdevel ThibG (~ThibG@81-64-13-28.rev.numericable.fr) |
07:36.27 | *** join/#openmoko-cdevel methril (~methril@189.27.135.133.dynamic.adsl.gvt.net.br) |
07:40.25 | *** join/#openmoko-cdevel ao2 (~ao2@2001:1418:117::1) |
07:44.32 | *** join/#openmoko-cdevel plotr (~name@m052d.studby.ntnu.no) |
08:01.04 | *** join/#openmoko-cdevel phh (~quassel@137.194.15.151) |
08:08.33 | *** join/#openmoko-cdevel morphis (~morphis@p5B3E6EC5.dip.t-dialin.net) |
08:20.49 | *** join/#openmoko-cdevel ThibG (~ThibG@81-64-13-28.rev.numericable.fr) |
08:46.30 | *** join/#openmoko-cdevel Heinervdm (~thomas@pD9E154B1.dip.t-dialin.net) |
09:16.48 | *** join/#openmoko-cdevel morphis (~morphis@p5489B868.dip.t-dialin.net) |
09:52.36 | *** join/#openmoko-cdevel NIN101 (~NIN101@p5DD2998F.dip0.t-ipconnect.de) |
10:13.03 | *** join/#openmoko-cdevel GNUtoo (~GNUtoo@host148-3-dynamic.244-95-r.retail.telecomitalia.it) |
10:13.38 | Lopi | grr I hate you baseband |
10:18.48 | GNUtoo | ask someone |
10:29.06 | *** join/#openmoko-cdevel dos1 (~dos@unaffiliated/dos1) |
10:38.22 | *** join/#openmoko-cdevel daniele_athome (~daniele@ppp-45-42.25-151.libero.it) |
10:45.47 | Lopi | GNUtoo: how can I test a compile of fsogsmd? |
10:48.08 | CIA-82 | freesmartphone.org: 03morphis 07aurora * r7937061e0eaf 10/aurora/systemmanager/controller.vala: aurora: handle backlight power status to deactivate touchscreen |
10:59.34 | CIA-82 | freesmartphone.org: 03morphis 07aurora * r1c4709a92089 10/aurora/ (configure.ac systemmanager/Makefile.am): aurora: even add libfsosystem as dependency |
11:09.02 | PaulFertser | tsting the bot |
11:09.07 | PaulFertser | s/tst/testing/ |
11:09.18 | PaulFertser | ~aptsnack |
11:09.47 | GNUtoo | Lopi, what do you mean? |
11:10.18 | Lopi | GNUtoo: ehh just running into a bunch of dependency errors whilst trying to compile |
11:10.58 | Lopi | http://pastie.org/private/92bgb2oheyq6dufrsmaw |
11:11.37 | Lopi | I understand what's wrong |
11:11.43 | GNUtoo | use bitbake |
11:11.45 | GNUtoo | and autorev |
11:11.56 | GNUtoo | and then |
11:12.04 | GNUtoo | you can use the devshell to compile it |
11:12.12 | Lopi | yeah, but how can I test if my plugin compiles with oe? |
11:12.19 | Lopi | I mean where can I put it |
11:12.23 | GNUtoo | I'll precise what I said |
11:12.31 | GNUtoo | bitbake fsogsmd (with autorev set ) |
11:12.35 | GNUtoo | and then |
11:12.44 | GNUtoo | source ~/path/to/devshell |
11:12.52 | GNUtoo | cd into your fsogsmd git dir |
11:12.57 | GNUtoo | and compile there |
11:13.10 | Lopi | how do you set autorev |
11:13.19 | GNUtoo | in local.conf add that: |
11:13.59 | Lopi | AUTOREV="yes" |
11:13.59 | Lopi | ? |
11:14.04 | GNUtoo | wait a sec |
11:14.27 | GNUtoo | require conf/distro/include/fso-autorev.inc |
11:14.46 | GNUtoo | if you want you can also add that: |
11:14.53 | GNUtoo | require conf/distro/include/shr-autorev.inc |
11:15.05 | Lopi | have them both already |
11:15.18 | GNUtoo | ok |
11:15.40 | Lopi | does fsogsmd end up as a package? |
11:15.50 | GNUtoo | which one? |
11:16.00 | GNUtoo | the one you compile with the devshell? |
11:16.02 | GNUtoo | no |
11:16.12 | GNUtoo | but it creates plugins for you |
11:16.14 | GNUtoo | for instance |
11:16.54 | leviathan | GNUtoo: did the phone arrive? |
11:17.00 | GNUtoo | not yet |
11:17.07 | leviathan | well |
11:17.20 | leviathan | I can go and track it next week |
11:17.27 | leviathan | if it doesn't arrive monday |
11:17.32 | GNUtoo | fsoaudiod/src/plugins/gsmvoice_alsa_cmtspeechdata/.libs/ |
11:17.36 | leviathan | I've sent it with high priority |
11:17.38 | GNUtoo | note the .libs |
11:17.40 | GNUtoo | ok |
11:18.06 | GNUtoo | thanks |
11:18.18 | *** join/#openmoko-cdevel GNUtoo-N900 (~root@host148-3-dynamic.244-95-r.retail.telecomitalia.it) |
11:18.26 | Lopi | mm |
11:18.36 | Lopi | so if it compiles, how do I test it on my device? |
11:18.49 | GNUtoo-N900 | you scp it |
11:18.56 | GNUtoo-N900 | the .so |
11:19.09 | Lopi | ah oaky |
11:19.17 | GNUtoo-N900 | opkg files fsogsmd will tell where to scp |
11:19.20 | Lopi | to /usr/lib/cornucopia/modules/fsogsm/ |
11:19.42 | GNUtoo-N900 | yes,something like that |
11:22.30 | Lopi | http://pastie.org/private/sifvdbwfjtsuu5qjelvezq |
11:25.12 | GNUtoo-N900 | sorry. I'll go |
11:31.07 | GarthPS | mickey|daddy: hi. can I ask you a confirmation , to see If i am stupid or what.., http://git.freesmartphone.org/?p=cornucopia.git;a=blob;f=fsodeviced/src/plugins/palmpre_quirks/powercontrol.vala;h=0e620ad40ef29637a4e7817238e6133c7c721dca;hb=HEAD line 77 is out of the previous else part ? |
11:32.07 | GarthPS | as this it will fail to bring the interface up if it actually fail to load the module right ? |
11:36.25 | mickey|daddy | garthps: morphis wrote that part, I'm not sure whether loading the interface is mandatory - it could be already loaded, so i guess we should try to bring up the interface no matter whether loading the module failed or succeede |
11:36.27 | mickey|daddy | d |
11:40.34 | GarthPS | mickey|daddy: hm yeah.. what will sirloin_wifi_mod.load() return is the modul is already loaded ? |
11:42.08 | GarthPS | then the is_active variable handling is unsufficient as it is never tested whether it is already active or not |
11:43.17 | GarthPS | plus then is it tryied to load the interface after failing to load the module then we have a crash |
11:43.35 | GarthPS | [CRITICAL] fsodeviced : GLib <fsodevice.palmpre_quirks>: file powercontrol.c: line 267: uncaught error: Could not bring interface eth0 up: No such device (fso_framework_network_error-quark, 0) |
11:44.09 | GarthPS | org.freedesktop.DBus.Error.NoReply: Message did not receive a reply (timeout by message bus) |
11:44.16 | GarthPS | sorry |
11:44.19 | GarthPS | root@palmpre2 ~ # mdbus2 -s org.freesmartphone.odeviced /org/freesmartphone/Device/PowerControl/0 org.freesmartphone.Device.PowerControl.SetPower true |
11:44.19 | GarthPS | org.freedesktop.DBus.Error.NoReply: Message did not receive a reply (timeout by message bus) |
11:44.31 | GarthPS | root@palmpre2 ~ # mdbus2 -s org.freesmartphone.odeviced /org/freesmartphone/Device/PowerControl/0 org.freesmartphone.Device.PowerControl.GetPower |
11:44.31 | GarthPS | org.freedesktop.DBus.Error.UnknownMethod: Method "GetPower" with signature "" on interface "org.freesmartphone.Device.PowerControl" doesn't exist |
12:00.08 | GarthPS | So for now I really think th ifup thing should be in the else part |
12:02.09 | *** join/#openmoko-cdevel fsteinel (~fsteinel@fedora/fsteinel) |
12:05.06 | *** join/#openmoko-cdevel oln (~oln@ws43177.studby.hig.no) |
12:33.15 | *** join/#openmoko-cdevel wolfspraul (~wolfsprau@mimi.q-ag.de) |
12:39.53 | *** join/#openmoko-cdevel valos (~valos@mar92-6-82-227-150-198.fbx.proxad.net) |
12:54.47 | mickey|daddy | GarthPS: yep. lets bring it up to morphis next time he's here |
12:58.43 | *** join/#openmoko-cdevel playya__ (~playya@unaffiliated/playya) |
13:01.22 | GarthPS | mickey|daddy: okey |
14:01.01 | *** join/#openmoko-cdevel ferluszt (~peter@p5DD2961D.dip0.t-ipconnect.de) |
14:38.49 | *** join/#openmoko-cdevel GNUtoo (~GNUtoo@host148-3-dynamic.244-95-r.retail.telecomitalia.it) |
14:48.47 | *** join/#openmoko-cdevel jluis (~jluis@2001:5c0:1400:a::87b) |
15:02.33 | *** join/#openmoko-cdevel fsteinel (~fsteinel@fedora/fsteinel) |
15:04.02 | *** join/#openmoko-cdevel alexxy (~alexxy@gentoo/developer/alexxy) |
15:21.15 | *** join/#openmoko-cdevel xrmx (~rm@151.69.107.74) |
15:24.05 | *** join/#openmoko-cdevel alexxy (~alexxy@gentoo/developer/alexxy) |
15:37.15 | CIA-82 | freesmartphone.org: 03morphis 07aurora * raf98757fecab 10/aurora/systemmanager/main.vala: aurora: don't drop priviledges in system manager for now |
15:41.40 | *** join/#openmoko-cdevel angelox (~angelo@189-18-205-179.dsl.telesp.net.br) |
16:06.14 | CIA-82 | freesmartphone.org: 03morphis 07aurora * r56f2c18efa89 10/aurora/systemmanager/controller.vala: aurora: handle backlight power update before the backlight is set |
16:34.35 | CIA-82 | freesmartphone.org: 03morphis 07aurora * r13ed1762ccac 10/aurora/systemmanager/controller.vala: aurora: handle power and audio mute keys in system manager |
16:37.02 | CIA-82 | freesmartphone.org: 03morphis 07aurora * r5bb5b2d8ea21 10/aurora/systemmanager/controller.vala: aurora: fix syntax error from commit before |
16:53.12 | *** join/#openmoko-cdevel Alex[sp3dev] (~alexander@86.110.163.19) |
17:00.38 | *** join/#openmoko-cdevel wiedi (~wiedi@smoot.w.fruky.net) |
17:05.50 | *** join/#openmoko-cdevel angelox|n900 (~user@201-95-128-37.dsl.telesp.net.br) |
17:05.54 | *** join/#openmoko-cdevel jluis (~jluis@2001:5c0:1400:a::77) |
17:17.08 | *** join/#openmoko-cdevel JesusMcCloud-lt (~JesusMcCl@84.119.83.234) |
17:18.26 | *** join/#openmoko-cdevel alexxy (~alexxy@gentoo/developer/alexxy) |
17:21.21 | *** join/#openmoko-cdevel MadTBone (~bruce@cpe-66-65-115-115.nyc.res.rr.com) |
17:30.39 | *** join/#openmoko-cdevel alexxy (~alexxy@gentoo/developer/alexxy) |
17:46.54 | *** join/#openmoko-cdevel angelox|n900 (~user@189-18-107-84.dsl.telesp.net.br) |
18:01.15 | *** join/#openmoko-cdevel GNUtoo (~GNUtoo@host148-3-dynamic.244-95-r.retail.telecomitalia.it) |
18:08.31 | *** join/#openmoko-cdevel gena2x_ (~gena@cpe-67-250-70-77.nj.res.rr.com) |
18:47.25 | *** join/#openmoko-cdevel max_posedon (~max_posed@178.124.160.59) |
19:01.09 | *** join/#openmoko-cdevel _ao2 (~ao2@2001:1418:117::1) |
19:07.10 | *** join/#openmoko-cdevel ferluszt (~peter@p5DD2961D.dip0.t-ipconnect.de) |
19:29.16 | *** join/#openmoko-cdevel dos1 (~dos@unaffiliated/dos1) |
19:57.38 | *** join/#openmoko-cdevel angelox|n900 (~user@201-13-208-30.dial-up.telesp.net.br) |
20:12.15 | *** join/#openmoko-cdevel GarthPS (~quassel@qrc29-1-82-245-206-103.fbx.proxad.net) |
20:14.33 | *** join/#openmoko-cdevel angelox (~angelo@201-13-208-30.dial-up.telesp.net.br) |
20:26.22 | angelox | GNUtoo: Hi,i'll try install u-boot,but i'm following a tutorial on "wiki.meego.com". That's secure? it shows me how to install u-boot from the phone... |
20:27.37 | GNUtoo | yes it is |
20:27.41 | GNUtoo | but it may not work |
20:27.48 | GNUtoo | basically it's as secure as flashing a kernel |
20:27.56 | GNUtoo | since uboot is flashed at the kernel place |
20:28.01 | GNUtoo | it's a chainloaded bootloader |
20:28.07 | GNUtoo | the original bootloader is still there |
20:28.14 | GNUtoo | which can be coldflashed btw |
20:28.45 | angelox | but why it may not work? |
20:33.02 | Alex[sp3dev] | GNUtoo: which kernel do you have on n900? |
20:33.59 | GNUtoo | the meego one |
20:34.12 | GNUtoo | the n900-adaptation one |
20:34.16 | GNUtoo | we update it often |
20:34.55 | Alex[sp3dev] | GNUtoo: when you have time, enable wl1251 debugging and make a log of successful association and pings from device to router and from pc to device |
20:35.16 | GNUtoo | I've something to do with Lopi right now |
20:35.21 | GNUtoo | but I'll do it right after |
20:35.24 | GNUtoo | if I don't go to bed |
20:50.05 | DocScrutinizer | the meego kernel is *cough* erhm... not optimized for N900. There are quite some bits missing of what genuine maemo kernel had for patches on top of mainline 2.6.28 that Nokia introduced to make for a proper power economic system that really works |
20:50.19 | GNUtoo | complain to mickey|daddy |
20:50.29 | GNUtoo | I fighted a lot for keeping 2.6.28 |
20:52.22 | DocScrutinizer | well, there's nothing bad really in going for [recent], but you need to take care about such things like sensor drivers, CPUfreq governor foo, whatnot - all the stuff that's been in maemo kernel and didn't make its way upstream for one reason or the other. Simply assuming meego folks will do it right is a bit over-optimistic |
20:53.15 | DocScrutinizer | basically my notion about meego-arm (N900) folks and their approach to kernel things: they got NFC |
20:53.40 | GNUtoo | mickey|daddy's reason was devtmpfs that was lacking on 2.6.28 |
20:54.00 | DocScrutinizer | they even deliberately admit they have no clue and go mere mainstream kernel, no matter what crap is or isn't there |
20:55.58 | DocScrutinizer | prominent example: LIS302 drivers - OM/OE has a proper driver, maemo has a similar proper driver, meego is using a *JOYSTICK* driver that happens to match regex >>.*lis.*3.*02.*<< and is available upstream |
20:56.18 | DocScrutinizer | this driver _evidently_ is fubar and bullshit |
20:56.48 | Alex[sp3dev] | why don't you port the needed stuff and mainline it? unlike other devices n900 is already well supported. you have to just do the boring integration work |
20:56.56 | DocScrutinizer | same pattern of operation applies to several other topics in meego I'd think |
20:56.58 | GNUtoo | Lopi, it just compiled |
20:57.24 | Lopi | GNUtoo: ohh nice ^^ |
20:57.35 | DocScrutinizer | Alex[sp3dev]: the problem is we can't get a proper lis302 driver upstream as there's that crap already |
20:58.07 | GNUtoo | http://gnutoo.homelinux.org/downloads/people/Lopi/plugin.vala |
20:58.08 | DocScrutinizer | getting good driver upstream would mean to nuke the shit, and nobody is going to fight that battle |
20:58.15 | GNUtoo | wget it and diff it with yours |
20:58.19 | GNUtoo | I commented a part tough |
20:58.24 | GNUtoo | because the function was not present |
20:58.34 | Alex[sp3dev] | DocScrutinizer: i see. so sitting and whining is better |
20:58.36 | GNUtoo | but it's pretty small |
20:58.39 | GNUtoo | it's only a check |
20:58.46 | DocScrutinizer | Alex[sp3dev]: you're free to do that |
20:58.52 | GNUtoo | I also surely messed up the identation |
20:58.59 | GNUtoo | also I didn't check pointers yet |
20:59.00 | DocScrutinizer | (whining, fighting, whatever) |
20:59.14 | GNUtoo | do you also want the binary? |
20:59.40 | Alex[sp3dev] | DocScrutinizer: sure. i just don't have n900.. but at least you can keep your own tree or integrate proper stuff to meego if mainline is a no-go |
20:59.48 | DocScrutinizer | for SHR however I strongly suggest to keep a local branch with platform adaptions that are local |
21:00.07 | DocScrutinizer | Alex[sp3dev]: exactly my suggestion |
21:00.25 | Lopi | GNUtoo: yeah give me the binary |
21:00.33 | Lopi | GNUtoo: I see what you changed, very minor |
21:00.42 | DocScrutinizer | meego has a policy of "mainline ONLY" so I don't care |
21:01.07 | GNUtoo | let me try to look of the & are correct |
21:01.11 | GNUtoo | and then I'll do |
21:01.16 | DocScrutinizer | Alex[sp3dev]: for SHR things are 'better' and I'm sure those smart guys here will eventually do "the right thing" |
21:01.21 | Lopi | okay sounds good ty |
21:02.15 | GNUtoo | ah my bad the & are correct |
21:02.28 | DocScrutinizer | embedded is always about fine tuning kernel stuff to the particular platforn, so in my book meego's "mainline only" approach never will work out |
21:03.37 | Alex[sp3dev] | DocScrutinizer: the thing is that no one likes to integrate stuff. because it is tiring and takes a lot of time. and if you take a commercial company, they'll use any opportunity to reduce development costs |
21:03.54 | GNUtoo | http://gnutoo.homelinux.org/downloads/people/Lopi/modem_singleline.so |
21:04.05 | GNUtoo | note that it may fail at runtime |
21:04.11 | GNUtoo | since you copy-pasted lot of C |
21:04.14 | DocScrutinizer | Alex[sp3dev]: sure |
21:04.17 | Lopi | okay |
21:04.18 | GNUtoo | I didn't extensively check everything |
21:04.25 | GNUtoo | I just tried to compile it |
21:04.27 | Lopi | we'll see what happens |
21:04.30 | GNUtoo | so you may need gdb |
21:04.31 | Lopi | at least it's a start |
21:04.46 | GNUtoo | to run it |
21:04.51 | GNUtoo | add DEBUG in fsogsmd |
21:04.54 | GNUtoo | config |
21:05.01 | GNUtoo | then do: |
21:05.05 | DocScrutinizer | Alex[sp3dev]: otoh the lis302 joysticj driver seems got introduced to mainline by uninspired unexperienced guys that were simply lazy and unaware of what problems they create for linux at large |
21:05.08 | GNUtoo | killall fsogsmd;fsogsmd |
21:05.15 | GNUtoo | and make your config use singleline |
21:05.24 | GNUtoo | btw I made a porting guide on SHR wiki |
21:05.32 | Lopi | oh nice |
21:05.34 | Alex[sp3dev] | DocScrutinizer: i can say the same thing about everything done by android developers |
21:05.39 | GNUtoo | http://www.shr-project.org/trac/wiki/Device%20Porting%20Guide |
21:05.44 | GNUtoo | it's a beginning |
21:05.47 | GNUtoo | we must add to it |
21:06.20 | Lopi | yeah, it's definitely needed tho |
21:06.34 | Lopi | that's the biggest problem with SHR |
21:06.41 | Lopi | our docs blow |
21:06.41 | Alex[sp3dev] | DocScrutinizer: anyhoo i had a look at the kernel source of asus transformer. they ignored pdata and decided to change gpio-keys, sdio.c and other files directly. so you're lucky with n900 |
21:06.43 | DocScrutinizer | Alex[sp3dev]: for SHR it basically doesn't mater *why* upstream aka meego is broken, it just needs to take the consequences and have a branch of local patches that "fix" those things |
21:07.27 | Alex[sp3dev] | DocScrutinizer: i think we both agree that code written by hardware vendors sucks and i don't want to continue this discussion |
21:08.08 | *** join/#openmoko-cdevel fsteinel (~fsteinel@fedora/fsteinel) |
21:08.09 | GNUtoo | ouch ignoring pdata???? |
21:08.17 | DocScrutinizer | you're free to stop answering to my comments directed to GNUtoo |
21:08.34 | GNUtoo | read it now |
21:09.10 | GNUtoo | and look at meego camera |
21:09.13 | *** part/#openmoko-cdevel fsteinel (~fsteinel@fedora/fsteinel) |
21:09.15 | GNUtoo | it's not so better |
21:09.20 | GNUtoo | the image quality...hmmm |
21:09.51 | Alex[sp3dev] | DocScrutinizer: ok, sorry for that |
21:11.45 | GNUtoo | so now wl1251_sdio |
21:12.02 | DocScrutinizer | GNUtoo: I think taking 2.6.28 maemo kernel and applying all the patches on top that lift it to 2.6.3x is the correct way to go. The other option was to take a plain 2.6.3x mainline and apply Nokia's patches there, but that may suck |
21:12.37 | GNUtoo | the "only" problem is that things change at a rate that is enormous |
21:12.41 | Alex[sp3dev] | GNUtoo: yeah, just make a dmesg of n900 wifi. i think it may be huge but i really need the last part of log about events received directly during pings |
21:12.47 | GNUtoo | s/enormous/very very very fast |
21:12.51 | DocScrutinizer | as you don't get the point what a particular nokia patch is meant to do, once it throws errors when applying it to 2.6.3x |
21:13.53 | GNUtoo | Alex[sp3dev], n900 is wl1251_spi |
21:14.16 | GNUtoo | and cleaning 2.6.28 takes a lot of time |
21:14.25 | GNUtoo | *2.6.28 patches |
21:14.42 | DocScrutinizer | GNUtoo: you're aware there's been quite some problems around wl1251 in maemo, and Nokia did some 'tricks' to make it behave? |
21:14.50 | Alex[sp3dev] | GNUtoo: i know it. i just want the logs of the events because i think tx part may be broken although i have no idea why it may work for n900 but not for our devices |
21:14.55 | GNUtoo | and I have huge issues with nervousness right now...... |
21:15.09 | GNUtoo | DocScrutinizer, uhhh no |
21:15.28 | GNUtoo | Alex[sp3dev], I'll compile for dream |
21:16.03 | GNUtoo | in some minutes |
21:16.11 | Alex[sp3dev] | GNUtoo: well, if it works there, then ok. but my device is almost the same like dream so i think you can face the same problems unless you use older code |
21:16.28 | DocScrutinizer | GNUtoo: one issue I know of is the chip's eating power like mad if not associated, even when "switched off". You need to unload the driver aka kernel module |
21:17.29 | DocScrutinizer | there have been other wicked screwed problems iirc |
21:18.03 | Alex[sp3dev] | GNUtoo: btw with the latest mainline wl1251 i can only scan with a dedicated irq. it won't work without it. |
21:18.41 | DocScrutinizer | GNUtoo: then there's been that mysterious chan6-only issue |
21:19.20 | DocScrutinizer | I never wraped my head around it |
21:19.32 | DocScrutinizer | just saying "watch out!" |
21:22.41 | DocScrutinizer | some really smart guy RE'd and improved the driver and "published" a way better implementation that also supports injection, monitoring and whatnot, but somehow breaks some API bits a little |
21:23.24 | DocScrutinizer | he asked for a donation of >=0ct for shipping the source, which is fine but caused lots of flames |
21:23.56 | Alex[sp3dev] | DocScrutinizer: unfortunately for me and GNUtoo on msm phones wl1251 doesn't work at all with power saving and doesn't always work on my device even without it |
21:24.31 | DocScrutinizer | Alex[sp3dev]: I can't comment on msm phones, never had a look at thiose |
21:25.12 | Alex[sp3dev] | DocScrutinizer: they're using sdio inteface, not spi. so maybe all the problems come from the way irqs are handled |
21:33.58 | DocScrutinizer | Alex[sp3dev]: quite possible. Ask wpwrak about it, he did "a master thesis" about the very topic |
21:34.48 | DocScrutinizer | on ar6000 though, but in the end it's been spi vs sdio IIRC |
21:35.14 | Alex[sp3dev] | ok, will do later. now i need to prepare for the exams. and after exams i'll need to analyze GNUtoo's dump first |
21:35.21 | DocScrutinizer | wpwrak: any comments? ^^^ |
21:37.40 | DocScrutinizer | I hope msm is using hardmac? as with softmac you're running into trouble you never get behind you, on embedded |
21:38.13 | Lopi | GNUtoo: I'm done with this for the day, but I made a lot of progress today |
21:38.16 | DocScrutinizer | (excuse my ignorance if wl1251 maybe has no hard- or softMAC |
21:38.18 | DocScrutinizer | ) |
21:40.54 | GNUtoo | DocScrutinizer, I think it's a softmac |
21:41.01 | GNUtoo | but it has some power saving features |
21:41.04 | GNUtoo | and it's pretty good |
21:41.07 | DocScrutinizer | OUCH |
21:41.18 | GNUtoo | it's an exception |
21:41.59 | DocScrutinizer | you're aware what softmac is. So you know your system standby will suck on WLAN with softmac |
21:42.47 | GNUtoo | ok |
21:42.54 | DocScrutinizer | waking main CPU for each random WLAN inbound pkg is for sure not helpful to get things working and saving power |
21:43.12 | GNUtoo | it has PSM tough |
21:43.21 | GNUtoo | but on msm I never got it working |
21:43.23 | DocScrutinizer | s/inbound pkg/random signal/ |
21:43.35 | GNUtoo | altough with the android driver it works |
21:43.45 | angelox | GNUtoo: do you know some link of a rootfs(and a kernel) to i test SHR quickly here on my N900? |
21:43.50 | Alex[sp3dev] | whatever, but the whole system with backlight, gsm and stuff takes the same amount of power as wl1251 alone without psm |
21:44.14 | GNUtoo | angelox, hi, go on SHR download page, download an n900 rootfs extract it on microsd |
21:44.24 | GNUtoo | angelox, did you flash uboot yet |
21:44.25 | GNUtoo | ? |
21:45.02 | angelox | i used that tutorial from meego,working uboot here already. |
21:45.13 | GNUtoo | ok |
21:45.16 | GNUtoo | is it flashed? |
21:45.32 | DocScrutinizer | Alex[sp3dev]: yes, PSM is about suspending the RX for arbitrary durations as advised by AP, and wake it up just in time via a hw alarm clock. This probably can't be done in softmac at all |
21:45.55 | GNUtoo | wl1251_sdio has psm |
21:46.00 | GNUtoo | it works on the pandora |
21:46.09 | DocScrutinizer | Alex[sp3dev]: and wl1251 is known to suck power for RX |
21:46.10 | GNUtoo | it also works on the ti driver |
21:46.20 | Alex[sp3dev] | GNUtoo: it even works in winmo |
21:46.55 | angelox | GNUtoo: i really don't know,how it makes uboot work here.i don't know if it self-flashed by installing the uboot-pr13(i don't think so) |
21:47.17 | GNUtoo | angelox, what do you get at boot? |
21:47.22 | DocScrutinizer | an additional problem is al lot of AP don't implement PSM correctly, so your PSM may fail despite you did everything 100% correct |
21:47.36 | GNUtoo | yes |
21:47.42 | GNUtoo | but I have hostapd |
21:48.13 | GNUtoo | and my laptop has that: |
21:48.15 | GNUtoo | <PROTECTED> |
21:48.23 | DocScrutinizer | hard to tackle that whole topic, without $$$k equipment to analyze OTA |
21:48.34 | GNUtoo | would an USRP help? |
21:48.38 | angelox | GNUtoo: i get uboot menu,and i can choose for a lot of options,there's also mmc boot option...so i think that it's working |
21:48.40 | DocScrutinizer | sure :-D |
21:49.13 | GNUtoo | because I suspect that the problem is not on the wireless transmit/receive, but rather on the kenrel code and with the IRQ handling |
21:49.21 | GNUtoo | angelox, ok |
21:49.35 | DocScrutinizer | GNUtoo: quite possible |
21:49.45 | GNUtoo | angelox, boot maemo with it for now |
21:49.50 | GNUtoo | angelox, give me some minutes |
21:49.52 | DocScrutinizer | for softmac to work you need "realtime!" |
21:50.01 | GNUtoo | I've to compile wl1251_sdio |
21:50.01 | angelox | GNUtoo: ok,thank you. |
21:50.15 | GNUtoo | I hope I don't go to bed before helping you |
21:50.18 | GNUtoo | but big warning |
21:50.24 | GNUtoo | don't stay too long in uboot |
21:50.26 | DocScrutinizer | as *all* is about timing in PSM |
21:50.29 | GNUtoo | else the battery discahrges |
21:50.41 | GNUtoo | and the battery charging is an userspace program |
21:50.51 | GNUtoo | so you have to be able to boot to charge it usually |
21:50.55 | GNUtoo | the details is |
21:51.05 | GNUtoo | that the bootloader has an emergency charge |
21:51.12 | GNUtoo | (not uboot but the one under it) |
21:51.20 | GNUtoo | so it can charge it enough just to make you boot |
21:51.28 | DocScrutinizer | not really |
21:51.36 | DocScrutinizer | but doesn't matter ;-) |
21:51.44 | Alex[sp3dev] | DocScrutinizer: which pretty much explains why it only works with a dedicated irq. sdio interrupts are too slow for it |
21:52.05 | DocScrutinizer | Alex[sp3dev]: yes, sounds like a story |
21:52.29 | GNUtoo | DocScrutinizer, what are the real details? |
21:52.34 | angelox | GNUtoo: hmm ok,thank you,i booted maemo by writing on uboot "run noloboot". |
21:52.46 | DocScrutinizer | Alex[sp3dev]: depending on your kernel you also got IRQ servicing latencies... |
21:53.04 | DocScrutinizer | GNUtoo: actually the chip does the emergency charging |
21:53.09 | GNUtoo | ah ok |
21:53.17 | GNUtoo | but the bootloader triggers it |
21:53.19 | GNUtoo | ? |
21:53.22 | DocScrutinizer | nope |
21:53.23 | GNUtoo | like trough i2c |
21:53.25 | GNUtoo | ah ok |
21:53.45 | DocScrutinizer | GNUtoo: look, if BL got loaded, the *hw* already 'booted' |
21:54.18 | DocScrutinizer | loading, BL, loading kernel. No diff for the hw |
21:55.04 | angelox | liked his N900 and don't want to brick it :) |
21:55.06 | DocScrutinizer | but the NOLO BL decides to immediately boot to kernel for further charging, and not allow flashing, if battery is too low |
21:55.18 | DocScrutinizer | angelox: you can't brick N900 |
21:55.38 | DocScrutinizer | well, it's at least incredibly hard to do |
21:55.39 | GNUtoo | angelox, indeed you can even reflash the bootloader trough usb |
21:55.42 | GNUtoo | I even did it |
21:55.46 | *** join/#openmoko-cdevel NIN101 (~NIN101@p5DD2998F.dip0.t-ipconnect.de) |
21:56.05 | wpwrak | DocScrutinizer: sdio interrupts eating power ? in general, they shouldn't. some mmc/sd/sdio host drivers use polling instead of real interrupts, so that would eat a bit more power. |
21:56.13 | DocScrutinizer | worst thing to do is nuke CAL with all the device specific data like serial# etc |
21:56.48 | DocScrutinizer | wpwrak: softmac eating power, compared to hardmac. That's my point |
21:57.26 | GNUtoo | /dev/mtd1 then |
21:57.27 | DocScrutinizer | also softmac is a bitch do get right, regarding strict timing of PSM |
21:58.01 | DocScrutinizer | GNUtoo: indeed. No reports on how to recover from that |
21:58.16 | DocScrutinizer | maybe it's irrelevant |
21:58.19 | DocScrutinizer | maybe not |
21:58.26 | GNUtoo | maybe reflashing a PR 1.3? |
21:58.31 | DocScrutinizer | nobody tested it, for obvious reasons ;-P |
21:58.51 | GNUtoo | lol |
21:58.52 | DocScrutinizer | flashing doesn't usually touch CAL |
21:58.57 | wpwrak | DocScrutinizer: you'd notice the biggest difference if you do things like wake-on-wlan. because softmac needs the cpu for that. so if the cpu isn't good at dealing with many short wakeups, then you have a problem. |
21:58.58 | angelox | if i flashed uboot,there's somehow to recover my warranty,or i didn't lose it yet? |
21:58.59 | GNUtoo | ah ok |
21:59.03 | DocScrutinizer | except for SW version etc |
21:59.17 | DocScrutinizer | angelox: irelevant |
21:59.24 | GNUtoo | wpwrak, there is a GPIO IRQ for PSM |
21:59.37 | GNUtoo | MSM_GPIO_TO_INT(29) if I remember well |
21:59.38 | DocScrutinizer | angelox: you don't lose warranty by flashing kernel |
22:00.13 | angelox | this is good... |
22:00.18 | DocScrutinizer | you have no warranty though on a working system with non-standard kernels :-P |
22:00.41 | DocScrutinizer | and flashing back to standard isn't covered by warranty |
22:01.28 | DocScrutinizer | so if you thoroughly screwed things, you better don't tell you messed with alterative kernels/OS ;-) |
22:01.53 | DocScrutinizer | otherwise they will charge you for restoring stock image |
22:02.39 | DocScrutinizer | though honestly you should be capable of doing that at home, if you dare to mess with alternatives ;-) |
22:02.59 | DocScrutinizer | again, you can't really brick N900 |
22:03.10 | GNUtoo | angelox, also beware with waranties, if you send it for repair you may get a N8 instead |
22:03.15 | GNUtoo | *get a N8 back |
22:03.23 | DocScrutinizer | just recovering gets harder still, if you mess up "better" |
22:03.54 | GNUtoo | coldflashing is not hard, but it require to use flasher, which is proprietary |
22:04.10 | DocScrutinizer | flasher is all you usually need to recover from whatever "bricking" might have happpened |
22:04.15 | GNUtoo | coldflash->flashing eMMC+NAND |
22:04.20 | GNUtoo | and you can recover |
22:04.27 | DocScrutinizer | yup |
22:05.09 | DocScrutinizer | CPU has a minimalistic USB bootloader that can't get nuked, so you can recover from all possible mess you might have caused |
22:05.31 | angelox | ok,i was afraid about brick it...thanks for explanation... about OS/Kernel/NAND/eMMC/Bootloader so we have no problems if "brick" it... |
22:05.37 | GNUtoo | another way to brick it is to overclock it |
22:05.40 | GNUtoo | don't overclock |
22:06.19 | DocScrutinizer | that's not "bricking" that's deliberately cooking hw to death |
22:06.40 | angelox | and if i brick some physical part of it? (for example,my touch stop working) |
22:07.10 | DocScrutinizer | you have several ways to do that, not all need sw tweaks: attach 50V~ charger for example |
22:08.14 | DocScrutinizer | angelox: if your ts stops working, the you either got a crappy driver and need to flash stock OS and kernel, or you killed the hw and nobody can help -> try to get warranty repair |
22:08.49 | DocScrutinizer | flasher won't glue broken hw parts |
22:09.00 | angelox | yes :) |
22:09.11 | DocScrutinizer | and while sw can't kill touchpanel, a nail or knife can |
22:09.29 | DocScrutinizer | please use common sense |
22:10.19 | GNUtoo | software can brick it that way: |
22:10.23 | GNUtoo | touchscreen not working |
22:10.26 | GNUtoo | because of software |
22:10.32 | GNUtoo | and user touching more hard |
22:10.35 | GNUtoo | *harder |
22:10.37 | angelox | i was talking about,if i flashed my phone and it starts with a lot of logs and i brick my ts and i send to warranty..what'll they say? that isn't on warranty by changed sw? |
22:10.39 | GNUtoo | which may damage it |
22:10.43 | GNUtoo | depending on the device |
22:11.17 | DocScrutinizer | angelox: nope |
22:11.32 | DocScrutinizer | angelox: first thing they do: reflash to known good stock OS |
22:11.44 | DocScrutinizer | angelox: then check if hw still acts up |
22:12.15 | DocScrutinizer | angelox: then if it's not evidently a OC caused CPU fry-out, they just will fix/replace |
22:12.52 | angelox | ah,ok,so i understood now,because i didn't know how they proceed to fix a phone problem |
22:14.13 | DocScrutinizer | angelox: and really, that's what you should do as well, prior to trying to go for warranty repair |
22:14.47 | DocScrutinizer | because if they find it's been a mere sw issue, they *might* decide to charge you for fixing your system sw |
22:15.29 | DocScrutinizer | and I'm all with them for ding so |
22:15.34 | DocScrutinizer | doing* |
22:16.12 | GNUtoo | Alex[sp3dev], now that I have my system setup |
22:16.16 | GNUtoo | what logs do you want? |
22:16.28 | GNUtoo | I can ping the router |
22:17.11 | DocScrutinizer | angelox: btw recently nokia repair sends out crappy N8 for replacement of nice maemo-based N900, as they don't have spare N900 anymore |
22:17.22 | GNUtoo | [ 339.542022] wl1251: ERROR elp wakeup timeout // when I iwconfig wlan0 power on |
22:17.26 | DocScrutinizer | :-/ |
22:17.31 | Alex[sp3dev] | GNUtoo: yeah, i think you can enable DEBUG_IRQ, DEBUG_MAILBOX, DEBUG_EVENT, DEBUG_TX, DEBUG_RX |
22:17.41 | GNUtoo | ok |
22:18.32 | Alex[sp3dev] | GNUtoo: then do it without ELP. i need the logs of it receiving events and doing tx. basically i think it fails with unaligned data and the code handling it might be inaccurate although it is 'fixed' by some patch already |
22:18.45 | angelox | DocScrutinizer: Hmm,i see,since i have bought my N900 a day before yesterday,they should have some at there :) (i hope,but i don't want brick it soon :D ) |
22:19.13 | Alex[sp3dev] | GNUtoo: so is the compat-wireless on gitorious the one you're using? |
22:19.13 | DocScrutinizer | your dealer may still have N900, but Nokia repair don't |
22:19.15 | GNUtoo | btw I used theses branches: |
22:19.44 | GNUtoo | master branch of git://gitorious.org/htc-msm-2-6-32/compat-wireless.git |
22:19.46 | angelox | DocScrutinizer: yes... |
22:20.06 | GNUtoo | remotes/htc-msm/android-msm-2.6.32-rebase+needs_compat_wireless+wifi-irq of the leviathan kernel |
22:21.01 | DocScrutinizer | (nokia repair don't) which is an incredible fail, as manufacturers are supposed to keep spare parts for years after EOL of product |
22:21.35 | angelox | as it is a great phone too |
22:21.52 | DocScrutinizer | and I really hope this has been a PEBKAC at service line of NOKIA, as otherwise the shall get sued to the ground |
22:22.54 | *** join/#openmoko-cdevel angelox|n900 (~user@201-13-208-30.dial-up.telesp.net.br) |
22:23.04 | GNUtoo | ELOP == EOLP(end of life products) |
22:23.09 | DocScrutinizer | you CAN NOT replace a linux device with a sybian based one, no matter how many more pixels the cam has |
22:24.03 | DocScrutinizer | hmm |
22:24.05 | DocScrutinizer | ~elop |
22:24.15 | DocScrutinizer | amazing |
22:24.30 | DocScrutinizer | ~tepco |
22:24.31 | apt | 'tepco' (1 of 2): sth especially stupid or messed up; "this construction is completely tepco"; "please stop behaving tepco" |
22:24.46 | angelox|n900 | dont know what do you think,but they also doesnt know that maemo is much more better than symbian one |
22:25.27 | Alex[sp3dev] | apt: ELOP is a microsoft spy and one of the most successfull corporate trolls |
22:25.28 | apt | okay, Alex[sp3dev] |
22:25.45 | DocScrutinizer | they switched to meego, then found this is a dead end adventure... |
22:25.50 | angelox | i think maemo 10x better than symbian for a phone/tablet device |
22:26.02 | DocScrutinizer | Alex[sp3dev]: thanks :-) |
22:27.09 | angelox | or meego too...great OS.. |
22:27.40 | GNUtoo | http://gnutoo.homelinux.org/downloads/people/Alex%5bsp3dev%5d/association.lzma |
22:27.47 | DocScrutinizer | well, as mentioned above, meego-arm still has not the maturity of maemo |
22:28.16 | Alex[sp3dev] | GNUtoo: thanks, downloaded. |
22:28.50 | GNUtoo | lzcat it |
22:28.59 | DocScrutinizer | curses space-bar of laptop started to grate |
22:29.02 | GNUtoo | basically I had connman running |
22:29.05 | Alex[sp3dev] | GNUtoo: luckily i know what lzma is |
22:29.06 | GNUtoo | so I killed it |
22:29.25 | GNUtoo | yes but I just discovered lzcat |
22:29.47 | *** join/#openmoko-cdevel valos (~valos@mar92-6-82-227-150-198.fbx.proxad.net) |
22:29.54 | GNUtoo | because there is bzcat zcat etc...so I tought maybe there is a lzcat |
22:30.09 | DocScrutinizer | *cough* |
22:30.31 | GNUtoo | so I was saying: |
22:30.41 | GNUtoo | I had connmand running and I killed it |
22:30.47 | GNUtoo | then I retried to associatae |
22:30.49 | GNUtoo | it worked |
22:31.25 | Alex[sp3dev] | ok, i'll reboot my pda into linux and try myself |
22:31.51 | GNUtoo | [ 388.391754] wlan0: associated |
22:31.54 | GNUtoo | anyway that should tell |
22:32.17 | Alex[sp3dev] | i can associate but my connection dies if i open browser or ping the device from another one |
22:32.31 | GNUtoo | is the log enough? |
22:34.37 | Alex[sp3dev] | looks like yes |
22:37.06 | DocScrutinizer | Alex[sp3dev]: sounds like dhcp problems |
22:37.30 | DocScrutinizer | or re-authentication problems |
22:37.42 | Alex[sp3dev] | DocScrutinizer: no it does not. i don't use dhcp at all. but i got the log so i can compare |
22:38.06 | DocScrutinizer | check wpa-supplicant, try open wlan |
22:38.27 | GNUtoo | connmand can block dhcp |
22:38.34 | GNUtoo | but I guess he got kernel issues |
22:38.44 | DocScrutinizer | hmm :shrug: |
22:38.49 | Alex[sp3dev] | maybe supplicant config needs to be tweaked though |
22:39.17 | GNUtoo | since I guess some stuff appeared in dmesg |
22:39.28 | Alex[sp3dev] | GNUtoo: i can upload my log if you want |
22:39.35 | GNUtoo | yes good idea |
22:39.37 | DocScrutinizer | when wpa-supplicant purges credetials and doesn't re-authenticate, it may look like "associated, but connection stalls/drops as soon as it gets used" |
22:40.09 | GNUtoo | Alex[sp3dev], I've open wifi btw |
22:40.14 | GNUtoo | not encrypted |
22:40.20 | DocScrutinizer | s/purges/keeps/ |
22:40.22 | Alex[sp3dev] | DocScrutinizer: very likely. if some pings are successful it is likely that supplicant is to blame |
22:40.36 | Alex[sp3dev] | GNUtoo: i have open wifi at the university and it didn't work there either |
22:40.42 | GNUtoo | ok |
22:41.47 | DocScrutinizer | hmmm |
22:41.53 | Alex[sp3dev] | GNUtoo: http://pastebin.com/CV9s4LTz |
22:42.01 | *** join/#openmoko-cdevel alexxy (~alexxy@gentoo/developer/alexxy) |
22:43.04 | DocScrutinizer | how are IP addresses associated to connection when there's no dhcp running/used? |
22:43.22 | Alex[sp3dev] | DocScrutinizer: well, actually i tried with both dhcp and manually setting address |
22:44.40 | DocScrutinizer | if your interface is missing proper address (in ifconfig list) and defalt route to AP, then your connection looks 'associated' but can't work |
22:45.02 | GNUtoo | the advantage the my compat wireless driver I put online is that one: |
22:45.04 | GNUtoo | Mode:Ad-Hoc |
22:45.28 | GNUtoo | so tethering,batmand and friends etc... could work |
22:45.58 | Alex[sp3dev] | DocScrutinizer: please, i'm not that naive. of course i have verified it. my situation is. i start wifi. get address via dhcp or manually. ping the router from the phone - works. ping the phone from the laptop - stops working. also stops working if i open browser and try any url, even router's address |
22:45.59 | DocScrutinizer | you're aware ad-hoc is incompatible with PSM |
22:46.06 | GNUtoo | yes of course |
22:46.23 | GNUtoo | but batmand is to call only for emergency cases..... |
22:46.52 | DocScrutinizer | Alex[sp3dev]: ok, bye then |
22:48.19 | Alex[sp3dev] | GNUtoo: could you please show the options in your supplicant config? |
22:48.35 | GNUtoo | none |
22:48.37 | GNUtoo | I do that |
22:48.39 | GNUtoo | *did |
22:48.42 | GNUtoo | ifconfig wlan0 up |
22:48.45 | GNUtoo | iwconfig wlan0 router |
22:48.48 | GNUtoo | udhcpc -i wlan0 |
22:49.16 | GNUtoo | I can make a new AP and try if you want |
22:49.20 | GNUtoo | but maybe not right now |
22:49.28 | *** join/#openmoko-cdevel alexxy (~alexxy@gentoo/developer/alexxy) |
22:49.31 | GNUtoo | I've added AP support for my bug devices |
22:49.42 | GNUtoo | (it was very easy, just rebase the patches) |
22:49.52 | GNUtoo | s/the/some/ |
22:52.15 | GNUtoo | Alex[sp3dev], btw in which country are you? |
22:52.25 | Alex[sp3dev] | GNUtoo: in Russia |
22:52.30 | GNUtoo | ah ok that's far |
22:52.40 | GNUtoo | because I've two htcdreams |
22:52.43 | Alex[sp3dev] | yeah, most likely i won't be able to use your AP |
22:53.20 | GNUtoo | it was rather for sending you an htcdream to help you, and you would fix the wifi on it while beeing at it |
22:53.40 | Alex[sp3dev] | that's a bribe |
22:53.50 | GNUtoo | something like that |
22:54.05 | GNUtoo | but I guess customs are problematic there |
22:54.51 | Alex[sp3dev] | well i guess if i fix wifi here it will work on dream. and i also wonder why ADSP doesn't respond when running from nand.. honestly don't want to disassemble the whole winmo sound driver |
22:55.15 | GNUtoo | ah? |
22:55.29 | GNUtoo | Maybe wrong nand partitioning(blind guess) |
22:55.33 | Alex[sp3dev] | GNUtoo: with LK bootloader we can now boot from nand on kovsky but sound won't work |
22:55.42 | GNUtoo | ok |
22:55.47 | GNUtoo | is there LK for htcdream? |
22:55.53 | Alex[sp3dev] | on rhodium [acl] even made a full nand rom and it works but they have different AMSS |
22:55.57 | GNUtoo | since LK has sources that interesses me |
22:56.11 | Alex[sp3dev] | GNUtoo: you can port but i don't think you need it, you have recovery and native bootloader |
22:56.13 | GNUtoo | ah ok...sigh with different AMSSes |
22:56.36 | GNUtoo | yes, but I guess soldering for the jtag is hard on that device |
22:56.44 | GNUtoo | so I won't port it |
22:56.54 | Alex[sp3dev] | why would you need jtag? jtag is cheating |
22:57.09 | GNUtoo | likes jtag, serial etc.... |
22:57.23 | Alex[sp3dev] | we just chain load it. we make a wrapper around LK that makes it look like wince image and is accepted by windows bootloader |
22:57.23 | GNUtoo | has some nervousness problems and having the right tools is so great |
22:57.32 | GNUtoo | ah ok |
22:58.13 | GNUtoo | tought of replacing totally fastboot with a free bootloader but: |
22:58.16 | GNUtoo | *that's risky |
22:58.28 | GNUtoo | *no one did it yet and I've no time to work on it |
22:58.42 | GNUtoo | * no way to get jtag easily |
22:58.58 | Alex[sp3dev] | well. it was done once for ipaq 4700 but it was done by paid developers with jtag who had documentation from hp. ahum |
22:59.09 | GNUtoo | ok |
23:05.14 | angelox | GNUtoo: it's possible run shr using bootmenu-n900 on n900 instead uboot? |
23:06.27 | GNUtoo | angelox, I guess you need to flash shr's uboot |
23:06.42 | GNUtoo | ah sorry |
23:06.45 | GNUtoo | I misunderstood |
23:06.48 | GNUtoo | I read too fast |
23:06.52 | GNUtoo | not currently |
23:06.57 | GNUtoo | long time ago it was possible |
23:07.06 | GNUtoo | we have two dual-boot solutions: |
23:07.11 | GNUtoo | * SHR's uboot |
23:07.28 | GNUtoo | I guess that one is broken currently(wrong /dev/mmcblk) |
23:07.43 | GNUtoo | if you find the right uboot for it it'll work |
23:07.46 | GNUtoo | mrmoku|italy, is not there |
23:07.53 | GNUtoo | he knew that stuff |
23:07.59 | GNUtoo | along with JaMa |
23:08.07 | GNUtoo | both are away for a long period |
23:08.09 | Alex[sp3dev] | GNUtoo: ok. i'm fairly sure wpa_supplicant is buggy now. and my power routine needs to be changed - does not survive and ifconfig down/up. although if i enable psm, writes and reads to/from elp are ok. if i ping from the laptop, i get events and i can scan for networks |
23:08.32 | GNUtoo | else I can give you something for kexecboot |
23:08.52 | GNUtoo | ok, wow |
23:09.09 | GNUtoo | where is your source for read/write elp? |
23:09.11 | Alex[sp3dev] | GNUtoo: does that look ok http://pastebin.com/A4tHzbCy? |
23:09.21 | angelox | GNUtoo: 'couse i wouldn't like ever time i boot the phone i need see uboot... |
23:09.23 | angelox | s/couse/cause/ |
23:09.42 | GNUtoo | Alex[sp3dev], at first sight that looks ok |
23:09.58 | GNUtoo | angelox, then there is kexecboot |
23:10.02 | GNUtoo | it can dual boot |
23:10.05 | GNUtoo | but it add one minute |
23:10.09 | GNUtoo | to the boot time |
23:10.17 | GNUtoo | basically you select SHR on the menu |
23:10.22 | GNUtoo | the screen becomes black |
23:10.30 | GNUtoo | during one minute |
23:10.35 | GNUtoo | and finally X start |
23:10.39 | angelox | even to boot maemo? |
23:10.45 | GNUtoo | no |
23:10.51 | GNUtoo | maemo is faster to boot |
23:10.52 | Alex[sp3dev] | GNUtoo: there we go https://gitorious.org/~ast/linux-on-qualcomm-s-msm/alex-linux-xperia/commit/61a60bba9cdb48fef326d713b2a94c988c23a568 |
23:10.58 | GNUtoo | but you need the power kenrel |
23:11.05 | GNUtoo | else you can't dual-boot it |
23:11.10 | GNUtoo | I've documented it here: |
23:11.28 | GNUtoo | http://www.shr-project.org/trac/wiki/Devices/NokiaN900/Booting |
23:11.50 | Alex[sp3dev] | GNUtoo: i stole that from 2.6.39 and hacked (added || card->sdio_funcs == 1) condition). that got me rid of -83 errors with CCCR_INTx |
23:12.09 | GNUtoo | ok nice |
23:12.23 | Alex[sp3dev] | GNUtoo: iirc that was your only problem, right? |
23:12.33 | angelox | GNUtoo: great,easy installation...let me try it |
23:13.16 | angelox | but..when i boot the phone after installed it,it'll show me something or go direct to maemo? |
23:13.48 | GNUtoo | Alex[sp3dev], yes |
23:14.03 | GNUtoo | angelox, it gives a menu |
23:14.08 | GNUtoo | a pretty menu |
23:14.15 | GNUtoo | but it add 1 minute to SHR boot time |
23:14.23 | GNUtoo | and it require you to follow a guide |
23:14.31 | GNUtoo | for preparing maemo for beeing kexec-booted |
23:15.04 | Alex[sp3dev] | GNUtoo: and this one https://gitorious.org/~ast/linux-on-qualcomm-s-msm/alex-linux-xperia/commit/e3634792a8d4dc1a6ece286f35c0ca4b94edf146 |
23:15.09 | Alex[sp3dev] | to actually enable that code |
23:15.33 | angelox | so only the shr-project.org/trac/wiki/Devices/NokiaN900/Booting how-to doesn't work only to boot maemo atm? |
23:16.40 | DocScrutinizer | GNUtoo: kexecboot, is that like kexec? and if yes, how comes everybody says kexec doesn't work on maemo? |
23:16.54 | GNUtoo | DocScrutinizer, it doesn't work with the normal kernel |
23:17.00 | GNUtoo | it works with the power kernel |
23:17.01 | DocScrutinizer | :nod: |
23:17.27 | DocScrutinizer | angelox: what's wrong with uboot? |
23:17.28 | GNUtoo | so it forces people to use the power kernel |
23:18.08 | DocScrutinizer | angelox: I think uboot is a clean nice solution, and adds only like max 5s to boot time |
23:18.27 | angelox | DocScrutinizer: i need ever when i'll boot maemo type "run noloboot" on it |
23:18.47 | DocScrutinizer | that's a definitely broken uboot config then |
23:19.10 | Alex[sp3dev] | GNUtoo: sorry for asking again but what kernel are you using? |
23:19.16 | DocScrutinizer | my uboot here boots maemo automatically after timeout of the ~3s to catch any keypress |
23:20.16 | angelox | my one too,but i got some *memory i guess* problems,since the gui goes crazy when i press 'power' button when i leave it boot maemo automatically |
23:20.28 | DocScrutinizer | the particularly fubar thing with uBoot n900 flavour is the missing env storage and edit methods |
23:21.21 | GNUtoo | Alex[sp3dev], leviathan-incommig |
23:21.24 | GNUtoo | from htc-msm |
23:21.31 | GNUtoo | I gave the branch not so long ago |
23:21.47 | DocScrutinizer | hmm, can't comment on that, as my setip here is PK-uboot with a massively tweaked PR1.2 based system |
23:22.08 | angelox | there's some uboot cfg file to i read it? |
23:22.12 | DocScrutinizer | never had any such problems though with it |
23:22.14 | GNUtoo | DocScrutinizer, there is the possibility to do some boot.cfg |
23:22.28 | DocScrutinizer | indeed? |
23:22.30 | DocScrutinizer | nice |
23:22.37 | angelox | never used u-boot since today |
23:22.42 | GNUtoo | not sure if it works well tough |
23:22.51 | GNUtoo | you may have to run a command for that |
23:23.16 | DocScrutinizer | I doubt any cmd will help, as uboot lives in kernel NAND partition |
23:23.27 | *** join/#openmoko-cdevel jeepingben (~jeepingbe@h69-131-52-12.swisme.dsl.dynamic.tds.net) |
23:23.29 | DocScrutinizer | either it accesses this file directly, or never cares |
23:24.32 | DocScrutinizer | I thought the parameters from any such config file get compiled into uBoot |
23:24.41 | Alex[sp3dev] | GNUtoo: i take it you never disable wifi power at all? |
23:25.11 | GNUtoo | Alex[sp3dev], ? |
23:25.22 | DocScrutinizer | which would be hard to establish anyway (disable wifi power) |
23:25.38 | Alex[sp3dev] | GNUtoo: with msm-wifi looks like you can only disable it by rmmod msm_wifi |
23:25.48 | GNUtoo | yes |
23:25.59 | GNUtoo | we should use runtime-pm API |
23:26.01 | Alex[sp3dev] | then it is the reason of my problems |
23:26.05 | GNUtoo | but I've an old kernel |
23:26.10 | GNUtoo | what problems? |
23:26.10 | Alex[sp3dev] | i've put power code into mmc vdd switch |
23:26.24 | GNUtoo | ok |
23:26.29 | GNUtoo | I've issues too with elp |
23:27.39 | Alex[sp3dev] | i wonder what the proper way to do it all is. looks like a circular dependency. we need to have the code in mmc vdd to make the card recognize but we don't need it there if we don't want it to die after ifconfig down |
23:27.58 | GNUtoo | ah I had that too |
23:28.14 | GNUtoo | rmmod msm_wifi froze the phone |
23:28.26 | Alex[sp3dev] | well. i hate modules |
23:28.40 | Alex[sp3dev] | i just want to control power via ifconfig or rfkill without hacks |
23:28.56 | GNUtoo | then learn about the runtime PM API |
23:29.07 | Alex[sp3dev] | i will |
23:30.27 | GNUtoo | [ 248.182495] wl1251: ERROR elp wakeup timeout |
23:30.33 | GNUtoo | I added your 2 patches |
23:30.36 | Alex[sp3dev] | that's not ok |
23:30.57 | Alex[sp3dev] | i have another little change but don't think it is relevant at all |
23:31.30 | GNUtoo | [ 324.482360] wl1251: ERROR sdio read failed (-110) |
23:31.46 | Alex[sp3dev] | strange |
23:32.19 | GNUtoo | I'll remove runtime PM |
23:34.50 | *** join/#openmoko-cdevel angelox|n900 (~user@201-13-208-30.dial-up.telesp.net.br) |
23:38.39 | Alex[sp3dev] | GNUtoo: ok, see you later. will fix wifi after the exams |
23:38.45 | GNUtoo | ok |
23:38.47 | *** part/#openmoko-cdevel Alex[sp3dev] (~alexander@86.110.163.19) |
23:38.47 | GNUtoo | thanks |
23:41.15 | angelox|laptop | device not booting.... bad |
23:42.00 | GNUtoo | but it boots maemo |
23:42.07 | angelox|laptop | <PROTECTED> |
23:42.17 | angelox|laptop | just reboot the device.... |
23:42.27 | GNUtoo | what did you do? |
23:42.37 | angelox|laptop | nothing,just restarted the phone... |
23:42.51 | angelox|laptop | removed the battery and put it again worked...why this is happening?? |
23:42.55 | GNUtoo | you forgot to do noloboot |
23:43.06 | angelox|laptop | i did,got the same |
23:43.14 | GNUtoo | retry then |
23:43.20 | GNUtoo | it worked some minutes ago |
23:43.41 | angelox|laptop | working maemo now,should i reboot to check if it is running fine? |
23:43.56 | GNUtoo | ? |
23:44.08 | angelox|laptop | i mean,about uboot,i think it's with some problem |
23:44.16 | GNUtoo | yes it has some issues |
23:44.20 | GNUtoo | at least for booting SHR |
23:45.11 | angelox|laptop | hmm,i think i'll remove u-boot and only install it when i really need use it... |
23:45.42 | GNUtoo | yes |
23:45.46 | GNUtoo | let's see that later |
23:46.33 | angelox|laptop | yes,let me enjoy my new linux device before hack it :) |
23:47.15 | *** part/#openmoko-cdevel jeepingben (~jeepingbe@h69-131-52-12.swisme.dsl.dynamic.tds.net) |
23:49.10 | GNUtoo | ok |