(02:58:37) thundercle1: miska?
(02:59:41) thundercles: have you tried the kernel patch Intel released a while back, I just got ahold of it off an old site, it claims it has the MMX wifi working
(02:59:58) thundercles: but it's a 2.6.9 patch, which kernel is the hackndev one?
(03:01:48) thundercles: ah appears you used 2.6.22
(03:02:56) thundercles: hrm well I'm gonna see if this patch even can get /dev/wlan0 up, if it can at least we have somewhere to work from rather then trying to develop a whole driver from scratch, which would suck since I don't think Intel's SDK supports that the PXAs anymore since they sold them to Maxell
(03:09:50) Sleep_Walker: sed -e 's#/dev/wlan0#wlan0#' -e 's/Maxell/Marvell/'
(03:18:32) thundercles: oh they have a marvell driver that will get that MMX thing up as wlan0?
(03:30:29) thundercles: oh I c
(03:31:07) thundercles: misunderstood your command, it's 3:30AM :-P
(03:35:28) thundercles: well hello again
(10:39:40) thundercles: miska, I found this commont on osdir, someone wrote a patch for glibc that fixes the fstat64 and mmap problems for those wireless MMX devices, the one that isn't working in the hackndev kernel for the TX. Also, apperently running through it as a coprocessor can speed up media applications, I read a thing on another mobile board where they develop for another device using the PXA270 about routing through it to speed up video playback, so
(10:39:40) thundercles: The iwMMXt is already actually working and recognized in the hackndev kernel, there is support for it in the 2.6.22 kernel, and my log shows it being detected, it's just a matter of finding a patch that we can grab code from or even directly patch the hackndev kernel with to get it to process wifi signals to wlan0
(10:42:27) thundercles: I know that it's being done, it's just a matter of finding a patch, although I think that the new intel driver in the .24 kernel may work for it, it may not, stuff is a little vague there, but it's worth a shot, if it does work we could go ahead and patch up the .24 kernel with whatever patches you guys used on your current .22 one and build the hackndev kernel into .24 with full iwMMXt support and wifi
(10:59:06) thundercles: alright, I checked it out apperently glibc 3.4 is when they added full support for Xscale and iwmmxt, so if you use glibc 3.4 or later then at least the code is being compiled to the full instruction sets
(11:00:16) thundercles: I meant gcc v 3.4 not glibc oops
(11:01:22) thundercles: I compiled with gcc 4.1.2 and the kernel detected the iwmmxt fine though
(11:02:03) thundercles: just doesn't have the support to run it to wlan0 yet:-/, but workin on it :)
(11:11:16) |miska|: /me has some problems with understanding to thundercles
(11:11:35) |miska|: /me didn't get highlighted because thundercles forgot ||
(11:12:28) thundercles: oh sorry, miska
(11:12:35) |miska|: First sentence I don't get at all
(11:12:59) thundercles: I found a post on a good open source website by someone who made new patches for glibc
(11:13:12) thundercles: they are patches for the coprocessor/wifi unit in the tx
(11:13:44) |miska|: Patched glibc makes wifi work?
(11:13:57) thundercles: they wont get the wifi working but they will help resolve that issue between the coprocessor/wifi and the actual main proc using different arch types
(11:14:04) thundercles: well not yet, but it's a start
(11:14:36) thundercles: but my main question for you was what version of gcc did you compile that kernel on your site in?
(11:15:01) |miska|: Something really old...
(11:15:05) thundercles: because I don't remember it finding the iwMMXt
(11:15:39) thundercles: yeah, try compiling it again in gcc 3.4 or later and the kernel will at least find the wifi card/coprocessor, it'll show up as iwMMXt
(11:16:05) |miska|: It will find wifi?
(11:16:11) thundercles: I can't get it tied to wlan0 yet, it still needs actual wifi drivers working for it, but I'm working on that, but the hardware is all ready
(11:16:19) thundercles: well the wifi card is a coprocessor
(11:16:26) thundercles: based on the pentium MMX technology
(11:16:45) |miska|: I tought what Marex found wifi already attached to PCMCIA...
(11:17:18) thundercles: yeah, he probably used a gcc 3.4 compiled kernel or just probed around the pcmcia bus
(11:17:46) thundercles: but it probably wont work without a gcc 3.4 compiled kernel because that one handles the issue of the different archetecture types
(11:17:54) thundercles: so you might want to upgrade your gcc
(11:18:09) thundercles: well 3.4 or above, I think mine is 4.something
(11:18:56) thundercles: but I think in the 2.6.24 kernel I may be able to get wifi to work, they put a ton of new wifi drivers in the 2.6.24
(11:19:41) |miska|: Why something as old as 2.6.24?
(11:20:04) thundercles: I think 2.6.24 is the newest kernel
(11:20:22) thundercles: well the newest stable I mean
(11:20:24) |miska|: nope, stable is 2.6.25
(11:21:45) thundercles: well, then we could use .25 doesn't really matter, I was just wondering what patches you applied to that .22 kernel source that hackndev uses, because I think that the TX could get a real boost from using the 2.6.24 (or 25) kernel, but that kernel that hackndev produced appears to be at least fairly heavily patched and you don't want to loose all that functionality either
(11:23:05) thundercles: but if all else fails, I got a kernel patch that intel developers produced way back when the TX very first came out
(11:23:06) |miska|: hackndev kernel is basically patched hh.org kernel ;-)
(11:23:23) thundercles: so hh.org has all the patches?
(11:23:31) |miska|: And what is that patch for?
(11:23:34) |miska|: No
(11:24:21) |miska|: hh.org is branch of kernel with a lot of patches and we added some other patches to it
(11:24:32) |miska|: And the result is in our git ;-)
(11:24:38) thundercles: the intel patch is for a lot of stuff on the TX, but it's for the 2.6.9 kernel so most of the patches are depreciated somehow
(11:25:07) thundercles: so if we ever did have to make any drivers for the TX, we could probably yank a LOT of code from that patch however
(11:26:15) thundercles: I know that the wifi works for the PXA270 in linux, I just am having problems finding a place that both claims they have it working and has good patches readily available, but chances are there is one.
(11:26:15) |miska|: And what does that patch for TX do?
(11:26:56) |miska|: Problem isn't wifi in general, but TX wifi ;-)
(11:26:59) thundercles: well if I can get ahold of my cousin and get access back to our webserver (I lost my RSA key in a hard drive failure) I can post some of that stuff up for you
(11:28:24) thundercles: yeah, but the problem is a lot narrower then you think, basically all the hardware is working, all the net drives are working, we just have to get a patch that will communicate to the hardware specifically for 802.11b and bind it to wlan0
(11:28:45) thundercles: hell it may even be working to the point where we could get it working just by doing some fancy stuff in the proc
(11:29:40) thundercles: cause like I said, my log on my TX shows the coprocessor that runs the wifi when I boot with the kernel I compiled in gcc 4.1
(11:30:12) |miska|: Don't think that WiFi is coprocessor
(11:30:18) thundercles: so it's recognized and working, it just happens to do more then just wlan stuff, so we need to get it mapped to wlan0 and possibly help it communicate a little bit
(11:30:21) thundercles: it isn't usaully
(11:30:25) thundercles: but it is in the TX it's weird
(11:30:35) thundercles: here I'll grab a link
(11:31:22) thundercles: from en wikipedia:
(11:31:22) thundercles: Intel's XScale processors starting with PXA270 include an extension to the ARM core called iwMMXt whose functions are similar to those of the IA-32 MMX extension. IwMMXt stands for "Intel Wireless MMX Technology". It provides arithmetic and logic operations on 64-bit integer numbers (the software may choose to instead perform two 32-bit, four 16-bit or eight 8-bit operations in a single instruction). The extension contains 16 data regis
(11:32:17) |miska|: Yeah, but that's not a WiFi
(11:32:30) thundercles: it handles the wifi though
(11:32:46) thundercles: the wifi is connected to that, not the ARM
(11:33:20) |miska|: This is extension which can speedup multimedias - mplayer use this
(11:33:37) thundercles: indeed it does
(11:33:44) thundercles: but it also handles the wifi, I'll find another link
(11:34:55) thundercles: hrm you might be right actually
(11:35:15) thundercles: but usaully they mention it right with the wifi capability though
(11:36:06) thundercles: but anyways I am thinking that the new intel drivers in the .24/.25 kernel will get it to work though
(11:37:01) thundercles: there are a bunch, including one intel that's just like a general intel wifi driver, they compiled support for mutiple intel interfaces into that one driver and I'm hoping it's just in there
(11:38:21) thundercles: it could be named as a marvell wifi controller too since they own that arch now, I'm just going to compile all the intel and marvell wifi modules in and see if one of them loads, move on from there at this point
(11:38:37) thundercles: but in .25 since they have a lot of drivers
(11:41:14) thundercles: but I'll try another lspcmcia in just a second when I have the thing booted up and see if I can find anything weird that could be a seperate wifi controller
(11:46:09) thundercles: oh yeah, and did you try compiling in the advanced power management as a module on the kernel and there was a problem? 'cause if not I'm going to see how it works on my next compile, just about the only bugs I have been getting have all been related to the power management
(11:49:17) |miska|: Don't know how am I compiling it, but I didn't had any error....
(11:49:43) thundercles: yeah it might be my cross-compiler, I found a better one for my next compile
(11:50:15) thundercles: one that actually will compile specifically for pxa270 and not for something similar and just work
(11:53:02) thundercles: But I got the newest kernel with the arm patch from the arm.com.uk or whatever site's git repo
(11:53:25) thundercles: are the hackndev patches available, I'd like to see how well they patch into the new kernel, they are for .22 so they will probably mostly patch in
(11:53:55) thundercles: and I'll go grab a bunch of the hh.org patches too and see what crazyness that cooks up
(11:54:22) thundercles: hopefully working wifi at least with that new intel wifi driver they introduced in .24
(11:58:56) thundercles: but really there are so many other devices running linux that also have the pxa270, we're bound to be able to grab a working module out of one of them that will bring up wlan0
(11:59:20) thundercles: if the new kernel compile/new cross-compiler doesn't work
(12:02:10) |miska|: Problem is that none of them has working driver for the same WiFi chip ;-)
(12:02:55) thundercles: yeah, well I need to compile my own lspcmcia then
(12:03:08) thundercles: because the one in my angstrom isn't pulling up anything
(12:04:08) thundercles: actually I had a really good tool I had to install seperatly but it just spouted off everything about the hardware on my PC, I wonder if they have an arm version I could stick on there
(12:05:27) thundercles: the tool was lshw
(13:30:21) Terinjokes: any progress on wifi?
(13:31:53) Terinjokes: hey sea1!
(13:32:25) sea1: hi
(13:32:39) Terinjokes: how's it going?
(13:32:52) sea1: good. )
(13:33:29) Terinjokes: cool! can l ask what device you use?
(13:33:41) sea1: sure PalmTX
(13:34:46) Terinjokes: cool, same here.... battery is going out, audio already did, nevertheless, still struggles by
(13:36:22) sea1: Do you mean that you are typing messages with it?
(13:37:29) thundercles: miska I'm compiling an xscale specific cross-compiler, the thing I'm using to make it is really good it makes cross compilers for anything check out the site:
(13:37:29) thundercles: http://kegel.com/crosstool/
(13:38:08) Terinjokes: No, the hardware in general (using upIRC in PalmOS atm)
(13:38:18) sea1: got you
(13:39:05) sea1: i'm not with tx right now at all - opera from windows, but i use it very often
(13:40:23) Terinjokes: i really like the linux on palm idea, but without a network connection, it's good only for show....
(13:41:25) sea1: totally agreed )
(13:41:40) sea1:  as i understand it is still possible to go to internet using BT and comp as gateway
(13:41:42) Terinjokes: imho, once it gets stable network support, the rest of the linu.
(13:42:08) Terinjokes: linux system will fall into place
(13:42:29) sea1: i think so
(13:42:39) Terinjokes: i tried it on my system.... didn't work
(13:42:48) sea1: BT?
(13:42:53) Terinjokes: yeah
(13:44:01) sea1: i managed to activate it and was able to see it from other computers, and to see others right in opie. but did not get any further yet
(13:44:13) Terinjokes: after network, i see lots of annoyed palmos developers help put the rest off the system together
(13:48:08) Terinjokes: sea1: i could scan bt in opie, but never could properly connect
(13:49:14) sea1: same here. though i read somewhere that you need something to do in linux running on peer
(13:49:39) Terinjokes: my bo.
(13:50:48) Terinjokes: my box said the palm was connected, and the network was being shared, but no network on palm
(13:51:28) Terinjokes: (to be fair, didn't work in PalmOS either)
(13:52:06) sea1: no way. winfilepro is the key )
(13:53:03) sea1: another annoying thing in opie - BT and network setting not kept, i had to enter everythime
(13:54:07) Terinjokes: sea1, i mean an internet connection on bt, a local subnet (palm and box) is avaibile, WFP sees this
(13:56:17) Terinjokes: bbl
(15:21:45) flyback: /me heads on outside to get his mind off the bullshit
(16:31:07) Marex: my my ... disassembling palmld bootloader is so much fun
(16:31:22) Marex: and I can clearly see what GPIOs do what 8)
(16:55:44) thundercles: Marex
(16:55:54) thundercles: do you know where I can find source code for lshw?
(16:55:56) Marex: no
(16:56:06) thundercles: okay then
(16:56:23) thundercles: it's pretty good for probing hardware, I'll find some source for it somewhere
(17:12:00) Lars_G: /me smashes his head on a wall
(17:19:52) flyback: /me is suffering a total emotional breakdown tonight, stand clear
(17:32:30) |miska|: Lars_G: Making some room for new window?
(17:37:33) flyback: /me decides to finally put the extra duron 850 cpu and mb with crappy caps together into a media pc so we can drown his sorrows in streaming movies and anime
(17:41:14) Marex: my my ... looks like we will have load of new correct stuff in -hh
(17:41:18) Lars_G: |miska|: a new archway
(17:45:28) |miska|: /me is starting to get hungry (2:45AM local time)
(18:02:12) flyback: /me brb
(18:02:36) Marex: woohoo ...
(18:02:52) Marex: got function call which can be considered "hw_init()"
(18:09:53) Marex: clock configuration ... wow
(18:10:13) Marex: maybe this way palmos can be convinced to stop whining ;D
(18:11:09) Marex: what clocks to enable at boot ... all available ... palm is really awesome
(18:36:35) |miska|: gn
(20:09:41) flyback: /me goes to find his inhaler
(20:41:33) flyback: /me clubs seal
(20:41:35) flyback: CANUCK
(20:41:35) flyback: CANUCK
(20:46:20) sea1: hi
(20:47:14) flyback: hi :P