00:00.22 | cr2 | what do you think these offsets mean ? |
00:01.35 | BabelO | knl for kernel, rfl ... |
00:01.45 | mxktmdude | cr2, what kind of cpu docs are needed? im not a kernel whiz or anything mostly just plain old apps etc |
00:02.04 | BabelO | cr2: nnd for nand sdr for sram |
00:02.56 | BabelO | cr2: something like that nnd =sdr = nand |
00:04.07 | cr2 | mxktmdude: the CPU manual describing subsystems and registers. or maybe the linux kernel source for it :) |
00:04.52 | mxktmdude | something similar to this: http://www.cdmatech.com/products/msm7500_chipset_solution.jsp |
00:05.41 | mxktmdude | i guess that doesnt list registers though hehe oops |
00:09.38 | cr2 | mxktmdude: something like that http://pxa-linux.sourceforge.net/documentation.htm |
00:09.39 | BabelO | cr2: rfl tell me something |
00:10.18 | cr2 | hehe. links are dead now. |
00:11.21 | mxktmdude | hehe yeah, i have the intel processor books here so i know what you mean |
00:11.29 | cr2 | mxktmdude: http://balloonboard.org/hardware/300/ds/PXA270-dev-manual.pdf |
01:37.54 | *** join/#htc-linux ltxda2 (n=ltxda@70.255.136.7) |
02:53.16 | *** join/#htc-linux Keizer (n=keizer@c-69-139-42-47.hsd1.fl.comcast.net) |
03:59.52 | regulate | sup |
03:59.57 | regulate | anyone awake to talk about hermes? |
05:55.22 | *** join/#htc-linux GPFerror (n=GPF@cpe-76-187-41-132.tx.res.rr.com) |
07:20.49 | *** join/#htc-linux goxboxlive (n=goxboxli@139.80-202-160.nextgentel.com) |
07:21.13 | *** join/#htc-linux the_sys0p` (i=nroberts@117.121.94.229) |
07:32.07 | *** join/#htc-linux sladen_ (i=paul@starsky.19inch.net) |
07:34.14 | *** join/#htc-linux keizer1 (n=keizer@c-69-139-42-47.hsd1.fl.comcast.net) |
08:02.12 | *** join/#htc-linux _dion_ (n=dion@inhex.net) |
08:46.28 | *** join/#htc-linux rmoravcik (n=rmoravci@adsl-d127.84-47-44.t-com.sk) |
09:10.02 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
09:12.38 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
09:17.01 | *** join/#htc-linux Marex-notebook (n=marex@gwfm10-3-250.802.cz) |
09:53.31 | *** join/#htc-linux pH5 (n=ph5@e178219091.adsl.alicedsl.de) |
10:08.38 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
10:09.10 | pH5 | good morning BabelO |
10:09.17 | BabelO | hi pH5 |
10:36.52 | *** join/#htc-linux rob_w (n=bob@X16f9.x.pppool.de) |
11:13.01 | *** join/#htc-linux ivucica (n=ivucica@217.14.216.214) |
11:21.34 | *** join/#htc-linux AC_Mast (i=s@gateway/tor/x-439bacb2407e3a2c) |
11:21.39 | ivucica | where can i check what kind of bluetooth device does my ipaq 3870 have? i'm trying to recompile kernel and i'm confused by: |
11:21.55 | ivucica | "Networking->Bluetooth subsystem support->Bluetooth device drivers" |
11:27.20 | *** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be) |
11:31.26 | ivucica | nevermind, i think i found info |
11:40.59 | ivucica | mkay, now i have another problem ;) |
11:41.09 | ivucica | perhaps cr2 can help |
11:41.27 | ivucica | i'm using 2.6 kernel from handhelds.org |
11:41.43 | ivucica | i'm using h3800_static_defconfig |
11:41.53 | ivucica | when punching in "make" i get |
11:41.54 | ivucica | cc1: invalid option `abi=aapcs-linux' |
11:41.54 | ivucica | make[1]: *** [arch/arm/kernel/asm-offsets.s] Error 1 |
12:41.04 | *** join/#htc-linux rob_w (n=bob@X0fbe.x.pppool.de) |
12:53.00 | *** join/#htc-linux TeringTuby (n=maarten@162-117-dyn.ipact.nl) |
13:23.39 | *** join/#htc-linux sladen (i=paul@starsky.19inch.net) |
14:52.53 | BabelO | cr2: i understand now how osm works ;) i just read the full wiki |
14:56.05 | cr2 | BabelO: i dislike the way how they do it, but like the result. |
14:56.20 | BabelO | :) |
14:56.44 | cr2 | ivucica: you need to disable the EABI if you care using the older toolchain. |
14:56.47 | BabelO | what you dislike in their method ? |
14:57.02 | ivucica | cr2: ok, i disabled it |
14:57.46 | ivucica | cr2: just a quick one; i have arm-linux-gcc-3.3 installed yet kernel build process complains that it's an unsupported compiler |
14:58.03 | cr2 | BabelO: web2.0_java_xml_solve_all_problems_GIS_ignorance_is_strength |
14:58.12 | ivucica | will setting CC be enough to switch it to arm-linux-gcc-3.3? |
14:58.52 | BabelO | cr2: ok i m agree with you, they don't use GIS then ? it is a big lack, need a converter ;) |
14:59.12 | cr2 | BabelO: they are wasting all kind of resources to no end, and reinventing the wheel every day. |
14:59.39 | cr2 | BabelO: they provide you with a 15GB XML file, and you are free to parse it yourself. |
14:59.59 | cr2 | that what i call 'GIS ignorance' |
15:00.14 | BabelO | ok |
15:01.16 | cr2 | and all the tools for parsing it are either perl of java. do i need to tell you they all do OOM ? :) |
15:01.36 | cr2 | or need 2 weeks to finish. |
15:01.53 | cr2 | there is one exception, though. |
15:02.00 | cr2 | the osm2pgsql |
15:02.06 | cr2 | written in c :) |
15:02.10 | BabelO | lol |
15:02.57 | cr2 | and that's the program that makes the google-map style tiles at openstreetmap.org |
15:03.39 | cr2 | you import this xml into postgis DB, and then the renderer code uses it. |
15:04.06 | cr2 | another funny thing about osm is that the whole thing runs on mysql |
15:04.43 | cr2 | but then they create the 15GB xml file out of mysql DB tables |
15:04.49 | BabelO | cr2: it is a good database mysql ;) |
15:04.52 | cr2 | only to import it into postgis later. |
15:05.09 | cr2 | this is web2.0 for me :) |
15:05.20 | cr2 | i have nothing against mysql. |
15:05.44 | cr2 | just the mysql -> xml dump -> postgis seems irrational for me. |
15:06.17 | BabelO | cr2: i ve a customer at work who say that database below 500Mb is for mySql and up is for Oracle ;) |
15:30.57 | cr2 | BabelO: i've looked at the ba bl, and it looks like that wdata is supported too. |
15:31.27 | BabelO | good, so i can have luck to try pH5 tools |
15:31.29 | cr2 | you may try to compile the htc-utils by pH5 |
15:31.36 | BabelO | :) |
15:31.54 | cr2 | i guess we need to add more saferguards there too. |
15:32.01 | cr2 | and full command logging |
15:32.38 | cr2 | i can't test it right now because of the 64bit. |
15:33.04 | cr2 | i'll try to add the ba initialization gpio settings to wiki. |
15:33.25 | cr2 | i guess we will unleash a lot of init bugs in this way. |
15:33.37 | cr2 | universal is already hit by such bugs ;) |
15:33.39 | BabelO | ok, i try this tool later in the evening |
15:33.44 | cr2 | ok. |
15:34.05 | cr2 | and we also can try to create the 'diagnostic' SD |
15:34.27 | cr2 | i have no other option on universal, because it's net-locked. |
15:34.59 | cr2 | btw, i guess i must buy an msm7x00 based phone as a true htc fan :) |
15:35.26 | cr2 | what do you think is the cheapest one ? |
15:36.23 | BabelO | the one without gpd ;) |
15:37.09 | cr2 | HTC S640 (HTC Iris 100) ? |
15:37.46 | BabelO | my next phone will be with gps and dvb-t receiver ;) |
15:37.49 | cr2 | its with kbd. |
15:38.12 | cr2 | i have a dvb-t usb stick. |
15:38.33 | cr2 | but since i hate tv, i need to write the linux driver for it i guess |
15:38.34 | BabelO | cr2: so you can use it as host on athena soon ;) |
15:38.55 | cr2 | and find out how to extract the firmware and upload it. |
15:39.06 | cr2 | yes, but it may be difficult. |
15:39.19 | cr2 | my SD cardreader does not work with athena. |
15:39.29 | cr2 | it wants too much current i guess. |
15:39.51 | cr2 | and shitty wince doesn't even create a popup saying that ;) |
15:39.51 | BabelO | MSM7500 based phone are Titan , libra, vogue, iris |
15:40.00 | cr2 | titan is CDMA |
15:40.24 | cr2 | iris has kbd, so it may be more expensive than iris |
15:40.37 | cr2 | somebody with libra was already here. |
15:41.19 | cr2 | libra also has the kbd. |
15:41.25 | cr2 | looks like hermes |
15:41.38 | cr2 | so i guess iris is the best choice. |
15:41.50 | BabelO | MSM72000 phone are Omni , Nike, sedna, erato, ultimate from i mate |
15:42.51 | cr2 | hmm. |
15:42.56 | cr2 | you are right. |
15:43.12 | *** join/#htc-linux TimRiker (n=timr@rikers.org) |
15:43.17 | cr2 | omni costs like athena i guess. |
15:44.20 | BabelO | 800x480pixe |
15:45.17 | cr2 | s420 is CPU: Samsung SC32442, 400 MHz |
15:45.19 | cr2 | i have hermes for that. |
15:46.23 | BabelO | there is htc sirius too |
16:04.13 | ivucica | cr2, BabelO: is this a valid way to force 2.6's compiling process to use gcc 3.3? |
16:04.15 | ivucica | ivucica@khaoticone:/opt/kernel/linux-2.6.x-hh$ export CC=arm-linux-gnu-gcc-3.3 |
16:04.15 | ivucica | ivucica@khaoticone:/opt/kernel/linux-2.6.x-hh$ export CXX=arm-linux-gnu-g++-3.3 |
16:04.40 | ivucica | and is there a way to get #error to print out a macro's value? e.g. value of __GNUC__ |
16:22.32 | ivucica | cr2 ? |
16:48.13 | *** join/#htc-linux ImCoKeMaN (n=me@rrcs-67-78-122-67.sw.biz.rr.com) |
17:04.43 | *** join/#htc-linux Marex-notebook (n=marex@gwfm10-3-250.802.cz) |
17:08.52 | ivucica | cr2: ever experienced this kind of error during compilation of kernel? |
17:08.55 | ivucica | --> /tmp/ccY3Llm1.s:3753: Error: .err encountered |
17:09.07 | ivucica | it happens on mm/page_alloc.c |
17:18.14 | *** join/#htc-linux scorpio16v (n=scorpio1@f050216157.adsl.alicedsl.de) |
17:35.05 | ImCoKeMaN | wow this is kinda funny, someone made a new login ID on xda for linux on mogul |
17:35.06 | ImCoKeMaN | http://forum.xda-developers.com/showthread.php?t=341903 |
17:35.41 | ImCoKeMaN | they basically got nowhere, and are just askig for help, but they say they are l33t |
17:37.08 | *** join/#htc-linux Zoolooc (n=fredsiba@p54950378.dip0.t-ipconnect.de) |
17:42.50 | ivucica | ImCoKeMaN: too bad there are not more self-pitiful people like me who directly say they are dumb and are asking for help without saying they're 1337 ;) |
17:47.00 | ImCoKeMaN | and myself |
17:47.22 | ImCoKeMaN | i've been in here a lot learning about it and definitely enough to realize that what they tried wasn't anyhting detailed |
17:48.16 | ImCoKeMaN | when Cr2 and Kevin both say it's a long way off i'm not gonna go and just take a 6700 image and then ask for generic help when it doesn't work |
17:51.36 | ivucica | hehe |
17:51.58 | ivucica | im trying to recompile the kernel so i can get built-in mmc support for my 3800 ;) |
17:52.12 | ivucica | now i'm getting some odd error that perhaps suggests a broken compiler |
17:52.25 | ivucica | /tmp/ccY3Llm1.s:3753: Error: .err encountered |
17:52.25 | ivucica | make[1]: *** [mm/page_alloc.o] Error 1 |
17:54.01 | *** part/#htc-linux scorpio16v (n=scorpio1@f050216157.adsl.alicedsl.de) |
18:00.41 | cr2 | ivucica: in which directory do you have the crosscompiler ? |
18:01.05 | ivucica | in /usr/bin i believe |
18:01.08 | ivucica | wherever emdebian placed it ;) |
18:01.46 | cr2 | ImCoKeMaN: IMO it's possible to hack some kernel image and boot on msm7x00, so you'll see the penguin on LCD. but is it really worth it ? |
18:02.08 | ImCoKeMaN | right |
18:02.30 | ImCoKeMaN | kinda putting the cart before the horse |
18:02.51 | ivucica | cr2: of course it's worth it |
18:02.56 | ivucica | penguins everywhere |
18:02.57 | ivucica | =] |
18:03.12 | ImCoKeMaN | how bout i make you a jpg instead? |
18:03.35 | cr2 | ivucica: then it may end like linux on xda1 aka wallaby |
18:03.52 | ivucica | eff |
18:04.14 | cr2 | ivucica: what is the path to arm-linux-gnu-gcc-3.3 ? |
18:04.15 | ivucica | regarding my issue, cr2, do you have any idea what's causing it? |
18:04.19 | ivucica | sec |
18:04.29 | cr2 | this -3.3 makes everything very complicated. |
18:04.40 | ivucica | --> /usr/bin/arm-linux-gnu-gcc-3.3 |
18:04.46 | ivucica | there's no problem with that |
18:04.49 | cr2 | can't you just download the toolchain from codesourcery ? |
18:04.56 | ivucica | hmm |
18:05.11 | cr2 | but i'm not sure if something will not break in the kernel. |
18:05.12 | ivucica | i hate installing non-debian stuff, tbh |
18:05.26 | pH5 | good evening, btw. |
18:05.26 | ivucica | see, i can replace 3.3 with some other things |
18:05.37 | cr2 | well, but then the kernel needs to be fixed. |
18:05.49 | ivucica | pH5: hopefully you're not referring to me? ;) |
18:06.04 | ivucica | i meant i can put 4.2 |
18:06.20 | ivucica | i tried 3.4 but it appears that neil (the emdebian guy) has messed something up |
18:06.28 | cr2 | ivucica: i don't know why some distro package authors like to name gcc gcc-3.3, moc moc-qt4 and so on. |
18:06.34 | ivucica | hehe |
18:06.44 | ivucica | well, actually i run /usr/bin/arm-linux-gnu-gcc |
18:06.50 | ivucica | that looks like a symlink |
18:06.57 | ivucica | i'll check if it is |
18:07.03 | cr2 | that'll be better. |
18:07.22 | ivucica | cr2: /usr/bin/arm-linux-gnu-gcc -> /etc/alternatives/arm-linux-gnu-gcc |
18:07.34 | cr2 | ivucica: i use PATH=~/stuff/tmp/cross/bin/:$PATH CROSS_COMPILE=arm-angstrom-linux-gnueabi- make |
18:07.48 | ivucica | and /etc/alt....blah that is a symlink |
18:07.52 | cr2 | that's how the kernel Makefile expect it. |
18:08.09 | ivucica | hmmm |
18:08.25 | cr2 | that the toolchain is in 1 directory, and all gnu tools are prefixed by CROSS_COMPILE |
18:08.27 | ivucica | i'll go and fetch the toolchain, then ... |
18:10.31 | ivucica | babelo@machine:~$ wine /usr/bin/start cooking |
18:10.33 | ivucica | =] |
18:11.25 | cr2 | pH5: i'm sick of wince2003 reboots on ba :( |
18:11.40 | cr2 | need to flash something better :) |
18:11.52 | pH5 | ivucica: lol, way no. of course I meant the l33tsomething guy :) |
18:12.28 | ivucica | hehe ;) |
18:12.40 | pH5 | cr2: I really, deeply understand that sentiment. |
18:12.42 | ivucica | cr2: ever had wince2002? |
18:13.05 | ivucica | i am absolutely disgousted with their startup "tutorial" |
18:13.15 | ivucica | i understand the calibration, ok, i get that ... |
18:13.29 | ivucica | .. but WHY THE HECK THE CUT'N'PASTE TUTORIAL!? |
18:13.32 | ivucica | =] |
18:13.38 | pH5 | ivucica: it's the same on wm2003se |
18:13.54 | ivucica | microsoft never changes, i suppose |
18:14.01 | pH5 | you get mad when you have to reboot many times due to initial kernel instabilities. |
18:14.38 | ivucica | ...or during your own knowledge (or lack of) during initial missteps in installation, haret-method? ;) |
18:16.09 | BabelO | cr2: that's why i flash wm5, but then you lose the nice widcomm bluetooth stack |
18:16.36 | pH5 | cr2: I have found and started to analyze the I2C code in herald's SPL. at least i2c_init seems quite similar to the i2c-omap driver in linux. |
18:17.58 | ivucica | BabelO: i find widcomm's stack on h3800 not-so-good |
18:18.02 | ivucica | no SCO support :( |
18:18.09 | ivucica | and more than that ... no HID support! |
18:18.22 | ivucica | how the hell am i supposed to play tetris on my ppc using my cellphone?! |
18:18.30 | cr2 | BabelO: i don't give a fsck about the bt stack, i use it as a linux dev platform ;) |
18:18.57 | cr2 | pH5: why should it be different ? it uses the omap-i2c port ... |
18:19.02 | BabelO | ivucica: o did not see the ms bt stack in wm5 is like the one in win xp , so bad , you can do nothing |
18:19.05 | ivucica | cr2: fsck.ext2 or fsck.vfat? |
18:19.38 | cr2 | ivucica: fsck.ms |
18:19.47 | ivucica | cr2: that will pass, too |
18:19.50 | ivucica | =] |
18:20.03 | cr2 | BabelO: i think i'll flash wince5. where can i get it ? |
18:20.07 | pH5 | cr2: there is some gpio setting going on, maybe that is a reset line for the CPLD. |
18:21.01 | cr2 | pH5: i didn't look for gpio in the herald spl. can you compare the 'triples' with artemis data ? |
18:21.04 | pH5 | also, I wasn't not sure if there are slight differences between omap850's i2c and generic omap i2c (besides the 32 vs 16bit registers). |
18:21.18 | pH5 | s/n't not/n't/ |
18:21.37 | cr2 | pH5: may be, but it's unlikely. |
18:21.58 | pH5 | which is good. I still have to find out why modprobing i2c-omap hangs the device though. |
18:22.10 | cr2 | i'm wondering about wizard,prophet and touch. |
18:22.34 | cr2 | if they are closer to artemis or herald. |
18:22.59 | ivucica | cr2: http://www.codesourcery.com/gnu_toolchains/arm |
18:23.02 | ivucica | is that it? |
18:23.09 | cr2 | BabelO said herald is the latest omap850 device ? |
18:23.46 | pH5 | hm. isn't touch the latest? |
18:23.48 | cr2 | ivucica: http://www.codesourcery.com/gnu_toolchains/arm/download.html |
18:23.57 | ivucica | yep, t |
18:23.58 | ivucica | ty * |
18:24.09 | cr2 | pH5: maybe. |
18:24.16 | BabelO | cr2: pH5 yes touch is the latest |
18:24.28 | ivucica | woah 91mb! |
18:24.29 | ivucica | cool. |
18:25.02 | BabelO | cr2: i got my wince5 from xda, but can't acces to wiki actually |
18:28.10 | pH5 | cr2: what triples in artemis data are you talking about? |
18:28.13 | ivucica | 2002 is the latest one for SA, right? |
18:29.43 | cr2 | pH5: the cpld calls have 3 parameters |
18:29.50 | cr2 | in both spls |
18:30.04 | pH5 | hm. linux i2c-omap doesn't seem to touch the OMAP_I2C_OA_REG register at all. |
18:30.27 | pH5 | cr2: in herald, first parameter is the i2c cpld's register number, second and third are bits to enable/disable. |
18:30.37 | cr2 | pH5: what is wince doing ? |
18:30.45 | cr2 | i guess it's the same on artemis. |
18:30.55 | pH5 | cr2: but artemis has it's cpld memory mapped? |
18:31.04 | pH5 | s/it's/its/ |
18:31.06 | cr2 | pH5: is it possible to flash magician with htc-tools ? |
18:31.39 | cr2 | yes, they are mmaped, but it does not matter for the higher level code |
18:31.48 | pH5 | cr2: sure, that is what I wrote it for. |
18:31.54 | cr2 | set_cpld(r0,r1,r2) |
18:32.01 | pH5 | I do not know how the cpld is handled in nk.exe |
18:32.36 | cr2 | pH5: i probably need to fix the 64bit issues in htc-tools first |
18:36.22 | pH5 | cr2: s/unsigned long/u_int32_t/ ? |
18:36.43 | pH5 | I never had a 64bit arch CPU, so no experience with that. |
18:39.01 | cr2 | pH5: amd64 is not that expensive. |
18:41.18 | *** join/#htc-linux rmoravcik (n=rmoravci@adsl-d127.84-47-44.t-com.sk) |
18:42.39 | ivucica | cr2: that codesourcery toolchain -- is it gcc 4.2? |
18:44.05 | BabelO | cr2: nice, but no gps icon :( http://www.iconlet.com |
18:47.12 | cr2 | BabelO: i've cloned the a780 icon, but didn't manage to clean it up yet. |
18:47.30 | cr2 | ivucica: 4.2.1 i think |
18:47.54 | cr2 | BabelO: which fm radio chip has artemis ? |
18:48.00 | ivucica | cr2: in that case, i found exactly what i need ;) |
18:48.10 | ivucica | an emdebian package exists |
18:50.12 | BabelO | cr2: that s the big question i hope it is one of chipset supported by kernel |
18:53.41 | pH5 | cr2: git clone http://linuxtogo.org/~ph5/htc-tools/.git |
18:54.05 | pH5 | could you check if it is ok 64bit? |
18:54.16 | pH5 | *with 64bit |
18:59.56 | cr2 | pH5: yes, i'll do it right now. even my old notebook is 64bit :) |
19:02.35 | pH5 | I can't be bothered to buy a new laptop until my old one falls apart and there are cheap ones with < 15" WUXGA screens, so I'm stuck with my olde P4 for now. |
19:05.03 | cr2 | i have acer aspire1501, it was the first 64bit laptop. very heavy, needs the power supply 20V 6A ;) |
19:05.50 | cr2 | btw, i can check htc-tools on linuxppc |
19:06.09 | ivucica | i can't recommend getting cheaper Dell models |
19:06.28 | ivucica | one i have is constantly on verge of dying |
19:06.35 | pH5 | oye, that's a lot of energy. mine only takes 4.5A at 20V max. |
19:07.07 | ivucica | and from what I've heard from some friends, HP is evil, a lot: they have started supporting only the greatest of all evils, exclusively |
19:07.19 | ivucica | XP drivers are not available anymore |
19:07.38 | ivucica | i'd rather use even (pfui) XP than Vista |
19:08.03 | ivucica | and so would almost everyone i know (i've so far met only two people who i failed to convince that vista is flawed, evil, and not to be trusted in any case) |
19:08.28 | ivucica | but HP forces people to use it by not only not publishing new drivers, but also by removing old ones off |
19:09.36 | *** join/#htc-linux rob_w (n=bob@Ma25d.m.pppool.de) |
19:13.19 | cr2 | pH5: compiled with -Wall |
19:13.33 | pH5 | good, thanks. |
19:18.06 | cr2 | using zimage from http://linuxtogo.org/~htcpxa/htcblueangel/O.P.I.E-II/ |
19:19.15 | cr2 | error opening tags.raw: No such file or directory |
19:21.38 | cr2 | pH5: 3 longs in setup.h |
19:22.14 | cr2 | do i need tags.raw ? |
19:22.34 | *** join/#htc-linux goxboxlive (n=goxboxli@139.80-202-160.nextgentel.com) |
19:23.21 | cr2 | <PROTECTED> |
19:23.29 | cr2 | unknown machine: "BLUEANGEL " |
19:25.00 | ivucica | it would appear that kernel is now being compiled successfully with arm-linux-gnueabi-gcc version 4.2 ;) |
19:25.10 | ivucica | WITH option CONFIG_EABI turned on ;) |
19:25.57 | cr2 | ivucica: great. |
19:26.07 | ivucica | cr2: indeed ;) |
19:26.12 | ivucica | weee! |
19:26.26 | goxboxlive | hi |
19:26.30 | cr2 | pH5: if i connect 2 phones : |
19:26.33 | cr2 | /tmp/htc-boot -l |
19:26.34 | cr2 | ECHO error: 'B' |
19:26.34 | cr2 | unknown machine: "UNIVERSAL " |
19:26.45 | cr2 | hi goxboxlive |
19:26.54 | cr2 | a flashing craze has begone. |
19:27.15 | goxboxlive | cr2: what? |
19:27.39 | goxboxlive | ahh you are going to test usb boot as pH5 made? |
19:27.56 | goxboxlive | does it also work on universal? |
19:28.23 | cr2 | pH5: htc-boot -l works only one time. then everything is messed. |
19:28.59 | cr2 | goxboxlive: trying it, but mine is locked. try it too. |
19:29.14 | cr2 | goxboxlive: don't forget to look for cplinux :) |
19:29.34 | cr2 | unknown machine: "UNIVERSAL " |
19:30.11 | goxboxlive | I have used it on magician, but never universal |
19:30.53 | cr2 | goxboxlive: i think it should work on universal too. |
19:30.57 | goxboxlive | cr2: whay do you want cplinux so much, is your broken? |
19:33.04 | pH5 | cr2: ok, this program is obviously still much too stupid. |
19:34.05 | cr2 | goxboxlive: it's autogenerated. but i'd like to compare. |
19:34.31 | cr2 | pH5: i'll add the uni and ba to the table. does the bl version play any role ? |
19:34.41 | pH5 | but anyway, usb boot won't work on anything but magician until somebody writes the ph5test.S code (it's magician specific right now) |
19:35.40 | pH5 | also the part where it writes 0x92000000 to 0x90032bf4 is highly magician specific. |
19:35.42 | ivucica | cr2, pH5: does 2.6 build process have a makefile target for creating an initrd image with all the modules? |
19:36.03 | pH5 | we have to look at the other machine's bootloaders to find the command tables and properly modify them depending on machine type. |
19:36.09 | cr2 | pH5: i'm more interested in the communication with bl now. and dumping the data. |
19:36.34 | cr2 | ivucica: initramfs, but you need to sort out the details. |
19:36.56 | ivucica | cr2: make: *** No rule to make target `initramfs'. Stop. |
19:36.56 | pH5 | cr2: good point, that is a prerequisite. |
19:36.56 | ivucica | :( |
19:38.42 | ivucica | ok, i'll try without initrd |
19:38.59 | ivucica | i think i built in mmc and stuff .... |
19:39.08 | ivucica | now, a backup ... |
19:39.18 | ivucica | plugging in bluetooth in my laptop -- and uploading zImage ;) |
19:57.17 | cr2 | pH5: hangs on athena |
19:57.20 | cr2 | /tmp/htc-boot -V -l |
19:57.21 | cr2 | Tx: "info 0" |
19:57.34 | cr2 | cu -l /dev/ttyUSB0 |
19:57.34 | cr2 | Connected. |
19:57.34 | cr2 | Cmd> |
19:57.52 | cr2 | Cmd>info 0 |
19:57.53 | cr2 | Cmd> |
19:58.17 | cr2 | hm. looks like echo off |
19:58.44 | pH5 | hm, should be possible to switch echo anyway. also, it has Cmd> instead of USB> or SER> |
19:59.40 | cr2 | yes. |
19:59.51 | cr2 | how do i switch echo on ? |
20:00.30 | cr2 | stty -F /dev/ttyUSB0 |
20:00.30 | cr2 | speed 115200 baud; line = 0; |
20:00.30 | cr2 | min = 1; time = 0; |
20:00.30 | cr2 | -brkint -icrnl -imaxbel |
20:00.31 | cr2 | -opost |
20:00.31 | cr2 | -isig -icanon -iexten -echo |
20:04.53 | cr2 | pH5: Cmd> is the only option |
20:05.27 | cr2 | there is a 'getdevinfo' command |
20:05.45 | pH5 | cr2: right, herald has that, too |
20:06.15 | cr2 | checksum, wdata, rmbc, lnbs, info |
20:06.16 | pH5 | I'm not sure what to do about the echo, htc-boot should enable echo on the tty. |
20:06.34 | cr2 | how ? |
20:07.02 | cr2 | erase, ls |
20:07.08 | cr2 | what is ls ? |
20:07.32 | cr2 | gpsf, gpsn |
20:07.39 | pH5 | cfmakeraw(struct termios *t) defaults to echo enabled I think |
20:08.49 | pH5 | hm, no. it _disables_ echo. so it has to be the bootloader that is sending the echo characters on magician and hx4700. |
20:10.41 | ivucica | cr2: can you check correct path for mmc card? |
20:10.52 | ivucica | _ /dev/mmc/ ...? what next? |
20:11.05 | ivucica | i need second partition |
20:12.01 | cr2 | <PROTECTED> |
20:12.09 | ivucica | works not |
20:14.02 | pH5 | cr2: I added Cmd> support, for the echo maybe I have to make the error checking less strict and/or add a buffer dump on echo errors. |
20:14.19 | pH5 | I think there is some possibility to change echo behaviour of the bootloader with the 'set' command.. |
20:14.50 | *** join/#htc-linux anYc (i=mario@hadince17.hadiko.uni-karlsruhe.de) |
20:15.19 | cr2 | pH5: updating |
20:16.10 | ivucica | cr2: is there any possibility that it's using some other device name? |
20:17.06 | cr2 | ivucica: angstrom ? |
20:17.18 | ivucica | custom hh.org kernel |
20:17.39 | ivucica | fetched from hh.org, made MMC built-in, and changed few other options |
20:17.57 | pH5 | ivucica: current 2.6? and do you have something like rootdelay=1 in your kernel commandline? |
20:18.05 | ivucica | in fact i have rootdelay=10 |
20:18.08 | ivucica | thought it might help |
20:18.48 | pH5 | ok, this is enough to make sure that this is not a race between rootfs mount and mmc init. |
20:19.18 | ivucica | i didnt bother plugging in the cradle and getting console onto serial port |
20:19.33 | ivucica | do you wnat me to set it up? it will take a minute or two |
20:19.59 | ivucica | (not at home -- i have dialup at home and that would annoy me greatly with files such as eabi toolchain sized 93mb) |
20:20.19 | cr2 | pH5: |
20:20.19 | cr2 | write(1, "Tx: \"info 0\"\n", 13Tx: "info 0" |
20:20.19 | cr2 | ) = 13 |
20:20.37 | cr2 | read(3, "\r", 1) = 1 |
20:20.37 | cr2 | poll([{fd=3, events=POLLOUT, revents=POLLOUT}], 1, 500) = 1 |
20:20.38 | cr2 | read(3, "\r\nCmd>", 17) = 6 |
20:20.38 | cr2 | poll([{fd=3, events=POLLOUT, revents=POLLOUT}], 1, 500) = 1 |
20:21.58 | pH5 | so athena can't be detected by "info 0" anymore. do I have to use getdevinfo? |
20:22.12 | pH5 | (and what is its output) |
20:22.19 | ivucica | cr2, pH5: any last thoughts? |
20:22.25 | ivucica | i'll be going now -- i'm getting hungry ;) |
20:22.48 | cr2 | pH5: does it need 'ocrnl' ? |
20:23.10 | pH5 | ivucica: I'd say setting up serial is a necessity as soon as you have kernel messages scrolling off the screen. |
20:23.27 | pH5 | I'd do that and recompile the kernel with mmc debug enabled. |
20:23.35 | ivucica | ack |
20:23.37 | pH5 | cr2: never seen that. |
20:23.37 | ivucica | will do |
20:24.19 | ivucica | odd |
20:24.27 | ivucica | i just turned on mmc debugging and now i get: |
20:24.31 | ivucica | arch/arm/mach-sa1100/usb/usb-char.c:61:27: error: asm/proc/page.h: No such file or directory |
20:24.48 | cr2 | pH5: stty -F /dev/ttyUSB0 nl |
20:24.56 | ivucica | even odder: make clean results in: |
20:24.56 | ivucica | scripts/Makefile.clean:17: /opt/kernel/linux-2.6.x-hh/arch/arm/fastfpe/Makefile: No such file or directory |
20:26.08 | cr2 | pH5: |
20:26.11 | cr2 | GetDevInfo: Get CID OK |
20:26.11 | cr2 | HTCSATHE1000TMB0HTCE |
20:26.11 | cr2 | Cmd> |
20:26.43 | cr2 | TMB0 looks like the crc32 ;) |
20:26.48 | pH5 | cr2: ok. no, it doesn't. It expects CR chars in the bootloader output. |
20:27.17 | cr2 | pH5: ok, but cu feels better with nl |
20:27.31 | cr2 | info 0 is empty. |
20:28.22 | pH5 | cr2: do you have getdevinfo on universal, too? |
20:28.32 | pH5 | magician and hx4700 don't have it yet. |
20:30.30 | cr2 | pH5: where did you get 'info 0' ? |
20:30.41 | cr2 | seems to be a very old command. |
20:30.49 | pH5 | well, hx4700 and magician are very old devices. |
20:30.52 | cr2 | there is no getdevinfo on uni |
20:30.56 | cr2 | ok :) |
20:31.09 | cr2 | here is the good list |
20:31.12 | cr2 | http://wiki.xda-developers.com/index.php?pagename=Hermes_BootLoader |
20:31.52 | cr2 | info 2 work |
20:32.09 | cr2 | info 3 works |
20:32.36 | cr2 | pH5: you probably really want 'info 3' output. |
20:33.03 | cr2 | lol |
20:33.07 | pH5 | cr2: info 3 is used for password authentication |
20:33.10 | cr2 | info 3 on athena says ... |
20:33.13 | cr2 | UNIVERSAL 1.06.0000 |
20:33.20 | pH5 | hehe |
20:33.37 | cr2 | after HTCSATHE10000T-MOB101 and hex. |
20:34.13 | cr2 | pH5: but ATHE1000 is a reliable data. |
20:34.30 | cr2 | info 4 works |
20:34.56 | pH5 | ok, so let's check whether we can get machine type info from "info 3" on all devices. |
20:35.09 | cr2 | info 5 empty |
20:35.20 | pH5 | info 4 on magician is only the vendor image version ("O2___102") |
20:35.37 | cr2 | info 6 works. |
20:35.51 | cr2 | ok, compare it with hermflasher |
20:35.56 | pH5 | info 6 just returns true or false here |
20:36.05 | cr2 | i may check hermes too. |
20:36.22 | cr2 | info 7 empty |
20:36.37 | cr2 | info 8 empty |
20:36.55 | cr2 | 7 and 8 work on hermes |
20:37.36 | cr2 | it seems htc worked hard to make them all incompatible. |
20:37.51 | cr2 | Cmd>wdata |
20:37.51 | cr2 | Invalid parameter number! |
20:38.52 | pH5 | arrgh, info 3 on hx4700 of course doesn't return any machine specific info at all |
20:39.49 | cr2 | ok, checking uni and ba. |
20:40.13 | pH5 | set 0 1 |
20:40.14 | pH5 | set 0 0 |
20:40.18 | pH5 | should turn echo on and off |
20:40.26 | pH5 | echo on seems to be default on mag and hx |
20:41.54 | cr2 | ba has info0,1,2,3,4 |
20:42.47 | cr2 | BLUEANGEL 2.05 |
20:43.25 | ivucica | cr2, pH5: nothing is printed out on serial |
20:43.55 | pH5 | cr2: and athena does only \r\n\r on info 0? |
20:44.08 | cr2 | UNIVERSAL 1.00 |
20:44.11 | cr2 | and |
20:44.14 | cr2 | UNIVERSAL 1.06 |
20:44.16 | pH5 | what happens on hermes? |
20:44.17 | cr2 | on athena |
20:44.24 | cr2 | pH5: wait :) |
20:44.35 | cr2 | there is sable too. |
20:44.59 | cr2 | pH5: yes, athena does \r on info 0 |
20:45.09 | cr2 | i used strace |
20:45.22 | ivucica | um, guys? |
20:45.41 | pH5 | ok, so I can issue "info 0" first, and when I get '\r' as first char, try echo 3 or getdevinfo or something like that. |
20:46.22 | pH5 | ivucica: kernel parameter console=ttySx <-- which one is the serial port on your device? |
20:46.24 | cr2 | ok. |
20:46.35 | ivucica | it worked with 0 in the past |
20:46.40 | ivucica | on "other" kernels |
20:46.52 | ivucica | ttyS0 and (iirc) ttySA0 |
20:47.04 | ivucica | the first worked on a 2.6 kernel, the second on a 2.4 kernel (iirc) |
20:47.52 | cr2 | pH5: hermes data is here UNIVERSAL 1.00 |
20:47.57 | cr2 | ah. |
20:47.59 | pH5 | hm. |
20:48.05 | cr2 | sorry, bad paste. |
20:49.15 | ivucica | is there a kernel option i have to turn on? |
20:50.12 | cr2 | pH5: info 3 on hermes |
20:50.15 | cr2 | HTCSHERM200VODAP001 |
20:50.21 | cr2 | then hex junk. |
20:50.33 | cr2 | info 0 is empty |
20:50.58 | cr2 | USB>getdevinfo |
20:50.58 | cr2 | Command error !!! |
20:51.17 | pH5 | ok. same on magician, only info 0 is still there. and on hx4700 I only have info 0, and info 3 only contains hex junk. |
20:51.25 | ivucica | ummmm.... yes. |
20:51.51 | ivucica | this has nothing to do with "serial i/o" right? |
20:51.59 | ivucica | (that's found under "input device support") |
20:52.22 | ivucica | ah! |
20:52.34 | ivucica | i think i found it and it's turned on |
20:52.38 | pH5 | oh wait, hx4700 has the "ROADSTER 1.01" in info 3, too. only it's inbetween the hex junk |
20:52.39 | ivucica | but it seems to suggest ttySA0 |
20:52.44 | cr2 | pH5: so hermes hangs too. |
20:53.10 | cr2 | ivucica: ttySA0 is old. |
20:53.17 | pH5 | cr2: does it still hang when you do "set 0 1" in cu before starting htc-boot? |
20:53.44 | ivucica | cr2: so, do you have any suggestion? |
20:54.12 | ivucica | Even if you say Y here, the currently visible virtual console |
20:54.12 | ivucica | (/dev/tty0) will still be used as the system console by default, but |
20:54.12 | ivucica | you can alter that using a kernel command line option such as |
20:54.12 | ivucica | "console=ttySA0". (Try "man bootparam" or see the documentation of |
20:54.12 | ivucica | your boot loader (lilo or loadlin) about how to pass options to the |
20:54.13 | ivucica | kernel at boot time.) |
20:54.41 | cr2 | pH5: trying |
20:55.22 | cr2 | pH5: i mean htc-boot hangs. |
20:55.25 | ivucica | indeed, documentation also suggests ttySA0 |
20:55.33 | ivucica | but again, it's from 2000 |
20:55.57 | ivucica | so what next? |
20:56.19 | cr2 | pH5: \r comes from echo. |
20:56.43 | cr2 | pH5: |
20:56.43 | cr2 | poll([{fd=3, events=POLLOUT, revents=POLLOUT}], 1, 500) = 1 |
20:56.43 | cr2 | read(3, "\nUSB>", 17) = 5 |
20:56.43 | cr2 | poll([{fd=3, events=POLLOUT, revents=POLLOUT}], 1, 500) = 1 |
20:56.47 | ivucica | giving several arguments for console is also valid, right? |
20:57.03 | cr2 | ivucica: yes, but only the last will be used. |
20:57.05 | ivucica | meaning console=tty0 console=ttyS0,115200n1 |
20:57.06 | ivucica | ah |
20:57.12 | ivucica | evil :/ |
20:57.23 | pH5 | cr2: really? |
20:57.45 | pH5 | I think I have both enabled on hx4700, and it gives messages to both the screen and serial port. |
20:58.11 | ivucica | cr2: if a device eg ttyS0 does not exists, it falls back to tty0? |
20:59.04 | cr2 | pH5: http://pastebin.ca/760224 |
20:59.22 | cr2 | ivucica: ttyS0 always exist on sa11x0 |
20:59.44 | ivucica | cr2: so, do you have any idea why do i not receive any traffic? :S |
21:00.31 | pH5 | cr2: argh, now you have echo enabled on both sides. |
21:00.55 | pH5 | I don't get it. whenever I run htc-boot, the tty is set to -echo afterwards. |
21:01.08 | pH5 | so here it disables local echo like it should. |
21:03.00 | ivucica | guys ... what could be the problem? or have you never experienced anything like this? |
21:03.23 | ivucica | the tty0 (correctly) receives no data |
21:03.30 | ivucica | but ttyS0 also does not receive anything |
21:04.19 | *** join/#htc-linux elysion (n=elysion@cd-a11.lkier.tontut.fi) |
21:07.44 | pH5 | ivucica: not even the uncompressor's output? |
21:08.13 | ivucica | pH5: not haret's output, not kernel's output, nothing |
21:09.36 | ivucica | i do get wince output |
21:10.00 | pH5 | as far as I can see, the uncompressor uses the first enabled uart (in the order 3, 1, 2) |
21:10.38 | ivucica | i really can't tell which one is that; i can only presume that's serial |
21:11.18 | pH5 | maybe the uncompressor output goes into the wrong uart because when booting from haret all uarts are enabled. |
21:11.37 | pH5 | I don't understand why there is nothing from the console though if it boots and you have console=ttyS0 set. |
21:14.00 | ivucica | and i'm certain that with one of the kernels i had it worked |
21:14.14 | ivucica | i've just tried with angstrom's kernel -- no luck |
21:38.04 | pH5 | cr2: updated htc-tools to use 'info 3' for identification. I hope all machines have device/version at 0x210 and 0x230 of the info 3 buffer. |
21:41.15 | cr2 | /tmp/htc-boot -V -l |
21:41.15 | cr2 | Tx: "info 3" |
21:41.15 | cr2 | Rx: "HTCS" |
21:41.15 | cr2 | Rx: "HTCE" |
21:41.15 | cr2 | Rx: "\r" |
21:41.15 | cr2 | unknown machine: "BLUEANGEL " |
21:44.08 | cr2 | /tmp/htc-boot -V -l |
21:44.08 | cr2 | Tx: "info 3" |
21:44.08 | cr2 | Rx: "HTCS" |
21:44.08 | cr2 | Rx: "HTCE" |
21:44.08 | cr2 | Rx: "\r" |
21:44.08 | cr2 | unknown machine: "UNIVERSAL " |
21:44.38 | pH5 | oh, sorry. I forgot to increase the array size in the for loop. |
21:45.13 | cr2 | i < 2 |
21:45.17 | cr2 | sizeof() |
21:46.17 | cr2 | 5 is probably redundant too. |
21:52.53 | ivucica | i will try this tomorrow |
21:53.01 | ivucica | got an entire day doing nothing ;) |
21:53.03 | ivucica | good night folks |
21:53.11 | pH5 | night ivucica |
21:53.50 | pH5 | two bug fixes in the last two commits for two bugs in the previous two commits |
21:54.04 | pH5 | --> good night :) |
21:56.02 | cr2 | <PROTECTED> |
21:56.02 | cr2 | <PROTECTED> |
21:56.02 | cr2 | <PROTECTED> |
21:56.02 | cr2 | <PROTECTED> |
21:56.17 | cr2 | /tmp/htc-boot -V -l |
21:56.21 | cr2 | Tx: "info 3" |
21:56.21 | cr2 | Rx: "HTCS" |
21:56.21 | cr2 | Rx: "HTCE" |
21:56.21 | cr2 | Rx: "\r" |
21:56.21 | cr2 | detected machine: HTC Universal |
21:56.21 | cr2 | unknown machine: "UNIVERSAL " |
21:56.58 | cr2 | /tmp/htc-boot -V -l |
21:57.02 | cr2 | Tx: "info 3" |
21:57.06 | cr2 | Rx: "HTCS" |
21:57.10 | cr2 | Rx: "HTCE" |
21:57.14 | cr2 | Rx: "\r" |
21:57.14 | cr2 | detected machine: HTC Blueangel |
21:57.14 | cr2 | Tx: "password 0~~U~~~0~~l|t000" |
21:57.14 | cr2 | Rx: "HTCS" |
21:57.14 | cr2 | Rx: "HTCE" |
21:57.15 | cr2 | <PROTECTED> |
21:59.24 | cr2 | pH5: still here ? |
22:00.02 | BabelO | lol, cr2 does it work now ? |
22:01.02 | cr2 | BabelO: in this part yes. |
22:01.13 | pH5 | cr2: hardly. |
22:01.37 | pH5 | any more stupid mistakes to fix quickly? |
22:02.41 | cr2 | pH5: don't see yet :) |
22:02.55 | cr2 | pH5: do i really need these asm blobs ? |
22:03.22 | cr2 | i'm looking at the SP and FB areas, so i can upload the zimage at some sane place. |
22:04.14 | cr2 | pH5: cool. |
22:04.18 | cr2 | /tmp/htc-boot -V -l |
22:04.22 | pH5 | cr2: only ph5test.S (trampoline code similar to what haret uses). |
22:04.22 | cr2 | Tx: "info 3" |
22:04.26 | cr2 | Rx: "HTCS" |
22:04.30 | cr2 | Rx: "HTCE" |
22:04.34 | cr2 | Rx: "\r" |
22:04.34 | cr2 | detected machine: HTC Universal |
22:04.34 | cr2 | Tx: "password ~;eg~0s~~]~\]000" |
22:04.34 | cr2 | Rx: "HTCS" |
22:04.35 | cr2 | Rx: "HTCE" |
22:04.35 | cr2 | <PROTECTED> |
22:05.08 | cr2 | pH5: i though the bl disables mmu before executing the uploaded code ? |
22:05.35 | pH5 | cr2: hehe, not this way. |
22:06.04 | pH5 | I'm just modifying the bootloader's command table to jump to a given memory location when a certain command is entered. |
22:06.22 | pH5 | That's why the htc-boot part is still highly magician 1.02 specific. |
22:07.27 | cr2 | pH5: looks like a wild hack :) |
22:07.32 | cr2 | /tmp/htc-boot -V -l |
22:07.36 | cr2 | Tx: "info 3" |
22:07.36 | cr2 | Rx: "HTCS" |
22:07.36 | cr2 | Rx: "HTCE" |
22:07.36 | cr2 | Rx: "\r" |
22:07.36 | cr2 | \r\ndetected machine: HTC Universal |
22:07.36 | cr2 | unknown bootloader version: "1.06.0000 " |
22:07.42 | pH5 | and that is exactly what it is :) |
22:07.42 | cr2 | got that on athena |
22:08.08 | cr2 | pH5: but there is some standard way to exec from ram. |
22:08.16 | cr2 | or is it SD-only ? |
22:09.15 | pH5 | cr2: if you help me to find it, I'd be more than happy to try it. maybe it is SD only, but maybe (with another wild hack) we can jump into the code there somehow. In any case, booting from bootloader command line will be highly bl specific. |
22:09.47 | cr2 | yes, i'm trying to remember how i wanted to boot off SD. |
22:10.54 | pH5 | great. good night for real now. I'll fix athena detection tomorrow. |
22:23.45 | Kevin2 | Hi |
22:24.41 | cr2 | hi Kevin2 |
22:25.40 | Kevin2 | What's new? |
22:26.59 | cr2 | many interesting developments, but nothing to present yet. |
22:30.43 | cr2 | Kevin2: i have some data for ATIDBG, so it should be possible to trace its power functions. |
22:31.10 | cr2 | then i've found some code for working with ossvcs.dll |
22:31.32 | cr2 | to enable/disable the bt/wifi/radio |
22:33.17 | Kevin2 | Sounds cool. |
22:35.50 | cr2 | here -> |
22:35.53 | cr2 | http://www.pocketpcdn.com/forum/viewtopic.php?p=27362&sid=b4738ab5d948cd0c197165912af628b8 |
23:00.31 | cr2 | BabelO: already sleeping ? |
23:01.42 | BabelO | no, looking at QTabWidget ;) |
23:02.16 | cr2 | :) |
23:05.39 | BabelO | cr2: any luck with bootloader for ba ? |
23:07.21 | BabelO | lol this is my new friend "Segmentation fault", it is same as blue screen a t work ;) |
23:08.20 | cr2 | i was tired of asm, so looking at g3.c now. |
23:08.29 | cr2 | need to mmap /deV/mem |
23:09.03 | BabelO | ok, i ve found a nice tux for my app welcome ;) tux viking |