| 00:28.20 | cp15 | re |
| 00:32.20 | pH5 | cp15: wb |
| 00:33.43 | pH5 | does pxa2xx_udc+g_ether work for blueangel? |
| 00:35.20 | cp15 | Yes, but for the pxa27 I think you need another module |
| 00:35.52 | pH5 | I compiled the handhelds.org 2.6.13-hh1 kernel for the magician because pxa27x_udc+g_ether from the xanadux kernel didn't work. unfortunately, the 2.6.13-hh1 pxa27x_udc segfaults :-/ |
| 00:36.46 | cp15 | What didn't work? |
| 00:37.44 | oxi | hey cp15 |
| 00:37.53 | pH5 | the module loaded correctly and g_ether printed its MAC to the console, but plugging in the usb cable only resulted in a usb reset message on the magician. nothing showed up on the host. |
| 00:38.43 | cp15 | Did you have dma enabled or disabled? |
| 00:39.38 | pH5 | disabled, I guess, if we're talking about your haret changes. |
| 00:40.04 | cp15 | There is a config option USB_PXA27X_DMA |
| 00:40.10 | pH5 | oh.. |
| 00:40.31 | cp15 | I would recommend to disable it for now |
| 00:41.28 | oxi | cp15: what was the addr or whatever you told me to check to find the framebuffer base addr? that thing that was changing constantly... you told me to check it 10 times and tell you the min value. |
| 00:42.21 | pH5 | cp15: ok then, it was disabled |
| 00:42.29 | cp15 | You can find the irc logs at http://ibot.rikers.org/%23htc-blueangel See http://ibot.rikers.org/%23htc-blueangel/20051216.html.gz |
| 00:42.40 | oxi | thanks |
| 00:43.09 | cp15 | USB_GADGET_PXA27X and USB_PXA27X are enabled? |
| 00:43.18 | pH5 | yupo |
| 00:43.23 | pH5 | s/yupo/yup/ |
| 00:46.22 | cp15 | Replace the #undef DEBUG and #undef VERBOSE in pxa27x_udc.c with #define DEBUG 1 and #define VERBOSE 1 |
| 00:50.15 | pH5 | heh. that's a good idea |
| 00:51.05 | pH5 | should ARM_FASTCALL be enabled or disabled for some reason? |
| 00:51.36 | cp15 | I have it on y, but don't know what it does exactly |
| 00:52.01 | pH5 | Ok, same here. Better not touch it. :) |
| 00:53.03 | cp15 | looks like it does nothing. Found no further references than its config value |
| 00:54.25 | pH5 | interesting. perhaps it's some handhelds.org artifact. |
| 01:14.07 | pH5 | gah. i don't want to see that wince cut&paste tutorial anymore. |
| 01:20.19 | cp15 | But you will have a master degree in cut&paste when linux development is finished :-) |
| 01:24.41 | pH5 | huh, still no debugging output. |
| 01:24.53 | pH5 | where/under what conditional is pr_debug defined? |
| 01:27.25 | cp15 | Still looking |
| 01:27.26 | pH5 | hm. kernel.h and when DEBUG is defined. but it is. |
| 01:28.18 | cp15 | Have you made a strings on the pxa27x_udc to see if the debug messages are really in it? |
| 01:30.04 | pH5 | actually no. but now I have, and they are in. |
| 01:30.59 | pH5 | after several minutes the screen always turns black. is this just the console screen blanking? |
| 01:31.31 | cp15 | Is the backlight still on? |
| 01:31.41 | pH5 | yes |
| 01:31.52 | lkcl | are you lot still up??? |
| 01:33.07 | pH5 | lkcl: how about you? |
| 01:33.28 | lkcl | no, i'm asleep. definitely :) |
| 01:33.56 | lkcl | just got back from olympia. |
| 01:35.24 | pH5 | hmm sleep. maybe that's not such a bad idea. |
| 01:35.57 | lkcl | oh darn, sf downloading gomunicator stopped the build, o poo. oh well, keep it going, shouldt take much longer |
| 01:35.58 | cp15 | pH5: It is probably the screen blanker. You can disable it when you set blankinterval to 0 in drivers/char/vt.c |
| 01:37.19 | lkcl | hey cp15: is there a blueangel program under linux for dumping the GPIO state, just like haret? |
| 01:37.31 | pH5 | cp15: it's not that important. it just bit me when I was trying to write down that oops message instead of taking a picture. |
| 01:37.56 | cp15 | lkcl: Not yet, but it probably would be a good idea to write one |
| 01:38.07 | lkcl | i wanted to compare things before a suspend, after a suspend, and in thingy. wince. |
| 01:38.09 | cp15 | I have several small utilities which could be added to it |
| 01:38.56 | lkcl | ack. |
| 01:41.08 | cp15 | BTW: I have written w4xy a mail asking him for administrator status for xanadux. I am curious whether I will get it |
| 01:45.56 | cp15 | pH5: Add some printk to pxa27x_udc_irq when entering and leaving |
| 01:46.31 | cp15 | We had a problem on the himalaya where the irq was in an endless loop |
| 01:46.49 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 01:47.27 | cp15 | So, going to bed now |
| 01:47.29 | cp15 | Good night |
| 01:47.36 | pH5 | good night. |
| 01:48.07 | *** part/#htc-blueangel pH5 (n=ph5@e178236159.adsl.alicedsl.de) |
| 01:52.43 | oxi | hey lkcl |
| 01:54.46 | lkcl | hiya oxi. |
| 01:58.34 | lkcl | ~botmail for cp15 ok, update again - hands.com. |
| 01:58.50 | lkcl | ~rob_w for cp15 ok, update again - hands.com. |
| 01:59.06 | lkcl | ~botmail for rob_w ok, update again - hands.com. |
| 01:59.14 | lkcl | that's better :) |
| 02:05.28 | lkcl | okay, am off zzs too. |
| 02:11.27 | *** join/#htc-blueangel _bob_w (n=bob@p85.212.143.46.tisdip.tiscali.de) |
| 02:17.45 | lkcl | new image at http://hands.com/~lkcl/blueangel - it's only 13mb this time because it's minus all the debug options. |
| 02:47.26 | *** join/#htc-blueangel dullard (n=jim@adsl-static-1-30.uklinux.net) |
| 03:58.09 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 06:09.24 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 07:02.27 | *** join/#htc-blueangel rob_w (n=bob@p85.212.143.46.tisdip.tiscali.de) |
| 07:02.27 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 08:20.40 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 10:04.58 | *** join/#htc-blueangel cp15 (i=iipnfi@p549A1416.dip0.t-ipconnect.de) |
| 10:04.58 | *** mode/#htc-blueangel [+o cp15] by ChanServ |
| 10:07.44 | cp15 | Good morning |
| 10:08.05 | *** join/#htc-blueangel oho_ (n=oho@calw-d9b907d6.pool.mediaWays.net) |
| 10:08.12 | oho_ | Hello |
| 10:09.28 | oho_ | cp15: sorry for not showing up much lately, but as I said before I was really bussy. I will get to work now |
| 10:15.06 | cp15 | Ok, no problem. But next time I think it would be better to publish your unfinished changes, so we can at least see on which files you are working and trying to stop incompatible changes |
| 10:15.38 | oho_ | agree, I will post the stuff later |
| 10:15.48 | cp15 | And: Welcome back |
| 10:15.53 | oho_ | cheers |
| 10:15.59 | oho_ | I had sooo much to do |
| 10:16.19 | oho_ | anyway, I will get on it right away .... |
| 10:16.34 | cp15 | I am trying to get administrator privileges on xanadux, so I can add users to the project and you can commit your changes |
| 10:17.01 | oho_ | that would be useful |
| 10:17.31 | oho_ | and I am going offline now, only dial up :-( |
| 10:17.35 | oho_ | speak to you later |
| 10:17.41 | cp15 | Ok, thanks |
| 10:31.55 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 10:37.52 | lkcl | heya |
| 10:38.35 | lkcl | oh, frak. my email server is down. |
| 10:39.21 | lkcl | hmmm... that's odd, it was actually switched off |
| 10:39.25 | lkcl | heya folks. |
| 10:40.46 | cp15 | Hi lkcl, I am just testing your new image. I like the size of it :-) |
| 10:45.41 | lkcl | heya cp15. hur hur |
| 10:46.09 | cp15 | Hmm... blueangel_battery crashes the device :-( |
| 10:46.27 | lkcl | oops. |
| 10:46.41 | lkcl | i think the cold weather crashes my email server. |
| 10:46.52 | cp15 | Probably because it is loaded before himalaya_ts3 |
| 10:47.10 | lkcl | it's in the garage, and it's an apricot. |
| 10:47.12 | lkcl | oops. |
| 10:47.30 | lkcl | oh - btw: been meaning to talk to you about modules and the asic3 module and stuff. |
| 10:47.47 | lkcl | in the himalaya, i made sure all the gpio asic setting stuff was functions. |
| 10:48.03 | lkcl | so i could load the asic3 stuff as a module. |
| 10:48.20 | lkcl | unfortunately, it's not possible to make the fb driver a module... |
| 10:48.33 | cp15 | Why? |
| 10:48.39 | lkcl | ... because of the chain of dependencies leading all the way back to the asic3 module. |
| 10:49.05 | cp15 | And if you compile in asci3? |
| 10:49.11 | lkcl | there are functions (macros?) not listed as being available |
| 10:49.23 | lkcl | then if you compile in asic3, you get _other_ functions not available. |
| 10:49.39 | cp15 | Hmm... Which ones? |
| 10:49.41 | lkcl | such as the blueangel_fb_init functions. |
| 10:49.52 | lkcl | there's a cyclic dependency, basically. |
| 10:50.52 | cp15 | I don't have a blueangel_fb_init at all in my kernel |
| 10:50.59 | lkcl | sorry, let me have a look... |
| 10:52.41 | lkcl | blueangel_fb_device... something like that... |
| 10:53.06 | lkcl | it was a long chain... |
| 10:53.22 | lkcl | want me to try find it again? |
| 10:56.51 | cp15 | Yes, please |
| 10:57.35 | lkcl | okie. i'll describe progress... |
| 10:58.06 | lkcl | making W100 and backlight a module: undefine reference to 'lcd_device_register, backlight_device_register, lcd_device_unregister' |
| 10:58.27 | lkcl | which means there's a Kconfig dependency missing, there... |
| 11:00.21 | lkcl | btw - when i uncheck HTC Blueangel, in "Intel PXA2xx implementations", the _himalaya_'s LCD and battery menu-sub-options are shown and not-shown. |
| 11:00.48 | lkcl | making himalaya lcd a module... |
| 11:01.15 | lkcl | ah! |
| 11:01.20 | lkcl | that's done the trick :) |
| 11:01.35 | lkcl | i must have made too many things to be modules, earlier :) |
| 11:04.26 | lkcl | let's just check by doing a make clean... |
| 11:04.57 | lkcl | so battery needs to be temporarily taken out? |
| 11:06.36 | *** join/#htc-blueangel eme (n=bosty@BSN-165-118-118.dial-up.dsl.siol.net) |
| 11:06.47 | eme | hi room |
| 11:06.54 | lkcl | elloo |
| 11:07.12 | eme | sunny morning here in si |
| 11:10.40 | lkcl | cp15: a suggestion - how about creating an alias for blueangel_kp so that its name really _does_ appear to have a z in it, and loads later? |
| 11:11.01 | lkcl | hey eme, we got a slight issue with a couple of modules loading in the wrong order |
| 11:11.33 | *** join/#htc-blueangel rob_w (n=bob@p85.212.143.46.tisdip.tiscali.de) |
| 11:11.34 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 11:12.25 | lkcl | allo rob. |
| 11:13.26 | cp15 | lkcl: Maybe a good idea, we need something similar for blueangel_battery :-( |
| 11:17.40 | eme | lkcl: I have ba an hmalaya and can test basicaly on both, otherone posses wife ;) |
| 11:17.54 | eme | otherone=secondone |
| 11:24.02 | eme | cp15: what is the blueangel_kp and where can I find descriptions of the other ba modules type please. |
| 11:25.37 | cp15 | Description currently only in the source. blueangel_kp is the the keyboard/button driver |
| 11:26.05 | eme | aha, where we need to fix some butons ;) |
| 11:26.20 | cp15 | No, it is a problem with the load order |
| 11:26.33 | cp15 | the touchscreen creates a device /dev/input/eventX |
| 11:26.38 | cp15 | the kp also |
| 11:26.48 | lkcl | module_conf_XXXXX = "alias somethingelse XXXXX" |
| 11:27.11 | cp15 | The X-Server assumes the touchscreen is located at /dev/input/event0 |
| 11:27.30 | cp15 | So if you load kp first, kp will be /dev/input/event0 |
| 11:28.06 | cp15 | lkcl: Don't think this will work. It makes modprobe zblueangel_kp load blueangel_kp, but not much more |
| 11:28.31 | lkcl | well, it says they're loaded in alphabetical order. i might as well try it. |
| 11:29.22 | lkcl | i ain't got anything better do to :) |
| 11:30.48 | cp15 | BTW: you have two typos in blueangel_leds |
| 11:31.11 | lkcl | burble... |
| 11:31.13 | lkcl | yeh. |
| 11:31.40 | lkcl | well spotted |
| 11:31.53 | cp15 | Just wondered why the leds stayed on |
| 11:32.15 | lkcl | oh. |
| 11:32.15 | lkcl | where's me phone??? |
| 11:32.24 | lkcl | *pat*, *pat*.... :) |
| 11:32.30 | eme | check for leds ;) |
| 11:32.49 | cp15 | Also there are some files missing in gomunicator to make it visible on the desktop |
| 11:32.55 | cp15 | I am still investigating why |
| 11:33.20 | eme | does apm somthing with battery and leds? |
| 11:34.03 | cp15 | Just with leds. It disables the phone button lights when going to suspend, for example |
| 11:34.33 | eme | now i make modprobe ba_leds and yellow led goes off |
| 11:34.50 | lkcl | hurrah |
| 11:34.59 | eme | lol |
| 11:35.17 | lkcl | yeh. i may have put it in the wrong group in gpe (yes, i know, not everyone wants gomunicator...) |
| 11:35.18 | eme | installe it joust yesterday now i'm discovering |
| 11:35.29 | lkcl | good, innit, eme? :) |
| 11:36.02 | eme | wat to get rid of M$ on this platform too |
| 11:36.11 | lkcl | tell me about it... |
| 11:36.20 | lkcl | ... but let me get some breakfast first :) |
| 11:36.31 | eme | u are the one ;) |
| 11:36.50 | eme | bon apetit |
| 12:19.46 | lkcl | bitbake irritating me with that cannot satisfy the following dependencies for task-bootstrap: libgpg-error thing. |
| 12:20.55 | cp15 | Is it a warning or an error? |
| 12:20.59 | lkcl | error. |
| 12:21.47 | eme | chk this out: writing with stylus on normal screen |
| 12:21.54 | eme | http://www.epos-ps.com/dwlfls/Mobile_flash_demo_4_site.swf |
| 12:21.54 | lkcl | hmmm, now, having deleted libgpg-error ican't get it to build |
| 12:23.19 | cp15 | libgpg-error is encryption related. Can't see why we would need it |
| 12:23.31 | cp15 | Does it say which package requires it? |
| 12:23.37 | lkcl | it's in there somewhere. task-bootstrap!!! |
| 12:24.46 | lkcl | gpgme and libgcrypt and meta-sdk all depend on it... |
| 12:25.48 | lkcl | gtk+ depends on libgcrypt... i think i know where this is going |
| 12:26.13 | lkcl | pretty much everything depends on gtk+ |
| 12:26.44 | lkcl | i _so_ don't want to do a complete rebuild. _again_. |
| 12:27.18 | cp15 | Try bitbake -c clean -b ../org.openembedded.dev/packages/libgpg-error/libgpg-error_1.0.bb |
| 12:27.19 | cp15 | and |
| 12:27.29 | cp15 | <PROTECTED> |
| 12:27.42 | lkcl | ah ha - -c clean, i was missing the -c |
| 12:28.11 | cp15 | BTW: I have found out why module_autoload_zblueangel_kp doesn't work |
| 12:28.20 | lkcl | task do_fetch stated, completed, do_unpack started: cannot open libgpg-error-1.0.tar.gz ??? |
| 12:28.24 | lkcl | oh-ah? what was it? |
| 12:29.12 | lkcl | that's better: i'd forgotten to delete the md5 file in sources. FRAK! |
| 12:29.19 | lkcl | bitbake is a _bitch_ to work with. |
| 12:29.52 | cp15 | The string after module_autoload_ has to be the name of a module |
| 12:30.27 | oxi | morning |
| 12:30.28 | lkcl | yers? |
| 12:31.08 | lkcl | mornin oxi |
| 12:31.20 | oxi | how would i go about mapping non-linear physical mem ranges fo fb? |
| 12:31.39 | lkcl | the name of a module, or the name of an alias? in kernel.classes it's not a module name, but an alias name |
| 12:32.24 | lkcl | heya oxi, there's usually some macros and functions that do that - it's been 18 months since i used them, though |
| 12:34.11 | lkcl | aw FRAK. this is pissing me off. delete time. |
| 12:34.16 | oxi | are they implemented by every fb driver(i.e. are they generic)? |
| 12:34.56 | oxi | better: can they be implemented* |
| 12:35.29 | cp15 | Yes, it can |
| 12:35.41 | cp15 | Take a look at arch/arm/mm/ioremap.c |
| 12:35.57 | oxi | okay thanks |
| 12:35.59 | cp15 | There is the function __ioremap which gets called from ioremap |
| 12:36.03 | lkcl | the way a framebuffer device works, you allocate an area of memory, and sometimes that area of memory can be directly read by your hardware - in some cases (like a driver i wrote for a handheld) i had to have a timer device which did a complete copy, looked for changes, and then issued LCD commands to upload line-by-line. |
| 12:36.44 | cp15 | lkcl: The problem is that on universal the screen is split into two different memory areas |
| 12:36.56 | cp15 | oxi: Basically you need to do the following: |
| 12:37.01 | lkcl | ah, joy. ... whatever for??? |
| 12:37.12 | oxi | lkcl: yup good question lol |
| 12:37.20 | lkcl | oxi: do you actually have a connection, over a serial line, then? |
| 12:37.44 | oxi | yes |
| 12:37.46 | cp15 | Request a free vm region with get_vm_area(complete size of your memory, VM_IOREMAP) |
| 12:37.50 | lkcl | EXCELLENT! |
| 12:38.03 | lkcl | i want one. waah! |
| 12:38.35 | cp15 | Then do a remap_area_pages(area->addr, physical address of first block, size of first block, flags) |
| 12:39.01 | cp15 | and remap_area_pages(area->addr+size of first block, physical address of second block, size of second block, flags) |
| 12:39.05 | cp15 | HTH |
| 12:39.21 | oxi | hth? |
| 12:39.27 | cp15 | Hope That Helps |
| 12:39.33 | oxi | oh :) |
| 12:39.39 | oxi | thank you |
| 12:39.40 | lkcl | damn drakkin bitbake, gotta do a complete rebuild. |
| 12:39.45 | cp15 | Why? |
| 12:40.11 | lkcl | that stupid libgpg-error0 dependency. can't work out how to get rid of it. |
| 12:40.19 | lkcl | figured sod it: just trash the lot |
| 12:40.43 | lkcl | it only takes a couple of hours, and if i'm friggin about for two hours... |
| 12:40.59 | cp15 | Hope not every little change requires a complete rebuild |
| 12:41.30 | lkcl | this started since i added gomunicator. |
| 12:41.49 | cp15 | But gomunicator was already in the image |
| 12:41.58 | lkcl | well - say since: the last time i had this problem, it started when i added gomunicator, and went away when i removed it. |
| 12:42.26 | cp15 | Hmm |
| 12:42.38 | lkcl | but a complete rebuild was clear - well... actually, a different error (alsa-lib missing, which it isn't). |
| 12:42.53 | lkcl | i can fix _that_ problem by doing a _separate_ bitbake alsa-lib and then carrying on. |
| 12:43.06 | lkcl | i _did_ say bitbake is awful, didn't i? :) |
| 12:43.15 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 12:43.22 | cp15 | You know something better? |
| 12:43.34 | lkcl | yes - debian dpkg!!!! |
| 12:44.25 | cp15 | Come on, it isn't as flexible as bitbake |
| 12:45.26 | lkcl | in cases like the wrt54g, where they've only 2-4mb of flashram, i can understand it - but even then, you simply write something like debootstrap which, afterwards, removes /var/lib/dpkg/ in order to save space. |
| 12:46.58 | lkcl | debian package building has been ongoing for what... 10 years. there's automated build tools, then there's also packages for doing cross-compiler unpacking... lots of stuff. |
| 12:47.16 | lkcl | i've never _used_ half of them, mind :) |
| 12:48.53 | lkcl | anyway. |
| 12:49.06 | cp15 | I am afraid we have to stick with bitbake |
| 12:49.10 | lkcl | got any suggestions for what i could do for the next 2 hours, other than go watch telly? :) |
| 12:49.41 | lkcl | just to be perverse i may end up trying to do a debian port, some day :) |
| 12:50.22 | lkcl | unfortunately, the arm gcc cross compiler package builder HOWTO doesn't work in building arm-linux-gcc-3.4.4 :( |
| 12:50.33 | cp15 | Hmm... Some items from my todo list: |
| 12:50.56 | cp15 | Find out why the shifted Fn Keys don't work |
| 12:51.49 | cp15 | I have thought it would be a good idea if the user presses Fn+R which gives a $ on the german keypad we could send a shift right + key 4 |
| 12:51.59 | cp15 | But it doesn't work. |
| 12:52.56 | cp15 | Next thing: Keypad illumination. Enable the light when the user presses a key and let it go out after a few seconds |
| 12:53.21 | cp15 | Next thing: Flash map |
| 12:53.43 | lkcl | i'll try running haret about the keypad one... |
| 12:54.03 | cp15 | Currently we have only one block in the flash. It would be highly useful to split this block in a bootloader and a os block |
| 12:54.34 | cp15 | It isn't a reverse engineering problem. The keys get transferred correctly to linux, it is a internal linux problem |
| 12:55.15 | lkcl | oh - right. OH. i know. |
| 12:55.22 | lkcl | that gpio program. |
| 12:55.54 | cp15 | Next thing: irda... Currently completely untested and most likely doesn't work |
| 12:56.36 | cp15 | And of course doc and wlan. I have made some progress in making a "shim" for using a motorola 2.4 driver on 2.6, but it is still incomplete |
| 12:56.59 | cp15 | wlan doesn't show up at all on my devel device. No idea why |
| 13:06.56 | cp15 | lkcl: Found a (ugly) solution to the kp and battery problem: |
| 13:08.25 | cp15 | Remove module_autoload_zblueangel_kp and module_autoload_blueangel_battery and set module_autoload_himalaya_ts3 = "himalaya_ts3 blueangel_kp blueangel_battery" |
| 13:14.13 | cp15 | Or maybe better: |
| 13:14.34 | cp15 | module_autoload_himalaya_ts3 = "himalaya_ts3 blueangel_battery" |
| 13:14.50 | cp15 | module_autoload_blueangel_kp = "blueangel_kp" |
| 13:15.20 | cp15 | and set TSLIB_TSDEVICE=/dev/input/event1 in packages/tslib/tslib/blueangel/tslib.sh |
| 13:34.22 | eme | how do I pass options parameter with modprobe? |
| 13:35.54 | cp15 | modprobe module option=value |
| 13:36.22 | eme | tks already find out that spaces are to much |
| 13:46.40 | cp15 | lkcl: please add alsa-conf to gomunicator rdepends |
| 13:56.01 | cp15 | And snd-mixer-oss and snd-pcm-oss to module_autoload_blueangel_audio |
| 14:09.03 | lkcl | cp15: i tried that (the add alsa-conf as a dependency) - it doesn't want to! |
| 14:09.08 | lkcl | ack about the uglies. |
| 14:09.50 | eme | scp give me error |
| 14:11.43 | lkcl | i get bitbake errors about alsa-conf not existing as a bitbake package / dependency. |
| 14:12.24 | lkcl | audio: done. |
| 14:12.35 | lkcl | ts3: done. |
| 14:34.06 | lkcl | cp15: there isn't an alsa-conf. there's alsa-utils, alsa-lib, alsa-driver, alsa-oss, alsa-tools. i've tried adding alsa-tools as a dependency, i've tried alsa-utils. about the only one i _haven't_ tried is alsa-oss. |
| 14:45.19 | eme | how can I restart Xsession when login over ssh? |
| 14:48.15 | cp15 | eme: Please explain |
| 14:49.19 | eme | i have srewed up grafical enviroment so half of screeni is gray. |
| 14:49.29 | eme | now I want startx again |
| 14:50.25 | lkcl | <PROTECTED> |
| 14:50.58 | eme | stop |
| 14:51.15 | eme | he dos not now restart ;) |
| 14:51.20 | lkcl | btw - anyone getting "can't find default mixer" when running gomunicator with the latest gpe-image i did? |
| 14:51.22 | eme | thank you |
| 14:51.27 | lkcl | start instead. |
| 14:51.44 | lkcl | alternatively just run gpe-dm & i think that's the name of the program. |
| 14:51.56 | eme | it is done. |
| 14:52.12 | eme | where have u dl ne gp-image on wiki? |
| 14:54.30 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 14:54.44 | eme | in new gpe-image on wiki? |
| 14:58.53 | cp15 | The snd-mixer-oss is missing |
| 15:01.22 | lkcl | http://hands.com/~lkcl/blueangel. |
| 15:01.50 | cp15 | Just a moment, I am trying to make a ringtone for gomunicator |
| 15:01.52 | eme | vy tnx |
| 15:02.13 | lkcl | yeh, i'm adding it to autoload as you suggested - but even when that module's loaded, it still says can't find default mixer on startup. |
| 15:02.15 | lkcl | v. odd. |
| 15:02.46 | cp15 | Maybe it is related to the missing alsa-conf? |
| 15:09.08 | lkcl | there isn't an alsa-conf. |
| 15:09.26 | lkcl | there's only alsa-utils, alsa-lib, alsa-driver, alsa-oss and alsa-tools. |
| 15:09.57 | lkcl | where alsa-lib is the only one i can get a complete build from (from scratch, only). |
| 15:15.57 | cp15 | do a find . -name alsa.conf in your build directory |
| 15:21.26 | lkcl | ack. |
| 15:22.14 | lkcl | you mean the tmp/ directory? |
| 15:23.03 | lkcl | FRAK that was a waste of time - getting the same stupid dependency build error! |
| 15:23.46 | *** join/#htc-blueangel TelePuzik (n=UNIX@80.81.39.150) |
| 15:24.01 | TelePuzik | Hi there |
| 15:24.09 | lkcl | oh. right. i get it. i'd added some modules that don't exist. damn. |
| 15:24.14 | lkcl | hiya TelePuzik. |
| 15:24.23 | TelePuzik | how are you? |
| 15:24.38 | lkcl | good, dude. |
| 15:24.55 | TelePuzik | whats up? |
| 15:25.18 | lkcl | am just building a new gpe-image for the blueangel |
| 15:25.20 | lkcl | again. |
| 15:25.35 | TelePuzik | :) |
| 15:26.49 | TelePuzik | when it will be ready for testing |
| 15:30.30 | cp15 | lkcl: I have something to change for you |
| 15:31.29 | cp15 | In gomunicator_cvs.bb add http://gnulinux.biz/files/ringtones/phone_ringing.raw to SRC_URI (seprarated with a space) |
| 15:31.36 | cp15 | and add the following: |
| 15:31.43 | cp15 | do_install_append() { |
| 15:31.43 | cp15 | <PROTECTED> |
| 15:31.43 | cp15 | } |
| 15:32.17 | cp15 | I have also fixed something to get the gomunicator icon on the desktop, but unfortunately it is not yet synced to the public cvs |
| 15:34.17 | lkcl | ack. |
| 15:36.23 | lkcl | oh well. |
| 15:36.29 | lkcl | well, let's try so far, anyway. |
| 15:37.12 | cp15 | Psyco is a joke |
| 15:37.39 | cp15 | The slowest "Handling BitBake files" step is just as slow as without it |
| 15:37.48 | lkcl | oops... |
| 15:38.31 | lkcl | install: too few arguments. oops. |
| 15:38.47 | lkcl | duh. |
| 15:38.59 | lkcl | got it. |
| 15:40.07 | cp15 | Buaahhh... It is even slower with psyco |
| 15:40.43 | lkcl | oh so i'm not missing out on anything, then |
| 15:40.54 | lkcl | by having an amd64 |
| 15:42.41 | cp15 | Ok, maybe it was a cache issue.. But it is only about 5% faster at best |
| 15:43.03 | *** join/#htc-blueangel sOCiETY (n=society@p549AE8A0.dip.t-dialin.net) |
| 15:43.07 | sOCiETY | hi |
| 15:43.16 | sOCiETY | i got one little question |
| 15:43.24 | lkcl | hello |
| 15:43.27 | sOCiETY | do the haret work on wm2005 on blueangel? |
| 15:43.30 | cp15 | Maybe java isn't that slow because it is interpreted, but because it is crap? |
| 15:43.48 | lkcl | don't get me started on java |
| 15:44.11 | lkcl | yeh, it does, society: it's the way to load / start linux. |
| 15:44.39 | cp15 | soCiETY: Partly... It is known not to work with WM2005 on the Blueangel, but works with WM2005 on Universal... Strange |
| 15:45.01 | lkcl | wm2005? oh. sorry. what's that? perhaps i should ask... :) |
| 15:45.13 | cp15 | Windows Mobile 2005 |
| 15:45.25 | lkcl | ah. |
| 15:45.31 | lkcl | what we all need, then. |
| 15:45.41 | cp15 | HTC somehow leaked a version for the blueangel |
| 15:46.31 | cp15 | sOCiETY: So the answer to your question is probably no, but maybe you can help to fix it? |
| 15:47.13 | *** join/#htc-blueangel sOCiETY2 (n=society@p549AD50C.dip.t-dialin.net) |
| 15:47.22 | sOCiETY2 | sorry git an disconnect |
| 15:47.29 | sOCiETY2 | somethings was wrote? |
| 15:47.42 | sOCiETY2 | kick the society ;-) |
| 15:48.29 | cp15 | I will repeat |
| 15:48.47 | sOCiETY2 | thx |
| 15:48.49 | cp15 | Partly... It is known not to work with WM2005 on the Blueangel, but works with WM2005 on Universal... Strange |
| 15:48.58 | cp15 | So the answer to your question is probably no, but maybe you can help to fix it? |
| 15:49.07 | sOCiETY2 | i can try it |
| 15:49.31 | sOCiETY2 | i want both wm2005 and linux *g* |
| 15:50.05 | cp15 | Haven't heard anything good about wm2005 except the permanent storage |
| 15:50.33 | sOCiETY2 | hmmm |
| 15:50.37 | TelePuzik | Nothing is good if done by MS |
| 15:50.55 | eme | true |
| 15:51.39 | cp15 | No, not true. Recently they used the mozilla rss icon for internet explorer, instead of inventing their own better one :-) |
| 15:51.47 | sOCiETY2 | hihihihiih |
| 15:51.51 | sOCiETY2 | lame |
| 15:51.54 | TelePuzik | :) |
| 15:52.18 | sOCiETY2 | i need to go now |
| 15:52.21 | sOCiETY2 | i will come back later |
| 15:52.49 | sOCiETY2 | cya |
| 15:52.53 | cp15 | cu |
| 16:07.49 | TelePuzik | cp15 in the INSTALLATION INSTRUCTIONS on the website |
| 16:08.13 | TelePuzik | can you add please "sync" after Execute tar xjvpf <path-to-it>/gpe-ba.tar.bz2 |
| 16:08.59 | cp15 | Sure I can, but you can also... It is a public wiki |
| 16:09.10 | TelePuzik | oh |
| 16:09.52 | TelePuzik | You are not allowed to edit this page. |
| 16:09.55 | TelePuzik | :) |
| 16:10.17 | cp15 | Tss... You have to register first |
| 16:10.44 | TelePuzik | no need then :) |
| 16:11.03 | lkcl | cp15: adding RDEPENDS="alsa-lib" stops the build of gpe-image with "ERROR: cannot satisfy the following dependencies for gpe-task-pim: alsa-lib". |
| 16:12.24 | lkcl | another rsync on way up... |
| 16:12.36 | TelePuzik | waiting on fire |
| 16:12.38 | TelePuzik | :) |
| 16:15.04 | lkcl | boot time.... wheeeee |
| 16:18.29 | TelePuzik | cp15 Last time you asked about the Board ID its Blue Angel Board ID 0x6 |
| 16:20.26 | TelePuzik | Hmm |
| 16:21.36 | TelePuzik | I just run the gomunicator it didn't even ask for the Pin Code and i could place a call %) |
| 16:22.22 | cp15 | If you entered it already with windows, the phone stayes unlocked |
| 16:22.29 | TelePuzik | yes |
| 16:22.32 | TelePuzik | thats how it was |
| 16:25.02 | TelePuzik | if you open Session Log in Log Display you get df:/dev/static/dev: permission denied |
| 16:27.48 | TelePuzik | also there is gpe-soundbite:3321 libgiade-WARNING could not load support for ;gnome libgnome.so; cannot open shared object file: No such file or directory |
| 16:27.59 | lkcl | phone, phone.... where's me house phone... |
| 16:28.39 | lkcl | ring! |
| 16:29.11 | lkcl | hmmm.. what's it doing on vibrate, with no volume?? |
| 16:29.13 | lkcl | hmmm.. |
| 16:31.20 | lkcl | okay, _again_ i need to run gpe-mixer to turn up the volumes from zero. |
| 16:35.42 | lkcl | damn. quitting the gpe-login session, the PCM volume isn't saved properly. |
| 16:36.24 | lkcl | volume, bass, treble and mic are all saved correctly; PCM volume (one of the important ones :) isn't. |
| 16:37.47 | lkcl | nope. remove blueangel_audio, and the gpe-mixer PCM volume gets set back to zero. |
| 16:39.38 | eme | try to call on gsm from other phone, with opened gomunicator, then when ringing open mixer settings. |
| 16:40.05 | eme | after that u will get blank square on bottom side of display where was before mixer |
| 16:40.30 | eme | that was the readon that I need to restrt gpe-dm |
| 16:43.09 | lkcl | ack. |
| 16:46.35 | lkcl | okay, when i run gconf sound settings, it's okay. |
| 16:46.43 | lkcl | but first, i have to run gpe-mixer. |
| 16:46.51 | lkcl | then gconf sound settings. |
| 16:53.17 | lkcl | ok. do i risk sending it to sleep, just to crash it? :) |
| 16:53.21 | lkcl | yehhhhhh |
| 16:53.32 | lkcl | zzzzs... |
| 16:55.05 | cp15 | Does it wake up? |
| 16:55.35 | eme | he keep us exiting |
| 16:55.36 | lkcl | yeh, it wakes up - just no screen. |
| 16:55.49 | lkcl | btw alsa.conf is in alsa-lib-1.0.10/src/conf |
| 16:56.01 | lkcl | in work/alsa-lib-1.0.10-r0/ |
| 16:56.26 | cp15 | But nowhere under image or install? |
| 16:56.50 | lkcl | oh yes, and in install. |
| 16:57.25 | lkcl | did i search image? i don't know. |
| 16:59.13 | immolo | alsa only saves sound if you open and close it via init doesn't it? |
| 16:59.44 | cp15 | There should be a install/alsa-conf/usr/share/alsa/alsa.conf under work/alsa-lib-1.0.10-r0 |
| 17:00.37 | lkcl | yep. that's it. |
| 17:01.46 | cp15 | Hmm... Strange |
| 17:03.28 | lkcl | hey, can i just write to 0x0c000098 and it will be GPIO_B address? |
| 17:03.46 | lkcl | or do i need to do the same trick as gsminit.c? |
| 17:05.49 | cp15 | You need to mmap /dev/mem |
| 17:05.51 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 17:06.04 | cp15 | There is a devmem program in the initrd which might help you |
| 17:07.15 | lkcl | gsminit.c does opening of /dev/mem. |
| 17:16.28 | lkcl | okay. need to be reminded of how these damn GPIO things are addressed... the ahh... it's a bit-field... |
| 17:35.04 | cp15 | lkcl: I have just extracted the bootloader power on sequence, and it is just like the one in blueangel_lcd_hw_init |
| 17:36.14 | cp15 | But the timing might be wrong: it is 2600 cpu cycles, 4436 cpu cycles and 18676 cpu cycles |
| 17:36.49 | cp15 | Try to increase the second mdelay(5) in blueangel_lcd_hw_init |
| 17:41.20 | cp15 | Aha, and between PWR5 and PWR1 the ATI gets initialized |
| 17:41.40 | cp15 | Maybe we have to split the initialisation up in a pre ati and post ati part |
| 18:33.00 | *** join/#htc-blueangel raz0r1911 (i=Peak@cm24.epsilon36.maxonline.com.sg) |
| 18:33.17 | raz0r1911 | Hi guys, I'm stucked. I'm not sure how to install it to my fone. |
| 18:47.18 | raz0r1911 | anybody here? |
| 19:08.07 | lkcl | ca15: yep - i did say it's likely to be like the thingy one - the himalaya one :) |
| 19:08.34 | lkcl | oh dear - someone came calling and dropped off. |
| 19:08.47 | lkcl | is there any way of telling people that this is a low-traffic irc? |
| 19:17.07 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 19:23.10 | lkcl | hiya bhima |
| 19:35.35 | lkcl | okay - created a patch for splitting lcd init into two halves... |
| 19:41.37 | lkcl | dang. |
| 19:52.55 | lkcl | hmmm... |
| 19:53.07 | *** join/#htc-blueangel oxipda (n=pocketir@213.4.169.86) |
| 19:53.24 | eme | ikcl: is the version u uploadet on hh at 02:04pm ok? |
| 19:53.25 | oxipda | hi |
| 19:53.38 | lkcl | trying to work out which bits are the lcd ATI init - they're nothing _like_ what i saw in himalaya code. |
| 19:53.44 | lkcl | eme - should be fine. |
| 19:53.48 | lkcl | hiya oxi |
| 19:53.50 | eme | tks |
| 19:53.55 | eme | will try it out now |
| 19:54.18 | lkcl | might need a couple of boots to get itself sorted (e.g. the usb0 only loads on the 2nd boot). |
| 19:54.24 | lkcl | don't let it go to suspend... :) |
| 19:54.57 | eme | till now he never suspend ;) |
| 19:56.33 | *** join/#htc-blueangel immolo (n=immolo@cpc1-cosh1-4-0-cust126.cos2.cable.ntl.com) |
| 20:01.59 | oxipda | cp15: what good is ioremap if you can only call remap_pages(or whatever, im in a tavern so) from ioremap.c? i guess its not a good idea to modify ioremap func itself... i thought of overloading it but that would mean changing all extern calls, so i ... |
| 20:04.04 | oxipda | ... concluded making an export symbol for remap_pages would be the best, but i dunno whether this fullfills security "requirements" in linux programming |
| 20:05.32 | oxipda | can anyone help? |
| 20:05.59 | oxipda | by giving me some directions... |
| 20:06.08 | oxipda | some advice |
| 20:07.07 | oxipda | i need to map "stuck" 2 memory regions |
| 20:11.11 | lkcl | http://bugs.openembedded.org/show_bug.cgi?id=540 |
| 20:11.21 | lkcl | ask also on #oe, oxi. |
| 20:12.17 | oxipda | thank you lkd :) |
| 20:18.39 | cp15 | oxi: An interesting post from cr2: http://forum.xda-developers.com/viewtopic.php?t=20670&start=250 |
| 20:21.26 | lkcl | okay - oxi - look at the sa1100fb source code, apparently they do the same "split" thing, there. |
| 20:21.40 | lkcl | A780 has a PXA27x SRAM driver, but the SRAM is too small. The pxafb driver will not use it anyway. |
| 20:21.40 | lkcl | A split framebuffer is nothing new: the SA1100 |
| 20:21.40 | lkcl | strongarm processor supported it in hardware, |
| 20:21.41 | lkcl | and the "sa1100fb' used it (for example on HP jornada820). |
| 20:21.41 | lkcl | BTW, the same 1MB+SRAM structure can be |
| 20:21.41 | lkcl | seen on all newer pxa27x wince devices. |
| 20:22.04 | cp15 | I have looked into it but can't see anything like a split memory here. |
| 20:23.52 | lkcl | damn. cp15: what _is_ the ATI initialisation code? what function am i looking at calling? |
| 20:24.41 | lkcl | and: are you sure it's between 5 and 1? |
| 20:24.47 | cp15 | Wrong way round: The ATI code calls enable LCD and disable LCD when something needs to be done |
| 20:25.46 | lkcl | yes... okay, so i've created an lcd resume_pre() and an lcd resume_post() |
| 20:26.08 | cp15 | Makes sense |
| 20:26.18 | cp15 | I can send you my emulator output, if you want |
| 20:26.20 | lkcl | so i have tg->resume_pre(); w100_hw_init(par); tg->resume_post(). |
| 20:26.25 | lkcl | yes please. |
| 20:26.53 | lkcl | i'm having difficulty tracking down what actually gets called in hw_init() - it's all far too tidy :) |
| 20:27.22 | lkcl | i like my code with numbers in it and a complete pile of hacked-up-but-working dog-poo. |
| 20:27.44 | cp15 | Maybe it would be easier to give resume and suspend a second parameter |
| 20:27.56 | lkcl | oh probably. |
| 20:28.17 | cp15 | Because now we need 3 functions: pre, post and a 3rd one when called from screen blank |
| 20:28.42 | lkcl | yep i've commented that one out, for now... |
| 20:29.21 | lkcl | yes i see what you mean. a bitfield - 0x1 is pre 0x2 is post 0x3 is both (called from screen blank) |
| 20:30.08 | *** join/#htc-blueangel rob_w (n=bob@p85.212.143.46.tisdip.tiscali.de) |
| 20:30.08 | *** mode/#htc-blueangel [+o rob_w] by ChanServ |
| 20:30.14 | lkcl | hiya rob_w |
| 20:30.20 | eme | ikcl: it is normal to boot in console? |
| 20:30.23 | rob_w | heya dudes |
| 20:30.37 | rob_w | how is all going ? |
| 20:30.46 | eme | vy fine |
| 20:30.56 | lkcl | ah, fine, after a kind, rob. |
| 20:31.11 | rob_w | cool , |
| 20:31.15 | lkcl | got the file, ta, cp15 |
| 20:31.30 | lkcl | eme: yes, it is - then things start up from there. |
| 20:31.38 | cp15 | Maybe you need a little explanation |
| 20:31.38 | rob_w | after a quick bath i will tune back in here ... to get my baby suspending again |
| 20:32.02 | eme | startx does not work ;) how then? |
| 20:32.31 | cp15 | The number in front of the register accesses is the delta of the cycle counter. It gives you a raw estimation how much time has gone by since the last register access |
| 20:32.36 | lkcl | you need a framebuffer /dev/fb0 - check that things go on-screen by doing echo dawskjdhasldhaslds > /dev/fb0 |
| 20:32.57 | cp15 | eme: Look for pivot_root: No such file or directory |
| 20:33.08 | lkcl | oh yes. mkdir oldroot |
| 20:33.25 | lkcl | you have to do mkdir oldroot on the ext3 partition. |
| 20:33.47 | cp15 | The lines without register addresses are the stuart output |
| 20:34.16 | cp15 | The ati_writes don't have a cycle counter. Yes, the whole thing is a mess |
| 20:34.57 | eme | mkdir oldroot and extract gpe-image to / or oldroot? |
| 20:35.20 | lkcl | extract gpe-image to / and then mkdir oldroot in /. |
| 20:35.37 | eme | tks |
| 20:35.47 | cp15 | Not to / of your linux host |
| 20:35.59 | eme | jeah |
| 20:36.02 | lkcl | in /mnt/sdcard/partition2 :) |
| 20:36.14 | eme | roger |
| 20:36.24 | lkcl | okay. need a lie-down again. |
| 20:36.29 | lkcl | large dinner :) |
| 20:36.29 | eme | playing games since 1984 |
| 20:36.33 | eme | :) |
| 20:37.22 | lkcl | back later... |
| 20:39.10 | oxipda | thanks for the help guys, i'll read the log when i get home. laterz |
| 21:26.39 | rob_w | lkcl , do u resume properly these days ? |
| 21:28.07 | rob_w | anyway .. |
| 21:28.22 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 21:30.14 | *** join/#htc-blueangel iclem (n=clement@iclem.net) |
| 21:31.38 | cp15 | Hi dullard, I have read in the xda-developers forum you want to help. Are you here? |
| 22:47.44 | cp15 | oxi: Are you here? |
| 22:50.58 | oxi | yup |
| 22:51.10 | cp15 | Do you have a sourceforge user? |
| 22:51.11 | oxi | just read forum |
| 22:51.18 | oxi | nope |
| 22:51.23 | cp15 | Then get one |
| 22:51.46 | cp15 | I have now administrator access to the xanadux project and can add you to the developers list |
| 22:51.54 | oxi | cp15: im not very sure i can do much at this, you know... |
| 22:52.01 | oxi | congrats :) |
| 22:52.49 | cp15 | This means that we can fork off a universal tree from the blueangel tree and you can publish your changes and others can contribute to it |
| 22:53.07 | oxi | id like to remain under the shadows until i feel confident enough i can help the project.... other than that im just testing how this fits on my pda and learning a lot |
| 22:53.40 | oxi | oof... |
| 22:54.22 | cp15 | Hmm... What a pity... But nontheless I would like to see your changes in cvs. If you don't want to, I will do it for you |
| 22:54.42 | oxi | yes sure |
| 22:55.10 | immolo | cp15, is the ba ready for some day to day usage yet? |
| 22:56.03 | cp15 | For me: Yes :-) For others... Hmm... Maybe... I hope it will when suspend code is debugged and the image is ready |
| 22:56.07 | oxi | but i guess you'll need to fix some things... for im not very familiarized with kernel structure and many many other things ... so... |
| 22:56.38 | immolo | suspend still not working? |
| 22:56.41 | cp15 | No problem |
| 22:57.06 | cp15 | Not 100%. Sometimes the screen stays black (0x5 BA) or white (0x6 BA) |
| 22:57.15 | cp15 | I hope we will get it fixed next week |
| 22:57.24 | immolo | yay I get a black screen :P |
| 22:57.46 | cp15 | Then you probably didn't load blueangel_suspend or blueangel_kp |
| 22:57.54 | cp15 | If it happens always |
| 22:58.06 | immolo | I meant if it did I would :P |
| 22:58.23 | cp15 | You would what? |
| 22:58.31 | immolo | doesn't matter |
| 22:58.46 | immolo | but I normally set the modules to autoload |
| 22:58.57 | immolo | it's annoying keep sshing in |
| 22:59.12 | cp15 | Hmm... |
| 22:59.31 | cp15 | oxi: Do you want to make the cvs checkin or should I do it? |
| 22:59.58 | oxi | to be honest i never used cvs before lol |
| 23:00.12 | cp15 | Then you have a good chance to learn something |
| 23:00.21 | cp15 | First, get a sourceforge account |
| 23:00.26 | cp15 | Isn't that difficult |
| 23:00.40 | oxi | okay |
| 23:00.42 | cp15 | BTW: immolo, do you also want developer access to Xanadux? |
| 23:09.55 | immolo | I can't do much kernel work |
| 23:10.13 | oxi | done |
| 23:11.13 | oxi | i add to immolo position |
| 23:11.51 | immolo | I'll be more help trying to help lkcl |
| 23:12.19 | cp15 | oxi: What is your sourceforge name? |
| 23:12.24 | oxi | oxi :D |
| 23:13.02 | immolo | and the second ba linux allows me to make calls and send sms I can switch and help beta |
| 23:14.06 | immolo | that would be the best xmas present I could think of :P |
| 23:14.12 | oxi | lol |
| 23:14.48 | cp15 | immolo: We will all do our very best :-) |
| 23:15.00 | cp15 | Ok, oxi, I have added you to the list |
| 23:15.19 | cp15 | Now create an empty directory on your machine |
| 23:15.37 | immolo | heh what currently does work, I mean battery life isn't much of an issue as long as it has 13 hours+ |
| 23:16.07 | oxi | k |
| 23:16.33 | cp15 | Yeah, I have my charger near my bed :-) But one time it happened it got empty. Got up too early, stayed up too long |
| 23:17.16 | immolo | heh battery monitor stilll isn't working then :P |
| 23:18.08 | oxi | cp15: what else... |
| 23:18.41 | cp15 | xport CVS_RSH=ssh |
| 23:18.41 | cp15 | <PROTECTED> |
| 23:18.41 | cp15 | cvs -z3 -d:ext:oxi@cvs.sourceforge.net:/cvsroot/xanadux co -ko -P -r K2-6-12-hh2-xda0-ba0 linux-2.6-xda |
| 23:19.12 | cp15 | e before export missing |
| 23:19.48 | oxi | should i include that env variable to have it available on reboot? |
| 23:20.04 | cp15 | This means that you want to check out the branch K2-6-12-hh2-xda0-ba0 of linux-2.6-xda |
| 23:20.04 | oxi | i guess i should |
| 23:20.30 | cp15 | Probably makes sense. But in most cvses from distributions it is already default |
| 23:20.44 | oxi | k2? |
| 23:21.01 | oxi | its asking me a pass |
| 23:21.13 | oxi | oh sourceforge pass i guess :) |
| 23:21.15 | cp15 | The one you gave your user |
| 23:21.44 | cp15 | You can also generate a key with ssh-keygen and upload the public part to sourceforge so you don't need a key every time |
| 23:21.58 | oxi | does it matter if im on amd64? |
| 23:22.02 | cp15 | No |
| 23:22.18 | oxi | okay |
| 23:22.31 | cp15 | amd64 seems to be the most liked developer cpu these days |
| 23:22.43 | cp15 | You have one, immolo and lkcl too |
| 23:22.57 | cp15 | Its just me who is working on old 32 bit |
| 23:23.05 | oxi | heheh |
| 23:23.10 | cp15 | Ok, the checkout will probably take some time |
| 23:23.28 | cp15 | When it is done, we fork of a new branch for universal |
| 23:23.47 | oxi | okay ... |
| 23:23.58 | cp15 | go into linux-2.6-xda when it is done |
| 23:24.32 | oxi | 1mbps dl speed |
| 23:24.41 | oxi | shitty connections in spain :) |
| 23:25.01 | cp15 | then issue a cvs tag -b K2-6-12-hh2-xda0-un0 |
| 23:25.05 | immolo | cp15, it's much easier in amd64 |
| 23:25.28 | immolo | but then you have pyscho to speed up bitbake |
| 23:25.35 | cp15 | The connection speed isn't the main problem. Sourceforge is slow even with the fastest link |
| 23:25.44 | cp15 | I have tried psyco today. Forget it |
| 23:25.50 | immolo | really |
| 23:26.07 | cp15 | The "Handling Bitbake" job is 5% faster at most |
| 23:26.12 | immolo | lol |
| 23:26.24 | immolo | well I get about 20% faster on 64bit |
| 23:26.49 | immolo | but my amd64 is 2ghz and my 32bit is 1.6 |
| 23:27.12 | oxi | athlon 3200 ? :) |
| 23:27.18 | immolo | 3400 |
| 23:27.19 | cp15 | But you can run 32 bit binaries on amd64 as well. Does it make any speed difference? |
| 23:27.25 | immolo | yes |
| 23:27.48 | immolo | the amd64 runs faster then normal x86 |
| 23:28.17 | lkcl | there's an emulation layer. |
| 23:28.42 | lkcl | also, the typical amd64 chipsets have 4-way interleave on memory (if you have 4 dimms) |
| 23:28.57 | immolo | we talking pure 32bit or 32bit in 64bit mode? |
| 23:29.53 | lkcl | doesn't matter: the motherboards have chipsets with 4-way interleave, so you just get... faster memory access, all round. |
| 23:30.26 | lkcl | that's that the 400/800/1200/1600 figures are all about - max of 400mhz DDR-2, times four. |
| 23:30.28 | immolo | really I just noticed a speed decrease in 32bit apps in 64bit |
| 23:30.56 | immolo | might be because they aren't optimaized like the native stuff |
| 23:31.02 | lkcl | yeh. emulation layer in s/w plus non-optimal compiled code. |
| 23:32.12 | lkcl | oxi - are _you_ complaining about 1mbps download speed???? i got to stick with damn 256mps unless i go for SDSL which is 10x more expensive (£300/mo)! |
| 23:32.41 | oxi | you mean 256kbps? |
| 23:32.49 | lkcl | oh. yeh. 256kbps :) |
| 23:32.56 | oxi | phew |
| 23:33.11 | oxi | well at least you got an option to upgrade, not in my country |
| 23:33.29 | immolo | lkcl, just go cable |
| 23:33.39 | immolo | 10mb/sec upgrade next year |
| 23:33.54 | cp15 | Then take Funklord on the other hand. 10 MBit Up- and downlink for 25 Euro/Month. And 100 MBit soon |
| 23:34.20 | oxi | the most you can get presently here is just 20mbps, thats adsl 2+, which i asked back in october... and still waiting, just cause the big one doesnt want to liberate its nodes |
| 23:34.26 | lkcl | oh. right. i get it. i'd added some modules that don't exist. damn. |
| 23:34.43 | lkcl | heya cp15: i got a message from koen, on handhelds.org. |
| 23:34.44 | lkcl | <PROTECTED> |
| 23:34.45 | lkcl | <immolo> suspend still not working? |
| 23:34.45 | lkcl | <cp15> No problem |
| 23:34.45 | lkcl | <cp15> Not 100%. Sometimes the screen stays black (0x5 BA) or white (0x6 BA) |
| 23:34.45 | lkcl | <cp15> I hope we will get it fixed next week |
| 23:35.04 | oxi | ? |
| 23:35.15 | cp15 | Wrong irc client... |
| 23:35.19 | *** join/#htc-blueangel lkcl (n=lkcl@bb-87-82-26-80.ukonline.co.uk) |
| 23:35.33 | lkcl | oops :) |
| 23:35.40 | lkcl | that was a bit of a fuckup for a cut/paste job |
| 23:36.13 | cp15 | What did koen say? |
| 23:36.28 | lkcl | well i was gonna cut/paste it in, but i clicked the wrong button |
| 23:36.43 | lkcl | and cut/paste an entire screen of irc - 60 lines - to freenode :) |
| 23:36.45 | lkcl | gimme 1sec... |
| 23:36.53 | lkcl | * blueangel.conf: the _kernel26 override isn't needed since a 2.4 kernel doesn't |
| 23:36.53 | lkcl | +exist for the ba |
| 23:36.53 | lkcl | * gpe-login: could also send it upstream? (gpe@handhelds.org) |
| 23:36.53 | lkcl | * netbase: could the usb0 use the same IP config as other handhelds |
| 23:36.53 | lkcl | +(192.169.0.201)? That would save |
| 23:36.54 | lkcl | us rewriting docs ;) |
| 23:36.56 | lkcl | * tslib.sh: hardcoding event0 will break everything Linus or Greg change socks. |
| 23:36.58 | lkcl | +See how h2200 and |
| 23:37.00 | lkcl | h3900 use detect-stylus to autodetect the correct pointer device |
| 23:37.05 | lkcl | note the bit about detect-stylus |
| 23:37.23 | lkcl | that'd save the problems we have with blueangel_ts3 and blueangel_kp |
| 23:37.25 | lkcl | hurrah. |
| 23:38.28 | oxi | cp15: i did: cvs tag -b K2-6-12-hh2-xda0-un0 , gave my pass and its now waiting or something |
| 23:38.42 | lkcl | ah _ha_. |
| 23:38.44 | cp15 | Yeah, I have tested detect-stylus. It also brings up the keyboard device if it is first |
| 23:38.51 | cp15 | oxi: May take some time |
| 23:38.57 | oxi | k |
| 23:39.02 | lkcl | TSLIB_TSDEVICE='tetect-stylus --device' |
| 23:39.05 | lkcl | tadaaaa |
| 23:39.41 | *** join/#htc-blueangel bhima (n=gf2e@i13pc168.ilkd.uni-karlsruhe.de) |
| 23:39.47 | cp15 | Yes, I know... But shutdown gpe-dm, do a rmmod himalaya_ts3 and rmmod blueangel_kp. Then insmod blueangel_kp and do a detect-stylus --device |
| 23:40.04 | cp15 | Don't think our keyboard driver is the problem... |
| 23:41.01 | cp15 | The IP address issue is probably valid. But I still want usb0 up per default |
| 23:41.20 | cp15 | It can save lifes when the touchscreen doesnt work :-) |
| 23:41.56 | lkcl | ack. |
| 23:42.21 | lkcl | oh. right. you want me to try it, first? |
| 23:42.31 | lkcl | or... it doesn't work? |
| 23:42.45 | cp15 | It doesn't work... I have already tested it today |
| 23:42.52 | immolo | hey cp15 you thought about adding hibrate to the ba? |
| 23:43.06 | cp15 | But your mileage may vary |
| 23:43.20 | immolo | damn I can't spell |
| 23:43.33 | cp15 | immolo: Yes, but not very useful looking at the ram/flash ratio on the BA |
| 23:44.11 | cp15 | It would only be possible with an SD card, and it will be slower than simply re-opening your apps |
| 23:44.14 | immolo | true, I would just be nice to switch the ba off :P |
| 23:44.36 | cp15 | You can still do it and won't loose any data |
| 23:44.59 | cp15 | You just have to login again after boot |
| 23:45.15 | immolo | not much of an issue |
| 23:45.37 | immolo | it's just when you have a little battery you want to save it for as long as possible |
| 23:46.10 | cp15 | Then the hibernate is the wrong way to go. Saving your ram to flash will consume lot of energy |
| 23:46.26 | oxi | cp15: finnished |
| 23:46.33 | oxi | finished even |
| 23:46.57 | cp15 | Ok, then you can integrate your changes into the checked out kernel |
| 23:47.10 | immolo | cp15, I guess I'lll just have to charge up more :P |
| 23:47.18 | cp15 | If you need to add a directory, do a cvs add directory |
| 23:47.33 | cp15 | If you need to add a file, do a cvs add file |
| 23:47.49 | oxi | phew i gues ill have to study cvs a bit before i can use it :) |
| 23:48.17 | cp15 | When you are done and think you have added and changed everything, go to the linux-2.6-xda directory and do a cvs commit |
| 23:48.42 | cp15 | It will ask you for a description of your changes and will then commit your changes into cvs |
| 23:49.15 | cp15 | Community, I have 2 things to discuss |
| 23:49.19 | lkcl | also, oxi, make sure you do a "cvs diff -u > foo; vim foo" or whatever your favourite editor is. |
| 23:49.24 | oxi | okay but they wont show up in ba branch right? last command was to create a new branch for uni right? |
| 23:49.38 | cp15 | First: cvs handling |
| 23:49.39 | lkcl | keep this open in a separate window whilst you are thingying the wotsit. doing the cvs commit. |
| 23:50.04 | oxi | phew |
| 23:50.45 | cp15 | My plan is to create a branch for every device after the scheme K2-6-12-hh2-xda0-xx0 where xx is a two letter code describing the device (ba=Blueangel, un=Universal etc) |
| 23:51.20 | cp15 | This makes it possible to keep conflicting changes in the repository |
| 23:51.55 | cp15 | After some time we might switch to K2-6-12-hh2-xda1 where we merge in all devices |
| 23:52.25 | cp15 | Any critics, comments, whatever? |
| 23:53.02 | cp15 | oxi: Yes, you will get your own branch, independent from BA, but with the same origin |
| 23:53.20 | oxi | i see |
| 23:53.37 | oxi | sounds like a good idea(for someone whos never used cvs before) :D |
| 23:53.54 | oxi | well gtg |
| 23:53.58 | oxi | gnite all |
| 23:54.02 | lkcl | yep. sounds good. the only thing: everyone keep in touch, keep an eye out for significant overlaps and stuff |
| 23:54.39 | lkcl | the _worst_ thing in the ... never mind. it's a long nasty story involving nearly 150,000 lines of additional code (to a project that initially was only 250,000 lines) |
| 23:54.45 | lkcl | this is kinda different. |
| 23:55.15 | lkcl | has anyone else noticed that the spring is a bit strong on the blueangel SD card slot? |
| 23:55.21 | lkcl | now where did that SD card go.... |
| 23:55.38 | cp15 | Second thing: I have made some changes to gpe programs. Namely bluez-utils-dbus (to let bluetooth work with full speed on BA), gpe-contacts (to make it possible to dial a number from the address book) and keylaunch (to get it much more flexibler) |
| 23:56.01 | lkcl | _great_. |
| 23:56.01 | cp15 | What do you think we should do: |
| 23:56.28 | cp15 | - Put it in Xanadux cvs and make bitbake use that repository |
| 23:56.50 | cp15 | - Create a patch and add this patch to bitbake |
| 23:57.06 | lkcl | i'll just have a word on the #oe list see if pb can commit it, quick |
| 23:57.22 | cp15 | - Try to get it in the official tree (of course..., but probably not a short term solution) |
| 23:57.45 | lkcl | or, are there going to be more? |
| 23:57.50 | cp15 | The thing is that we probably still need more modifications |
| 23:58.07 | lkcl | ack, then definitely xanadux cvs - no wait, how much active work is there on these programs? |
| 23:58.15 | lkcl | aside from what you're up to? |
| 23:58.49 | cp15 | The bluez-utils have to be extended to get our BT scripts going, and also to react to USR1/USR2 when suspending/resuming |
| 23:59.14 | cp15 | The gpe-contacts have to be extended to make it possible to also write an sms from addressbook |
| 23:59.32 | cp15 | And the keylaunch also needs some improvement in a few details, I think |
| 23:59.40 | lkcl | phil says that gpe-contacts has very small traffic; bluez has a lot. |