IRC log for #htc-linux on 20070326

00:24.30*** join/#htc-linux Foxdie (i=Foxdie@77-98-48-193.cable.ubr05.harb.blueyonder.co.uk)
01:27.43*** join/#htc-linux rob-ix (n=rob@unaffiliated/rob-ix)
02:04.33*** join/#htc-linux drbombay[lap] (n=drbombay@pdpc/supporter/active/drbombay)
02:46.27Kevin2psokolovsky / bd2 : Something seems to be wrong with microsd on the apache.  It doesn't look to be just a bad sector, so testing is going to take some more time.
02:48.40Kevin2mmcblk0: error 2 transferring data
02:48.43Kevin2and then
02:49.04Kevin2mmcblk0: error 1 sending read/write command
03:01.13bd2Kevin2, no problem. When Paul will commit ADC stuff, and if it will have any defect in ad7877, I'll gladly assist you with fixing it, if needed.
03:11.29Kevin2Ughh.  second microsd card also fails.  It's either a kernel problem or I something on the hardware side isn't right.
04:31.35*** join/#htc-linux TeringTuby (n=maarten@188-116-dyn.ipact.nl)
05:41.54*** join/#htc-linux TeringTu1y (n=maarten@188-116-dyn.ipact.nl)
05:57.44*** join/#htc-linux _law_ (n=law@mail.stiftadmont.at)
06:02.32*** join/#htc-linux psokolovsky__ (n=psokolov@82.193.98.7)
06:08.31*** join/#htc-linux rob_w (n=rob_w@p549ba14d.dip0.t-ipconnect.de)
06:17.35*** part/#htc-linux manveru (n=manveru@pq163.opt2.point.ne.jp)
06:38.40*** join/#htc-linux _law_ (n=law@mail.stiftadmont.at)
06:48.18*** join/#htc-linux psokolovsky (n=psokolov@82.193.98.7)
06:50.58*** join/#htc-linux psokolovsky (n=psokolov@82.193.98.7)
07:10.32*** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk)
07:13.49*** join/#htc-linux pH5 (n=ph5@p5485ecf8.dip.t-dialin.net)
07:18.51*** join/#htc-linux TripleDE1 (n=sergio@89.129.155.39)
07:32.21mndctrlExcept the universal and blueangel, wich of the htc phones with expandable keyboards runs linux most stable these days?
07:32.25mndctrlany of them at all?
07:51.35cr2mndctrl: hermes still does not work, and athena does not work yet.
07:56.10*** join/#htc-linux rmoravci1 (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk)
07:58.02*** join/#htc-linux TBBle (n=tbble@202.55.155.91)
07:59.21pH5cr2: is there some information about the multiplexer protocol of the blueangel in the wikis? I'm looking for the channel numbers besides 0x16 (at cmd)
08:03.50cr2pH5: no. afaik only for wallaby
08:04.52pH5ok
08:05.54*** join/#htc-linux TripleDE1 (n=sergio@89.129.155.39)
08:07.00*** join/#htc-linux TBBle (n=tbble@202.55.155.91)
08:14.26psokolovskyHi!
08:14.45psokolovskypH5: I wonder, did you test bd2's adc classdev patches?
08:16.34mndctrlcr2: how about the wizard or apache?
08:16.40pH5hi psokolovsky!
08:17.01pH5no, I still need to figure out why the new ts driver doesn't work to get more motivated there.
08:17.08pH5I'm currently playing with the GSM modem.
08:17.50cr2mndctrl: apache is a cdma phone, and it does not work. wizard is an omap850 phone, it will be difficult to develop a full-fledged linux installation for it (no CPU docs)
08:18.16mndctrlah... so that sucks ;)
08:18.42mndctrlso basically your saying: go get a universal?
08:19.23psokolovskypH5: well, I see TS handling regressions with it. I'm really stumped what to do about them - Anton insist that they are right and should be committed, but they are barely tested by someone but him. Any suggestions?
08:19.29cr2yes. universal will be the best HTC free software phone for a long time. IM(not very H)O
08:20.05mndctrlhehe...
08:20.32*** join/#htc-linux TripleDE1 (n=sergio@89.129.155.39)
08:20.38cr2i mean, if you need 100% hardware support.
08:21.05cr2and a modern umts device.
08:21.51mndctrlwell I don't, but would be nice to have some hardware working ;) the fact that the wifi is up and running is a real plus...
08:22.05mndctrlcr2: isn't it a bit big to carry around though?
08:22.14cr2yes, and all the hardware is documented.
08:22.32cr2it is big.
08:23.19mndctrlhmm... how is work on hermes progressing?
08:24.35pH5psokolovsky: I'm a bit worried because we have so many structural changes lately and I don't really keep up, currently. What regressions do you see?
08:26.25pH5psokolovsky: btw, did you see my last mail about defconfigman?
08:26.26psokolovskypH5: on h4000, TS press causes IRQ loop - infinite, but not causing noticeable performance degrade, and sometimes stoppable by new clicks
08:26.59psokolovskypH5: yes, thanks much, will process it. just now all cycles take this adc merge ;-(
08:27.00cr2mndctrl: it's a substantially different CPU, so it's not so easy as i would wish.
08:27.34pH5psokolovsky: should I test current cvs on hx4700? because on magician ts-adc-debounce is currently unusable anyway.
08:28.14cr2mndctrl: but the chances are not bad (not talking about the 2mpix cam)
08:28.16psokolovskypH5: and I'm worried too. My idea was to introduce few new APIs, which were in big need for long time, and STOP introducing new. but yes, we appear to be caught in positive feedback loop - more and more deep changes come
08:29.19psokolovskypH5: yes, I would appreciate if you test hx4700. But for me it works. h4000 doesn't ;-(. Also, maybe you could try and see what behavior magician will exhibit with it still - would it be even "worse" than before?
08:30.28pH5psokolovsky: do you see some way to push (maybe half-baked) parts of our current infrastructure to mainline? I fear we might end up in this ond interdependency hell again, if we're not very careful.
08:30.39pH5s/ond/old/
08:33.03psokolovskypH5: actually, that's one of the root of the problem: bd2 seems to got idea that time has come. so, he wants to start pushing something, say h5000. adc classdev appears to be part of this plan, "doing coding right so it was accepted into mainline". and not the last, as I understand
08:33.29psokolovskyProblem, we all have different aims, and this appears to start unconverge
08:33.44mndctrlcr2: alright... tnx for the update ;) I think I'll actually get a universal for now... Try it out...
08:33.49psokolovskySo, I wanted to discuss what to do in such situation
08:35.22psokolovskypH5: Actually, not CVS. I started to commit Anton's patches, but saw problems and reverted it. Latest set of patches is available from him - he just doesn't want to spam k-d with large repetative patchsets. I guess, I'll forward it to you.
08:35.44pH5psokolovsky: ah, ok. that'd be great.
08:36.21psokolovskypH5: sent
08:38.00pH5psokolovsky: Thanks. What I'd love to see upstream first would be the battery classdev or something similar. but previous discussions on lkml didn't have a real closure and the OLPC guys' api looks a bit different from ours. Any idea how to advance that?
08:39.51psokolovskypH5: my plan always have been: work on better support for all devices, and on converging them (common drivers). After that, we could show off relatively compact code supporting bunch of machines.
08:41.16psokolovskypH5: as for mainline, I wouldn't even start with such high-level stuff as battery classdev. There're base stuff like extendable PXA clk API not yet in mainline. Next, SoC base drivers - core of our machine support not in mainline, etc.
08:41.33psokolovskybattery stuff is minor comparing with this
08:42.43psokolovsky- and all above is just illustration that we all think differently and have different priorities. we should somehow think out common plan (in the sense that everyone else knew what you're working on), or we may have bad things like forks, etc
08:47.03cr2psokolovsky: i doubt there will be forks. you are the only one person really maintaining linux on mainstream pda. and zaurus is evil in many respects.
08:48.31pH5I just expect that things like the battery api require much more coordination, as there are vastly different machines that have to use it, too (ACPI, ugh) and I don't see us interacting too much with development outside the little ARM/embedded community.
08:48.51psokolovskycr2: sorry, mistake. I just commit patches from folks who don't have commit access. in other respect I'm just like other devels who has commit access.
08:48.57pH5I hope things like SoC (once it works) won't be too controversial since we are the only real users.
08:49.33pH5damn, I should rewrite my CPLD and PASIC3 drivers.
08:49.57psokolovskycr2: and bd2 is prolific contributor, and it's really his own choice not to request cvs commit. if he'll grow dissatisfied with cvs progress, we'll have another hx2750
08:50.17psokolovskyin thsi respect, I don't want to be one, who made him make such choice
08:50.30cr2pH5: other CPLDs are coming, so do it good
08:51.40psokolovskyOk, so please think what plan *we* have, after all, if any. And I will ask bd2 to post his masterplan first before coding anything controversial.
08:51.54cr2psokolovsky: bd2 is a very smart guy :) not being able to commit directly forces one to write much better code ;)
08:53.16psokolovskycr2: possibly. current situation is that his better code adds regressions to machine support. so, what do now?
08:53.39cr2pH5: http://wiki.xda-developers.com/index.php?pagename=AthenaCPLD1
08:54.24pH5cr2: just gpios?
08:54.40cr2psokolovsky: a hard question.
08:54.43pH5that's a case for gpiodev :)
08:55.00cr2pH5: the GPIOA is for acking the irqs.
08:55.30cr2it's empty for now. there is a chained irq on the PXA.
08:55.42pH5ah ok. is it on a xilinx coolrunner too?
08:56.32cr2pH5: probably. i don't have the device or the motherboard pictures. it's pure RE :)
08:56.35pH5ah :)
08:57.48cr2pH5: hehe. http://wiki.xda-developers.com/index.php?pagename=AthenaResearch
08:59.48cr2DPRAM is coming. it is used for AT communication (!)
09:01.09cr2http://wiki.xda-developers.com/index.php?pagename=HTC_Serial
09:01.35pH5heh, that's a first
09:03.20psokolovskypH5, cr2: on completely another topic: did toy guys tried recent Angstrom-x11-image's? I find them regressing too - buttons don't work, mouse cursor is visible, BL control issues; plus, on hx4700 I was getting suspend/resume hangs with Xw100 or what is used there.
09:04.16pH5psokolovsky: bl control issues are the ipaq-sleep/"bl" program not working correctly (it switches the backlight off during login)?
09:04.37psokolovskypH5: yep, exactly
09:04.41pH5buttons don't work is probably missing keylaunch, I had that problem on my latest magician image
09:04.49pH5not sure where that got lost
09:04.59pH5but the suspend/resume hangs I didn't see yet.
09:05.30psokolovskypH5: I saw buttons notw orking some time ago on hx4700, thotught were due to another Xserver impl. now confirmed same on h4000
09:05.51pH5psokolovsky: keylaunch?
09:06.13psokolovskypH5: suspend/resume only on hx. and actually some time ago, not recently
09:06.37psokolovskypH5: I just checked that xmodmap is loaded with proper buttonmap (common)
09:06.48psokolovskypH5: thanks for hint, will check as time permits
09:24.40*** join/#htc-linux psokolovsky_ (n=psokolov@favoured-food.volia.net)
10:33.59*** join/#htc-linux TripleDE1 (n=sergio@89.129.155.39)
10:45.38*** join/#htc-linux tudenbart (n=willi@xdsl-213-196-251-225.netcologne.de)
11:07.32*** join/#htc-linux TripleDES (n=sergio@89.129.164.41)
11:39.15*** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-83-233.dynamic.mnet-online.de)
12:42.25*** join/#htc-linux apt (i=ibot@pdpc/supporter/active/TimRiker/bot/apt)
12:42.25*** topic/#htc-linux is HTC Linux Channel: Find logs at http://apt.rikers.org/%23htc-linux/ | please check http://handhelds.org/moin/moin.cgi/HTC_2dPhones | http://wiki.xda-developers.com/index.php?pagename=Xanadux | <cr2> let's define a common setup.
13:08.19*** join/#htc-linux ljp (n=lpotter@203.94.178.46)
13:11.58*** join/#htc-linux PlayaX (n=PlayaX@cpe-76-187-37-171.tx.res.rr.com)
13:13.23*** part/#htc-linux PlayaX (n=PlayaX@cpe-76-187-37-171.tx.res.rr.com)
13:18.50*** join/#htc-linux ljp_ (n=lpotter@203.94.178.46)
13:40.39*** join/#htc-linux ljp (n=lpotter@203.94.178.46)
14:01.18*** join/#htc-linux drbombay[lap] (n=drbombay@pdpc/supporter/active/drbombay)
14:37.55*** join/#htc-linux jeanseb (n=jeanseb@gazypan.dyndns.org)
14:44.14*** join/#htc-linux TBBle_ (n=tbble@2001:388:f000:0:0:0:0:2d)
15:03.07*** join/#htc-linux Ralith (n=ralith@66.150.12.86)
15:21.59*** join/#htc-linux goxboxlive (n=goxboxli@206.80-202-161.nextgentel.com)
15:24.04*** join/#htc-linux crashmatrix (n=crashmat@s5590785f.adsl.wanadoo.nl)
15:35.35*** join/#htc-linux |Ralith| (n=ralith@66.150.12.86)
15:38.29*** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-83-233.dynamic.mnet-online.de)
15:42.26*** join/#htc-linux _law_ (n=_law_@213.173.86.202)
15:45.15*** join/#htc-linux crashmatrix (n=crashmat@s5590785f.adsl.wanadoo.nl)
15:48.29*** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be)
16:02.05*** join/#htc-linux rob_w (n=bob@p85.212.74.34.tisdip.tiscali.de)
16:14.47*** join/#htc-linux crashmatrix (n=crashmat@s5590785f.adsl.wanadoo.nl)
16:36.52*** join/#htc-linux TeringTuby (n=maarten@188-116-dyn.ipact.nl)
16:47.15*** join/#htc-linux g3gg0__ (n=g3gg0@ppp-82-135-12-32.dynamic.mnet-online.de)
16:50.54*** join/#htc-linux goxboxlive (n=goxboxli@206.80-202-161.nextgentel.com)
16:56.06*** join/#htc-linux cyrill62 (n=cyrill62@ble59-5-82-233-205-36.fbx.proxad.net)
16:58.52goxboxlivecr2 are you there/here ?
17:06.04goxboxlivepsokolovsky: you asked cr2 about angstrom-x-image. I have no idea, it wouldnt build, it failes when doung rootfs of the image. It's something about the matchbox-panel that wouldnt install
17:06.35psokolovskyfor me it builds
17:06.45goxboxlivehave you build openmoko also?
17:08.05goxboxliveI think openmoko is using a newer matchbox-panel version, whiles x-image is using an older one. I have mention it for koen, but he said he had no idea why it happend.
17:10.06pH5goxboxlive: matchbox-panel-2 provides matchbox-panel, so if you have mp2 built and depend on mp, it will try to use mp2 (which clashes with the lib versions somehow)
17:10.41pH5workaround: build x11/openmoko-image in different dirs or remove matchbox-panel-2 before building the x11-image
17:10.48goxboxlivepH5:yes i have noticed, any way to override it?
17:10.52goxboxliveok thx
17:11.12pH5goxboxlive: or you could try to remove the PROVIDES="matchbox-panel" from mp2.
17:12.04goxboxliveok
17:14.27cr2goxboxlive: 'git pull' does not work today too. installing something manually defeats the purpose of oe :)
17:15.05cr2NOTE: package xcalibrateext-0.0+git20060814-r0: task do_fetch: started
17:15.05cr2Failed to run command 'pull': Permission denied
17:15.26cr2git pull git://anongit.freedesktop.org/xorg/proto/calibrateprotrateproto failed
17:21.40pH5cr2: that whole package looks bogus to me. this is the same as calibrateproto with a wrong SRC_URI. psokolovsky?
17:23.40pH5cr2: what depends on xcalibrateext?
17:24.56psokolovskypH5: ?
17:25.47pH5psokolovsky: you added xcalibrateext. what does depend on it. can it be removed? the correct package is xorg-proto/calibrateproto.bb
17:26.03pH5(or at least that's what monotone thinks)
17:26.43psokolovskypH5: what I added was working for more than half-year already. if something changed recently, then...
17:28.56pH5psokolovsky: so you have no idea?
17:29.18psokolovskypH5: for me x11 image building works
17:29.59psokolovskypH5: tell me ipk names, I can check their dates I have if that helps
17:31.36*** join/#htc-linux rob__w (n=bob@p213.54.29.134.tisdip.tiscali.de)
17:33.27*** join/#htc-linux rob__w (n=bob@p213.54.29.134.tisdip.tiscali.de)
17:35.38pH5uh-oh. I just noticed that xcalibrateext is still used instead of calibrateproto if you compile X11R7 xserver. I guess we should clean that up when we upgrade our xorg packages.
17:35.53pH5I now see that you just replaced the old xcalibrateext_cvs package with the git one.
17:36.14pH5cr2: even though the SRC_URIs are different, both packages fetch just fine for me.
17:52.38*** join/#htc-linux RoEn_PC (n=roen@p54A64FC1.dip.t-dialin.net)
17:54.45psokolovskypH5: yes, that what my memory tells me too. I wish that OE upgraded to R7.2 soon and we had new round of cleanup on X.
18:02.47*** join/#htc-linux FossiFoo (n=Fossi@e176099120.adsl.alicedsl.de)
18:04.20goxboxlivecr2: still here
18:06.15goxboxliveIf it is a long time since you checked out OE you might consider do it again. Because sometimes it's not enough to just pull. I dont know why, but i have noticed it.
18:09.18goxboxliveBut, forget about OE, i do the images you do the kernel drivers. Let's try to fix the asic3 keys for luminar sensor and rotation.
18:11.29goxboxliveand then update the rest of the drivers so we are at same status as HX4700
18:12.25*** join/#htc-linux drbombay[lap] (n=drbombay@pdpc/supporter/active/drbombay)
18:12.57pH5psokolovsky: I was waiting for epoch support in bitbake, but I guess now that we have it there is no excuse :)
18:15.16psokolovskypH5: was it added? maybe it's not in OE then? because I got full epoch support depends on openmoko guys being ready for that, blah-blah, blah-blah ;-I
18:16.31_law_hi all
18:16.41_law_my tornado boots linux via haret :-)
18:16.52pH5psokolovsky: bitbake trunk supports it, so we could start adding PE="1" to packages. it *shouldn't* be a problem for openmoko because new packages with smaller version numbers just won't be selected if bitbake doesn't support PE.
18:17.11_law_incl. working usb_ether :-)
18:18.05*** join/#htc-linux drWyrM (i=c88e3026@gateway/web/cgi-irc/ircatwork.com/x-5f1f6e9218fd57d2)
18:18.05pH5and as soon as a distro has set its version in stone aka PREFERRED_VERSION, it doesn't matter anyway (until people start putting epochs in preferred versions).
18:18.09pH5_law_: congrats!!
18:19.54_law_pH5: thanks
18:20.13psokolovskypH5: yeah, IIRC, problem was exactly that OM is not ready to switch to bb1.8, because ..., etc. RP posted about that. I just hope that someone will go for that anyway ;-)
18:37.16cr2pH5: manual git pull does not work for me.
18:38.29cr2goxboxlive: luminosity sensor and rotation were implemented as "switches", not "keys". i know that opie will break if we change it, but is there any compelling reason to use switches at all ?
18:38.50cr2hehe
18:39.55cr2psokolovsky: are the switches necessary ? what is the logic behind switch/key separation ?
18:45.16psokolovskycr2: I have no idea. your call. zaurus use them. there limitation is that they're binary, so bad abstraction for light sensor. h5000 implements it using adc classdev for example
18:47.42pH5cr2: you could ask RP about that. I think the reasoning behind it was that switches are expected to be '1'/'0' for a long time vs keys that are only pressed for a short time. He said that this was 'controversal', so there should be some discussion on lkml about it.
18:48.44cr2ok. i think it's easier to assign then to some bizarre keys and fix/change the userspace.
18:49.38_law_http://law.hostsen.net/oe/pictures/tornado/DSCN4323.JPG :-)
18:50.18cr2i guess real "switches" are treated by the kernel. as irqs for (near-)realtime performance. userspace is more flexible.
18:50.59cr2_law_: great. why do you have ext2 errors ?
18:51.42_law_cr2: dunno, but perhaps the ramdisk is do small
18:55.04*** join/#htc-linux goxboxlive (n=goxboxli@206.80-202-161.nextgentel.com)
18:55.34*** join/#htc-linux goxboxlive (n=goxboxli@206.80-202-161.nextgentel.com)
18:57.29goxboxlivehi cr2 i am back
18:57.40goxboxliveI read the logs
18:58.13goxboxliveBut do you think that we shall drop the switches ?
18:58.51goxboxliveI mean both of them are realy necessarry
18:58.51cr2ok. i see the problem as follows. does openmoko deal with the screen rotation ? i guess not.
18:59.06goxboxlivewell, you can rotate using xrandr
18:59.31cr2are there any openmoko guidelines how to deal with auto-rotation ?
18:59.46goxboxlivehold on let me chwck
19:00.52goxboxliveBut, couldt we use xrandr to handle it, by a signal from the rotate switch?
19:01.16goxboxliveor am i talking tooooo electrical curcuits now :-)
19:01.27cr2neo does not support it, but the "official" word is that openmoko on other devices is welcomed. so there should be a guideline.
19:02.05cr2the x server gets a key event and should deal with it. so it's the problem of the x server (config?)
19:02.14goxboxliveNo, you right, Neo doesnt support it, it is on the HW wish list, also luminance sensor
19:02.41goxboxliveok
19:03.03goxboxliveBut, isent it easy as :
19:03.09goxboxliveif
19:03.11cr2wish list is for the FIC hardware. we need a software guideline today. also for wifi, btw. and for the camera(s).
19:03.16pH5cr2: xrandr. it's the job of a daemon to watch for those events and rotate the screen. zaurusd does this, for example.
19:04.00cr2pH5: ok.
19:04.29goxboxliveyes it shouldnt be that dificult. All we need is the key_switch to talk with the xrandr-panelapp.sh file and then it would be fully automatic
19:05.28goxboxliveAnd, why cant we make or own deamon to help us out with the keylight/luminance sensor
19:05.53cr2ok, so the question reduces to a convention what key should be used for the lum. sensor and screen rotation event.
19:06.22cr2i'll add these two codes to asic3-gpio init, and the kernel side is finished.
19:06.41goxboxlivegood
19:07.45cr2KEY_MSDOS :)
19:08.34cr2or KEY_SPORT
19:08.51cr2KEY_SWITCHVIDEOMODE ?
19:08.59goxboxlivedoesnt mather, just make it work :-)
19:09.10pH5KEY_DOWHATIWANT
19:09.11cr2and KEY_KBDILLUMTOGGLE
19:09.22cr2it's what is in linux/input.h
19:10.14cr2KEY_VIDEOPHONE, not bad.
19:10.18goxboxliveok, i'll look to
19:11.08cr2so KEY_SWITCHVIDEOMODE and KEY_KBDILLUMTOGGLE
19:11.26goxboxlive#define KEY_ZENKAKUHANKAKU :-)
19:11.42cr2KEY_VIDEOPHONE can be used too. at the appropriate place.
19:12.50goxboxlivedefine KEY_ANGLE ?
19:13.17cr2i think KEY_SWITCHVIDEOMODE is quite good.
19:13.56cr2KEY_KBDILLUMTOGGLE is what really happens on the universal :)
19:14.03goxboxlive#define SW_TABLET_MODE ??
19:14.17cr2SW_* is for switches.
19:15.18goxboxlivei know
19:16.35cr2pH5: what has happened with your BOTHEDGE asic3 patches ? i've lost the track.
19:17.59cr2_law_: is you actual ramdisk size equal to the kernel commandline parameter ?
19:18.04pH5cr2: I committed them to hh.org cvs and forgot about them. If there is any strange edge flipping behaviour left in drivers outside of asic3, you should try to convert them to just use BOTHEDGE. If it doesn't work for some reason, I'd be interested.
19:18.46cr2pH5: ok, i was interested if asic3-gpio supoorts BOTHEDGE
19:19.41pH5cr2: it should. there is only one issue know with the hx4700 ac/usb plug irqs, but I couldn't reproduce that so far.
19:19.49pH5s/know/known/
19:20.08cr2ok.
19:21.51cr2haha. you don't have a choice.
19:22.14cr2hmm. screen_cover. what event is that ?
19:23.43cr2KEY_FINANCE
19:23.45goxboxlivei think it's lid
19:23.51goxboxlivenot cover
19:24.37cr2KEY_SCREEN ?
19:24.45goxboxliveyes there is one
19:24.50goxboxliveKEY_SCREEN
19:25.06goxboxlive#define KEY_SCREEN              0x177
19:33.57cr2ok. done. KEY_F10 for the backlight...
19:36.45cr2so it's a userspace problem now. next one should be the external headset/BT headset on/off event.
19:37.25cr2look for the appropriate names :)
19:39.28cr2external headset should be easy, because it's an asic3-gpio too.
19:39.57cr2ah, the remaining general problem on the universal. what is BT irq...
19:40.26cr2pH5: do you have... hmm. you don't have TI BT.
19:40.46pH5cr2: wakeup? that should be somewhere early in the nk.exe code.
19:40.59pH5yeah, I have some csr bluecore
19:41.37cr2the devices with TI BT chipsets have something that wince calls "BT irq".
19:42.26goxboxlivegreat cr2
19:42.33cr2it can be observed at least when the firmware update is sent. i have no idea what it is doing, so we just ignore it ;)
19:43.16pH5cr2: for perhaps for wakeup? I have a wakeup bt irq on the cpld.
19:43.32pH5It fires (at least) when the device is at sleep and somebody tries to connect.
19:43.51cr2connect from where ?
19:45.34pH5via bluetooth. if I suspend the magician and hcicontrol cc <btaddr> it from my laptop, it wakes up.
19:45.56*** join/#htc-linux SmallR2002 (n=SmallR20@88-108-236-102.dynamic.dsl.as9105.com)
19:46.26cr2hmm. maybe. i need to check it. but it happens also under other circumstances...
19:47.22cr2no. it is masked on sleep.  GPIOB4 in http://wiki.xda-developers.com/index.php?pagename=UniversalASIC3
19:50.21SmallR2002i just saw the ameo
19:50.39SmallR2002since when has mobile electronics had such an 'i want one' factor for me
19:51.32goxboxliveSmallR2002 WHat is the "ameo"
19:51.57SmallR2002tmobile athena
19:52.24SmallR2002well
19:52.28goxboxliveCr2: My status table is turning very green now :-) . Now, lets do the cam while where on it :-)
19:52.29SmallR2002tmobile version of the athena
19:52.31goxboxliveol
19:52.36goxboxliveok
19:52.57goxboxliveAthena is a great device, but it's too expensive and to big
19:53.42cr2SmallR2002: i've done a lot of RE for athena.
19:54.03SmallR2002it seems to be something like the mda pro, just a bit better
19:54.05cr2but there is 1 small problem. or maybe 2 :)
19:54.32SmallR2002like it not being much of a phone?
19:54.53cr2no. from the "porting linux" angle
19:54.58cr2http://wiki.xda-developers.com/index.php?pagename=AthenaResearch
19:55.28psokolovskypH5: cool! what BT chip is in magician?
19:55.40SmallR2002cr2, i thought because it's an intel processor it's likely to be easier to use existing code maybe?
19:55.52cr2goxboxlive: somebody should create the pxa27x-cif platform_device from pxa_camera.
19:56.06SmallR2002(this is totally abstract due to total lack of knowledge about the architecture of it)
19:56.27cr2SmallR2002: yes, it will be very easy to boot linux from the ramdisk on athena.
19:56.43cr2look at the hardware list at the link above.
19:56.55SmallR2002i am
19:56.56cr2there are 3 "HDDs" in the device.
19:57.15SmallR2002boot, os, docs?
19:57.22cr2mdoc G4, miniSD and the microdrive
19:57.30SmallR2002hmmm
19:57.40SmallR2002let me guess, the microdrive could be hardhacked?
19:57.45cr2G4 is a tough nut, no linux driver available.
19:58.31cr2miniSD is not on the PXA :( but on ATIW2284. completely undocumented. needs some clever RE approach.
19:58.39SmallR2002any chance of a io spy driver?
19:58.57cr2Kevin's hw debug trace will be very helpful.
19:59.24cr2and it may be very helpful to document the ATI ace_ddi API.
19:59.42cr2so it needs a new SD host driver.
19:59.55cr2the microdrive:
20:00.16cr2it depends on the board_id <= 6 how it is connected.
20:00.23SmallR2002so it's a more likely device than the hermes?
20:01.00cr2there is an Philips ISP1583 USB2.0 chip, and Epson USB2.0/ATA100
20:01.46cr2both have the detailed datasheets, but it needs to be debugged how they are connected, which chip do you have, etc.
20:02.09cr2so it's not just Plug&pray to have a HDD on athena.
20:02.44cr2hermes SD driver support 2 SD controllers: the S3C2442 onboard and the ATI W2282.
20:03.26cr2but (imho, fortunately for us) only the S3C2442 is used :) with a working linux driver.
20:04.51cr2SmallR2002: we have an excellent "io spy driver" on pxa cpus :) it is called haret with pxa27x support.
20:05.03SmallR2002just over four hundred nicker for it feels a little steep, but if it's a better lot to go with i might start saving
20:05.17pH5yay! unmodified gsmd runs on magician.
20:05.35SmallR2002incidentally, does the hermes have some level of gps support somewhere?
20:05.40pH5psokolovsky: some csr bluecore
20:05.45cr2pH5: what about other htc devices with TI BB ?
20:06.25cr2SmallR2002: no, afaik. some small hardware parts are missing, grounded, etc.
20:07.22cr2i didn't follow the research recently, but they meant that the input LNA is grounded.
20:07.29SmallR2002ah
20:08.10cr2athena has sirf3, that one is easy.
20:08.57cr2it's on the ffuart, so you can just 'cat /dev/ttyS0' to get the NMEA datastream.
20:09.32cr2or use the sirfmon, but it's only for the hardcore people :)
20:11.23cr2another problem with athena is that it uses ATI W2284 also for the sound, and i have no idea of the audio/BT routing on it. And the W2284 sound capabilities, or course.
20:13.27goxboxlivepH5: Is there a way to deamonize xrandr, or make a deamon for it? Do you have any example wich is allmost the same?
20:14.05SmallR2002hmmm
20:14.17goxboxliveI am thinking about making it start from init.d
20:14.20SmallR2002so, cr2, the message is wait for athena version two?
20:16.07cr2SmallR2002: get one and hack :) or get the universal if you are lazy and don#t need HSDPA and sirf3.
20:17.45cr2the version two may be even more complex.
20:21.16_law_hmm i think the problem are the MMC1: Command timeouts, COMD0
20:21.29_law_s/COMD0/CMD0
20:21.40_law_amdn CMD7
20:22.03SmallR2002i need 3G
20:22.13_law_could it be that kernel 2.6.20 works better with mmc then 2.6.16 ?
20:22.13cr2_law_: i have no clue about the omap.
20:22.13SmallR2002is the universal the mda pro?
20:22.23cr2SmallR2002: yes.
20:22.46cr2i've said you only get HSDPA and sirf3 on athena.
20:22.53SmallR2002i thought that wasn't 3g compatible
20:23.16cr2and the 3mpix CAM, which is also unsupported ;)
20:23.26goxboxliveSmallR2002: Yes get a universal, look at out status page, where allmost done with the most important ones: http://www.handhelds.org/moin/moin.cgi/UniversalStatus?action=show
20:23.40goxboxlives/out/our/
20:24.12goxboxliveWe only need sound to let it work, then we can use the phone in daily use
20:24.19goxboxliveand suspend/resume
20:26.01SmallR2002now why did tmobile not let me get that phone...
20:26.07cr2goxboxlive: what's wrong with suspend/resume ?
20:26.18goxboxliveusbnet and pxafb
20:26.23goxboxliveusbnet doesnt resume
20:26.51goxboxliveand pxafb has prblem with it's refresh rate when it's resumed
20:27.32goxboxlivedont you remember cr2? The lethers flickers when it is resumed
20:29.29cr2ah, ok.
20:29.42cr2a pxafb problem.
20:30.15cr2and the usbh gpio problem.
20:31.57SmallR2002tmobile wouldn't allow me to get that because it wouldn't handle high speed data if i remember correctly
20:32.07goxboxliveos that something we can fix?
20:32.33goxboxlivethats's right, no HSDPA, onlu UMTS
20:32.36SmallR2002i'm a lucky bastard who pays out 50gbp/month for 10gb/month unfiltered
20:33.30goxboxliveBut cr2, why doesnt the hx4700 have the problem with resuming usbnet?
20:33.34cr2goxboxlive: i can add LCCR4 save/restore to htcuniversal_lcd_suspend/resume, but it's a really braindead way to fix it ;)
20:34.08cr2hm. but better than nothing.
20:34.10goxboxliveI think you should, even though it is braindead solution it will at last work
20:34.19cr2ok.
20:34.43cr2well, i can just set it in resume.
20:34.56cr2because we support only this mode.
20:35.01goxboxliveok
20:35.07cr2SmallR2002: that's a lot of money.
20:35.15SmallR2002aye lad, it is
20:35.27cr2400 for the device are peanuts in comparison.
20:35.43SmallR2002it's good value for the cash
20:35.58SmallR2002what you're getting, in comparison with other offers, etc
20:36.08SmallR2002and it unties me from my desk
20:37.57cr2ok.
20:38.05cr2goxboxlive: compiling.
20:38.05pH5cr2: re TI BB: that's why I asked about the blueangel protocol. are there some blueangel/alpine/apache serial dumps available somewhere?
20:38.11SmallR2002does mean that i'm living off a pittance for food
20:38.21goxboxlive:-)
20:38.31cr2pH5: i can do it for the blueangel.
20:38.49pH5goxboxlive: did you look at zaurusd? xrandr itself is a dump cmdline app, so no daemonization.
20:39.05cr2apache is cdma, forget it. Kevin2 has the traces of this qualcomm chipset.
20:39.08SmallR2002eighty pence for a loaf, grab some cheese or something like that and some butter... that's lunch for a week
20:39.50goxboxlivepH5 ok, i take a look
20:40.47cr2<PROTECTED>
20:40.58cr2pxa-regs is incomplete.
20:41.17psokolovskypH5: re: some csr bluecore: ah, I see, I bet it's not any device's functionality. bit really cool. wireless wake up - something I dreamed of all the time
20:41.27pH5cr2: the ba use '0x2 0x16 [data] 0x2 0x16' frames? because on magician it's just '0x2 0x16 [data] 0x2'.
20:42.37pH5psokolovsky: yup. there is no way to keep the hx4700's bt chipset running in suspend, unfortunately.
20:42.48pH5but all pda/smartphones I know can do it.
20:43.01psokolovskynice
20:44.18cr2pH5: isn't it the same ?
20:44.39SmallR2002hmmm
20:45.11cr2psokolovsky: a pipe dream for the bt virus writers :)
20:45.16cr2#if 1
20:45.16cr2#ifndef LCCR4
20:45.17cr2#define LCCR4           __REG(0x44000010)  /* LCD Controller Control Register 4 */
20:45.17cr2#endif
20:45.17cr2<PROTECTED>
20:45.18cr2#endif
20:45.24cr2this makes nice kernel code
20:45.33pH5cr2: here the frames end just with 0x2. I read somewhere that on the blueangel the frames end with 0x2 + channel number, not sure if that is true.
20:45.46cr2pH5: why do you care ?
20:46.06cr2i can check it with haret, anyway.
20:47.06cr2goxboxlive: i'll commit it, but it makes me feel dirty ;)
20:48.40cr2done.
20:50.31goxboxlive:-) thx
20:50.31SmallR2002my device randomly decided to loose all it's charge by just gone lunchtime
20:50.40SmallR2002not a clue why
20:51.02SmallR2002generally lasts me a full day of random ircing etc
20:51.22goxboxlivecr2: Realy progressing today cr2, what about usbnet? psokolovsky claimes it works on hx4700
20:52.02goxboxlivei am looking at the zaurusd stuff. I am gonna try making it work on universal
20:53.12SmallR2002do you have an estimated time of arrival for a working hermes linux?
20:53.24cr2goxboxlive: i don't see anything wrong with the UDC part. it maybe something in the USBH part. i've just copied the code from magician, and it's a copy from zaurus :) wince debugging is required.
20:53.42goxboxliveok
20:54.12cr2SmallR2002: if it will boot, everything will move really fast.
20:55.52SmallR2002hmmm
20:56.16*** join/#htc-linux TeringTu1y (n=maarten@188-116-dyn.ipact.nl)
20:56.27SmallR2002it'd be nice to get native ssh support which doesn't result in very fast battery usage
20:56.59cr2i can't use the IR debug console, because the PCI-LPC bridge on my notebook is not supported ;)
20:57.22SmallR2002hmmm
20:57.59SmallR2002i've got an asus a6u, is that supported
20:58.08*** join/#htc-linux bd2 (n=cbou@89.113.85.104)
20:58.43cr2cpu ?
20:59.17SmallR2002os[Linux 2.6.20.3 i686] distro[Debian 4.0] cpu[1 x AMD Turion(tm) 64 Mobile Technology MT-30 @ 1.60GHz] mem[Physical : 1900MB, 76.1% free] disk[Total : 51.17GB, 1.90% Free] video[Silicon Integrated Systems [SiS] 661/741/760 PCI/AGP or 662/761Gx PCIE VGA Display Adapter] sound[]
20:59.38cr2ah, the notebook.
20:59.47cr2you run in the 64bit mode ?
20:59.51SmallR2002nope
21:00.02cr2my IR works in 32bit mode, but i hate 32bit mode.
21:00.03SmallR2002too many 32bit stuff i want to run to be faffed with that
21:00.20SmallR2002boot in 32bit just to get the data?
21:00.39cr2to get the LPC bus support.
21:01.04cr2suse intentionally did not write the code for ISA support on amd64
21:02.07cr2there is no technical problem with that. only political.
21:02.36SmallR2002oh great, more suse annoyances
21:03.28cr2i agree that it's a support hell, but nevertheless.
21:04.05cr2the argument was "no amd64 machine" uses ISA.
21:04.39cr2i bought my amd64 notebook 3 years ago, and complained on the suse-amd64 mailing list :)
21:05.11cr2they have said, too exotic, no time or no real use. something like that, i don't remember exactly.
21:06.19SmallR2002<insert wintel conspiracy theory>
21:06.59SmallR2002:p
21:07.14SmallR2002i have noticed something similar myself
21:07.20cr2<insert suse greed> :)
21:08.25cr2ok, amd64 notebook were exotic 3 years ago. this was the first one afaik.
21:08.58SmallR2002maybe set up a very basic copy of 32bit linux on a usb drive or something, so you can get the data?
21:09.46cr2hm. can't boot from usb.
21:10.05SmallR2002ouch :(
21:10.18*** join/#htc-linux psokolovsky_ (n=psokolov@82.193.98.7)
21:10.31cr2i've bought pcmcia 2 port sata card yesterday.
21:10.48cr2but the raid1 is already half-full :)
21:11.33cr2goxboxlive: does this patch help ?
21:13.36cr2pH5: switching on the wifi router to connect the blueangel.
21:16.22goxboxlivecr2: I havent tried it out yet. Have to test it tomorrow. I am trying to let zaurusd work with uni now.
21:18.03cr2ok.
21:19.25goxboxlivecr2: i'll try it out now
21:19.28goxboxlivehold on
21:24.36goxboxlivecr2: Yes, now the screen looks good when resuming, at least in the shell.
21:25.31goxboxliveyes, also in Openmoko
21:25.47goxboxlivegood job even though it is dirty trick.
21:26.59cr2ok.
21:41.18SmallR2002has anyone else noticed that mobile phone interfaces in general are big, clunky and very bad at using screen space?
21:44.50*** part/#htc-linux rmoravci1 (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk)
21:48.22pH5cr2: great, I'll post the ldisc tomorrow after cleaning it up a little.
21:48.44pH5(I would prefer if the blueangel protocol is the same, my code is ugly enough as it is, currently)
21:48.45cr2pH5: i've done the log, will decode it right now
21:49.25cr2it seems like 02 16 ... 02 02
21:50.01_law_tomorrow i?ll try do get the minisd working...
21:50.04_law_n8 all
21:50.26cr2pH5: no, other commands are just 02 16 ... 02
21:50.47*** part/#htc-linux _law_ (n=_law_@213.173.86.202)
21:51.51cr2ok. decoded by an awk script :)
21:52.37SmallR2002how much could i pick up an old universal for?
21:52.56cr2pH5: sorry. it is 02 16 ... 02 02
21:53.03cr2i'll pastebin the output
21:53.17goxboxliveSmallR2002 check ebay
21:54.07SmallR2002not eighteen for a few months mate, can't really do online payments and my mum will go spare if i ask if i can put it through her card again
21:54.19pH5cr2: this looks very good.
21:55.08goxboxlivecr2: the keys are working too
21:55.17goxboxlivenot with zurusd though
21:55.23SmallR2002she's put up with a fair bit :p running my own business and all
21:55.42cr2pH5: http://www.pastebin.ca/411235
21:56.53cr2hm. the line 10 has only 1 02. but it may be a bug.
21:57.29pH5Perfect. Except for that it looks just like it does here. I'll send you the ldisc tomorrow for review/testing.
21:57.45pH5good night
21:58.31goxboxlivenight all
23:07.50*** join/#htc-linux tudenbart (n=willi@xdsl-213-196-255-40.netcologne.de)
23:15.45*** join/#htc-linux TeringTuby (n=maarten@188-116-dyn.ipact.nl)
23:29.35*** join/#htc-linux Ralith (n=ralith@soggy202.drizzle.com)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.