IRC log for #htc-linux on 20090905

00:03.25*** join/#htc-linux Reactor16 (n=Reactor1@41.105.45.139)
00:03.29Reactor16hi all
00:22.16*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
00:29.40*** join/#htc-linux dilinger (n=dilinger@66.92.71.179)
00:30.31dilingerhi.  is anyone here familiar w/ the AT+PHSM commands that libhtc_ril.so seems to send when importing contacts from the sim?
00:36.12dilingeri'm noticing that the way ofono imports the phonebook results in "0000000" for each contact name; android's ril-daemon doesn't appear to have this problem, but it uses AT+PHSM to do it rather than AT+CPBR
00:39.27*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
00:47.35*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
00:50.23Reactor16anyone tried to play with RIL ?
01:12.23*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
01:29.05*** join/#htc-linux rafyvitto (i=rafyvitt@24.54.253.34)
01:33.37rafyvittodzo,cr2 any of you guys there?
01:33.41rafyvittotmzt?
01:36.30cr2rafyvitto: going back to bed :)
01:36.51rafyvittooh ok xD
01:36.56rafyvittogood night then xD
01:39.31*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
01:48.12*** join/#htc-linux druidu (n=druidu@78.97.155.94)
01:49.32rafyvittohey druidu
01:49.55dzohi rafyvitto, which ril are you using?
01:51.40rafyvittoi pm you
01:54.17druiduhey
01:54.22druiduok
02:01.37druiduwhat's new? :)
02:02.05rafyvittonot much, enjoying data connection on my raph110 xD
02:02.17rafyvittojust missing my HW keyboard xD
02:03.58rafyvittoi tried changing the mtype(to the raph110 one) on my default.txt, but when android is about to boot , the device turn's off
02:28.19tmztdzo: there's a commit that needs reverting
02:28.30tmztfor that lvd problem
02:28.39tmztwell, fixedbwouldbbe better
02:30.51tmzt18:56 < cr2> 12:18IcaptnoordIcommit
02:30.51tmzt43287807ca0dfb3ceac6a272ee90d6554ce0ebec
02:30.51tmzt18:56 < cr2> 12:18IcaptnoordIbreaks the lcd stuff
02:39.07*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
02:44.03druidu`use some more of the stacked memory` stuff?
02:45.13tmztoh
02:45.46tmztwell that should onot break interbank setup is different
02:46.01druidu43287807ca0dfb3ceac6a272ee90d6554ce0ebec
02:46.06druiduthat's what you said :)
02:47.07tmzts/obot/only/ s/inter/if the/
02:47.17tmztyeah
02:47.29tmztao it probably works fine for dzo
02:47.38tmztjust not others
02:48.17tmztdid you try banks=1
02:48.42druiduI have no problems with that, works for me
02:48.47druiduI have two banks anyway
02:49.25tmztok
02:49.33tmztdwarfed:
02:49.38tmztdwa:
02:50.44*** join/#htc-linux marcin88 (n=marcin@chello089078160132.chello.pl)
03:09.20miknixwho wants a beer?
03:10.52miknixhum.. last one, sorry :D
03:13.18druidujust had some few hours ago, Chicane had a concert in Bucharest
03:18.59*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
03:24.39darkstar62mmm, beer
04:34.49*** join/#htc-linux droid001 (n=g1@p4FDCE376.dip.t-dialin.net)
05:34.24*** join/#htc-linux stickboy (n=anonymou@128.153.211.18)
05:52.48druidudoes anyone know anything about linux kernel 2.6.29 having a different boot protocol? (talking about linux bootloaders, tags, initrd, stuff like that)
05:53.51*** join/#htc-linux AgentBlair (n=Andrew@adsl-69-225-127-88.dsl.sndg02.pacbell.net)
05:54.52tmztddrug: don't think so
05:54.59tmztdruidu:
05:55.33druiduit's strange, 2.6.29 doesn't recognize my initrd
05:55.43druiduthink it's bullshit
05:56.00druidu*thinks
05:56.02tmztin what way?
05:56.27druiduit doesn't start with the magic gzip chars
05:56.27tmztcpio.gz?
05:56.35tmztah
05:56.37tmztok
05:56.44druidui've been generating it with the same script, can't be wrong
05:56.50tmztso it thinks it's an initrd
05:57.09druiduyet after printing some bytes @ initrd_start, it looks bad
05:57.33druidunope, it doesn't, tries to unpack it and fails => panic
05:58.02tmztno CONFIG_BLOCK_INITRD?
05:58.43druiduit's not activated, but checked the code -- wouldn't do anything good or bad
05:58.56druiduactually sorry, it would do bad, but it's disabled
05:59.03druiduI get the message on the #else branch
05:59.45druiduhttp://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=blob;f=init/initramfs.c;h=d9c941c0c3cacd26fb051059540673b5e011d967;hb=refs/heads/htc-msm-2.6.29#l577
05:59.50tmztbad?
06:00.01tmztchecking if it is an INITRD fs
06:00.15tmztchecking if it is an initramfs...
06:00.19tmztno
06:00.29tmztthen it will try to mount root
06:00.35tmztthen panic
06:01.05druiduI get the `Unpacking initramfs...` message, then a panic with the message in http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=blob;f=lib/inflate.c;h=1a8e8a97812817f12460dbc880ce8fd4c53cf1cf;hb=refs/heads/htc-msm-2.6.29#l1190
06:01.24druiduif (magic[0] != 037 || ((magic[1] != 0213) && (magic[1] != 0236))) {error("bad gzip magic numbers");
06:02.03druiduit *should* be an initrd anyway :)
06:02.21druidumaybe I fucked up something in default.txt, can't figure out what
06:02.31druiduI just copy-pasted stuff from older scripts that worked
06:03.31druiduonly answer would be that I mapped something over that area and fucked it up...
06:03.41tmzt-H newc ?
06:03.50druiduyepp, that's not a problem
06:03.55tmztcat |file -
06:04.01tmztzcat |file -
06:04.29druiduit's ok, I tell you :)
06:04.50druidumanually checked first 3 bytes and they match the sig
06:04.57druiduthis must be it, it was overwritten
06:05.11druiduroot.img: gzip compressed data, from Unix
06:10.37druiduinitrd is at 0xc0508000, at least that's what the kernel is getting
06:10.57druiduinitrd_size is a bit bigger than the actual file, but I guess haret probably padded it
06:10.59tmztwhat are you doing?
06:11.48druiduplaying with 2.6.29 branch :)
06:12.07druidumanaged to get my console driver up
06:12.37druiduthey changed some stuff and had to tweak a few things
06:13.29tmztah ok
06:25.31druiduAllocated 691 pages (tags=59000000/14b81000 kernel=59001000/14b80000 initrd=591B5000/149cc000 index=592AF000/148d2000)
06:25.41druiduthat's weird, the address looks pretty strange
06:25.50druidudoesn't match the one printed in the kernel
06:26.16tmztin haret?
06:26.35druiduso initrd_start was 0xc0508000
06:26.59druiduin haret.log is initrd=591B5000/149cc000 (which one is virtual, which physical?)
06:27.53druidu2nd value in each pair is physical address
06:28.37druiduso, would 0xc0508000 map out to 0x591B5000?
06:28.57druidusorry, to 0x149cc000 (physical address used in haret)
06:31.28tmztnot sure which is virt
06:36.15druiduLaunching to physical address 148cf010
06:36.28druiduso second addr in pair is physical
06:41.18druiduoops... I think I activated the wrong option in config... `Initial RAM filesystem and RAM disk (initramfs/initrd) support`
06:41.45druiduI thought it was about the other one, CONFIG_BLK_DEV_RAM
06:42.15druiduactually no, it seems ok
06:42.51druiduCONFIG_BLK_DEV_RAM is disabled, BLK_DEV_INITRD is enabled
06:49.25*** join/#htc-linux Tinyboom (n=nahh@41.80-202-154.nextgentel.com)
06:55.49*** join/#htc-linux dcordes (n=luke-g@unaffiliated/dcordes)
06:56.02dcordesmorning
07:22.41*** join/#htc-linux kiozen (n=kiozen@rgnb-5d87c78f.pool.mediaWays.net)
07:24.25druiduhello
07:32.25druiduI've got a strange problem with initrd being apparently overwritten...
07:32.45druidubut well, it's some test on 2.6.29 branch
07:35.41*** join/#htc-linux toi (n=toi@d54C2AAB7.access.telenet.be)
07:35.45*** join/#htc-linux marex (n=marex@thor.hackndev.com)
07:48.04*** join/#htc-linux rafyvitto (i=rafyvitt@24.54.253.34)
08:14.40*** join/#htc-linux unholy (n=IceChat7@24.54.253.34)
08:16.53*** join/#htc-linux x29a_ (n=x29a@unaffiliated/x29a)
08:17.42*** join/#htc-linux marcin88 (n=marcin@chello089078160132.chello.pl)
08:22.52dcordesdruidu, hi. do you work in htc-msm-2.6.29 ?
08:24.55druiduI'm trying to get it working
08:25.02druidugot my console driver working
08:25.10druidubut this initrd problem is *really* strange
08:25.19druiduthere's nothing that could overwrite it...
08:25.31druiduall the haret-part is identical to the one for 2.6.27, which works
08:25.52druidubut for 2.6.29 somewhere, sometime, the memory gets overwritten with bullshit
08:43.51druidubastard!
08:44.13druiduI like transcribed the hexdump from the display to a file... when read as ascii it starts like: .<6>rev_crc_err_count %d, INT %x..<6>rev: got crc error without pending read.
08:44.45tmztram console?
08:45.01druiduwhich is definitely kernel code (matches a string msg in mddi.c), notice the %d %c -- it's definitely kernel code, not logs
08:45.02druidunope
08:45.02tmztare you using dzo's patch?
08:45.02druidu%d %c wouldn't show up in ram console
08:45.05tmztah
08:45.07druidunope, it's plain 2.6.29
08:45.12druidu+ some hacks to make it working
08:45.25druiduconclusion: haret doesn't support latest kernel?
08:45.57druiduor kernel is buggy
08:46.57druiduvmlinux is like 37 MB
08:46.59druiduhmmm
08:47.28druidu<PROTECTED>
08:47.28druidu#define PHYSOFFSET_INITRD (PHYSOFFSET_KERNEL + 0x500000)
08:47.33druiduthat's from haret's linload
08:55.42tmztkernel size?
08:56.05druidunot really, compressed size = 1.7 MB
08:57.08druiduit's either a bug in the arm kernel bootstrapper, or haret doesn't follow the guidelines and under certain conditions might cause the bootstrap code to overwrite stuff
08:57.43druidu(lucky me :P)
09:02.52druiduok, so at virtual address 0xc0508000 (which is mapped to physical 0x10508000) I have those strings, which are part of the vmlinux image @ offset 0x00343d8c
09:06.38*** join/#htc-linux mrmoku|away (n=mrmoku@ppp-62-216-214-29.dynamic.mnet-online.de)
09:12.09druiduthat means the kernel image was decompressed at physical address 0x101C4274...
09:12.57druiduand according to haret logs compressed kernel image started at 0x10008000 (hardcoded) and ended at 0x101BAE34
09:13.22dcordesdruidu> nope, it's plain 2.6.29
09:13.22dcordes<druidu> + some hacks to make it working
09:13.26dcordesvery useful approach imho
09:13.37druidu:P
09:14.12dcordesdruidu, did you try very small initrd?
09:14.15druiduhacks I mean disabling the debug_timer irq, using default sapphire board + everything disabled except the timer
09:14.56druiduactually it doesn't matter, the first byte gets overwritten
09:14.57*** join/#htc-linux x29a__ (n=x29a@unaffiliated/x29a)
09:15.22druiduI could write a 10 byte initrd and still won't boot
09:15.30dcordesk
09:15.47druiduit's a gzipped stream... and the kernel is like 37 MB
09:17.27druiduit overflow into the initrd area by 33 MB
09:17.30*** join/#htc-linux phh (n=quassel@2a01:e35:2e4b:b2b0:0:0:0:4)
09:17.31druidu*overflows
09:18.07druidugonna see if latest haret version changes the offset and if not patch it myself
09:23.44druiduwhat toolchain do I need to crosscompile for arm/wince?
09:25.17druidu(in linux)
09:25.20tmztcegcc
09:28.32tmztmingwce for haret actually
09:35.58*** join/#htc-linux osilvan (n=o@172.122.222.87.dynamic.jazztel.es)
09:38.40*** join/#htc-linux tsdogs (n=tsdogs@tsdogs.metalit.net)
09:40.39*** join/#htc-linux Squarc (n=Squarc@82.217.32.29)
09:42.37druidufound it, thanks
09:43.38druidunow I'll be able to script linload versions too :P
09:45.01*** join/#htc-linux toi (n=toi@84.194.170.183)
09:57.36tmztyeah
09:57.46tmztand graphical haret?
10:01.57*** join/#htc-linux dcordes_ (n=luke@unaffiliated/dcordes)
10:02.28druiduI'll compile that too, but a linux-in-a-exe version looks better :)
10:02.40druidubtw
10:02.58druiduyou wanna try it, I'm curios if it only happens to me
10:03.09druiduit's probably because of the large kernel size, lots of debug info I guess
10:03.19*** join/#htc-linux j0b0 (n=jobo@5ED416CF.cable.ziggo.nl)
10:11.06*** join/#htc-linux ellisway (n=ellis@94.30.45.128)
10:17.04*** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net)
10:20.31tsdogshi, is audio also working on raph?
10:23.57tmztgsm?
10:24.02tmztI think so
10:25.31tsdogstmtz 2.6.27 ?
10:29.40*** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net)
10:30.42tsdogstmzt: this is in 2.6.27 branch right?
10:32.15tmztyeah
10:33.34tsdogsok, what's the difference between htcraphael_defconfig and htcraphael_android_defconfig ?
10:35.00tsdogswtf shappire failed build (why is it even tryin to build that!)
10:40.51*** join/#htc-linux cr2 (n=cr2@ip-77-25-40-242.web.vodafone.de)
10:42.22tsdogscr2 hi
10:42.33tsdogstrying to boot androind on raph :)
10:44.41cr2hi tsdogs
10:44.50cr2tsdogs: you got raph100 ?
10:45.04tsdogsnope raph120 I believe
10:45.16cr2what's the difference ?
10:45.25tsdogsnot much I believe
10:45.43cr2do we have it in wiki ?
10:45.50tsdogsyep
10:45.55cr2ok
10:47.36tsdogshttp://pdadb.net/index.php?m=specs&id=1362&view=1&c=o2_xda_diamond_pro___xda_serra_htc_raphael_100
10:48.05tsdogshmm, it sais it's a raph100
10:49.11*** join/#htc-linux cr2 (n=cr2@ip-77-25-40-242.web.vodafone.de)
10:49.34cr2damn dhcp
10:50.14tsdogsis it in the boot loader that you see the version=
10:50.15tsdogs?
10:50.28*** join/#htc-linux xperia (n=chatzill@77-56-89-109.dclient.hispeed.ch)
10:52.51tsdogsreally have no idea on what's the difference, though I believe it was a 120
10:53.30cr2tsdogs: you can look in the haretlog. or under the battery
10:53.59tsdogsyeah raph120
10:54.13cr2i'm looking in my archives
10:54.28cr2don't boot the latest kernel
10:54.41tsdogshmm, just tryed it :)
10:54.54cr2got WSOD ?
10:55.10tsdogsbut something is wrong with the init
10:55.24tsdogsno, it dropped to a shell. keyboard was working :)
10:56.04cr2use initrd compiled by dcordes
10:56.19tsdogswhere is it?
10:56.34tsdogsI downloaded android for diamond from the wiki...
10:56.49cr2initrd-angstrom-2009-07-06
10:57.07cr2android is not really good for debugging
10:57.16tsdogsho, ok. http://www.megaupload.com/?d=O7Q39LOZ
10:57.33cr2btw, something for our wiki
10:57.34tsdogssure, but I wanted to start it up :)
10:57.37cr2http://www.linuxjournal.com/article/2204
10:58.11cr2were tested on only on diamond. afaik
10:58.48tsdogsok, but to make something work on raph I must start from somewhere :)
10:58.53tsdogsbrb
11:02.16cr2http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commitdiff;h=a1870cb0bcd126e3a3ba3a84f55a7cf936e7f242
11:02.21cr2this patch is problematic
11:02.30tmztcr2: what is that?
11:02.43cr2tmzt: ?
11:02.56tmztwe need a better way to do banks
11:03.02tmzt<PROTECTED>
11:03.06tmztthe
11:03.55cr2the article is about RPC
11:04.11tmztok
11:04.25tmztwe aren't really using ce-rpc though
11:04.33tmztdce
11:05.05cr2oncrpc
11:05.26cr2we should learn its wince API
11:05.42cr2to shutdown the router
11:05.45tmztah
11:05.57cr2in haret
11:07.35tmztyeah
11:07.45tmzteasier that coding it
11:16.47xperiatmzt: are you here ?
11:20.22tmztyes
11:20.30tmztwhy?
11:23.41*** join/#htc-linux SOG (n=SOG@n219079205042.netvigator.com)
11:23.48tsdogscr2: yeah latest kernel won't boot :(
11:24.23xperiado you have any recent news about microp from olipro
11:25.14tmztask him?
11:25.46xperiahe dont answer somehow my questions :-)
11:27.53tmztdid you see dzo supprt for backlig
11:28.05tmzton klt
11:34.43xperiai have seen some commits. one is the initial blinking of the leds before the system boot however i dont believe it will works.
11:35.04xperiafor the htc-kovsky device.
11:35.56xperiai am asking me why the two microp version of raph/kovsky differ so mutch.
11:37.25xperianearly everything is different.
11:38.21xperiaas a example. raphael has this here
11:38.26xperia#define MICROP_KLT_ID_VERSION           0x30
11:38.37xperiahtc kovsky need this here
11:38.50xperia#define MICROP_KLT_ID_VERSION           0x07
11:39.19xperiaraphael has microp versions like this
11:40.04xperia0105, 0205, 0a05, 0b05, 0d02
11:40.06xperia8182, 8185
11:40.07xperia0c82, 0c85
11:40.35xperiahtc kovsky has a version of this
11:40.37xperia0787
11:41.08*** join/#htc-linux cr2_ (n=cr2@ip-77-25-25-19.web.vodafone.de)
11:41.29tsdogspwd
11:41.34*** join/#htc-linux cr2_ (n=cr2@ip-77-25-25-19.web.vodafone.de)
11:41.55tsdogscr2_: should I revert that commit ?
11:42.02tsdogsin my local repo i mean
11:42.08xperiaraphael works with the MICROP_KLT_ID_LED_STATE
11:42.31xperiahtc-kovsky works with some hardcoded led tables
11:42.47xperiait is everything different
11:43.33cr2_tsdogs: i think yes
11:43.40cr2_tsdogs: maybe even more
11:44.00cr2_xperia: the tables are hardcoded on raph100 too
11:44.00tsdogsok, going to try this first. should I keep other comiits?
11:44.46cr2_druidu: wtf is ram console ? and why is it at 0xe00000 ? this area is protected by MPU, because it's amss .text there
11:45.32cr2_tsdogs: i'll look at others
11:46.07*** join/#htc-linux marex (n=marex@thor.hackndev.com)
11:46.48tmzttgata google's
11:46.56tmztthats
11:47.36tmztI guess its shadowed
11:47.49cr2_ok, but ten point it to some decent location
11:48.33xperiacr2: in this case just the MICROP_KLT_ID_LED_STATE could be wrong. As i know this value is crucual for the working of the leds on all htc-kovsky devices.
11:48.35xperiabut how can i find out the right MICROP_KLT_ID_LED_STATE for this device.
11:48.36xperiaI have done a brute force allgorithm that set MICROP_KLT_ID_LED_STATE from 0x0 to 0xff
11:49.03xperiathe value 0x18 for MICROP_KLT_ID_LED_STATE shutdown the LCD Backlight
11:49.13tmztcr2: it surviea reboot
11:49.56xperiaand on the value 0x128 i am getting as a feedback on the console some heavy strang 3 lines of rpc stuff
11:49.56tmztonly that value?
11:50.15tmztcan you type them
11:50.29cr2_tmzt: you cant write over amss ;)
11:50.52tmztyou might have found a
11:51.03xperiayeah i can just need to boot it. one moment please
11:51.31tmztvreg
11:51.37tsdogscr2_: didn't work WSOD again
11:51.50tmztcr2_: right, but it could be shadowed by ram
11:52.00cr2_tsdogs: ok, then we need to revisit all recent patches
11:52.15cr2_tmzt: what can be shadowed ?
11:52.44tmztdzo said there's a second bank of smem at 0x20000000
11:52.48cr2_itt's a protected physical ram location.
11:52.50tmztthe amss area
11:52.56cr2_tmzt: not on raph
11:53.03tmztok
11:53.05tmztyeah
11:53.08tmzton diam
11:53.09cr2_tmzt: 288MB means 256+32
11:53.17cr2_32 is 1 sram bank
11:53.30cr2_yes, 192=128+64
11:53.38cr2_64 is two banks
11:53.47tmztbut split
11:53.57cr2_but 0xe00000 is in the first bank, and belongs to amss
11:54.07tmztat 0x10000000 and 0x20000000
11:54.10cr2_split ?
11:54.15xperiaokay i have boote now in linux on the htc-kovsky device and the brute force algortithm is counting and probing MICROP_KLT_ID_LED_STATE from 0x19 till 0xff.
11:54.17xperiaat the moment i am at 0x44
11:54.21tmztdevided
11:54.22cr2_the SDRAM is split on raph
11:54.33cr2_sram is 32+32
11:54.37cr2_if you have 64
11:54.44tmztdivided
11:54.51tmztyeah
11:54.52tmztok
11:55.01cr2_but the only reliable place for this info is manufacturer nand section
11:56.09xperiaokay 0x80 is the value
11:56.50tmztwhat does 128 do?
11:56.54cr2_tsdogs: compile the rev which is known to work. but you will miss sound :)
11:57.04tmztwhat's the rpc message?
11:57.38xperiatmzt: it prints the lines to fast. have to rebuild the kernel with this value becouse the algorithm dont stop on this value and the screen is flooded with the other values
11:58.11tmztyeah that
11:58.15tmztor dmesg
11:58.36tmztcr2_: did you email cyanogen ?
12:02.43cr2_tmzt: ?
12:03.07tmztyou must have left
12:03.20tmztabout dumping MSM_CLK
12:03.31tmzton g1
12:03.51cr2_yes, such sumb is a must :)
12:04.01cr2_s/sum/dum/
12:04.08cr2_lol
12:04.45cr2_hehe. debian updates oo packages
12:06.36xperiatmzt: 128 is 0x80 or at least nearly. 0x80 is the right value.
12:07.01tmztnearly?
12:07.34xperiatmzt: what value is 0x80 in decimal
12:07.42tmztcr2_: what were you trying to say then?
12:07.54tmzt128 I thnk, 8*16
12:08.22cr2_tmzt: can some g1 owner provide us with this data ?
12:08.47cr2_tmzt: the 0x400 byte block starting at 0xa8600000 (physical)
12:09.01tmztcyanogen said he would, wanted you to email a patch
12:09.33tmzthe's the on doing donut images
12:09.37cr2_a patch ?
12:09.51cr2_does he have a working /dev/mem ?
12:10.21tmztI don't know
12:10.43tmztseems google blocks the lowest 32k
12:11.10cr2_of msm_clk ?
12:11.27tmztof ram
12:11.46cr2_512k on wince. i think
12:11.47tmztin the mmap/misc drivers
12:12.10cr2_MPU for spl/resume code path
12:12.48cr2_tmzt: where is the MSM_CLK remapped ?
12:13.09cr2_google messes with the gp_clk, so they must have mapped it somewhere
12:13.11tmztiomap.h?
12:13.49cr2_.27 kernel ?
12:14.00tmztyeah
12:14.09tmztthat might be the wrong file
12:14.21cr2_msm_io*
12:14.25tmztmaybe iomem, but it's in mach-msm include
12:15.20cr2_haha
12:16.48cr2_msm_iomap.h
12:17.07cr2_MSM_CLK_CTL_BASE
12:17.48cr2_just dump the 0x400 bytes behind it.
12:19.12cr2_if /dev/mem does not work
12:20.12tmzt[right
12:20.49Bushmanave
12:28.06tsdogscr2_: this is the revision with problems (have done a bisection) http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commitdiff;h=43287807ca0dfb3ceac6a272ee90d6554ce0ebec
12:28.17tsdogsnow I'll try reverting this.
12:29.39cr2_revert the board-htcraphael.h changes
12:29.49cr2_we don't have the second sram bank
12:30.14tsdogsI did a revert on the commit, now trying to boot
12:30.53cr2_the board-htcdiamond.c part is ok
12:31.02cr2_at least for diam100
12:31.14*** join/#htc-linux ptitjes (n=didier@93.2.7.143)
12:31.18cr2_i don't know if all diam* have 64MB sram
12:31.22tsdogsok, but just want to find the raph first
12:31.29cr2_ok
12:31.55cr2_but the lcd init patch needs to be reworked too
12:32.17tsdogsok, reverting that boots up :)
12:32.34cr2_yeah. the memory map was fscked
12:32.53Bushmanyay!
12:33.02Bushmanmemory map was fscked!
12:33.04cr2_i see that the last patch sets pmdh_clk @ 0xa04
12:33.05Bushman:D
12:33.32cr2_it's not true for raph. check MSM_CLK wiki page
12:33.34tsdogscr2_: no idea on what that means
12:34.06cr2_but with the correct value it does not boot too
12:34.18tsdogslol
12:34.19cr2_ok
12:34.50cr2_dcordes_: still here ?
12:36.06tsdogsandroid crashed though :/
12:37.29tsdogscr2_: is there an angstrom image somewhere for raph ?
12:37.47tsdogsdon't want to go all the hassle to build one :)
12:37.50cr2_tsdogs: i have it locally, but we need a way to recompile it
12:37.56cr2_yeah
12:38.01tsdogshmm...
12:38.03*** join/#htc-linux pH5 (n=ph5@g229229115.adsl.alicedsl.de)
12:38.32tsdogsok, I'll give it a try.
12:40.13cr2_tsdogs: 7.2MB
12:40.52tsdogsthx
12:41.29cr2_tsdogs: dcc ?
12:41.49tsdogshmm, no idea using konversation
12:42.23cr2_accept the transfer
12:42.53*** join/#htc-linux miknix (n=miknix@gentoo/developer/miknix)
12:44.06tsdogscr2_: I did accept :)
12:44.38cr2_times out. may be the firewall
12:44.49tsdogsyeah, try this one: http://195.32.70.240/tsdogs/openupload/index.php?action=u
12:45.12tmztdzo: + htcraphael_process_mddi_table(client_data, mddi_lcm_init_table,
12:45.13tmzt+ ARRAY_SIZE(mddi_toshiba_init_table));
12:45.28tmztwrong size?
12:47.33cr2_tsdogs: uploading :)
12:47.37tsdogsok
12:47.53cr2_tsdogs: no progress bar ?
12:48.06*** join/#htc-linux x_coRe_x (i=coRe@95-37-154-107.dynamic.mts-nn.ru)
12:48.10tsdogshmm I think I miss some extension on that server ...
12:48.19cr2_ok
12:48.22tsdogsphp isn't very nice for upload traking
12:50.48cr2_complete upload
12:53.30tsdogsdid u set some description or so and finished it?
12:54.28cr2_http://195.32.70.240/tsdogs/openupload/www/?a=d&i=BUGMcv
12:54.45tsdogsthx
12:55.24cr2_tsdogs: is it free software ?
12:55.40tsdogsyeah: http://openupload.sf.net/
12:55.58*** join/#htc-linux pH5_ (n=ph5@e179170144.adsl.alicedsl.de)
12:56.00cr2_i'm dreaming of disabling ftp upload at work ;)
12:56.05tsdogslol
12:56.27tsdogsit needs some more working but it's usable.
12:56.51cr2_if there is an upload excryption option, that'll be nice.
12:57.03tsdogsafterwards u mean?
12:57.23tsdogsor during?
12:58.19cr2_you may use https
12:58.35cr2_ok, i've looked at the site. looks nice :)
12:58.36tsdogsyes sure for encryption during upload
12:58.44tsdogswasn't designed by me :)
12:59.12cr2_ok
12:59.54tsdogslike expected :) ... perl-native_5.8.8.bb, do_configure) failed
12:59.57tmztcr2_: could that have been the problem with lcd init?
13:00.07tsdogsok, gg now bbl
13:00.18tmztor I guess dzo doesn't detect the model
13:00.39cr2_tmzt: it's completely broken
13:01.03cr2_i don't know why he has patched board-htcraphael.h with diamond map
13:01.17tmztyeah
13:01.33tmztbut if devices are detectable it shouldn't matter
13:01.36cr2_tmzt: i don't know of any raph with 64MB sram (smi) and 256MB sdram (ebi)
13:02.10cr2_we need a full device table for known devices.
13:02.20tmztI mean lcd device
13:02.27tmztebi?
13:02.46cr2_documenting and parsing the nand may have been even better
13:02.53cr2_but it may be tricky in wince
13:03.15cr2_maybe with the code from itstutils
13:03.36tmztcould we read from spl area?
13:03.43tmztit must know
13:03.46cr2_yes
13:03.54cr2_but it's not spl
13:04.12cr2_with itsutils you can actually read everything
13:04.47cr2_by disabling nand mpu
13:05.40cr2_we need somebody to correct the raph memory layout in git
13:05.49cr2_dzo: ping
13:06.54*** join/#htc-linux cr2 (n=cr2@ip-77-25-25-19.web.vodafone.de)
13:10.56cr2tmzt: apropos ram console
13:10.59cr2+#define MSM_LOG_BASE 0x0e0000
13:11.00cr2+#define MSM_LOG_SIZE 0x020000
13:11.16cr2e0000, not e00000
13:11.24cr2druidu got it wrong
13:11.53cr2e0000 is just behind the spl (MPU'd)
13:12.00tmztah
13:17.50dcordes_cr2, yep
13:18.21cr2dcordes_: we need to fix the raph memory map
13:20.23dcordes_cr2, what needs changing?
13:21.22cr2i'm looking if it will hurt diamond
13:23.35tmztdiamond_fixup should only affect diam mtype I would think
13:23.57cr2tmzt: it's the hardcoded layout in board-htcraphael.h
13:24.20cr2#define MSM_PMEM_MDP_BASE       MSM_SMI2_BASE
13:24.22cr2#define MSM_PMEM_MDP_SIZE       0x800000 /* 8mb */
13:24.24tmztcan that be lit?
13:24.33tmztsplit
13:24.35cr2this is completely wrong
13:25.09cr2but there is no board-htcdiamond.h
13:25.34cr2either we create it, or do some #ifdef spaghetti
13:25.58tmztwill that break multiple mtype though?
13:26.10tmztmaybe we should do per-cpu
13:26.24tmztsince the stacked is different
13:26.30cr2raph100 is different from raph800
13:26.31tmztwell
13:26.40tmztdiam/raph is same
13:26.42tmztyeah
13:26.53tmztbut those are msm7501a
13:27.01cr2look at arch/arm/mach-msm/board-htcraphael.h
13:27.04tmztraph500 is msm7500a
13:27.06cr2has nothing to do with the cpu
13:27.24tmztsoc
13:27.50tmztsmi is fixed for 7201a?
13:28.39cr2nand+sram is a separate samsung chip
13:28.39tmztanyway, I mean having these defines breaks things
13:28.48tmztoh
13:29.16cr2yeah, we may copy this file to board-htcdiamond.h
13:29.23cr2and fix it.
13:29.29tmztright
13:29.50tmztbut then the same zImage won't be usable on raph/diam etc.
13:29.53cr2there is arch/arm/mach-msm/board-htcblackstone.h already
13:30.08tmztunless baord.c is split also
13:30.15tmztincluding cdma's
13:30.16cr2and all defines are for raph100 anyway.
13:30.50cr2cdma is done more cleanly via pdev
13:31.01cr2no really hardcoded ifdefs
13:31.15tmztI don't understand whhy google used these defines for pmem driver anyway
13:31.22tmztwell
13:31.33cr2arch/arm/mach-msm/board-htcdiamond.c includes
13:31.40cr2#include "board-htcraphael.h"
13:31.43tmztI'm thinking this is a kernel-wide header
13:31.50tmztfor some reason
13:31.55cr2dcordes_: still here ?
13:32.09cr2tmzt: it was done by googel
13:32.10dcordes_yes
13:32.10tmztright on cdma
13:32.25tmztbut smem layout/size is different
13:32.29cr2for allocating large dma'able areas
13:32.30tmztI know
13:32.42cr2or other non-transparent closed-source stuff
13:33.09cr2i'd be happy to get rid on pmem* for non-android.
13:33.35tmztwe still have fixed fb
13:33.38cr2because it's stupid design
13:34.10cr2yeah, fb is pita now.
13:34.26cr2don't know if .29 may have been better.
13:34.33cr2need to wait for druidu for that
13:34.58tmztfor what?
13:35.01cr2dcordes_: can you copy board-htcraphael.h to board-htcdiamond.h ?
13:35.16dcordes_yes
13:35.34cr2dcordes_: and make arch/arm/mach-msm/board-htcdiamond.c #include "board-htcdiamond.h" ?
13:35.45dcordes_ok
13:35.57cr2then  we can revert the board-htcraphael.h changes.
13:37.43cr2i'm curious if this patch by dzo will work on raph :)
13:37.46cr2-       { .clk_id = PMDH_CLK, .offset = 0x8c, .ns_only = 0xa00, .name="PMDH_CLK"},
13:37.48cr2+       { .clk_id = PMDH_CLK, .offset = 0x8c, .ns_only = 0xa04, .name="PMDH_CLK"},
13:38.35tmztseems it neds to be able to save those values
13:39.01tmztoh, not the pm part
13:39.05tmztjust clk
13:40.40cr2clk/dev  g1  raph100  blac100  raph800
13:41.01cr2mddi_clk  ?  0xa19  0xa19  0xa41
13:41.22cr2PMDH_CLK is mddi_clk
13:41.42cr2the ? for g1 needs the MSM_CLK dump for g1 ;)
13:42.19cr2tmzt: btw, what is te short name for topaz ? TOPA100 ?
13:43.20xperiatmzt: this here is the output  of the rpc lines
13:43.21xperiaINFO:  task rpcrouter: 11 blocked for more than 120 seconds.
13:43.23xperia"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disable this message
13:43.24xperiarpcrouter D c01f4418 0 11 2
13:43.26xperia.....
13:43.27xperiasome kernel stamp like
13:43.29xperia.... (shedule+.... from ... (rr_read ....
13:46.55tmztxperia: that happens every time?
13:47.24tmztcr2: don't know topaz
13:48.17xperiawith my brute force algorithmus that test MICROP_KLT_ID_LED_STATE from 0x0 to 0xff that happen allways
13:48.25tmztcr2: xperia found something on kovs microp that intereres with rpc
13:48.58*** join/#htc-linux ellisway (n=ellis@genkt-048-006.t-mobile.co.uk)
13:49.54xperiai you want i can post the algorithm so you can veryfi it on your device
13:50.22*** join/#htc-linux cr2 (n=cr2@ip-77-25-25-19.web.vodafone.de)
13:51.39tmztis it writing that value or an uninterruptable loop in the kernel that's causing it?
13:53.53dcordes_cr2, unfortunately I cannot push due to the linuxtogo.org breakage
13:54.55xperiatmzt: this is my code here in the microp-klt.c file
13:54.56xperiafor( ; x<0x81; x++)
13:54.58xperia{
13:54.59xperiaprintk(KERN_WARNING "MICROP_KLT_ID_LED_STATE = %d\n", x);
13:55.01xperia//buffer[0] = MICROP_KLT_ID_LED_STATE;
13:55.02xperiabuffer[0] = x;
13:55.04xperiabuffer[1] = 0xff & state;
13:55.05xperiabuffer[2] = 0xff & (state >> 8);
13:55.07xperiadata->led_states = state;
13:55.08xperiar = micropklt_write(client, buffer, 3);
13:55.10xperiamdelay(1500);
13:55.11xperia}
13:55.56cr2dcordes_: pastebin a diff
13:56.19tmztmdelay?
13:56.43xperiafor a more easy reding and watching on the screen
13:58.08xperiashould i remove this as this is maybe the problem ? i will try a new algorithm outside of the mutex locking now
14:01.09dcordes_cr2, http://pastebin.ca/1554944
14:01.35Bushman0xf00d
14:01.59dcordes_Bushman, is this the cheeseburger register?
14:02.07cr2dcordes_: thanks
14:02.15Bushmannods
14:02.38xperiatmzt: you need somehow to call the function micropklt_set_led_states i am doing this allways on the keyboard slider event
14:02.59dcordes_brb
14:03.35xperiathis way i can run the algorithm allways when i need it
14:05.27xperiacr2: do you know where this string here can be found. is this a linux kernel output ?
14:05.29xperia"INFO: task rpcrouter: 11 blocked for more than 120 seconds."
14:16.36xperiahmmm strange git grep "task rpcrouter" dont give any results. this output is not in the kernel.
14:18.49phhmm, sorry for disturbing, i'm trying last (from the diamond's android thread on xda) TaC's package of android for diamond, and it seems the detected SD size is false (with every FAT related error). It seems to think the partition is only ~ 900kB. Any idea ? already fixed ? pebkac ?
14:21.31*** join/#htc-linux Marajin_ (n=marajin@87-194-102-189.bethere.co.uk)
14:23.46phhand last release from connect-utb "works" (means android do starts), and it's a diam100 if that can help
14:24.25*** join/#htc-linux dcordes (n=chatzill@f054180100.adsl.alicedsl.de)
14:36.45xperiahave done now also a binary search on the Generic WinMo 6.1 Kovsky Rom.
14:36.47xperiafor the string "rpc" i am gettting nearly a file with 1 Mb search results
14:36.48xperia"rpcrouter" however give me zero search results.
14:36.50xperiaThis string must be on the hardware or comming from the hardware.
14:36.51xperiaTrying now my new kernel with the new brute force allgorithm
14:43.17*** join/#htc-linux luc_ (n=luc@89-115-128-35.cl.ipv4ilink.net)
14:46.01*** join/#htc-linux zmc-mobile (n=zmc@m3c5e36d0.tmodns.net)
14:46.55zmc-mobileHas anyone else had experience with interrupts not working on omap hardware in 2.6.30-series kernels?
14:48.24*** join/#htc-linux dcordes (n=luke-g@unaffiliated/dcordes)
14:56.02ali12341yes i have :)
14:56.13zmc-mobilelol.
14:56.25ali12341but seriously, all the opem people hang out at linwizard :)
14:56.28ali12341*omap
14:56.32zmc-mobileI was hoping that someone outside of linwizard had seen the problem. :)
14:59.16zmc-mobileidk why but I can't seem to fetch from miknix
15:00.10miknixzmc-mobile, no worries, do you want me to send you a patchset?
15:00.49phhthe exact detected size for my bug is 981120
15:04.10*** join/#htc-linux Gnutoo (n=gnutoo@ABordeaux-152-1-84-218.w86-196.abo.wanadoo.fr)
15:11.24xperiait looks like that i have crashed this hardware based linux rpc router on my htc-kovsky device
15:11.25xperiaINFO: task rpcrouter: 11 blocked for more than 120 seconds.
15:11.27xperia"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disable this message
15:11.28xperiarpcrouter D c01f4418 0 11 2
15:19.48*** join/#htc-linux LTxda (n=anon@c-98-194-107-10.hsd1.tx.comcast.net)
15:41.40cr2awk '{print 768*1/4/4}'
15:41.42cr248
15:42.03cr2freq=48000000 MD=0x1fffb NS=0xfffc1b59 M=1 N=4 D=2 P=4 S=1 C=2 s0=r4=0x3 s1=r5=0x2 s2=r6=0x1 s3=r7=0x11
15:42.16cr2with the missing bits fixed, it looks good.
15:44.02*** join/#htc-linux x_coRe_x (i=coRe@95-37-154-107.dynamic.mts-nn.ru)
15:44.13*** join/#htc-linux paran (n=paran@cust.fiber-lan.snet.lk.212.214.112.112.visit.se)
15:45.36cr2awk '{print 768*5/32/2}'
15:45.45cr260
15:46.02cr2freq=60000000 MD=0x5ffdf NS=0xffe40b49 M=5 N=32 D=16 P=2 S=1 C=2 s0=r4=0x1 s1=r5=0x2 s2=r6=0x1 s3=r7=0x1
15:46.35cr2<PROTECTED>
15:46.36cr2<PROTECTED>
15:46.38cr2<PROTECTED>
15:46.40cr2<PROTECTED>
15:46.57cr2yeah. looks good :)
15:47.24dcordescr2, where is 768 from ? is that usb clock ?
15:47.26cr2PLL1
15:47.32cr2S=1
15:47.53cr2source=1 is PLL1, source=4 is PLL0
15:48.35cr2so usb_hs is 48MHz and usb_otg (host) is 60MHz
15:49.14*** join/#htc-linux Zoolooc (n=fredsiba@p5495571F.dip.t-dialin.net)
15:52.28cr2M=225 N=2048 D=1024 P=1 S=4
15:52.35cr2uh. big numbers
15:52.58cr2pll0 was ...
15:54.04cr2<PROTECTED>
15:54.15cr2druidu has 245 on his diam
15:54.53cr2awk '{print 196608000*225/2048/1}'
15:54.54cr221600000
15:54.58cr2a bit little
15:56.04cr2awk '{printf("%d\n",245000000*225/2048/1)}'
15:56.05cr226916503
15:56.19cr2that looks better. 27MHz for tvout.
16:00.25paranhi. does your kernel have drivers for audio,gps,camera,wifi etc that works with newer HTC devices like dream, sapphire, hero?
16:00.58*** join/#htc-linux Gnutoo (n=gnutoo@ABordeaux-152-1-84-218.w86-196.abo.wanadoo.fr)
16:01.36paranI have a sapphire ("HTC magic") and is getting a bit depressed after realizing how much of the hardware requires HTC proprietary files
16:02.45cr2paran: gps is processed by the modem firmware
16:03.20cr2we have the interface to it, but it does not yet work.
16:04.13paranok, I am trying to build and run the latest android from source, with limited sucess.
16:04.23cr2for the audio proprietary is only the mixer, filtering et al. nobody worked on that, but with some patience you can do it.
16:05.01cr2the camera is more complex.
16:05.17parangot your vouge RIL (android-htc-ril) working after a little work
16:05.26cr2paran: you can get the binary libraries from htc itself.
16:05.37parancan make and recieve calls, send/rec SMS, but no data traffic yet
16:05.56cr2works for me in non-android setup
16:06.02cr2umts data
16:06.07parancr2: problem is that all the HTC libs are for android 1.5
16:06.31cr2btw, do you have /dev/mem ?
16:07.00cr2you will (may) have a problem anyway, because you have different modem firmware.
16:07.20cr2and the interfaces are different.
16:07.27Gnutooparan, I've liberated the g1(audio(with the help of cr2's playwav2.c) and ril(pin don't work yet))
16:07.35Gnutoothe project is called replicant
16:07.54paranGnutoo: cool, that will likely work on sapphire as well. do you have the code online somewhere?
16:08.01cr2Gnutoo: why does the pin not work ? it's really beyond my understanding ?
16:08.03Gnutooparan, yes on replicant wiki
16:08.09paranGnutoo: url?
16:08.13cr2Gnutoo: do you have /dev/mem ?
16:08.31parancr2: not now, but I am running my own 2.6.29 kernel, so that can be arranged
16:08.36Gnutooparan, that's because of a race: AT+CPIN=>+CREG: 2
16:08.53Gnutooso for the generic ril the response of AT+CPIN is CREG:
16:08.56Gnutooand not:
16:09.01cr2Gnutoo: the codeaurora git has msm alsa driver.
16:09.07Gnutoo+PB_READY
16:09.11Gnutoowow
16:09.29parancr2: is all your work currently 2.6.27 based?
16:09.30cr2Gnutoo: i don't know how good it is.
16:09.36Gnutooparan, http://trac.osuosl.org/trac/replicant/wiki/BuildAndroid
16:09.54parancr2: your htc-msm-2.6.29 seems identical to googles android-msm-2.6.29 branch
16:10.07cr2paran: we use 2.6.27, but will switch to 2.6.29 eventually
16:10.11*** join/#htc-linux druidu (n=druidu@78.97.155.94)
16:10.13paranGnutoo: I got pin working by a big kludge
16:10.13Gnutooparan, the patch is apache 2.0 but I don't know where copyright should go on a patch
16:10.21Gnutooparan, how?
16:10.28cr2hey druidu
16:10.31Gnutooby removing singleline?
16:10.48Gnutoodiffing openmoko's ril and the generic one shows that
16:11.08*** join/#htc-linux swc|666 (n=infidel2@unaffiliated/swc666/x-4934821)
16:11.46paranGnutoo: I always return RIL_E_SUCCESS :)
16:12.02cr2lol
16:12.12Gnutooparan, basically I'll explain the problem
16:12.22Gnutooat first you got PASSWORD_INCORRECT
16:12.51druiducr2: 0xe0000 survives a reboot
16:12.57Gnutoothen you dig into it and see that the     err = at_send_command_singleline(cmd, "+CREG:", &p_response); got error -6
16:13.00druiduif you get a crash, you dump it with haret and you have panic logs :)
16:13.33Gnutoothen you look what error -6 means into the atchannel.h
16:13.38Gnutoo: #define AT_ERROR_INVALID_RESPONSE -6
16:13.50Gnutoothen you look when there is an invalid response
16:14.02Gnutoothen you learn that there are no intermediate response
16:14.07druidubtw, about what tsdogs, htcraphael_android_defconfig won't work by default because BOARD_SAPPHIRE is on by default (and it's incompatible with our current boards files), so because I didn't update that option when I synced with upstream... by default it'll get activated
16:14.13cr2druidu: yes. it was 0xe0000 vs 0xe00000
16:14.33cr2druidu: there is 1 more bug.
16:14.47Gnutooand then you see that there is no intermediate response because of that:
16:14.59cr2druidu: dzo added 64MB sram to raph , which does not exist
16:15.14cr2druidu: here is the patch http://pastebin.ca/1554944
16:15.21paranGnutoo: looking at logcat I see that the modem replies: +CREG: 2\n+CREG: 2\nOK
16:15.34cr2druidu: but dcordes could not commit it to linuxtogo git
16:16.30*** join/#htc-linux TravelerRom (n=traveler@s55918972.adsl.wanadoo.nl)
16:17.07paranGnutoo: so instead of figuring out why that was not accepted I just hard coded a success, I wil fix it up later :)
16:17.18Gnutooif                 && strStartsWith (line, s_responsePrefix) => add intermediate response
16:17.26Gnutoook
16:17.42cr2M=1 N=75 D=37 P=1 S=0
16:17.50Gnutooso if it does not starts with the response prefix...no intermediate response
16:17.55Gnutooand here are some debug logs:
16:18.07paranGnutoo: which RIL code do you use?
16:18.17Gnutoothe free one
16:18.24Gnutoothe generic ril
16:18.36Gnutoos/generic/reference
16:18.43Gnutoohere's debug code:
16:18.47cr2awk '{printf("%d\n",19200000/75)}'
16:18.48cr2256000
16:18.54cr2256kHz for sdac ?
16:18.55GnutooE/AT      (   33): REMOVE_THAT: line : +CREG: 2;E/AT      (   33): REMOVE_THAT response : +PB_READY
16:19.17GnutooREMOVE_THAT is my debug string
16:19.28Gnutooline is what we get
16:19.34TravelerRomnewbee question: I can't seem to find a how-to or install guide for raphael, does someone have a URL for me?
16:19.35Gnutoo+PB_READY is what we expect
16:20.09dcordesTravelerRom, what do you want to run?
16:20.38TravelerRomdcordes, an android test version for htc raphael
16:21.33cr2<PROTECTED>
16:21.34cr2<PROTECTED>
16:21.35paranGnutoo: Have you tried the vouge ril? android-htc-ril git
16:21.38cr2wow. cool
16:21.44*** join/#htc-linux Squarc1 (n=Squarc@82.217.32.29)
16:22.04Gnutooparan, no I've only tried the generic one...vouge has pin working ? and is it for msm modems?
16:22.47paranGnutoo: that is the one I have based my work on... PIN only works with my ugly kludge
16:22.55Gnutoook
16:23.04paranGnutoo: but I can make/recive calls, send/recieve SMS
16:23.08Gnutoome too
16:23.20Gnutoobut only if it's provisioned
16:23.33dcordesTravelerRom, don't know where android builds for raphael are sorry
16:23.35Gnutoo(for receiving calls)
16:24.12dcordescr2, I'm going to build some new images. do we need anything new for testing?
16:24.25TravelerRomdcordes, thanks anyway!
16:24.54druiduparan: you can extract htc proprietary files from Android Dev Phone's system image
16:24.58cr2768/=256
16:25.22cr2druidu: developer.htc.com
16:25.33druiduno, they don't have all the files there
16:25.35parandruidu: or from my own sapphire imge. but my goals is to not use proprietary stuff
16:25.45Gnutoodruidu: me too
16:25.49Gnutoo(same goal)
16:26.04GnutooI've used only the wifi firmware
16:26.24Gnutooril and audio are free on my phone
16:26.44druiduhttp://www.htc.com/www/support/android/adp.html => get signed-dream_devphone_userdebug-img-150275.zip, extract system.img with unyaffs, then modify the script vendor/htc/dream-open/extract_files.sh to get files from the extracted image, instead of adb
16:27.15parandruidu: main problem is that the proprietary stuff only works with android 1.5, I am trying to run android from git
16:27.36cr2Gnutoo: we need to compare  the wince wifi firmware with googels
16:27.39druiduI'm not sure if you have a choice about the firmware stuff
16:27.49druiduand you'd have to write all the HAL libraries
16:28.05druiduI'm running android from git, latest version
16:28.10druiduand works with the proprietary stuff
16:28.17parandruidu: firmwares is ok... :)
16:28.36cr2Gnutoo: i really need the MSM_CLK(+0x400) dump from g1 :)
16:28.52Gnutoocr2, what's that?
16:28.55parandruidu: hm... how did you manage that? with android from git I get no gps, no compass, no RIL, no wlan.
16:29.16druiduno point rewriting all the hardware abstraction libraries for now, we can just adapt our drivers to provide same interface, then use the proprietary libs
16:29.50druidulol, sorry, I don't have an original android-phone, I run diamond with minimal stuff working for now...
16:30.47cr2Gnutoo: do you have /dev/mem on g1 ? can you read from it ?
16:30.54Gnutoocr2, I'll look
16:31.47parandruidu: I thought the problem was that the interfaces had changed in android git, so proprietary libs from 1.5 no longer works
16:32.14Gnutoocr2, seems that I don't have /dev/mem
16:32.17parancr2: any information you need, I can rebuild a kernel with /dev/mem.
16:32.18Gnutoobut I have /dev/pmem
16:32.21druiduI see, then we have a problem
16:32.34druiducr2: compiled 2.6.29 kernel and tried to boot it, made my console driver work but initrd gets overwritten when uncompressing the kernel...
16:33.04cr2druidu: use a sane initrd ?
16:33.44cr2druidu: how did you manage to compile the 37MB kernel ? too much debugging symbols ?
16:35.06Gnutoobtw still no Xorg working on msmfb?
16:35.39dcordestmzt, said he'd try to implement the raster ideas on it
16:35.59cr2M=1 N=8 D=4 P=4 S=7
16:36.03dcordesGnutoo, also curious about that. would be nice to have it working properly
16:36.17Gnutooabout pin?
16:36.23cr2Gnutoo: pmem is some other thing.
16:36.28Gnutoook
16:36.48cr2Gnutoo: you may enable /dev/mem in the kenrel .config, and recompile the kernel
16:36.58Gnutoocr2, yes I know
16:37.14Gnutoocr2, but should I do it if I've somewhat random access at internet?
16:37.41cr2Gnutoo: anyway, i need a dump of the 0xa8600000-0xa86003ff physical address space area.
16:37.41Gnutoosomeone is searching in the koolu forum for me on how to fix the pin
16:38.04Gnutoo(because I need registration....)
16:38.30cr2M=1 N=8 D=4 P=4 S=7 for vfe
16:38.39cr2S=7 is something new
16:39.12cr2don't know its frequency. hehe
16:39.39cr2not in the 7230 list too
16:39.43cr2<PROTECTED>
16:39.44cr2<PROTECTED>
16:39.46cr2<PROTECTED>
16:39.47cr2<PROTECTED>
16:39.49cr2<PROTECTED>
16:39.50cr2<PROTECTED>
16:41.39cr2so vfe is 1/32(1/(8*4) of this S=7
16:41.52Gnutoowow
16:41.58Gnutooour research found that:
16:42.00Gnutoohttp://www.koolu.org/releases/beta7/Changelog
16:42.05cr2druidu: does linuxtogo git commit work for you ?
16:42.09Gnutoo<PROTECTED>
16:42.29Gnutooso we should use openmoko's code
16:42.32Gnutoofor the pin
16:42.39Gnutooand indeed that's what I thought
16:42.48Gnutoothe diff shown that
16:42.55Gnutooat_send_command_singleline -> at_send_command
16:43.08Gnutoobut I didn't know how it worked to check the readyness
16:44.29Gnutooperfect I should get it working soon
16:44.41*** join/#htc-linux Zinbolic (n=zinbolic@84.238.80.215)
16:44.45Gnutooif onUnsolicited is the same
16:44.50cr2<PROTECTED>
16:44.52cr2<PROTECTED>
16:45.38cr2CLK_BRIDGE(AXI_LI_VFE,  GLBL_CLK_ENA_SC,        B(23),  AXI_LI_APPS),
16:46.07cr2<PROTECTED>
16:48.53cr2M=1 N=15 D=7 P=4 S=4 ecodec
16:49.36cr2awk '{printf("%d\n",245000000*1/15/4)}'
16:49.38cr24083333
16:49.46cr2~4MHz
16:50.50cr2or
16:50.54cr2awk '{print 196608000*1/15/4}'
16:50.55cr23276800
16:53.30druiducr2: 37 MB uncompressed, 1.7 compressed
16:53.58druidudunno, probably a lot of debug info but nothing excessive... msm_defconfig minus some drivers
16:54.36cr237MB is huge
16:55.01cr2you have probably compiled the debug symbols into the kernel. check this option in .config
16:55.31cr2M=1 N=5 D=2 P=2 S=4 icodec rx
16:55.56druidujudging by the fact that it compresses down to 1.7 MB... yes, it's a lot of text there
16:56.24druiduwhat's your vmlinux size for 2.6.28?
16:56.37druidu(i don't wanna recheckout the branch and recompile everything :P)
16:56.38cr2M=1 N=30 D=15 P=4 S=4 icodec t
16:56.53cr26 times slower. weird.
16:57.04druiduwell, kernel is compiled with debug info
16:57.09druiduCONFIG_DEBUG_INFO
16:57.21cr2yeah, 31MB
16:57.24cr231320094  3. Sep 22:25 vmlinux
16:57.43cr2we should remove it some time ;)
16:57.45druidunot much of a difference
16:57.56druiduyeah, like when we finish with the debug :))
16:58.11Gnutoommm how do I set the sim status: there is no setSIMStatus
16:58.18druiduhttp://www.linuxjournal.com/article/2204 -> not that useful, it's general rpc stuff, mostly the tcp/udp protocol
16:59.07GnutooI'll hack arround it
16:59.08Gnutoobye
16:59.56druiduGnutoo:  I get that too, although phone is connected to gsm network :)
17:01.58parancr2: what is at 0xa8600000-0xa86003ff ?
17:02.08cr2paran: MSM_CLK
17:02.13cr2control
17:02.24cr2http://www.htc-linux.org/wiki/index.php?title=MSM_CLK
17:02.37druiduanybody knows if arm-mingw32ce toolchain is compatible with 64bit linux?
17:02.54cr2druidu: ??
17:02.55druiduI get this: ..... /mingw32ce/bin/../libexec/gcc/arm-mingw32ce/4.4.0/cc1plus: error while loading shared libraries: libmpfr.so.1: cannot open shared object file: No such file or directory
17:03.09druiduthere's no 32bit package for libmpfr.so.1
17:03.23druiducr2: ?
17:03.30cr2you need an appropriate libmpfr.so.1
17:03.41cr2hehe.
17:03.46druiduI straced cc1plus, it looks for it under /usr/lib32
17:03.50cr2complile it with -m32
17:04.04druiducompile what? the lib?
17:04.05cr2from source
17:04.11cr2yes
17:04.19druiduah, too bad ubuntu guys didn't package the lib :)
17:04.29cr2of link it to /lib/libc.so
17:04.31parancr2: ok. I'l compile a new kernel later, have some other stuff to do first...
17:04.40cr2paran: thanks.
17:05.08cr2a19 vs a41 va a0c vs a04
17:05.10cr2weird
17:06.32cr21.1001, 100.0001, .1100, .0100
17:07.03cr2quite a difference ;)
17:07.30druiduanyone knows flags to dpkg-buildpackage to build 32-bit lib?
17:09.26cr211.001 1000.001 1.100 0.100
17:10.02parandruidu: you can set up a 32-bit pbuilder
17:11.34parandruidu: or why not simply download the 32 bit deb package, unpack and grab the files you need.
17:12.09druiduah, from x64 repo
17:12.38druidu*x86
17:15.39paranright. http://packages.ubuntu.com/jaunty-updates/i386/libmpfr1ldbl/download
17:21.55cr2druidu: need this patch in git :) http://pastebin.ca/1554944
17:22.59druiduone patch coming right up
17:26.42cr2don't see it
17:27.07druidufatal: git apply: bad git-diff - expected /dev/null on line 17
17:27.20druiduit's not done yet, I had to switch branches
17:27.32druiduI gate git apply, it never works...
17:28.00cr2ok
17:32.53druidudone
17:33.14*** join/#htc-linux rafyvitto (i=rafyvitt@24.54.253.34)
17:34.48cr2druidu: diamond should use it ?
17:34.49druiducheck if it's ok, gotta go soon http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=3e9999048397251a33fecd73480e434c37dcf3bf
17:35.02druiduwhat do you mean?
17:35.12druidushit
17:35.19druiduforgot to add .c to commit
17:35.25cr2-#include "board-htcraphael.h"
17:35.27cr2+#include "board-htcdiamond.h"
17:35.29druiduI know
17:35.45druidugit apply really sucks, did it semi-manually with patch -p1
17:36.05cr2you may also revert the board-htcraphael.h patch by dzo
17:36.22cr2i need to get some food, before it's too late ;)
17:36.49druiduhttp://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=ea2eecc036471903f416cb6599e4e99ad6e2c3a6
17:36.58druiduI amended the patch, do a pull with --force
17:38.27druidu+       mi->bank[0].size = (107 * 1024 * 1024); // Why 107? See board-htcraphael.h
17:38.30druiduthis stays right?
17:38.45druidu(board-htcdiamond.c)
17:41.32cr2should be "See board-htcdiamond.h"
17:41.36cr2#ifndef CONFIG_MACH_SAPPHIRE
17:41.38cr2<PROTECTED>
17:41.39cr2#endif
17:41.46cr2why is that in the raph board file ?
17:42.24druiduhttp://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=commit;h=539f890dde3b08e88c8f25e889d6f2d58be224c6
17:42.28druiduthis should do it
17:43.06druiduwait wait
17:43.10druidudon't pull :)
17:43.20cr2wait. the diamond can have  mi->bank[0].size = (107 * 1024 * 1024); // Why 107?
17:43.37cr2only the board-raphael.h patch should be reverted.
17:43.40druidudidn't revert that part
17:43.41cr2bbl :)
17:43.47druidubut don't pull yet
17:44.00cr2i'm away to get food.
17:45.15*** join/#htc-linux dreamer_ (n=dreamer@c-71-239-79-222.hsd1.il.comcast.net)
17:45.31dreamer_what file do I need to modify to correct the keymap on my apache?
17:45.38dreamer_none of the fn buttons work
17:46.45druiduok, patch done, you can pull now (--force maybe)
17:47.08druidui'm out, bye
17:47.19*** part/#htc-linux druidu (n=druidu@78.97.155.94)
17:52.05*** join/#htc-linux darkstar62 (n=darkstar@97-126-104-157.tukw.qwest.net)
17:52.42*** join/#htc-linux balans (n=Administ@212-123-149-239.ip.telfort.nl)
17:55.05*** join/#htc-linux Squarc (n=Squarc@82.217.32.29)
18:31.33cr2ok, i can boot the git head
18:33.36*** join/#htc-linux balans (n=Administ@212-123-149-239.ip.telfort.nl)
18:34.02cr2[    5.568389] +mddi_client_nt35399_init()
18:34.07cr2hm
18:46.13cr2weee
18:46.26cr2playwav2 plays sound :)
18:46.59cr2root@htcraphael:~# playwav2 KDE_Startup.wav
18:47.13cr2mickey|sports: ping
18:48.47cr2dcordes: here ?
18:49.42cr2found 39 snd endpoints
18:49.44cr2<PROTECTED>
18:49.45cr2<PROTECTED>
18:49.47cr2<PROTECTED>
18:49.56cr2...
18:49.59cr2Select device 1.
18:50.00cr2rpc_snd_set_device(1, 1, 1)
18:50.02cr2Set master volume to 5.
18:50.03cr2rpc_snd_set_volume(13, 1, 5)
18:50.05cr2playwav: 1 ch, 22050 hz, 16 bit, PCM
18:50.06cr2prefill
18:50.08cr2start
18:51.01cr2for mp3:
18:51.04cr2cannot open mp3 output device: No such device
18:51.56cr2open("/dev/msm_mp3", O_RDWR)            = -1 ENODEV (No such device)
18:52.22cr2root@htcraphael:~# ls -als /dev/msm_mp3
18:52.23cr2<PROTECTED>
18:52.35cr2hmm
18:53.15cr2playwav2 -rec foo.wav
18:53.24cr2root@htcraphael:~# playwav2 -rec foo.wav
18:53.25cr2found 39 snd endpoints
18:53.32cr2Select device 1.
18:53.33cr2rpc_snd_set_device(1, 1, 1)
18:53.35cr2Set master volume to 5.
18:53.36cr2rpc_snd_set_volume(13, 1, 5)
18:53.38cr2buffer size 2048 x 8
18:53.39cr2*** RECORDING * HIT ENTER TO STOP ***
18:54.56*** join/#htc-linux stickboy (n=anonymou@128.153.210.13)
18:55.34cr2record works, but does not record anything ;)
18:56.00cr2mixer problem.
18:57.25cr2<PROTECTED>
18:57.58*** join/#htc-linux g55 (n=root@rgnb-5d87ca52.pool.mediaWays.net)
18:59.21cr2<PROTECTED>
18:59.22cr2<PROTECTED>
18:59.24cr2<PROTECTED>
18:59.25cr2<PROTECTED>
18:59.27cr2<PROTECTED>
19:00.13cr2<PROTECTED>
19:00.15cr2//  do_route_audio_rpc(0,SND_MUTE_UNMUTED,SND_MUTE_UNMUTED);
19:02.21cr2[ 1858.555145] adsp: module AUDRECTASK: READY
19:02.23cr2[ 1858.564880] [RPC] REPLY to xid=41 @ 0:00000001 (24 bytes)
19:02.24cr2[ 1858.576232] audpre: CFG ENABLED
19:02.26cr2[ 1858.589172] adsp: bogus task id 17
19:02.27cr2[ 1858.597839] adsp: bogus task id 17
19:02.29cr2[ 1858.606292] adsp: bogus task id 17
19:02.30cr2[ 1858.614715] audrec: PARAM CFG DONE
19:02.32cr2bogus task id 17 does not look good
19:10.32rafyvittohey cr2,do you know how i can enable the hardware keyboard, using one of the latest diamond kernel's?, i have a raph110, ive been looking at the raph default.txt from jun trying to for something that might enable the hardware keyboard
19:10.52rafyvittofind*
19:11.09cr2rafyvitto: it works for me. but i'm not using android.
19:11.36cr2probably you need to change something in android libs/interface
19:17.02xperiacr2: do you know for what CONFIG_MSM_BATTCHG is good ?
19:17.28xperiai have CONFIG_MSM_SMEM_BATTCHG too ?
19:18.18cr2i think so
19:18.52*** join/#htc-linux balans (n=Administ@212-123-149-239.ip.telfort.nl)
19:19.05xperia<PROTECTED>
19:19.16cr2xperia: do you have working sound now ?
19:19.44cr2as the name says: battery charger monitor.
19:20.19xperiai am trying to reboot the kernel. can tell you the result in about 1 hour. P.S. thanks for the answer.
19:20.31xperiarebuild the kernel
19:20.47cr2ok
19:30.56*** join/#htc-linux surge (i=surge@pool-98-118-154-21.bflony.fios.verizon.net)
19:33.38cr2hm. the dsp firmware support for .mp3 may be missing in out amss
19:34.13cr2at least wince does not use it too
19:34.15*** join/#htc-linux crakou (n=crakou@fon38-1-82-233-110-43.fbx.proxad.net)
19:34.36*** part/#htc-linux crakou (n=crakou@fon38-1-82-233-110-43.fbx.proxad.net)
19:38.45cr2<PROTECTED>
19:38.46cr2<PROTECTED>
19:38.48cr2<PROTECTED>
19:38.49cr2<PROTECTED>
19:39.25*** join/#htc-linux balans (n=Administ@212-123-149-239.ip.telfort.nl)
19:39.48cr2int adsp_init_info(struct adsp_info *info)
19:39.49cr2{
19:39.56cr2<PROTECTED>
19:41.12cr2/* Table of modules indexed by task ID for the QTV_LP image */
19:41.13cr2static qdsp_module_type qdsp_qtv_lp_task_to_module_table[] = {
19:41.41cr2<PROTECTED>
19:41.46cr2this is =17
19:41.57cr2sizeof()=20
19:42.07cr2hmm. puzzled now.
19:42.12*** join/#htc-linux leaigor (n=laigor@188.134.36.14)
19:43.12cr2but 17 should be ok, since 18,19 and 20 are dummies anyway
19:46.36cr2#define QDSP_RTOS_MAX_TASK_ID  19U
19:48.04dcordescr2, back now
19:48.15dcordestook a nap
19:48.35xperiacr2: what for a amss version is right in raphael_android exist CONFIG_MSM_AMSS_VERSION=7000 in raphael it is CONFIG_MSM_AMSS_VERSION=5200
19:49.02*** join/#htc-linux balans1 (n=Administ@212-123-149-239.ip.telfort.nl)
19:49.29cr2xperia: where did you see 7000 ?
19:50.20xperiahtcraphael_android_defconfig in arch/arm/configs
19:51.39dcordescr2, wow congrats on the sound!
19:52.04dcordescr2, gotta try it on xperia
19:52.25cr2xperia: it's too old
19:52.30cr2<PROTECTED>
19:52.36cr2looks like a bug to me
19:52.57cr2should be  QDSP_RTOS_MAX_TASK_ID probably
19:53.15cr2tmzt: can we ask San about it ?
19:55.15xperiadcordes: making a fresh git repo at the moment on my hard disk. and analyzing all the configs at the moment.
19:55.31dcordescr2, do you have a patch to try sound?
19:55.33dcordesxperia, nice
19:56.03cr2dcordes: you don't need a patch. it works as is with the latest git head
19:58.52xperiacr2: for what is this here good. i dont have it in my kovsky config. CONFIG_DEFAULT_MMAP_MIN_ADDR=4096
19:58.56*** join/#htc-linux cnossa (n=cnossa@a213-22-39-95.cpe.netcabo.pt)
19:59.42cr2excludes mmaping of the zero page
20:00.09xperiais this good. should i enable it ?
20:00.18cr2why not
20:00.32xperiaokay
20:05.47cr2[130858.068878] audpre: CFG ENABLED
20:05.48cr2[130858.080902] audrec: PARAM CFG DONE
20:05.50cr2[130858.088562] audpre: err_index 0
20:05.51cr2[130858.095977] audpre: unknown event 1
20:05.53cr2[130858.103363] audpre: unknown event 2
20:05.54cr2[130858.110656] audpre: err_index 2
20:05.56cr2[130858.117828] audpre: unknown event 1
20:06.10cr2bogus task is gone, but new came in ;)
20:08.52dcordescr2, what about in call audio?
20:08.53cr2dcordes: what about adding startgps to initrd ?
20:09.53dcordescr2, where's stargps ?
20:10.10*** join/#htc-linux ph5 (n=ph5@g229145139.adsl.alicedsl.de)
20:10.43dcordesph5, hi
20:10.52ph5hi dcordes
20:11.32dcordesph5, is it expected that (some) ltg accounts/fileservers don't work?
20:13.32cr2dcordes: i think i have sent it to you.
20:14.35ph5dcordes: I am not aware of any such thing.
20:15.25cr2dcordes: startgps -> https://privatepaste.com/dd1hQU6VVQ
20:15.52*** join/#htc-linux leaigor (n=laigor@188.134.36.14)
20:19.46dcordescr2, can you give me the md5sum of your current initrd?
20:20.00*** join/#htc-linux ph5 (n=ph5@g229162124.adsl.alicedsl.de)
20:20.25cr271138b2d61ff186321eaaa4ff3c732603536e1faa96091163c8d8537a64aed3a7941d4fffe707682178c044f13fc655b413d176af457fb2a0c387572ca575df1  initrd-angstrom-2009-07-06
20:20.48cr2sorry. it was sha512 :)
20:20.51cr2426a4ea53f6a46af440781de80f421f0  initrd-angstrom-2009-07-06
20:25.12dcordescr2, don't have that one locally
20:25.27*** join/#htc-linux leaigor (n=laigor@188.134.36.14)
20:34.09dcordescr2, http://www.angstrom-distribution.org/~lgorris/
20:46.30*** join/#htc-linux rafyvitto (i=rafyvitt@24.54.253.34)
20:55.27*** join/#htc-linux Squarc1 (n=Squarc@82-217-32-29.cable.quicknet.nl)
21:08.30*** join/#htc-linux dwaradzyn (n=chatzill@chello089079067084.chello.pl)
21:09.04*** join/#htc-linux marex (n=marex@thor.hackndev.com)
21:10.33dwaradzynwith the last commit (Revert "Use some more of the stacked memory.") does it boot correctly on diams and raphs?
21:11.06cr2dwaradzyn: yes
21:11.33dwaradzynfor blackstone there is no white screen now. it simply hangs. without "boot console for HTC phones" in config it only vibrates and nothing
21:11.57dwaradzynwith "boot console" it goes to console handover and hangs
21:12.20cr2are you using ramdisk or android ?
21:12.55dwaradzynremoving "use console only during boot" makes it go to "io scheduler anticipatory registered (default)"
21:13.23dwaradzyncurrently android. i will try angstrom with latest kernel
21:13.45cr2ok
21:25.34dwaradzynwith angstrom i have the same symptoms - hangs on scheduler
21:26.29dwaradzynit looks like that: http://tinypic.com/r/6h6e4i/3
21:37.27parancr2: if you was interested in some sapphire info I now have /dev/mem
21:39.15*** join/#htc-linux kw (n=nnscript@adsl-178-60-36.gnv.bellsouth.net)
21:39.55cr2paran: can you try
21:40.26kwAny progress with CDMa diamond, etc?
21:40.40cr2dd if=/dev/mem of=g1clk bs=4 skip=$((0xa8600000/4)) count=0x100 ?
21:40.59cr2kw: with the sound ? not yet.
21:41.04kwAnything my friend
21:41.13kwMy cdma diamond doesn't turn on radio lol
21:41.53dcordes<PROTECTED>
21:42.27cr2dcordes: can you skip= with hexdump
21:42.44cr2dcordes: otherwise you will fail on reading the first page.
21:43.24dcordescr2, sorry, what do you mean, skip= ?
21:43.40cr2mickey|zzZZzz: zzz ?
21:43.53mickey|zzZZzzgoing to bed now...
21:44.02cr2dcordes: hexdump /dev/mem reads from 0
21:44.10mickey|zzZZzzzzZZzz ... you know ... like in comics, when someone sleeps
21:44.19cr2mickey|zzZZzz: playing sound works for me with angstrom
21:44.37mickey|zzZZzzcool. so do you think i can boot a kernel tomorrow?
21:44.39cr2mickey|zzZZzz: i meant it's wrong time for zzz :)
21:44.45mickey|zzZZzzah, well, almost midnight here
21:44.52cr2mickey|zzZZzz: yes. i've done it 2 hours ago.
21:44.54mickey|zzZZzzand tomorrow i'm playing squash, need to get up early
21:45.11cr2ok, just fyi.
21:45.40*** join/#htc-linux phnom (i=simomn@pub98-78.pub.luth.se)
21:45.54dcordescr2, what's the problem about reading from first page?
21:46.23mickey|zzZZzzthanks, I'll try to boot then
21:46.29mickey|zzZZzzthe defconfig is pretty trashed though
21:46.31ali12341dcordes: null pointer dereference?
21:46.36mickey|zzZZzzwill take me some boots to get it back on track
21:47.05cr2dcordes: it will take some time waiting for the output at 0xa8600000
21:50.10cr2paran: does reading from /dev/mem work for you ?
21:50.32dcordescr2, what about 0xa8600000 ? it's not shown
21:50.49kwanyone get the radio to work on diam500
21:50.50kw?
21:51.31cr2kw: you need the raph800 setup for it, i guess.
21:52.45kwAh, which build should i use>
21:52.55kwwhen i run that TAc build i get some ramaddr error
21:53.29cr2your device is probably not detected by haret
21:54.14kwAh
21:54.31kwTry diff version of haret?
21:54.36parancr2: I just get "dd: /dev/mem: Bad address"
21:54.59cr2paran: yeah. something is broken in the g1 kernel
21:55.21cr2paran: i get the same message on raph100
21:55.35parancr2: this is not a g1, and its my own kernel :)
21:56.07cr2paran: the easiest thing to do is to white a loop into the kernel itself
21:56.20cr2in trout_init or something like that
21:56.51dcordesbrb
21:56.54ali12341i don't think /dev/mem likes dd
21:57.01kw"Please set start of ram
21:57.04kwis the error i get
21:57.15ali12341i use this instead: http://al.robotfuzz.com/~al/wizard/devmem3.c
21:57.17cr2ali12341: always worked for me even on sa1100 ;)
21:57.37cr2ali12341: can you link it statically ?
21:57.50ali12341yeah. not that it uses anything exotic
21:58.23cr2kw: this message means that haret did not detect your phone as having the MSM750cA cpu
21:58.29cr2kw: this message means that haret did not detect your phone as having the MSM750xA cpu
21:58.30*** join/#htc-linux dcordes_ (n=luke@unaffiliated/dcordes)
21:59.05kwi have a 7501A
21:59.05kwHmm odd.
21:59.32cr2ali12341:  This software has been developed for the LART computing board
21:59.45ali12341right, it's based on devmem2
21:59.45cr2ali12341: lol. it was a sa1100 based machine.
22:00.10ali12341it's the same as the devmem in busybox pretty much, except it takes a length param and writes to stdout
22:00.33ali12341the static version is 0.5mb
22:01.51cr2paran: do you have it in .config ? CONFIG_DEFAULT_MMAP_MIN_ADDR=4096
22:02.16cr2i'll change it to 0
22:02.38ali12341http://al.robotfuzz.com/~al/wizard/dm3
22:03.26ali12341cr2: we're having irq problems with omap and 2.6.31-rc8 - any idea how to debug it? basically it seems like no irqs work at all
22:04.09ali12341eg serial ports work in polled mode until the kernel switches to interrupt mode, nothing else works at all
22:05.06cr2ali12341: debugging irqs is the most complex task at all
22:05.51ali12341it's really annoying me
22:06.10ali12341the kernel boots all the way up but nothing works
22:06.41parancr2: devmem3.c worked, I got a dump
22:07.55ali12341paran: beware of the bug: if the start address isn't page aligned it gets rounded down silently
22:08.24*** join/#htc-linux ximon (i=simomn@pub98-78.pub.luth.se)
22:09.12cr2paran: can you upload it somewhere ?
22:09.33cr2ali12341: 0xa8600000 is page aligned.
22:09.47ali12341sure
22:10.20parancr2: http://pastebin.ca/1555403
22:10.23cr2paran: it may be also interesting to dump the RPCCALL fifos. before starting gps, and after starting gps.
22:11.22cr2paran: can you redo the dump in 4byte mode and up to 0x400 ?
22:11.43cr20x100 was in 4byte chunks=0x400
22:15.20paransure
22:16.01cr2od -t x4 -A x instead of hexdump
22:16.54paranhowever do dump before/after gps I have to flash and run the HTC original image. in my curreny build the gps don't work at all
22:18.01cr2paran: just dump the shared mem (1MB@0x01f00000) some time later
22:18.14cr2right now i'm interested in the MSM_CLK
22:19.12cr2but i'm pretty sure that we will have a free implementation of non-assisted GPS really soon.
22:21.06parancr2: http://pastebin.ca/1555415
22:22.46cr2paran: that looks very good.
22:23.11cr2paran: i'll paste it into MSM_CLK wiki.
22:23.22dcordes_strip: Unable to recognise the format of the input file `startgps.o'
22:23.42cr2hm. ugly *
22:24.00cr2od -v -t x4 -A x for the future :)
22:24.03dcordes_oh I should use the arm one
22:24.28cr2dcordes_: or better binutils
22:25.00parancr2: I like that, you dont have to actually read all the zeros...
22:26.12cr2paran: the table includes them already.
22:26.29cr2fixed length for the table is also betetr
22:38.14dcordes_cr2, http://www.angstrom-distribution.org/~lgorris/startgps http://www.angstrom-distribution.org/~lgorris/initrd-angstrom-2009-09-06
22:39.34cr2dcordes_: does it include i2ctools ? devmemX ?
22:41.46dcordes_it's the current initrd with /usr/bin/startgps and kovsky keymap file
22:42.32dcordes_there are /dev/mem , /dev/kmem
22:43.31cr2the raph/g2 side-by-side CLK comparison. very useful. http://www.htc-linux.org/wiki/index.php?title=MSM_CLK
22:43.43cr2now if somebody did it for g1.
22:44.16cr2we know the clock frequences from g1/g2 source, so the register settings can be backtracked.
22:46.36cr2ec/f0 is different
22:49.14dcordes_cr2, I get the same freeze on 'ioscheduler' message during boot like dwaradzyn http://i28.tinypic.com/6h6e4i.jpg
22:51.01cr2dcordes_: ok, then something is missing, or is behind #ifdef raphael/diamond
22:51.49dcordes_cr2, diamond works with HEAD ?
22:56.46cr2dcordes_: it should
22:57.16dcordes_cr2, I'm browsing the commit log since kovsky initial commit
22:58.28cr2ok
22:59.53cr2hehe. ec/f0 is vdc
23:00.15parancr2: would a CLK dump from a Hero help in any way? I could probably get that quite easy
23:00.54cr2is it g1 ?
23:01.15cr2if it's a different linux kernel, then yes.
23:03.43xperiai cant compile the kernel for kovsky !
23:04.30xperiaFor some reason saphire things are builded. why that ?
23:04.32xperiaarch/arm/mach-msm/built-in.o: In function `config_gpio_table':
23:04.33xperia/mnt/sdb1/XPERIA/KERNEL/kovsky-dev/kernel/arch/arm/mach-msm/board-sapphire.c:998: undefined reference to `msm_proc_comm'
23:04.35xperiaarch/arm/mach-msm/built-in.o: In function `sapphire_mddi_power_client':
23:04.36parancr2: I would boot my friends Hero with my 2.6.29 kernel... the stock HTC kernel does is 2.6.27 and does not have root or /dev/mem
23:04.36xperia/mnt/sdb1/XPERIA/KERNEL/kovsky-dev/kernel/arch/arm/mach-msm/board-sapphire-panel.c:418: undefined reference to `msm_proc_comm'
23:04.38xperia/mnt/sdb1/XPERIA/KERNEL/kovsky-dev/kernel/arch/arm/mach-msm/board-sapphire-panel.c:431: undefined reference to `msm_proc_comm'
23:04.39xperia/mnt/sdb1/XPERIA/KERNEL/kovsky-dev/kernel/arch/arm/mach-msm/board-sapphire-panel.c:441: undefined reference to `msm_proc_comm'
23:04.41xperia/mnt/sdb1/XPERIA/KERNEL/kovsky-dev/kernel/arch/arm/mach-msm/board-sapphire-panel.c:451: undefined reference to `msm_proc_comm'8
23:05.21parancr2: and Hero is not a g1.
23:05.28cr2paran: ok. you may also at the same time switch on the camera, gps, bluetooth, wifi, bt headset and such weird things :)
23:05.59cr2paran: which kernel is it using. i know only g1 aka trout, and g1 aka sapphire
23:06.45cr2xperia: these are some merging artefacts. i have some SAPPHIRE defines in the raph board file where they make noe sense, and will never be reached,.
23:07.28cr2xperia: the Kconfig should prevent compiling other board files: board-sapphire-panel.c
23:08.23xperiabut they must be new. could it be that they are related to druidu has commits. he has implemented some morse vibra code
23:09.46parancr2: the stock hero rom most likely runs 2.6.27, as it is adroid cupcake. or what do you mean?
23:10.14cr2paran: the kernel source
23:10.15xperiafor the raphael build. as i can remember he needed msm_proc.
23:11.27cr2M=1 N=2 D=1 P=4 S=1 vdc on g2
23:11.49cr2runs at 768*1/2/4 MHz
23:12.11cr296MHz
23:12.15parancr2: wouldn't surprise me if it is the same as sapphire, the hardware is very similar. but I don't know
23:12.27*** join/#htc-linux MrPippy (n=pip@adsl-75-11-160-45.dsl.sndg02.sbcglobal.net)
23:13.28tmzt16:10 < cr2>  info->max_task_id = 16;
23:13.29tmzt16:10 < cr2> looks like a bug to me
23:13.29tmzt16:10 < cr2> should be
23:13.29tmzt<PROTECTED>
23:13.29tmzt<PROTECTED>
23:13.31tmzt16:11 < cr2> tmzt: can we ask San
23:13.34tmzt<PROTECTED>
23:14.05tmztcr2: #android is just AOSP now and very quiet
23:18.02dcordes_hrmm
23:18.05dcordes_no worky
23:18.14Bushmanyay! i has tarantula!
23:18.20Bushmanhttp://en.wikipedia.org/wiki/Brachypelma_albopilosum
23:18.40cr2tmzt: do they have some kernel bugtracker ?
23:19.38cr2tmzt: i've fixed that one, but hit some other weirdness
23:19.55Bushmancr2: hi, wha'choo fixing? ;]
23:20.00tmztjust b.android.com
23:20.41Bushmanandroid for one of the htc's?
23:22.37tmztcr2: by g2 you mean maguc/sapphire/sapp?
23:22.57cr2yes
23:23.18tmztyou should know there are two boardids for sapp with different smi size, but radio can be flashed on either
23:23.25cr2tmzt: the clock-msm7230.c from codeaurora is an incredible source of data
23:23.51cr2i'm wondering if i'll find such hardcoded clock tables also in our amss.
23:24.00parantmzt, cr2: mine is the pvt32a
23:25.19tmzt19:16 < cr2> hehe. ec/f0 is vdc
23:25.25tmztwhat is that?
23:27.25cr2the clocks are used differently, and you can see the register setting differences in the dump
23:28.26cr2<PROTECTED>
23:28.28cr2<PROTECTED>
23:28.49cr2<PROTECTED>
23:28.51cr2<PROTECTED>
23:28.52cr2<PROTECTED>
23:28.54cr2<PROTECTED>
23:29.21cr2so these are PCLK too, although their name does not say it.
23:29.36tmztwhat do you mean by used differently?
23:29.49tmztand they have different numbers?
23:30.01dcordes_I tried everything that touches only raphael and diamond
23:30.06cr2one device uses a different clock source, another does not use it at all.
23:30.21dcordes_it must be something general that breaks kovsky and blackstone
23:30.26tmzthow can it not use it?
23:30.28cr2you make an rpc call, and still see the result in the MSM_CLK dump
23:30.43cr2tmzt: it may be disabled.
23:30.54dcordes_few lines before io scheduler there is clock-wince printk
23:31.20tmztyes, but if there is nobsdc pclk how does mmc work?
23:31.21cr2dcordes_: which one ?
23:31.29dcordes_clk set rate
23:31.31tmztor does it enable it when needed
23:31.46dcordes_I will put an older version of clock_wince.c
23:31.53cr2tmzt: there is SDC1_CLK and SDC1_PCLK "clock" in the g1 api
23:31.58tmztit's not iosched, that's normal
23:32.10cr2<PROTECTED>
23:32.14tmztI know
23:32.25cr2the B(7) here is for the PCLK
23:32.29tmztso what's the issue?
23:32.57cr2and the MD/NS settings (in S) are for the _CLK
23:33.18cr2we don't treat them separately so far.
23:33.39cr2<PROTECTED>
23:33.42tmztwe use the google way I thought
23:33.54tmztbut not proc_comm
23:33.55cr2the B(25) is USBH_PCLK
23:34.13tmztbit?
23:34.35cr2GlblClkEna  GlblClkEna  GLBL_CLK_ENA
23:34.46cr2this is for the PCLK bits
23:35.01tmztbit 25?
23:35.10cr2some clocks (which are not Peripheral) do not have such bits.
23:35.26cr2they must have some other source
23:35.50cr2graphics is
23:35.54cr2<PROTECTED>
23:36.13tmztany documentation anywhere on axi?
23:37.44cr2i doubt it
23:38.05tmztis it a qualcomm thing?
23:38.13cr2i think we may actually find such binary table in the amss
23:38.49tmzthave you talked to viperbjk?
23:39.48cr2no
23:40.21cr2and the Md/Ns pairs are very good visible in the binary register dump.
23:41.47cr2tmzt: i've already decoded 95% or the clock setup
23:42.01tmztin amss
23:42.10dcordes_http://pastebin.ca/1555500 this is how kovsky boot freeze looks
23:42.36cr2tmzt: no, just for msm7201A
23:43.02dcordes_exactly the same as blac
23:43.03tmztce then?
23:43.26cr2clock-wince: pc_clk_set_rate: id=7 rate=19200000
23:43.27cr2clock-wince: set mdns: 7, 19200000; bitidx=0, offset=5c, ns=a06
23:43.32tmztdcordes_: it's just htcfb handover isn't it?
23:43.41cr2it's the LCD mddi bridge clock
23:44.10cr2dcordes_: probably somethig in the LCD setup is missing, or makes them unhappy
23:44.42dcordes_ok I'll review log
23:44.50tmztwhat lcd do they have?
23:44.58cr2gp_clk   +0x5c
23:45.16tmztcan you put microp led flashes or something in there
23:45.22cr219.2MHz for gp_clk ?
23:45.31dcordes_tmzt, the wvga ones. both use board-htcblackstone-panel.c
23:45.32tmztone of dzo's commits had a patch for that
23:46.44cr2i think that we will return to the switch(clock_id)
23:46.49paranjust got config.gz from a hero, it has CONFIG_MACH_SAPPHIRE=y, CONFIG_MACH_HERO=y
23:46.52cr2otherwise it's too difficult
23:47.04*** join/#htc-linux Reactor16 (n=Reactor1@41.105.117.128)
23:47.10Reactor16hi all
23:47.26cr2or we must inherit the clock-msm7x30.c structure and functions.
23:48.48Reactor16is gsm work on linux ?
23:49.21tmztReactor16: on?
23:50.21Reactor16htc elfin ?
23:50.31tmztI think so
23:50.36tmztno sound yet
23:50.42tmzt#linwizard
23:51.03Reactor16what project here ?
23:51.25tmztnon-omap htc's
23:51.28tmztmostly
23:51.47tmztexcept BabelO
23:57.16dcordes_cr2, don't know what to try regarding lcd
23:57.54tmztdisable reset/init?
23:58.30tmztmost of the commits haven't been very bisectsble unfortunately

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