irclog2html for #htc-linux on 20061021

00:08.29*** join/#htc-linux psokolovsky_ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com)
00:19.28*** join/#htc-linux psokolovsky_ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com)
02:08.10*** join/#htc-linux asylumed_ (n=insanity@196.211.28.91)
03:03.57*** join/#htc-linux bb (n=bbaniste@w160186.wireless.fsr.net)
03:58.09*** join/#htc-linux asylumed (n=insanity@196.211.116.6)
05:15.04*** join/#htc-linux vnevoa (n=vasco@a83-132-50-19.cpe.netcabo.pt)
05:15.05*** join/#htc-linux rejon (n=rejon@64-121-195-22.c3-0.sfpo-ubr4.sfrn-sfpo.ca.cable.rcn.com) [NETSPLIT VICTIM]
05:15.05*** join/#htc-linux derekS (n=DerekS@unaffiliated/dereks) [NETSPLIT VICTIM]
05:15.05*** join/#htc-linux iggy (n=iggy@gentoo/developer/iggy) [NETSPLIT VICTIM]
05:15.06*** join/#htc-linux gw280 (i=authdeni@vm.gwright.org.uk) [NETSPLIT VICTIM]
05:15.06*** join/#htc-linux Kevin2 (n=Kevin@207-237-52-30.c3-0.avec-ubr12.nyr-avec.ny.cable.rcn.com) [NETSPLIT VICTIM]
05:16.23*** join/#htc-linux lpotter (n=ljp@c210-49-127-23.rochd1.qld.optusnet.com.au)
06:04.45*** join/#htc-linux RoEn_PC (n=roen@p54A6792F.dip.t-dialin.net)
06:39.32*** join/#htc-linux Ro_En_PC (n=roen@p54A6558E.dip.t-dialin.net)
07:22.37*** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com)
07:25.44*** join/#htc-linux Ro_En_PC (n=roen@p54A6558E.dip.t-dialin.net) [NETSPLIT VICTIM]
07:25.44*** join/#htc-linux lpotter (n=ljp@c210-49-127-23.rochd1.qld.optusnet.com.au)
07:25.45*** join/#htc-linux vnevoa (n=vasco@a83-132-50-19.cpe.netcabo.pt)
07:25.45*** join/#htc-linux rejon (n=rejon@64-121-195-22.c3-0.sfpo-ubr4.sfrn-sfpo.ca.cable.rcn.com) [NETSPLIT VICTIM]
07:25.45*** join/#htc-linux derekS (n=DerekS@unaffiliated/dereks) [NETSPLIT VICTIM]
07:25.45*** join/#htc-linux iggy (n=iggy@gentoo/developer/iggy) [NETSPLIT VICTIM]
07:25.45*** join/#htc-linux gw280 (i=authdeni@vm.gwright.org.uk)
07:25.46*** join/#htc-linux Kevin2 (n=Kevin@207-237-52-30.c3-0.avec-ubr12.nyr-avec.ny.cable.rcn.com) [NETSPLIT VICTIM]
08:14.38*** join/#htc-linux pleemans (n=peter@d51A5E421.access.telenet.be)
08:51.38*** join/#htc-linux Ro__En_PC (n=roen@p54A670AC.dip.t-dialin.net)
09:01.40*** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com)
09:36.10*** join/#htc-linux AlphaPi[jr] (n=a@host173-154-dynamic.58-82-r.retail.telecomitalia.it)
09:36.23*** join/#htc-linux senaqueue (n=scarface@154.16.240.220.dynamic.dsl.rns02-kent-syd.comindico.com.au)
09:36.31AlphaPi[jr]hallo
09:36.45goxboxlivehi
09:37.16AlphaPi[jr]hi goxboxlive
09:37.42AlphaPi[jr]can you tell me if nbfutils can also manage ipaq firmware file?
09:39.09goxboxlivei I have no idea
09:40.45AlphaPi[jr]another question
09:40.55AlphaPi[jr]osloader image can be uploaded via jtag?
09:41.19goxboxliveMan, i have no idea. This is way over my knowledge
09:41.26goxboxliveWich distro are you using?
09:42.50AlphaPi[jr]no distro
09:43.00goxboxliveok
09:43.13AlphaPi[jr]i 've fucked up an ipaq 1940 trying upgrading firmware
09:43.25AlphaPi[jr]by hp upgrade firmware tools
09:43.55AlphaPi[jr]it runs the bastard-inside-exe "EnterBootLoader.exe"
09:44.21AlphaPi[jr]and now it is blocked onto "ipaq 1.07" bootloader screen
09:45.06goxboxliveok
10:07.53*** join/#htc-linux pH5 (n=ph5@p5485F734.dip.t-dialin.net)
10:14.59*** join/#htc-linux lkcl (n=lkcl@bb-87-82-5-55.ukonline.co.uk)
10:15.51lkclipaq serial cable arrived.  hurrah.
10:27.24AlphaPi[jr]lkcl why u neeed the ipaq serial cable?
10:39.35*** join/#htc-linux jeanseb (n=jeanseb@88.164.32.155)
10:39.53AlphaPi[jr]lkcl....
10:42.55*** join/#htc-linux LunohoD_ (n=alex@e180119100.adsl.alicedsl.de)
11:55.43*** join/#htc-linux AlphaPi[jr] (n=a@host182-116-dynamic.60-82-r.retail.telecomitalia.it)
11:59.42*** join/#htc-linux florian (n=fuchs@p5083981A.dip0.t-ipconnect.de)
12:31.14*** join/#htc-linux skodde (n=skodde@unaffiliated/skodde)
13:03.08*** join/#htc-linux rob_w (n=bob@p85.212.170.192.tisdip.tiscali.de)
13:04.36*** join/#htc-linux RoEn_PC (n=roen@p54A670AC.dip.t-dialin.net)
13:07.22*** join/#htc-linux asylumed_ (n=insanity@196.211.28.91)
13:30.07*** join/#htc-linux RoEn_PC (n=roen@p54A670AC.dip.t-dialin.net)
14:13.15*** join/#htc-linux rob_w (n=bob@p85.212.170.192.tisdip.tiscali.de)
14:13.54Kevin2Hi
14:29.19Kevin2Anyone awake?
14:45.50pH5hey Kevin2
14:56.09Kevin2Hi pH5.
14:59.20Kevin2Can you try the latest haret on the magician?  It is in svn: svn://www.shadowmite.com/home/svn/ppc6700/haret
15:01.07pH5gharet-20061021.patch?
15:01.52pH5is there something special you'd like me to try?
15:02.40Kevin2Booting linux. You can grab the exe if you don't want to compile it.
15:04.16Kevin2I rewrote the linux booting code.  It works on apache, but I can't test other platforms.
15:18.20*** join/#htc-linux pH5_ (n=ph5@p5485E6B4.dip.t-dialin.net)
15:19.57pH5Kevin2: I tried to boot linux with boot2 and it reliably locks up
15:20.20Kevin2Ughh.  Thanks.
15:20.30Kevin2Can you forward a log of the event?  Do you see colored bars on the screen?
15:21.16pH5I see some green and black in the middle. I'll have to look for a cable to get to a log
15:21.58Kevin2You can create a log by adding "log x" to your start script.  Or, you can create a file "earlyharetlog.txt" in the same directory as haret.
15:22.08Kevin2Did it say "detected machine magician" at startup?
15:22.44pH5yes
15:25.05pH5Kevin2: Go Go Go... is the last thing I see via the haret telnet connection
15:25.40pH5http://en.pastebin.ca/213652
15:26.14*** join/#htc-linux rob__w (n=bob@p85.212.144.65.tisdip.tiscali.de)
15:26.19pH5hi rob__w
15:28.39rob__whi pH5
15:39.25Kevin2ph5: Is the magician a pxa27x cpu?
15:39.54Kevin2Does it have 64M of ram?
15:40.07pH5Kevin2: yes. pxa272 with 64M.
15:51.39*** join/#htc-linux booba (n=booba@AAnnecy-256-1-213-73.w90-9.abo.wanadoo.fr)
16:14.08*** join/#htc-linux WizMaui (n=WizMaui@62.112.90.179)
16:17.21Kevin2ph5: Can you grab the latest binary from svn and try again?
16:56.47Kevin2WizMaui: Can you grab the latest haret from svn (svn://www.shadowmite.com/home/svn/ppc6700/haret) and try booting linux again on the Wizard?  If it doesn't work, try creating a file "earlyharetlog.txt" in the same directory as the haret exe, and then post "haretlog.txt" to pastebin.ca
16:57.57pH5Kevin2: I see yellow, blue, red, cyan and a black line, and it hangs.
16:59.57pH5http://en.pastebin.ca/213786
17:02.37Kevin2pH5: Thanks.  Dumb question - are you sure your zImage is good?  We got much further this time.  The only thing left after that color sequence is the jump to Linux.
17:04.17pH5Kevin2: Yes, I'm quite sure about that. It boots fine with xanadux cvs gnuharet.
17:05.40Kevin2Okay.  Thanks.  I'll look again.
17:06.40WizMauiok...will try it too
17:06.40WizMauionce i can find my wizard ;)
17:06.42WizMauitime to tidy my room up
17:11.48*** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com)
17:12.31Kevin2Hi goxboxlive:  Can you run a haret test?  (Grab latest haret from svn at svn://www.shadowmite.com/home/svn/ppc6700/haret and try to boot linux.)
17:13.56goxboxliveHi Kevin2. I have to do that later. Have to put my child to bed first
17:17.32*** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net)
17:23.51pH5btw, is anybody actively working on htc hermes?
17:41.00lkclAlphaPi[jr] - just in case you check the logs: i need the serial cable because when trying to boot linux on the ipaq hw6915, i get nothing on-screen.  the serial cable allows me to run the console on serial - and thereby get some lovely debug info.  or actually start writing debug info out - by hand - if i get really desperate
17:41.19lkclback later darlings..
17:44.22*** join/#htc-linux jeanseb (n=jeanseb@88.164.32.155)
17:46.03Kevin2pH5: I put a new binary in svn.  This latest version disables the BTB.  I'm not very hopeful it will fix your problem, however.
18:16.41cr2Kevin2: hi. your patch is really huge. what do you think about applying it in pieces ?
18:17.34cr2lkcl: you should find out how and which serial ports are used, before you'll be able to debug over serial :)
18:21.44pH5hey cr2, how is the universal phone coming along?
18:23.29cr2pH5: it's an extremely complex thing to control ;-) i'm already thinking about looking at the radio control code itself. given some more time...
18:24.44cr2i can probably clone the bootloader behaviour and will be able to talk AT, but it will only shift the problems one step further.
18:25.30cr2btw, you can grab the latest haret, and check your IR and RIL code too.
18:26.16cr2hmm. do you know your RIL_devspecific functions ? i fear that 51 is universal-specific
18:26.35cr2http://wiki.xda-developers.com/index.php?pagename=rilgsm.dll
18:26.38cr2at the bottom
18:29.08cr2goxboxlive: i have tried to send you the opie image yesterday, but you don't accept 25MB e-mails ;) we need to find another way.
18:29.44cr2asylumed: what about putting the blueangel patches in the hh.org CVS ?
19:06.15pH5Kevin2: as you expected, your last binary still shows the same problem.
19:06.26pH5cr2: how do I use the new IR / RIL code in haret?
19:10.31goxboxlivecr2: I have tested autodim and sleep. Non of them are working even with the new backlight driver. The sleep function did go to sleep, but it wouldnt wake up. Had to do a soft reset.
19:16.34*** join/#htc-linux rob_w (n=bob@p85.212.182.112.tisdip.tiscali.de)
19:17.10asylumedcr2: i'm keen... where? and who could i get to have a look over them? (since i'm rather new to this?)
19:18.35asylumedrob_w: seems you can be flicker to stop in the new w100 driver by tweaking some values...
19:18.52asylumeds/be/get/
19:20.02pH5asylumed: If you want more people to look at them, you can always send to them kernel-discuss@handhelds.org
19:20.39asylumedthx... will do...
19:24.08cr2pH5: you open two telnet connections, run the commands 'irctl 1' 'irctl 0' 'tstril' and run wgpio in the other session.
19:24.56cr2goxboxlive: i have added the sleep (htcuniversal_pm.c) driver to hh.org CVS, but it's not linked into the kernel insfrastructure.
19:25.14cr2so it will not really wake up.
19:25.40rob_wasylumed, which values ?
19:26.22cr2asylumed: you can post them somewhere. are   you planning to do some more hardware tracing ?
19:27.50cr2goxboxlive: i have traced the both camera gpios. the big camera is relatively simple. the small one has a flipping gpio, it will probably require some disassembling  to understand it's function.
19:29.03asylumedrob_w: .pll_freq and .pixclk_divider
19:29.31rob_wstrangly   i didnt found values differing from the .12 driver to .16
19:29.32asylumedrob_w: not totally gone, but *greatly* reduced...
19:30.28rob_wi was able to remove it totally by backporting the w100fb to .12 .. so i assumed it is a combination on how the w100fb calculates the lcd sizes
19:30.30asylumedrob_w: yeah, noticed it too but i've see a discussion about it somewhere and tried the values that they suggested
19:31.31rob_wbut thats great news maybe now we can find the difference between .12 and .16 better and make it nativly work.  16 then could bring us 3d support !!
19:32.01asylumedrob_w: keef
19:33.18rob_wpeople improved the .16 w100fb massivly  .. i would love to leverage that .. maybe then we could great a nice fancy ui on such devices
19:33.26asylumedrob_w: although, since the resume functions have been merged, i'm puzzeled on how to get the chips to enable in order properly... :(
19:34.04rob_wsorry i need to go for today
19:36.11cr2asylumed: what are the capabilities of the w100fb driver ?
19:36.35asylumedcr2: haven't looked into it fully i'm afraid...
19:36.49cr2is it capable of hw portrait/landscape switching ? IDCT ? video grab ?
19:37.04cr2what 3D does rob_w mean ???
19:37.23asylumedcr2: hehe, not entirely sure/convinced yet...
19:37.43goxboxlivecr2: Ok camera sounds fun, and the sleep driver, why dont you link it?
19:38.23cr2btw, you can try to catch the reads/writes to the control registers with haret's 'wi'
19:38.29pH5cr2: do I need an inserted sim card for the ril test? it fails with "DSRIL: error 183".
19:39.08cr2goxboxlive: because i'm completely overloaded with other things ;)
19:39.24goxboxlive:-)
19:40.31cr2pH5: does ril_initialize cause some gpio action ? DSRIL means an error from the "devspecific" function to send and arbitrary AT command to the phone. do you have this function ?
19:41.41cr2goxboxlive: the "linking" is rather generic, it should be the same as in the h4000 code.
19:41.42Kevin2cr2: I can break the patch up, but if the Linux boot doesn't work, there isn't much point in doing it.  (Work on other platforms, it works fine for me.)
19:42.17Kevin2In order to get the easy stuff in, I can break the patch into two parts - the cleanup stuff, and the linboot stuff.
19:42.21goxboxliveok, wich files do i need to compare?
19:43.24Kevin2cr2: Can you try the latest haret from svn on Universal?  I'm curious if it works like Apache, or hangs at the sample place as Magician.
19:44.06pH5cr2: there is no gpio activity, and I didn't spend enough time reading the rilgsm disassembly to know about the devspecific function yet
19:44.27pH5Kevin2: I think it's not your patch - I just noticed that current cvs hangs, too.
19:44.38pH5My working version was from sept.16
19:45.04cr2Kevin2: ok. i'll try
19:45.44cr2pH5: if you phone is not (just) working, you can't avoid studying the rilgsm.dll...
19:49.30cr2pH5: have you seen that the demo IDA version works in wine ?
19:52.23pH5cr2: right about that rilgsm.dll. I wish it wouldn't take such an awful amount of time to understand anything.
19:53.06pH5cr2: IDA, did you try it yet?
19:54.20Kevin2pH5: Your last checkout from CVS was on 20060915?  Or is that the version from the exe?
19:54.31cr2Kevin2: dead after GOGOGO. the "color stripes" look very strange.
19:54.57Kevin2What colors do you see?
19:55.22cr220061021112*
19:56.09pH5Kevin2: the 20060916 build (and I think I always updated before building) works, the 20061021 one doesn't.
19:56.27Kevin2cr2: You used gnuharet-200610211341.exe?
19:56.45cr2i have the narrow magenta+something at +32000 pixels as usual, then the black wide stripe in the middle ? with some small pixels inside.
19:57.19cr2the dpi is very fine on universal, so i can't tell you the single pixel color.
19:58.05cr2yes. gnuharet-200610211341.exe
19:59.18cr2pH5: i have tried, and it works. but you can't save the output, maybe only make a screenshot ;) and there is a time limit, don't know of what nature. the output looks very good.
19:59.41cr2i have tested it on the nk.exe
20:00.58pH5cr2: that sounds tedious but promising :)
20:01.08Kevin2pH5 - the only thing noteworthy I can see in that time frame is the '-O' flags was added to the build.
20:02.52cr2pH5: maybe cut and paste works, but i'm not a windows expert ;)
20:03.28Kevin2cr2: The black wide stripe is probably the preloader.  It gets copied to the top of ram which is where your fb is.  Does the magenta stripe come anywhere near the black one?
20:04.25cr2no. the screen is 480x640, so you can guess where is the middle.
20:05.45cr2i don't know if my fb is really contiguous on the cpu address bus.
20:06.25cr2one part is in the pxa SRAM and the other one at the end of RAM.
20:06.53cr2lkcl was speculating about some possible asic bus translator, but i don't know.
20:06.58Kevin2Is your screen 16bit color?
20:07.05cr2yes.
20:07.11cr2i think all of them are.
20:07.25cr2otherwise m$ will have problems ;)
20:07.42Kevin2640*480*2 == 614400 which should fit in the 1M page set aside for your fb..
20:08.08cr2SRAM works faster.
20:08.35cr2so they decided to use a "combined" fb.
20:08.50cr2SRAM alone is too small for 480x640.
20:09.07cr2but it is used on all 240x320 and 240x240 phones.
20:10.49pH5Kevin2: thanks for that hint, after removing -O from gnuwince.smak, the kernel boots again
20:14.16Kevin2pH5: That is a bit scary..
20:14.42Kevin2cr2: I've pushed a new binary to svn.  This one has bigger lines.  :-)  Now they go for 2500 pixels.
20:20.55cr2now it works :)
20:21.43Kevin2*boggle*
20:42.02Kevin2pH5: I dropped a non optimized binary on svn if you want to try that one.
20:42.15pH5Kevin2: thanks, I'll try
20:48.58pH5hm.. gnuharet-200610211611.exe has the long coloured stripes, but it still hangs
20:51.50Kevin2Did you try: gnuharet-noopt.exe ?
20:53.23pH5whoops, I missed that one. one min..
20:53.29Kevin2To confirm - you see yellow, blue, red, cyan, black?
20:53.46pH5yes
20:54.29pH5ok, gnuharet-noopt boots. there is a very short delay visible after red now.
20:57.23Kevin2After red is when we relocate the kernel, so the slower memcpy without optimization makes sense.
20:57.43Kevin2I'm really confused why optimization would change the boot for you..
20:59.23pH5Kevin2: I remember being equally confused by the erratic boot behaviour before haret learned to stop pxa27x dma properly.
20:59.59Kevin2I was just thinking that.  Maybe we aren't ensuring that all DMA has completed.  The slower bootup might be hiding this.
21:18.08*** join/#htc-linux asylumed (n=insanity@196.211.28.91)
21:49.05*** join/#htc-linux RoEn_PC (n=roen@p54A670AC.dip.t-dialin.net)
22:21.27Kevin2pH5: I dropped a new binary in svn: gnuharet-200610211819.exe -- this one alters the dma shutdown to confirm the descriptors have stopped before continuing.  It also has -O back on.
22:26.16pH5Kevin2: unfortunately that didn't help :-/
22:33.24Kevin2Still get all the colors?
22:34.30Kevin2Oh well, I can't think of anything that would work on my cpu but somehow fail on yours.  I guess we could put a one second sleep in there just to see if it is timing related.
22:43.52Kevin2Hrmm.  I reverted my dma changes and added a new version that adds a delay (long - could be 10 seconds).  It is in svn.  You should see the green bar for a good long time, and then see the rest.
22:44.27Kevin2pH5: If this doesn't work, you may have to live without optimized binaries.
22:44.45pH5Kevin2: yeah, I still saw all the colors. I'll try the delayed version now.
22:46.55pH5no way. locked up again.
22:48.55Kevin2You saw all the colors?  There was a notable delay after green?
22:50.28Kevin2pH5 - you don't use an initrd, right?
22:51.19pH5Kevin2: yes, yes, and yes.
22:56.39pH5btw, the new binaries don't work on the hx4700 either because of a missing library, most probably RIL.dll
23:02.29Kevin2pH5: cr2 did that, I'm not sure what it involves.
23:15.30lkclthe full version of IDA pro works under wine, as well.  i can probably let one person ssh in to my computer, with ssh forwarding, to run it, if you like.
23:17.11lkclthat's assuming that any of you egg-heads use linux...
23:32.37lpotterwhats linux?
23:45.41*** join/#htc-linux NickBee (n=NickBee@dialpool-210-214-80-105.maa.sify.net)
23:53.05pH5~linux
23:53.06apthmm... linux is the cure for cancer, AIDS and slavery to corporations
23:54.33lkclhey are you mr trolltech lpotter?
23:57.49*** join/#htc-linux senaqueue (n=scarface@adsl-020.isp.net.au)

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.