IRC log for #htc-linux on 20080428

00:05.09*** join/#htc-linux aLeSD (n=alex@227.Red-83-44-153.dynamicIP.rima-tde.net)
00:30.55BabelOgood night
00:56.52*** join/#htc-linux aLeSD (n=alex@227.Red-83-44-153.dynamicIP.rima-tde.net)
01:24.24*** part/#htc-linux nx5_off (n=pocholo@84.78.40.245)
01:43.34*** join/#htc-linux zwox (i=zwox@par69-7-82-243-234-94.fbx.proxad.net)
04:18.15*** join/#htc-linux dzo (n=dzo@121.98.128.127)
04:28.37*** join/#htc-linux goxboxlive (n=goxboxli@195.159.97.196)
04:41.24*** join/#htc-linux jeebster (n=kanakana@a88-112-94-153.elisa-laajakaista.fi)
05:07.46*** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz)
05:32.49*** join/#htc-linux cr2_ (n=konversa@crpl6.physik.uni-wuppertal.de)
05:49.10*** join/#htc-linux kiozen (n=oeichler@p5492A059.dip0.t-ipconnect.de)
05:57.01*** join/#htc-linux kiozen (n=oeichler@p5492A059.dip0.t-ipconnect.de)
06:06.56*** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net)
06:13.05BabelOhi
06:25.58*** join/#htc-linux rmoravcik (n=rmoravci@gtsgw.ttc.cz)
06:27.17*** join/#htc-linux tsdogs_ (n=tsdogs@62.123.180.130)
06:47.02*** join/#htc-linux rob_w (n=bob@M9551.m.pppool.de)
07:08.12*** join/#htc-linux LunohoD_ (n=alex@e180070084.adsl.alicedsl.de)
07:09.23*** join/#htc-linux BabelO_ (n=fcr@unaffiliated/babelo)
07:23.17BabelO_hi kiozen, do you have a banner or a splash for installer ? ;)
07:23.44kiozenmorning BabelO_
07:23.48kiozen:)))) sure
07:24.37kiozenBabelO_: src/pics/splash.png
07:25.03kiozenif that is too big you have to tell me the wanted size
07:25.09BabelO_ok, i connect to home :)
07:26.03BabelO_x kiozen: there is two option, a banner splash like on your website at the top , i need a size of 150x57
07:26.47kiozenBabelO_: ok, that size, wait ..
07:33.54kiozenBabelO_: will take some time, have to find a nice pic source without strict copyrights :)
07:34.19dcordeshi
07:34.27BabelO_kiozen: no problem, i copy the licence ;)
07:34.37kiozenlol
07:39.09BabelO_kiozen: what do you want for licence ? GPLV3 ?
07:39.42kiozencurrently it's GPL2, don't know what to think of GPL3
07:41.04BabelO_kiozen: ok
07:54.20*** join/#htc-linux patoche (n=patoche@77.197.65.152)
08:01.23*** join/#htc-linux areq_ (n=areq@q.areq.eu.org)
08:01.54dcordesBabelO_: how is artemis progress?
08:08.48*** join/#htc-linux JEEB (n=kanakana@a88-112-94-153.elisa-laajakaista.fi)
08:09.46dcordesBabelO_: what does htc-egpio make?
08:15.35dcordeswonders what .flags = IORESOURCE_MEM, in loox.c is
08:16.03dcordesalso whats N560_EGPIO_BASE?
08:19.52dcordescr2_: you around?
08:21.16BabelO_dcordes: good, i make the cpld working yesturday night ;)
08:21.44BabelO_dcordes: dcordesonly thing you need to change is the baseaddr
08:22.08BabelO_then add it to your platform init
08:22.24dcordesBabelO_: I show you what I have done so far in kaiser.c
08:22.36dcordescr2_: Serial port RS-232 , 115200bit/s <- from some kaiser spec site
08:22.42BabelO_dcordes: htc-egpio on artemis is some externded gpio power, for wifi, bt , gps and lcd
08:23.13dcordesI see egpio=cpld?
08:23.18dcordesextended gpio?
08:24.26dcordesBabelO_: http://rafb.net/p/i3hMSj38.html can you look?
08:27.46BabelO_dcordes: replace the n560_egpio by kaiser :)
08:27.54dcordesjust the name?
08:28.02BabelO_dcordes: and edit in include/asm/arch/kaiser*.h
08:28.48BabelO_dcordes: and in your static struct platform_device *devices[] __initdata = { n560_egpio too
08:29.50BabelO_dcordes: then when you start kaiser, you should see a lne in dmesg about EGPIO something.. this tell you egpio driver is loaded
08:30.20dcordesI haven't anything done with *.h yet
08:30.49dcordesonly aded the htc-egpio.c with Kconfig and Makefile change and edited board now
08:33.03kiozenBabelO_: http://www.qlandkarte.org/logo_300.png and http://www.qlandkarte.org/logo_150.png
08:33.45BabelO_dcordes: just test like this to see if driver is properly loaded
08:33.50BabelO_dcordes: it is first step
08:34.11dcordesBabelO_: also if addresses are shity it should load?
08:38.12BabelO_kiozen: ok
08:39.01BabelO_kiozen: you are the king for logo :)
08:39.36BabelO_dcordes: driver should load and set all egpio to zero, so if you have LCD power gpio on it then LCD goes black ;)
08:40.33BabelO_dcordes: why you have .gpio_base   = N560_EGPIO_BASE,  0x98 ? what is the 0x98 ?
08:41.04BabelO_dcordes: and you need to edit your board.h file to add #define KAISER_EGPIO_BASE GPIO_BASE_INCREMENT
08:41.13BabelO_dcordes: don't know how many gpio kaiser have
08:41.21kiozenBabelO_: gimp -> Xtns -> Logos -> Simple1
08:42.23kiozenBabelO_: it would be nice if you could use the bigger one, 150x57 it pretty small
08:42.52BabelO_kiozen: ok
08:45.09dcordesBabelO_: on the cpld there are 8 gpio banks
08:45.21BabelO_dcordes: ok
08:45.35dcordeshttp://wiki.xda-developers.com/index.php?pagename=Kaiser_CPLD
08:46.49BabelO_dcordes: if you are sure about the baseaddr, you can try like this
08:47.09BabelO_dcordes: and then look for in dmesg | grep egpio -i
08:48.05BabelO_dcordes: does it compile like this ?
08:48.37dcordesdidn't try I'm sure what I did so far is wrong
08:49.39dcordesI have 8 banks with 8 gpios each
08:49.53dcordesinfernix: you there?
08:50.03dcordesI need the config for kaiser
08:51.49dcordesRaven_coda removed it
08:51.56dcordesRaven_coda: ping
08:57.18BabelO_dcordes: i think your egpio setup is ok, once you check your dmesg, there is some other change to do
09:00.24dcordesBabelO_: KAISER_EGPIO_BASE, 0x.... is undefined and what is , 0x..?
09:01.09BabelO_dcordes: edit asm/arch/board.h and add #define KAISER_EGPIO_BASE GPIO_BASE_INCREMENT
09:02.41BabelO_dcordes: and set only .gpio_base   = KAISER_EGPIO_BASE,
09:03.23dcordeswait I'm confused
09:05.13BabelO_dcordes: http://pastebin.ca/1000160
09:07.23dcordesok I had it similar
09:10.58dcordesok added define line in board.h
09:11.15dcordesI will try to make a .config now
09:23.12BabelO_dcordes: you use make menuconfig ?
09:26.55dcordesBabelO_: no I'm lazy and use xconfig
09:27.04dcordesdid a defconfig for msm and then xconfig
09:27.51BabelO_hum ok, long time i did not use xconfig, since i see it miss some config option i have in console mode
09:27.58dcordesarch/arm/mach-msm/board-htckaiser.c:239: error: variable ‘egpio_data’ has initializer but incomplete type
09:28.01dcordesarch/arm/mach-msm/board-htckaiser.c:240: error: unknown field ‘gpio_base’ specified in initializer
09:28.05dcordesarch/arm/mach-msm/board-htckaiser.c:240: error: ‘GPIO_BASE_INCREMENT’ undeclared here (not in a function)
09:28.08dcordesoh no this should have gone to pastebin
09:28.10dcordesarch/arm/mach-msm/board-htckaiser.c:240: warning: excess elements in struct initializer
09:28.14dcordesarch/arm/mach-msm/board-htckaiser.c:240: warning: (near initialization for ‘egpio_data’)
09:28.18dcordesarch/arm/mach-msm/board-htckaiser.c:241: error: unknown field ‘nrRegs’ specified in initializer
09:34.08BabelO_dcordes: did you #include <linux/mfd/htc-egpio.h> ?
09:36.49dcordesBabelO_: no
09:37.02dcordeshold on
09:37.02BabelO_;)
09:37.26dcordesyou see my 10 years of professional C and kernel experience there ;p
09:37.35BabelO_lol
09:38.04BabelO_dcordes: you see my delphi experience with include file :)
09:41.35dcordesBabelO_: there is no mfd in kernel/include (msm-git)
09:42.56BabelO_dcordes: hen create one
09:43.12dcordesok
09:43.14BabelO_you already add drivers/mfd ?
09:43.20dcordesit was already there
09:43.28BabelO_dcordes: what was in it ?
09:43.31dcordesI only put egpio.c in and edite Makefile and Kconfig
09:43.35dcordeswait
09:44.29dcordeshttp://git.android.com/?p=kernel.git;a=tree;f=drivers/mfd;h=4ad5df0ad1f502caefea0ca6b74f786c6126a74e;hb=HEAD
09:45.47BabelO_dcordes: ok they put .h in same directory :(
09:45.58dcordesok I copy h there then?
09:46.41BabelO_dcordes: no because you need to include it from your board file
09:46.53BabelO_dcordes: create the include/linux/mfd
09:49.51dcordeskernel/include/linux/mfd/htc-egpio.h
09:49.57dcordescopied it there from hh now
09:54.58dcordesBabelO_: any change to that file required?
09:55.17BabelO_dcordes: no
09:55.36BabelO_dcordes: try to compile now, you should have an error message
09:55.36dcordesI see Kevin2 made that general
09:56.35dcordeshttp://rafb.net/p/u1U3Sy54.html
09:57.35BabelO_dcordes: you add the include in your board file ?
09:59.02dcordesBabelO_: which include?
09:59.29BabelO_dcordes: #include <linux/mfd/htc-egpio.h>
09:59.46dcordesah no wait
10:01.05dcordesok 1 error
10:01.07dcordesarch/arm/mach-msm/board-htckaiser.c:241: error: ‘GPIO_BASE_INCREMENT’ undeclared here (not in a function)
10:02.15BabelO_dcordes: this one is normal
10:02.36BabelO_dcordes: now you need to get include/linux/gpiodev2.h and add it to your msm tree
10:02.48dcordesok
10:02.53dcordesedit?
10:03.20BabelO_dcordes: no, but you need to do something else
10:03.46dcordeswhat?
10:03.54BabelO_dcordes: i m looking for the msm gpio funciton,do you know the name ?
10:04.50dcordesin egpio-htc.c?
10:05.28BabelO_dcordes: ok i found edit file arch/arm/mach-msm/generic_gpio.c
10:07.10BabelO_dcordes: then add this two line
10:07.13BabelO_#include <linux/gpiodev2.h>
10:07.14BabelO_struct gpio_ops gpio_desc[16];
10:08.32dcordesok great
10:09.30BabelO_dcordes: does it compile now ?
10:10.31dcordesis struct gpio_ops gpio_desc[16]; one line only to add?
10:14.48dcordesBabelO_: ?
10:15.17BabelO_dcordes: with the include two lines
10:15.23dcordesok
10:15.30dcordesis the position variable?
10:15.32BabelO_dcordes: it is first step to see if your kaiser still boot
10:16.03BabelO_dcordes: it is just to compile, i show you once you boot the other change to add
10:16.52dcordesok but where in generic_gpio.c must I add struct gpio_p?
10:19.20BabelO_dcordes: add it at the top, don't care
10:20.40dcordesarch/arm/mach-msm/board-htckaiser.c:241: error: ‘GPIO_BASE_INCREMENT’ undeclared here (not in a function)
10:20.45dcordes-_-
10:27.26*** join/#htc-linux Marex (n=marex@vasut.kolej.mff.cuni.cz)
10:29.32dcordesBabelO_: same error as before
10:29.34dcordeshi Marex
10:30.22Marexhi dcordes
10:39.40Marexcr2_, Im going to tear apart the N560 loader
10:48.18dcordesBabelO_: any idea what's wrong?
10:51.12BabelO_dcordes: #include <linux/gpiodev2.h> in your board file, i ll look how to fix it better later
10:53.06BabelO_dcordes: did you add the include to generic_gpio.c too ? or you forgot ?
11:05.16*** join/#htc-linux p3t3r__ (n=p3t3r__@wlan.net-you.de)
11:07.38BabelO_~ping dcordes
11:07.40aptpong dcordes
11:21.50dcordesBabelO_: back
11:22.14BabelO_:)
11:22.42dzohi dcordes, how is sd going on kaiser?
11:22.54dcordesdzo: we are working on accessing the power gpio
11:23.35dcordeshi dzo
11:24.21dzohi, is that the vreg?
11:24.35dcordespower gpio is on the cpld
11:24.40dcordesdon't know what vreg is
11:24.53dzonor do I
11:25.11dcordesdzo: do you know whether or not vogue has a cpld?
11:25.44dzoyes it does (cpld3), some buttons and leds are on the cpld.
11:25.59dcordesnot more?
11:26.30dzoI traced writes to it and found lots of stuff but i don't know about SD.
11:26.50dcordesI see. maybe it is much similar to kaiser cpld: http://wiki.xda-developers.com/index.php?pagename=Kaiser_CPLD
11:27.32dzoI haven't had time to try sd on vogue yet. the cpld probably is similar.
11:27.57dcordesI could bet you also have the SD power gpio on the cpld
11:28.10dcordesI will ping you when htc-egpio driver works on kaiser ;)
11:28.28dzois kaiser also CPLD3? how did you find out that sd power is on the CPLD?
11:29.05dcordesI think it is cr2_'s disassembly and partly my haret trace work
11:29.19dzoif you boot while accessing sd perhaps the power will be on.
11:29.23dcordesyou can take a lot from the SPL disassembly
11:30.06dcordesdzo: that would be a try! any idea how I could cause sd writes while booting?
11:30.08dzook, if i have time i'll have a look a the vogue spl.
11:30.12dcordesor reads
11:30.25dcordesdzo: bug cr2_ he's really advanced on all this
11:30.53dzotry booting while watching a movie stored on SD.
11:31.01dcordesok
11:31.47dcordesdzo: you can go online with vogue I read in your changelog?
11:31.57BabelO_dcordes: did you compile now ?
11:32.01dzook, it's late here, good night.
11:32.10dcordesdzo: good night
11:32.17dcordeswas nice to hear from you
11:32.21dcordesBabelO_: no will do now
11:35.19dcordesBabelO_: I'm getting non egpio related problems now.
11:35.53BabelO_dcordes: what ?
11:36.20dcordesI will fix it
11:36.21dcordesirq stuff
11:36.52dcordesthink it's my config. have two irq c files building
11:38.15dcordescan somebody hand me their .config for kaiser against Raven_coda patched tree?
11:39.32dcordesI'm getting duplicates
11:40.43dcordesBabelO_: lol I should have tried to boot with this tree before I added egpio
11:40.48dcordesbut it seems to build now
11:42.29dcordesthis will never boot ^^
11:43.13BabelO_dcordes: to desactivate egpio you just need to remove one line, not a big change ;)
11:44.43dcordesok
11:46.25dcordesBabelO_: you know what aDSP is?
11:48.39dcordesBabelO_: when I compiled before I had the display turn off after a time. is that a kernel thing or a userspace thing?
11:48.54dcordesbut backlight stayed on. just turned black
11:49.29BabelO_dcordes: it is console blank, your console go black
11:49.53dcordesok so not kernel?
11:50.38BabelO_dcordes: no it is normal, it is the blank console, you have it on computer too
11:51.15BabelO_dcordes: normally it go back when you press a key
11:51.30dcordesyes I know. it was like that in the images before
11:51.39dcordesI just wondered if that's a kernel feature
11:51.45BabelO_dcordes: about aDSP maybe audio dsp
11:51.45dcordesI will disable PM
11:51.55dcordesok disable it
11:52.46dcordesI need a bootup logo
11:53.00dcordes224 color tux
11:54.10dcordesI wonder why android needs kernel level things
11:54.12BabelO_dcordes: does it boot with the egpio stuff ?
11:54.28dcordesmaybe so android can communicate with arm9..
11:54.34dcordesBabelO_: still configuring.
11:54.44dcordesI must make sure it will boot at all
11:56.49dcordesBabelO_: when I boot from initial ramdisk what filesystem is that?
11:56.55dcordesIs it a file system at all?
11:58.50BabelO_dcordes: yes it is, mount your ramdisk under linux and check whzt filesystem it is with mount command
12:04.11dcordestmpfs
12:04.30infernixjust use marbalons .config
12:04.30infernix?
12:04.53dcordesinfernix: does that work with Raven_coda's seperate kaiser files?
12:05.01infernixshure
12:05.08infernixor ican send you my .config
12:05.21infernixive got some usb stuff in there too
12:05.27infernixto no avail, though
12:05.30dcordesyes please
12:06.02dcordesit's not so easy as I thought turning the msm defconfig to a working config
12:06.04infernixhttp://dx.infernix.net/kaiser/
12:06.21infernixits easy if you build kernels since 1996 :)
12:06.37dcordesI never even build one kernel for an x86 machine
12:06.44dcordesbut plenty for arm and powerpc
12:08.14infernix<dcordes> dzo: that would be a try! any idea how I could cause sd writes while booting?
12:08.31dcordesinfernix: you could do that while I try and build egpio
12:08.31infernixuse resco and copy a large file from the network to the sd *whilst* booting into haret
12:08.50dcordesno time now
12:08.55infernixi have to do some work so no time either
12:09.06infernixmaybe later
12:09.49dcordesok
12:20.42dcordesinfernix: initrd-angstrom-x11-test.zip is that your stripped down x11?
12:21.17dcordesinfernix: wow the zImage from your config is a full 1,4 M
12:23.51infernixyep:)
12:24.07infernixand uhm no, angstrom x11 test is older
12:24.35infernixuploading initrd-xdadev-x11, 17mb
12:25.03dcordesdid you try to playback a video with mplayer? :D
12:27.03infernixnope
12:27.12infernixwatched a movie with my girlfriend instead
12:27.13infernix:)
12:27.34dcordeshow lame :P
12:27.40infernixnu-uh
12:29.57dcordesoops my fb just faded white
12:30.02dcordesafter booting
12:32.05dcordeshad to take out the battery
12:35.33dcordesinfernix: something in that tree/conf which my kaiser doesn't like at all
12:35.42infernixnot likely
12:35.46dcordesall I get is a white screen with fading lines
12:35.46infernixmore likely your gpio hack:p
12:35.59dcordesyou're right
12:36.08dcordessomebody said it could affect such things
12:36.11dcordesBabelO_: ping
12:36.18BabelO_dcordes: yes
12:36.27dcordesBabelO_: my screen turns white with egpio enabled
12:36.30dcordesafter boot
12:36.46dcordesBabelO_: actually it is white with a few funny lines
12:36.51BabelO_dcordes: ok, can you comment it just the &kaiser_egpio, line
12:37.11dcordesin the common gpio file?
12:37.20BabelO_dcordes: it is strange cr2_did not identify lcd power on cpld
12:37.44dcordesBabelO_: could it be that it also affects non-cpld gpios?
12:38.19BabelO_dcordes: no don't think
12:38.42dcordesBabelO_: oh. lcd power IS on cpld
12:38.47dcordes<PROTECTED>
12:38.49BabelO_dcordes: yes i just see it :)
12:38.58dcordesok.. now?
12:39.03BabelO_dcordes: try to comment the line to be sure it is that
12:39.15dcordescan't I disable in menuconfig?
12:39.28BabelO_dcordes: no
12:39.34dcordesok
12:39.35BabelO_dcordes: just comment the like i told you
12:39.38dcordesin which file?
12:39.45BabelO_your board file
12:39.51BabelO_&kaiser_egpio,
12:40.14dcordesdisable the init right?
12:40.23dcordesI put it right at the beginning.
12:41.01dcordesbuilding..
12:41.09dcordesoh that was quick
12:41.35dcordesBabelO_: is it normal that it's that fast after small change?
12:41.52BabelO_dcordes: yes it just redo the board.o file
12:42.00dcordesok
12:42.26dcordesBabelO_: strange is: after white screen, I can't reset. It looks like kernel was able to break the _full_ system :)
12:45.06dcordesarm9 has direct access to the cpld iirc
12:45.34dcordesBabelO_: booted!
12:45.40dcordesworks. I see the console.
12:46.01BabelO_dcordes: ok good, so it seem it is the lcd power gpio
12:46.31BabelO_dcordes: but strange it hang your device i think there is other gpios too
12:46.41dcordesBabelO_: yees
12:46.47BabelO_so you need to edit generic_gpio.c now
12:47.08BabelO_dcordes: how much gpio msm have ?
12:47.14dcordests, led, sd power, bt, wifi pwr, lcd pwr etc
12:47.21dcordesalso kbd
12:47.48dcordesBabelO_: strange is that dzo said he has buttons and keyboard on cpld
12:47.55dcordesand he didn't have any driver
12:48.00dcordesmaybe that's the smd thing
12:48.25BabelO_dcordes: the msm gpios, should be < 0x100, no i just check it it is ok
12:48.38BabelO_you have 105 gpios
12:49.43BabelO_dcordes: you have to edit generic_gpio.c function void gpio_set_value(unsigned gpio, int on) line 232
12:50.07dcordesBabelO_: I don't understand why egpio writes to the cpld although nobody told it to write
12:50.20BabelO_dcordes: ?
12:50.39dcordesI mean why does it change the lcd gpio?
12:51.07BabelO_dcordes: because the htc-egpio driver by default init all the gpios of cpld to 0
12:51.14dcordesahh I see
12:51.34BabelO_dcordes: so you have to declare a struct with default gpios or set them by hand
12:51.44BabelO_dcordes: do you open generic_gpio.c ?
12:51.52dcordesyep it's in front of me
12:52.10BabelO_line 232 or close you should see gpio_set_value function
12:52.17dcordesyep
12:52.57dcordesint gpio_get_value(unsigned gpio)
12:52.57dcordesand
12:53.00dcordesvoid gpio_set_value(unsigned gpio, int on)
12:54.38BabelO_dcordes: look at this example http://pastebin.ca/1000306
12:56.11dcordeswhat's difference btween int on and no int on?
12:59.28BabelO_dcordes: where do you see that ?
12:59.43BabelO_dcordes: ah sorry, on / off gpio
13:00.12BabelO_dcordes: did you do it ?
13:00.18dcordesyour paste?
13:00.23BabelO_yes
13:00.31dcordesyes it's in my code now
13:00.34dcordesthe else function thingy
13:00.41dcordeswhat does it do?
13:00.45BabelO_good with the if too
13:01.06dcordesyes
13:01.34BabelO_dcordes: now edit your board.h file where you add the #define KAISER_EGPIO GPIO_BASE_INCREMENT
13:01.39dcordeswhat kinf of value is GPIO_BASE_INCREMENT?
13:01.50dcordesis it the bootup gpio state?
13:02.06dcordess/kinf/kind/
13:02.17BabelO_dcordes: GPIO_BASE_INCREMENT is just a varable set to 0x100 which mean all egpio are set > to this
13:02.48BabelO_dcordes: < 0x100 it call the standard msm gpio funciton and > 0x100 it call the egpio stuff
13:03.02BabelO_dcordes: open the board.h
13:03.05dcordesok..
13:03.45BabelO_dcordes: add #define KAISER_EGPIO(reg,bit) (KAISER_EGPIO_BASE + 16*(reg) + (bit)) just under the line you add previously
13:04.27BabelO_then add your cpld definition like this
13:04.46dcordesok
13:04.46dcordes#define KAISER_EGPIO_BASE GPIO_BASE_INCREMENT
13:04.47dcordes#define KAISER_EGPIO(reg,bit) (KAISER_EGPIO_BASE + 16*(reg) + (bit))
13:04.50dcordesin threre now
13:05.13BabelO_#define EGPIO_2_0_SDPWR                           KAISER_EGPIO(2, 0)
13:05.37dcordesalso in it?
13:05.41BabelO_yes
13:05.50BabelO_this is definition of each EGPIO of your CPLD
13:05.59BabelO_do you understand how to do the others ?
13:06.15dcordesnot yet
13:06.17BabelO_2_0 mean gpio 0 of bank B
13:07.20dcordescool
13:07.21dcordesunderstood
13:07.44dcordesbut what means on, what means off?
13:08.11BabelO_dcordes: i show you after, fill the header file with the wiki cpld definition
13:08.32BabelO_be back in few minutes ( i m at office, have some work to do ;) )
13:08.44dcordesalso should do stuff
13:08.53dcordesnow I know what needs to be done, I will do it later
13:09.04dcordesBabelO_: thank you very much so far. will be back later
13:09.26*** part/#htc-linux p3t3r__ (n=p3t3r__@wlan.net-you.de)
13:19.37BabelO_dcordes: ok after you have to add more line to the egpio struct def in you board file
13:43.05dcordesBabelO_: what does cr2 mean with init=1? does that mean that it is on in the SPL?
13:46.19dcordesBabelO_: can you give me your email?
13:47.44*** join/#htc-linux darkenCZ (n=michal@rb3g230.net.upc.cz)
13:48.14darkenCZHi, I would like to ask what is current state of porting linux to Hermes
14:21.27dcordesmsm board.h with kaiser definitions: http://pastebin.ca/1000383
14:22.59BabelO_dcordes: sorry, now i m back
14:23.32BabelO_dcordes: never do such thing #define EGPIO_2_6_+init,bt
14:23.37BabelO_or the ?
14:24.12BabelO_instead of ? set UKN, but please no special char in #define
14:24.32dcordesBabelO_: I commented all the strange stuff out
14:24.47BabelO_dcordes: rename the init thing we need it
14:25.04dcordesthe not commented things are all clear capital letters
14:25.09dcordesBabelO_: which? there are some
14:25.21BabelO_dcordes: can you repaste your .h ?
14:25.33dcordesok...
14:25.50dcordeshttp://rafb.net/p/yEoGai70.html
14:26.33BabelO_dcordes: set back the ini egpio
14:26.48dcordessorry?
14:27.14BabelO_dcordes: #define EGPIO_8_3_INIT  KAISER_EGPIO(8, 3) and not #define EGPIO_8_3_init=0  KAISER_EGPIO(8, 3)
14:27.32dcordes3_1 and 5_4 are init
14:28.03BabelO_same for all init
14:28.35BabelO_dcordes: be back in few minutes
14:29.05dcordesI name all gpios that have init in them INIT
14:41.01dcordesBabelO_: http://rafb.net/p/4EDcU085.html
14:41.03dcordesbbl
14:53.42BabelO_dcordes: good
14:54.04*** join/#htc-linux TimRiker (n=timr@68.27.117.245)
15:03.23*** join/#htc-linux patoche (n=patoche@lns-bzn-23-82-248-66-22.adsl.proxad.net)
15:13.14kiozencr2_: loading a map in referencing mode will center the map, please svn up
15:17.36dcordesBabelO_: still here?
15:22.12BabelO_dcordes: yes
15:22.20BabelO_dcordes: next step now ? :)
15:22.41dcordesyes please.
15:28.33dcordes~ping BabelO_
15:28.35aptpong BabelO_
15:29.28Raven_codadcordes, you amaze me, I mean you have been working on this none stop
15:29.41Raven_codahi, btw
15:29.56Raven_codaIt's like you never rest :)
15:34.01*** join/#htc-linux comicinke2 (n=simeon@R380f.r.pppool.de)
15:34.20comicinke2hi there. the openmoko guys told me to go here for new images for htc-magician of openmoko. how do I compile openmoko for htc-magician? is there a makefile?
15:34.33kiozenRaven_coda: haven't you noticed the matches in his eyes and the grin ;)
15:35.10dcordesRaven_coda: thanks
15:35.41dcordesBabelO_: bb in an hour or two. maybe you can give me an example of how to gon
15:39.26infernixRaven_coda: hi
15:39.39Raven_codahello
15:39.44infernixi've been meaning to ask you whether if you incorporated any of the msm_sdcc.c code into your patch
15:40.12BabelO_dcordes: ok i do the change from your pastebin
15:40.32Raven_codathere are small changes to msm_sdcc in there
15:40.46Raven_codawhat code were you thinking of?
15:40.55infernixlemme find it
15:41.32infernixhttp://dx.infernix.net/kaiser/ravencoda-1-extrasdcc.diff
15:44.30Raven_codasome of that is there but not all of it
15:44.40Raven_codadoes that make it work?
15:44.48infernixno
15:45.08Raven_codasome of that is old code
15:45.26Raven_codathe google guys updated msm_sdcc a little while ago
15:45.43Raven_codaand some of there old stuff is in that patch
15:48.21infernixok
15:51.37Raven_codaI'm looking at the changes made to android-msm on the 22nd and it looks like the patch you sent me would just revert all Googles work on that day
15:53.27*** join/#htc-linux AlGe (n=alge@chello080109231226.4.uni-klu.teleweb.at)
16:07.22*** join/#htc-linux patoche (n=patoche@lns-bzn-23-82-248-66-22.adsl.proxad.net)
16:29.26dcordesBabelO_: back
16:32.11BabelO_dcordes: :)
16:32.38BabelO_dcordes: open arch/arm/mach-pxa/htcapache/htcapache.c
16:33.45dcordesBabelO_: ok
16:34.07dcordesopenned
16:34.52BabelO_dcordes: look at htc_egpio_pinInfo
16:35.03BabelO_you see it ?
16:35.46dcordesyep
16:36.07BabelO_dcordes: good, this structure let you define the initialisation of cpld pin on load
16:36.51BabelO_you just have to add .pins = pins, to egpio_data ( the structure below )
16:37.20dcordesit already has that
16:37.25BabelO_dcordes: and you can fill ths pins with your init pin set to 1 lcd pwr set to 1 and sd power set
16:37.35BabelO_dcordes: can you paste bin our file ?
16:37.42BabelO_just to check
16:37.43dcordesthe board.h?
16:37.55BabelO_dcordes: no your main board file .c
16:37.59dcordesok
16:38.43dcordeshttp://pastebin.ca/1000527
16:46.29*** join/#htc-linux Marex (n=marex@vasut.kolej.mff.cuni.cz)
16:47.46BabelO_dcordes: when i say change your boarc c file, i mean like this http://pastebin.ca/1000539
16:48.33BabelO_there is also usbpen to set, have to ask cr2
16:48.50Marexcr2_, hi, are you around ?
16:49.42BabelO_dcordes: i think you do a mistake in then lcd pwr, it is 4_4 not 3_4 according to http://wiki.xda-developers.com/index.php?pagename=Kaiser_CPLD
16:51.17Marexcr2_, well check your email and tell me what do you think about it
16:51.18dcordesBabelO_: ok thanks I have that in
16:52.27dcordesBabelO_: lcd_pwr1 is on bank C
16:52.59BabelO_dcordes: yes and bank C is 0x4 cpioc4
16:53.45dcordesBabelO_: looks like I have to change whole header
16:55.09BabelO_dcordes: yes, douvble check it
16:55.17dcordesBabelO_: is the first line #define EGPIO_0_0_BOARDID0 KAISER_EGPIO(0, 0)?
16:55.33BabelO_yes
16:55.35dcordesok
16:57.49dcordesBabelO_: do you know what usb puen is?
16:59.12BabelO_dcordes: it is to setup your usb connection
16:59.19BabelO_i thinkit is what you miss
17:00.22dcordesI'll have that as USB_PUEN
17:00.41dcordesis ist just power or something else?
17:00.47dcordesclock?
17:01.41BabelO_no it is usb power enable
17:03.01dcordesshouldn't we call it USB_PWR then?
17:03.37BabelO_dcordes: no puen is the correct name, it is not the main chip power
17:03.43dcordesalright
17:03.55BabelO_cr2_ tell you better what it is
17:04.42BabelO_once you set the correct header and pins default in c, don't forget to uncomment &kaiser_egpio,
17:08.06dcordesBabelO_: where is ARRAY_SIZE defined?
17:08.23BabelO_dcordes: don't know, why ?
17:08.35dcordesjust wonder what it is
17:08.45dcordesI corrected the value and uncommented the init
17:08.51dcordesnow compile?
17:09.08BabelO_yes, you check what i add in the structure
17:10.46dcordesyes they are correct now with the bank -1 shift
17:10.57dcordesBabelO_: but one question what does INIT=1 do?
17:11.31BabelO_dcordes: it mean the cpld is initialized to 1 on boot, don't know what it do
17:11.50dcordeswe just know it's on in SPL right?
17:11.59BabelO_dcordes: it is not only a -1 shift you miss some registers in cpld
17:12.04BabelO_yes
17:12.40*** join/#htc-linux JEEB (n=kanakana@a88-112-94-153.elisa-laajakaista.fi)
17:13.16dcordesBabelO_: hm I doubble checked and it looks ok now: http://rafb.net/p/KbAcPf67.html
17:13.24BabelO_oula, your cpld is different from artemis, it seem there is more setup to do, maybe Kevin2 now
17:14.07dcordesBabelO_: ie it won't work when I compile now?
17:16.55dcordesat least we have access
17:17.06*** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz)
17:17.16*** join/#htc-linux pH5 (n=ph5@e178246020.adsl.alicedsl.de)
17:20.51dcordeshi pH5
17:21.48pH5hi
17:22.26BabelO_hi pH5
17:22.41BabelO_pH5: how can we configure the kaiser cpld ?
17:22.55*** join/#htc-linux tsdogs (n=tsdogs@net-84-18-144-39.c.bkom.it)
17:22.59dcordesBabelO_: it worked: I have booted with a working display.
17:23.14dcordesdidn't fade to white this time
17:23.14MarexpH5, hi ;)
17:23.15BabelO_pH5: each bank have a shift of +4
17:23.39BabelO_dcordes: and you reactivate the éegpio_init ? can you do a dmesg ?
17:23.48dcordesyes can do
17:23.54dcordesbut I have mmc0 pritnk shit.
17:24.09dcordesit's spamming dmesg
17:24.39BabelO_hmm
17:24.44dcordesBabelO_: I still can't reset (while linux is running also)
17:24.56BabelO_dcordes: dmesg -s 128
17:25.14dcordesI recompile w/o the sd printk. moment
17:25.47pH5BabelO_, dcordes: for kaiser you should use the mainline version of htc-egpio
17:26.06pH5platform data has .reg_width and .bus_width parameters
17:29.25*** part/#htc-linux comicinke2 (n=simeon@R380f.r.pppool.de)
17:31.04*** join/#htc-linux goxboxlive (n=goxboxli@208.84-48-176.nextgentel.com)
17:32.04BabelO_pH5: strange, it seem it works for dcordes like this i dont understand how
17:32.13BabelO_pH5: can you give a link ;)
17:32.25BabelO_pH5: does it still need the gpiodev2 ?
17:32.39dcordesRaven_coda: I don't know whether or not that is used with tha kaiser specific setup but the msm_sdcc.c in your tree adds sdc3 and 4 which we do not have
17:32.50dcordespH5: I can tell. egpio is in charge of display power.
17:34.17pH5BabelO_: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=a1635b8fe59de2c5223cda5ca8397b875c901904
17:34.26BabelO_pH5: thanks
17:35.14Raven_codadcorded, what lines are you looking at?\
17:35.17dcordesBabelO_: can you look if that makes a drastic change? would be better if we continue with new version in that case
17:35.26dcordesRaven_coda: just grep for sdc3
17:35.55BabelO_dcordes: better is you upgrade to the pH5 link only two file is minor, then you need to add two parameters ;)
17:36.08Raven_codathere is sdc3 stuff in there but it's in IFdef staments
17:36.15Raven_codait won't bother you
17:36.54dcordesnot all the spots. don't know if that matters. those at the beginning have no ifdef
17:37.03Raven_codathat's fine
17:37.33dcordesgd gd
17:37.56*** join/#htc-linux JohnnyK (n=johnnyk@nfx-nat-229.pilsfree.net)
17:38.53dcordesBabelO_: ok will compare the mainline v. with my files
17:39.12BabelO_dcordes: don't compare, just replace ;) it is generic now
17:39.14dcordesbut now I boot and try to find egpio in the logs
17:39.20dcordescool
17:39.30dcordeswhich files? the driver.c and .h?
17:39.46BabelO_dcordes: yes
17:41.26BabelO_pH5: sorry, but what are difference between my artemis cpld and the kaiser cpld ? artemis regiser are 0x0 0x1 0x2 ... and kaiser are 0x0 0x2 0x4 so 16 bits for kaiser, and me how many ?
17:41.26dcordesok I'm in the kernel
17:41.43BabelO_dcordes: dmesg | grep -i egpio
17:41.51dcordescan't make no pipe
17:41.58BabelO_lol
17:42.01dcordes:)
17:42.07dcordesshould remap matrix keyboard
17:42.35BabelO_ok so it is just before LCD init you should see EGPIO PHYS=.. VIRT=....
17:42.40dcordesvi /var/log/messages
17:43.00pH5dcordes: also the platform data / pin configuration changed a bit. look at magician for an example of how it is configured:
17:43.00pH5http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;f=arch/arm/mach-pxa/magician.c;h=3b7e8bac7344706672740b34cc2c58b703d58c33;hp=0160cd8f6c653eac614e966b1fff19856083edc9;hb=70e357f8426c54e1500ac4fdb6b1172df16d8b93;hpb=a1635b8fe59de2c5223cda5ca8397b875c901904
17:43.15pH5BabelO_: is the Kaiser CPLD info in the wiki somewhere?
17:43.31BabelO_pH5: http://wiki.xda-developers.com/index.php?pagename=Kaiser_CPLD
17:43.38pH5thanks
17:43.46dcordesBabelO_: my /var/log is empty
17:43.52BabelO_pH5: and artemis http://wiki.xda-developers.com/index.php?pagename=ArtemisCPLD
17:43.59BabelO_dcordes: yes sure
17:44.02pH5so it is reg_width=8, bus_width=16
17:44.06BabelO_vi /etc/syslog.conf :)
17:44.23BabelO_pH5: ok and mine ?
17:44.27dcordesI have no <esc>
17:44.41pH5hm. do those registers contain input and output pins mixed?
17:44.50BabelO_pH5: no only output
17:45.00BabelO_pH5: i test it yesturday and it works :)
17:45.00pH5BabelO_: yours is 8bit both, I think
17:45.11pH5does wince set it with strb?
17:45.21dcordesIs it possible to select cached logging in the kernel? So that I get a /var/log/messages ootb?
17:45.35BabelO_pH5: ok, no with strh i think have to check once i m at home
17:46.01BabelO_dcordes: /etc/syslog.conf to get messages stuff
17:46.18pH5BabelO_: if it is strh then you should use 16/16bit (which i think is more probable than 8bit packed) and combine A+B, C+D, etc. into one register each
17:46.48BabelO_pH5: ok i check in one hour, when back to home
17:47.27BabelO_bbl
17:48.16dcordespH5: can you setup logging from console image OE buildtime?
17:48.17BabelO_pH5: i was able to set VIBRA and LCD power which are on F bank in the clean way
17:48.18BabelO_bbl
17:48.21*** part/#htc-linux BabelO_ (n=fcr@unaffiliated/babelo)
17:51.36*** join/#htc-linux noelcragg (n=noel@166.129.254.115)
17:51.57*** part/#htc-linux noelcragg (n=noel@166.129.254.115)
17:54.30pH5dcordes: I don't know, maybe by adding syslog packages to the image
18:11.04BabelOback
18:12.27*** join/#htc-linux patoche (n=root@lns-bzn-23-82-248-66-22.adsl.proxad.net)
18:15.48*** join/#htc-linux pikapika (n=pikapika@mar75-8-88-164-227-147.fbx.proxad.net)
18:15.52*** join/#htc-linux marbalon (n=balon@aefy18.neoplus.adsl.tpnet.pl)
18:16.04pikapikahi
18:16.55*** join/#htc-linux Nickbee (n=dheeraj@122.160.64.226)
18:17.34dcordeshi marbalon
18:17.43marbalondcordes: hi
18:17.49dcordesany news?
18:19.11BabelOdcordes: did you replace with the .c and .h from pH5 ?
18:19.28dcordesBabelO: no I was on the phone
18:19.37marbalondcordes: unfortunately no :(
18:19.46BabelOdcordes: ok so no dmesg too
18:19.50dcordesmarbalon: could we replace windows key with replace key
18:20.02dcordesBabelO: doing dmesg now.
18:20.05BabelOdcordes:  i think once you turn sdpwr on it shoudl work
18:20.34dcordeswell I didn't have sd clock yet.
18:20.48marbalondcordes: windows key with ... ?
18:21.11dcordesmarbalon: lol sorry. <ctrl>
18:21.12BabelOdcordes: but usbpuen can help you too
18:21.54marbalonI have a idea hov to use Fn key as ctrl
18:22.14dcordesmarbalon: but we need Fn keys :(
18:22.23BabelOpH5: ok it is strh
18:22.39marbalondcordes: yes but Fn as hosd meen ctrl
18:22.45marbalon*hold
18:23.27BabelOpH5:  but still cannot understand how can it works for me and dcordes with same driver but different bank shift
18:23.39marbalondcordes: if you press and hold Fn key + C = ctrl+C
18:23.44dcordesmarbalon: that's a good idea
18:24.14marbalondcordes: but if you press and relese witout any  key  - Fn
18:24.15dcordesand one press Fn and after C = :
18:24.41pH5BabelO: ok, so 16bit registers then. the driver only supports 8 or 16bit registers, 16bit or 32bit aligned
18:25.36BabelOpH5: anyway these last day i make good progress on artemis ;)
18:26.10pH5that's good to hear. does i2c work?
18:26.39BabelOpH5: just some trouble to understand howtsc2046 and spi bus driver works together, i ve made my own spi bus driver for omap850 whithout irq handling
18:27.04BabelOpH5: i did not check the i2c will do it now ;) just need to activate my FM receiver
18:36.23BabelOpH5: last time i check i2c i was able to use i2cdump tool two time
18:37.26*** join/#htc-linux noelcragg (n=noel@66.54.159.26)
18:51.16*** join/#htc-linux infernixx (i=gerben@unaffiliated/infernix)
18:53.12infernixxdcordes: ping
18:54.23BabelOpH5: FYI
18:54.26BabelOcat /proc/iomem
18:54.38BabelO04000400-04000410 : htc-egpio
18:54.42BabelO10000000-13bfffff : System RAM
18:54.47BabelO<PROTECTED>
18:54.50BabelO<PROTECTED>
18:54.55BabelOfffb0000-fffb0015 : serial
18:54.59BabelOfffb0800-fffb0815 : serial
18:55.03BabelOfffb3800-fffb383f : i2c_omap.1
18:55.07BabelO<PROTECTED>
18:55.11BabelOfffb4000-fffb40ff : omap_udc
18:55.15BabelO<PROTECTED>
18:55.19BabelOfffb4800-fffb485f : omap_rtc
18:55.23BabelO<PROTECTED>
18:55.27BabelOfffb7800-fffb7836 : mmci-omap.1
18:55.31BabelO<PROTECTED>
18:55.43BabelOfffc0800-fffc08ff : omap1_spi100k.1
18:55.47BabelO<PROTECTED>
18:55.51BabelOfffc1000-fffc10ff : omap1_spi100k.2
18:55.55BabelO<PROTECTED>
18:55.59BabelO~ $
18:55.59apti heard $ is a dollar sign
18:56.03BabelOoups sorry bad paste
18:56.03BabelOpH5: http://pastebin.ca/1000716
18:56.25BabelOlol
18:56.54infernixx~ *
18:56.55aptextra, extra, read all about it, * is asterisk
19:00.39*** join/#htc-linux kiozen (n=kiozen@rgnb-4db1de8b.pool.einsundeins.de)
19:01.07*** join/#htc-linux noelcragg (n=noel@m320e36d0.tmodns.net)
19:01.32dcordesinfernixx: pong
19:01.42*** part/#htc-linux noelcragg (n=noel@m320e36d0.tmodns.net)
19:01.43dcordesdo you get a /var/log/messages on kaiser console-image?
19:01.47dcordeshttp://pastebin.ca/1000733 that's my conf
19:01.48infernixxprogress with gpio yet?
19:01.53dcordesI'm on it
19:02.02infernixxno you have to edit /etc/syslog.conf and change "buffer" to "file"
19:02.08dcordesgreat thanks
19:02.20infernixxi would also set rotatesize to 2048
19:02.21dcordesI think that was in minimal where I gor it to file on default.
19:02.26dcordesok
19:02.28dcordesthx
19:02.48infernixxand use less /var/log/messages in a regular login shell
19:02.58dcordesI guess once we are in the kernel we can abuse the ram to the fullest
19:09.54dcordesinfernixx: I have an interesting finding: I think when we have the boothang, the kernel _is_ loaded but the fb does not work
19:10.17dcordesso I suspect bigger ramdisks to overlap with FB area
19:10.30dcordesmaybe Kevin2 has an idea on that
19:10.37infernixxif it would overlap, screen would corrupt
19:10.57dcordesI'm sure the kernel took over. I was not able to reset.
19:11.15dcordesthat is in the new cpld age.
19:12.08infernixxah
19:12.14dcordesyou get what I mean?
19:12.23dcordesthere must be a takeover
19:12.27infernixxso we should somehow exclude the FB range
19:13.13infernixxmaybe with the badram optin
19:14.04infernixxdo we have any idea where the framebuffer area is in ram?
19:14.09dcordesaw shit my log doesn't work
19:14.23dcordesno can't tell
19:14.27infernixx/etc/init.d/syslog start
19:14.31dcordesmust be in the kernel since haret doesn't define it.
19:15.15dcordesinfernixx: I think it is the mount thing
19:15.33dcordessyslog initscript responds: nothing in /proc - notmounted?
19:17.52dcordesnow that I did a touch /etc/mtab I have no rights
19:19.50dcordesminimalist-image is my friend
19:20.48dcordes3mb minimalist boots _always_
19:21.56dcordesinfernixx: it has to do something with the kernel. I can't read any log in the minimal either
19:22.12infernixxnonsense
19:22.31dcordessuper you have any idea why I had a log before in angstrom then?
19:22.33infernixxbuild the devimage
19:22.37dcordeswith this exact image?
19:22.49infernixxsyslogging has nothing to do with the kernel
19:23.09infernixxits merely a matter of mounting proc and dev and running syslog
19:23.27dcordesI have a kernel with different mountstuffs then.
19:23.37infernixxkernel doesn't do mounting
19:23.39infernixxinit does
19:24.00dcordesyou have any idea what in your .conf or Raven_coda tree could cause this?
19:24.17dcordesor in general
19:24.18infernixxadd task-devimage
19:24.31dcordesmy OE is screwed, can't buil
19:24.42infernixxok 1 sec then
19:24.46dcordesok thanks.
19:25.00dcordesgetting some food in the meantime. can you upload to your server?
19:25.05dcordesbrb
19:25.38infernixxput this in /init: http://rafb.net/p/CJXs2m10.html
19:25.49infernixxboot with rdinit=/init
19:26.32infernixxand make sure /etc/syslog.conf is configured as i mentioned
19:32.17cr2_hi
19:32.33BabelOola cr2
19:32.37cr2_BabelO: you have strange data for ARTE CPLD in the wiki
19:32.58cr2_it's a usual 8bit 16offset CPLD
19:33.04cr2_pH5: hi
19:33.24cr2_pH5: i've documented the WIZA CPLD
19:34.07BabelOcr2_: yes, but how you can explain it works like this ? ;) i test kbdbacklight , bibra, lcd power
19:34.11cr2_pH5: there are 2 SD power gpios, and one is on the CPLD. if the i2c feels unhappy, then there is an omap gpio to kick it
19:34.24pH5hi cr2_
19:34.49pH5BabelO: because you can strb/ldrb into/from 16bit registers just fine :)
19:34.57cr2_BabelO: your CPLD is the same as on n560, that's why it works as is.
19:35.04BabelOcr2_: ok
19:35.12cr2_pH5:  http://wiki.xda-developers.com/index.php?pagename=WizardCPLD
19:35.25BabelOcr2_: did you see that dcordesgot it working with my explanation ? ;)
19:35.54cr2_yes, and it's how it should be :D
19:36.05BabelOcr2_: but i did not finish change, he need to do same for gpio_get_values
19:36.06cr2_i mean the development speed :)
19:36.19BabelOcr2_: lol
19:36.44BabelOcr2_: not too speed, take an afternoon , and i did not found time to do the gt installer
19:37.14cr2_pH5: how do we find out the i2c_id for CPLD ? can you trace the i2c port ? otherwise i'll look deeper into the driver.
19:38.28cr2_Marex: are you still here ?
19:38.39Marexcr2_, yea
19:38.56cr2_Marex: i can send you the ida .asm for n560_ipl
19:39.23*** join/#htc-linux pleemans (n=peter@d51A5E76A.access.telenet.be)
19:39.55Marexcr2_, do I need it? ;)
19:40.01Marexbtw my N560 wakes up to wince too :(
19:40.15cr2_you have compiled the kernel ?
19:40.23Marexyea
19:40.41Marexbut still ... why does it wake up to wince ?
19:40.52Marexcheck my disassembly ,that should be the codeflow to wake up
19:40.56Marexat least in qemu it did this
19:41.00cr2_Marex: i think we should check and document the G4 raw loader
19:41.31cr2_maybe compare it to the docg3.c driver.
19:41.38cr2_ok, i'll have a look
19:41.41*** join/#htc-linux aLeSD (n=alex@227.Red-83-44-153.dynamicIP.rima-tde.net)
19:42.05cr2_Marex: 2f0:e3520b01 cmpr2, #1024; 0x400// compare R2, 0x400
19:42.15Marexhm ?
19:42.18cr2_this is a check of the G4 sig
19:42.28Marexfeel free to add it there :)
19:42.44Marexbtw I also wrote simple resolver for addresses in the disassembly dump
19:43.13Marexhttp://www.pastebin.cz/show/5295
19:43.37Marexit takes two params - the disassembly file and binary file
19:48.13cr2_lol, in bash
19:48.32Marexit'd work in shell too I guess
19:48.35Marex*sh
19:48.43*** join/#htc-linux kiozen (n=kiozen@rgnb-4db1de8b.pool.einsundeins.de)
19:49.15kiozenisn't made for telephone customer service
19:50.37BabelOola kiozen
19:50.48kiozenola BabelO :)
19:51.14kiozentried to explaine my dad how to replace the qlandkarte binary in windows :/
19:51.17BabelOkiozen: i just discover a new virtual box function, desktop integration
19:51.47kiozenu mean the on that ket's you place the windows eveywhere?
19:52.26*** join/#htc-linux infernixxx (i=gerben@91.141.251.23)
19:52.43BabelOyes wait i do a screenshot
19:53.24cr2_hi kiozen
19:53.29cr2_BabelO: lol http://forum.xda-developers.com/showthread.php?t=388809
19:53.42kiozenhi cr2_
19:54.06BabelOcr2_: lol
19:54.12kiozencr2_: it's missing qlandkarte :)
19:55.37cr2_LOL!!
19:56.13cr2_kiozen: BabelO had once compiled qlandkarte for qtopia
19:58.16cr2_kiozen: GT looks more userfriendly now.
19:59.08BabelOkiozen: look what i mean http://linuxbox2.free.fr/vbox-suse-win1.png
19:59.25BabelOkiozen: two task bar on linux :) the top is the windows virtual machine
19:59.51kiozenBabelO: yes i know
20:00.10BabelOkiozen: oh ? i just discover it yesturday ;)
20:00.22kiozenhehe:)
20:00.24BabelOkiozen: this is so nice to work, with the two os at same time
20:03.15dcordesBabelO: sorry for the delay. I have dmesg now in front of me
20:03.31kiozencr2_: lol, user friendly, just had a talk with my dad ^^
20:03.33BabelOdcordes: cool, did you see egpio ?
20:03.37dcordesinfernixxx: thanks I have the odd less and more scroll throug problem
20:03.48infernixxxdcordes: exec /sbin/init
20:03.53infernixxxlog in and try again
20:04.03cr2_kiozen: QPoint(2575,3628) QPoint(608,93) what's that ?
20:04.28kiozen?? debug output?
20:05.14dcordesBabelO: yes. EGPIO phys=c80000000 virt=980000000
20:05.42BabelOdcordes: uh, problem, th reverse you mean ?
20:05.45dcordesinfernixxx: no vi logread works. I wonder why logging worked ootb before *_*
20:05.58BabelOphys=980000000 ?
20:05.59dcordesBabelO: the reverse?
20:06.23infernixxxdcordes: just a messy userspace, thats all
20:06.40dcordesBabelO: no
20:07.49dcordesBabelO: it must work though. I still can't reset
20:08.17BabelOdcordes: strange
20:08.38BabelOi think sdpower is set , do you set usbpuen too ?
20:09.35dcordesBabelO: wait
20:10.11kiozencr2_: svn up, to remove the debug output
20:10.18dcordesINIT=1 SDPWR LCD PWR
20:10.38dcordescr2_: hi. I have htc-egpio running on kaiser thx to BabelO
20:12.16dcordeshow would I try if sd is powered now?
20:12.56dcordes~ping SanMehat
20:12.57aptpong SanMehat
20:14.03dcordesI see a mesg  "mmc0: clock 144000Hz busmode 1 powermode 2 cs 1 Vdd 16 with 0 timing 0"
20:14.06dcordeson kaiser
20:14.14dcordesinfernixxx: can you compare this to what you see?
20:16.11cr2_kiozen: what's the difference between LSR and ASR ? ASR sets the carry bit ?
20:17.49cr2_BabelO: do you have a "real" qtopia on artemis screenshot ?
20:18.20BabelOcr2_: no not now, why ? :)
20:18.50Marexcr2_, so what are wr going to do about n560 PM ?
20:18.50kiozencr2_: afaik yes
20:20.32cr2_BabelO: to post an answer to "suse on artemis"
20:21.02cr2_Marex: we should document the G4 loader.
20:21.14dcordesI have seen that fake before. it is old
20:22.32dcordeshttp://blog.linuxoss.com/2007/04/29/pics-my-htc-artemis-runnig-suse-linux/
20:22.56cr2_dcordes: why do you have such weird EGPIO addresses ? is it a copypasted output ?
20:23.33dcordescr2_: no how would I copy from kaiser w/o any write access.. I've written it down manually
20:25.36dcordescr2_: although you both say it is strange, it still works
20:25.47dcordesthe driver can definetly access gpios
20:25.55cr2_dcordes: they look swapped
20:26.03cr2_i guess so.
20:26.13dcordesdid you read the log?
20:26.17cr2_yes
20:26.30dcordesto my understanding this must work.
20:26.32cr2_i'd like  to look at the vogue spl
20:26.35cr2_yes.
20:26.44dcordesyea dzo told me he will do.
20:26.56dcordesand that he could address you with spl needs
20:29.15dcordescr2_: I doubble checked it: EGPIO phys=c800... virt=9800....
20:29.33cr2_dcordes: what you need now is the 4 param gpio function
20:29.34dcordessome bug in the driver.c?
20:29.57dcordescr2_: ok. how can I get it :D
20:30.01cr2_0x98000000   1   KaiserCPLD
20:30.12cr2_phys is 0x98000000
20:30.39dcordesI'm personally not so concerned about this as long as it works properly
20:31.05cr2_ok.
20:31.23cr2_and you need to learn more about VREG
20:31.26infernixxxdcordes: im online through my kaiser right now so testing is abit hare
20:31.28infernixxx*hard
20:32.15dcordesinfernixxx: ok. did you see the mmc0: message I copied? power = 2 sounds to me like there is power
20:32.30infernixxxiirc i had powermode 0
20:32.36dcordes:)
20:32.38infernixxxi pasted here, look at logs
20:32.52infernixxxdo you still get command timeouts?
20:33.19dcordesinfernixxx: they should appear in dmesg repeatedly all the time right?
20:33.27infernixxxup to a minute
20:33.33infernixxxthen they stop
20:33.37dcordeshm well I have seen some at the very beginning
20:34.10kiozencr2_: lol, so much to say about GPS height measurement. This was recorded on a lake: http://www.qlandkarte.org/shot2.png
20:36.00*** join/#htc-linux dcordes_ (n=lukas@92.227.184.138)
20:36.09*** join/#htc-linux dcordes_ (n=lukas@92.227.184.138)
20:36.19dcordes_is there any way to scroll left/right in vi viewmode w/o left right keys?
20:36.59infernixxxjust use less
20:37.04infernixxxless will auto-wrap
20:37.41cr2_kiozen: it picks the wrong satellites for calculating the position
20:37.58cr2_kiozen: good for long lat, and bad for h
20:39.28kiozenyes but +-15m is imho pretty bad
20:40.06dcordes_cr2_: the timeouts will also appear when printk is off iirc, right?
20:40.11kiozenand the receiver used the wrong hight
20:40.18dcordes_s/cr2_/infernixxx/
20:40.27infernixxxyep
20:40.36infernixxxwell
20:40.44dcordes_looks like I see none of them
20:40.46infernixxxactually i did enable some more debugging output
20:40.57infernixxxany chance you can send me a diff?
20:41.02dcordes_do you have axx do you tree later?
20:41.09infernixxxi have access
20:41.15*** join/#htc-linux p3t3r__ (n=p3t3r__@wlan.net-you.de)
20:41.21infernixxxjust bitbaking all over again remotely
20:41.38dcordes_could you give me your msm_sdcc.c so I can lookup how you do the debug?
20:42.35infernixxx#define VERBOSE_COMMAND_TIMEOUTS        1
20:42.35infernixxx#define MSMSDCC_DEBUG_DMA               1
20:43.32infernixxxalso make sure .config has CONFIG_MSMSDCC_TRACE set
20:43.56dcordes_ok thanks
20:44.35dcordes_cr2_: is that 4 parameter gpio function already present somewhere? I have seen something with gpio parameters in the msm tree
20:49.07dcordes_cr2_: can you look at proc_comm.h?
20:52.06Marexcr2_, was that disassembly any use for you?
20:52.14Marexcr2_, btw as you can see, it jumps to 0xa0100000
20:56.20cr2_Marex, ok, if you are so sure, change the address to 0xa0100000
20:56.31Marexcr2_, I tried ;)
20:56.35Marexstill reboots to wince
20:56.47Marexthere has to be something ... else
20:57.35Marexcr2_, I have an idea where you got that 0xa0201000 from, but that's not the correct address I guess ;-)
20:58.44cr2_Marex: not from ipl.
20:59.20Marexcr2_, you think it uses something else to wake up ?
21:00.41marbaloni found aroow keys in kaiser ;)
21:00.52dcordes_marbalon: great
21:01.00marbalonCPLD+6
21:01.35dcordes_marbalon: how did you find it?
21:02.24marbalonread info in wiki, and write simple driver to read values from all cpld regs
21:03.05Marexcr2_, Im really sure about the codeflow ... qemu shows the same ;)
21:03.30Marexmoreover the IPL should be enough to wake up ... so ...
21:04.05marbalondcordes_: but i don't know how can I use it in driver, pooling mode is ugly
21:05.21marbalontoday I go sleep but tomorrow I try to write something
21:05.35infernixxxmarbalon
21:05.40infernixxxdid you get the i2c-tools image?
21:05.50marbalonyes thanks
21:05.54cr2_Marex: imho the jump to 0xa0100000 is a jump to the spl which was loaded from G4 with the code before ;)
21:06.06dcordes_marbalon: bye
21:06.07cr2_marbalon: have you fixed the wiki bits ?
21:06.25Marexcr2_, nothing gets loaded from anywhere .... the code just sets GPIOs and memory, that's all
21:06.37Marexcr2_, just ... look at the file I sent you :)
21:07.53*** join/#htc-linux noelcragg (n=noel@adsl-68-121-164-151.dsl.pltn13.pacbell.net)
21:08.00*** part/#htc-linux noelcragg (n=noel@adsl-68-121-164-151.dsl.pltn13.pacbell.net)
21:09.53cr2_Marex: hm. then a jump to 0xa0100000 should resume linux ?
21:10.30Marexcr2_, isnt 0x4b4 clear enough ?
21:10.50marbaloncr2_: no, tomorrow, today sleep
21:10.55marbalonnight
21:11.08Marexcr2_, but still it doesnt resume and I have no idea why
21:11.26Marexcr2_, btw what's that 2 meg offset while loading kernel on N560 ?
21:12.50cr2_Marex: can you do 'pd 0x201000 4' ?
21:13.11Marexcr2_, after I find N560 ... sure
21:16.50cr2_Marex: btw, do you know where to get a G4 datasheet ?
21:17.15cr2_i need to check somehting.
21:17.40Marexprobably from msys ... I doubt it's publically available
21:18.47cr2_it was.
21:18.59Marexhuh? :O
21:19.01cr2_in a crippled form ,but it was available
21:19.20cr2_i'm interested in the address map.
21:19.24Marexcr2_, if you get G3 or G4, can you send me ?
21:19.34cr2_afair G4 has embedded ram
21:19.40MarexIm interested too because of palmtt5 and treos
21:19.42Marexugh
21:20.02cr2_Marex: check my g3 page in the xda-dev wiki
21:20.15cr2_many key registers are not documented there.
21:20.36cr2_so this datasheet is mostly useless, because it#s so crippled.
21:21.31cr2_mdocg4_spec.pdf
21:23.10dcordes_how can marbalon work with the cpld without htc-egpio?
21:23.11Marexwell there is register map ... at xda-devs
21:24.05cr2_hehe. ram page command register :)
21:24.49Marexcr2_, where did you get that file ?
21:24.54cr2_this wiki page should be amended for g4.
21:25.24cr2_i have many datasheets later removed by their producers :)
21:26.15Marexthat's good, but ...
21:27.46Marexok, got that file from archive.org
21:27.56Marexhttp://web.archive.org/web/20061111025052/http://www.tri-m.com/products/msystems/files/specs/mdocg3_spec.pdf also :)
21:28.02cr2_30 70 and 80
21:28.32cr2_it's marketing junk that you have there :)
21:28.44cr2_i have a 86 pages manual
21:29.01Marexyea ... damn
21:31.19cr2_still does not tell me what's at 0x201000
21:31.29cr2_let me know if you'll find more.
21:32.04Marexmaybe some hardware that's mmaped there
21:32.26cr2_aliased
21:32.41Marexthat strb indicates it works in 8 bit mode
21:32.53Marexalso it stores data there, so that has to be some setup
21:33.46cr2_0x1000 should be the register area
21:33.57Marexso that's DoC
21:34.14cr2_but not 0x201000
21:34.15Marexmaybe that's second DoC ?
21:34.23Marexor second part of it
21:34.30cr2_sable has it at 0x1000
21:34.36Marexhmm
21:34.39cr2_hmm. interesting idea
21:34.51cr2_btw, the manual has the docs for stacking DoCs
21:35.09Marexwell 128m DoCG3 were two 64m sticked together iirc
21:35.14cr2_did you get it ?
21:35.16Marexhmm ...
21:35.22Marexjust a sec
21:37.12Marexyup, got it
21:37.37Marexhmm ... 0x200000 is 2megs :S
21:37.49cr2_0x201000
21:37.53cr2_128K
21:38.17cr2_check the stacked config
21:38.48cr2_btw, how much flash does n560 have ?
21:39.41Marex128 megs
21:39.52Marex0x1000 + 0x0c ... DoC control register
21:40.09Marex0x1000 + 0x72 ... DoC control confirmation register
21:40.25Marex0x2cc and 0x2d4 in the disasm I sent you :)
21:40.33Marexdo you believe me now ? :)
21:40.41Marexit kind of makes sense ;)
21:40.58cr2_pd 0x201000 4
21:41.08cr2_then i'll believe you.
21:41.41cr2_the stuff above is for the 0x1000, not 0x201000
21:42.07cr2_or you have the second DoC.
21:42.32cr2_at this address, but you can check it in the manual where one finds the control area for a stacked DoC
21:42.56cr2_would be a really wicked config :)
21:43.09cr2_no other devices uses stacked DoCs afaik.
21:44.58Marexdamned ... my haret is trashed
21:45.07Marexgimme a sec
21:45.23MarexI cant connect to the device
21:45.46cr2_my dev info says 51MB free
21:46.04cr2_md8832-d1G
21:47.00Marexstorage - total - 78.20 megs
21:47.19Marexwhich is more than 64 => 128MB DoC
21:47.44MarexMD8832-d1G
21:48.00Marexhmm ... total size 51208 KB
21:48.02cr2_51208KB total say fsc info
21:48.11Marexthat's weird dammit
21:48.28Marexthat wm memoryinfo shows 78 megs of storage :/
21:48.57cr2_hm.it's probably the RAM - 51MB
21:49.02Marexfilestore is 80 megs
21:49.06cr2_80MB for ROM
21:49.08cr2_yes.
21:49.27Marexcr2_, I cant do that PD for you, sorry :/
21:49.37MarexIm stuck with it
21:49.43cr2_i'll try myslef
21:49.55BabelOgood night
21:50.16Marexcr2_, well I put it into default.txt, but I see nothing in the output
21:53.12cr2_let's assume it's the second chip
21:53.18cr2_the 0x400 is the g4 sig
21:53.41cr2_so you'll not jump to 0xa0100000
21:54.03cr2_at this place
21:54.27MarexI see
21:54.56MarexI will look further then
21:54.59cr2_+1000   short   +   RO   chipid_0 (G3=0x200, G4=0x400)
21:56.14Marexlets use my might disassembly resolver :)
21:57.15Marexwell it configures DoC ...
21:57.18Marexat 0x500
21:58.04dcordes_cr2_: is that 4 variable gpio setup required to use SD?
21:58.55cr2_dcordes_: yes.
21:59.28dcordes_cr2_: can you point me in a direction?
22:00.02cr2_Marex: btw, the 0xa0100000 is entered with an r0 param
22:00.06Marexcr2_, looks like the DoC configure routine is pretty obfuscated there
22:00.09cr2_Marex: it can be 1 or 0
22:00.42cr2_Marex: it's the SPL loader to 0xa0100000, see docg3.c driver
22:00.59cr2_dcordes_: it's too difficult.
22:01.31Marexcr2_, but still the IPL has to take care of wakeup, isnt it so ?
22:03.30cr2_yes.
22:03.43cr2_but it then loads the SPL at 0xa0100000
22:03.47Marexfine then ... I will disassemble it whole :E
22:03.51cr2_then black magic happens.
22:04.26cr2_i'm interested to understand the r0 on spl entry meaning.
22:04.27MarexI wont let that little shit play with me :E~
22:04.29dcordes_cr2_: do you know why I can't use reset since kaiser has egpio?
22:05.30dcordes_is that because reset is controlled by the cpld?
22:06.16cr2_dcordes_: no idea.
22:07.21cr2_Marex: r0=0 means this "fast boot" without loading spl. r0=1 means the spl was loaded from doc
22:11.03Marexcr2_, I have to convince qemu that 0x201000 is 0x400
22:12.09Marex<PROTECTED>
22:12.13cr2_yes.
22:12.32cr2_hmm. r0 is immediately killed by the spl anyway.
22:12.59cr2_we should do a better analysis of the ipl and spl gpio setup.
22:13.20cr2_maybe they mess too much with it.
22:13.45Marexyou can just check the ipl.asm I sent you .. .everything is there
22:13.52MarexI will check the new codeflow soon
22:13.55Marextomorow probably
22:14.08dcordes_cr2_: do you think we can get the 4 variable stuff working at all?
22:14.18cr2_dcordes_: yes.
22:15.07dcordes_ok ping me when there is time/action required
22:18.45cr2_dcordes_: do you know perl ?
22:19.01dcordes_cr2_: no
22:19.10cr2_ok
22:21.13cr2_Marex: moved a bit forward
22:21.25Marexhow come ?
22:21.27dcordes_cr2_: is it a good language for beginners?
22:21.33Marexcr2_, what happened ?
22:21.56cr2_at the spl start it checks for RCSR bit 4 set
22:22.13cr2_and there are multiple gpio configs depending on the setup.
22:22.20cr2_Marex: do you know perl ?
22:23.24Marexjeez ...
22:23.29cr2_25300 and 25304
22:23.41Marexcr2_, just the real basics of perl (and I mean basics )
22:24.01Marexcr2_, do you think it jumps to SPL and does the wakeup there ?
22:24.03Marexno way
22:24.33cr2_yes.
22:24.51cr2_omg, it wants to turn the mmu on.
22:25.00cr2_no, not yet.
22:25.08dcordes_cr2_: are you in artemis SPL?
22:25.08Marex:D
22:25.32cr2_i'm bad in mrc cp15 decoding, need to write a memo into wiki
22:25.36cr2_dcordes_: n560
22:25.37Marexcr2_, I will look at the IPL code ... it has to wake up from there ... I never saw SPL to be used for wakeup
22:25.55cr2_Marex: me too
22:26.04Marexthen it has to be in IPL somewhere
22:26.09cr2_but n560 "thinks different"
22:26.13Marex:D
22:26.21Marexat least you wont be bored ;)
22:26.49cr2_heh. it turns mmu on with the wince config ;)
22:27.23cr2_yeah, we are toast
22:28.09cr2_nw i need to check where is the stack, and if the SPL does bad things later
22:28.31cr2_otherwise we can turn the mmu off somewhere later.
22:29.40Marexin qemu I hang at 0x564
22:32.57*** join/#htc-linux zwox (i=zwox@par69-7-82-243-234-94.fbx.proxad.net)
22:33.34cr2_SP is at 0xa0004000
22:33.37cr2_it's ok.
22:33.48cr2_no.
22:34.14cr2_the page table is at +0x4000, the stack is 64K up. +0x14000
22:34.46cr2_that's ok, the first MB may be trashed without any harm to linux kernel et al.
22:37.42cr2_0x14da48
22:37.55cr2_in the spl
22:40.14cr2_hehe. it's somewhere outside.
22:42.17Marex:S
22:42.22MarexN560 is perverted :S
22:43.17cr2_Marex: one opcode hack to spl may be the best solution
22:43.48cr2_but it will kill the wince resume
22:44.06Marexthere has to be a way to wake up with stock bootloader too
22:44.42cr2_the MMU and stack in the first MB is ok.
22:44.56cr2_the second MB is free for weird things too.
22:45.09cr2_the spl will not hurt itself i guess.
22:45.45cr2_it seems the the splash means 'resume failed'
22:46.04cr2_so it's necessary to backtrack from show_splash
22:46.09dcordes_infernixxx: ping
22:46.11Marexcr2_, I will disassemble the rest tomorow, ok ?:)
22:47.01cr2_and it's not much, i've already found the splash entry.
22:47.09cr2_it depends on 1 functions.
22:47.20Marexwell it does some weird delay loop calibrations or something
22:47.21cr2_on its return value :)
22:47.51cr2_if we will resume on n560, that will be the most neat hack i've done so far :)
22:48.57cr2_hmm.
22:49.17cr2_it checks 253ac and 253a8
22:49.20Marexcr2_, you will ;)
22:49.40cr2_no, they are zeroed
22:50.37cr2_i guess it calculates some checksums
22:50.50cr2_for 0x25000 +0x1000 page
22:52.35infernixxxdcordes_: pONG
22:53.01cr2_41000000 and 41500000
22:53.05dcordes_infernixxx: any luck with your OE images?
22:53.06cr2_what's that on pxa ?
22:53.23infernixxxsomehow after rm -rf tmp, openssl fails to build
22:53.37infernixxxhavent investigated, watching some linux.conf.au footage
22:53.52Marexcr2_, 41300000 are clock
22:53.53Marex;D
22:54.13cr2_ssp ?
22:54.14dcordes_infernixxx: you in australia?
22:55.05dcordes_you can try to mtn pull && mtn update
22:55.12dcordes_xora pushed our changes btw
22:55.38infernixxxdcordes_: no. off the website
22:55.45*** join/#htc-linux surgex0 (i=surge@pool-71-186-234-216.bflony.fios.verizon.net)
22:55.51cr2_kpkdi and sscr0
22:57.09cr2_Marex: no, it's too complex ;)
22:57.32cr2_imho it's better to take a risk and check how to reflash the DoC
22:58.06Marexcr2_, Im looking at the codeflow with DoC ID in place
22:58.10Marexthat's really something
22:58.38dcordes_infernixxx: I will prepare a diff of my tree when I cleaned it up a bit
22:59.04dcordes_bbl
22:59.37infernixxxcool
22:59.57cr2_Marex: a lot of bitstuffing goes on there. looks worse than aes256 :)
23:02.07Marexthe IPL is totally screwed :E~
23:02.15Marexit jumps totally randomly in the code
23:05.24cr2_#define RCSR            __REG(0x40F00030)  /* Reset Controller Status Register */
23:06.45cr2_enter cold reset
23:07.52cr2_Marex: did i write something about 25304 and 25314 in _pm.c ?
23:08.32Marexyup
23:08.44Marex25304 is there
23:09.01cr2_rcsr is checked again.
23:09.39Marexjeez
23:10.17cr2_1a2b3c4d leads to cold reset here.
23:10.30cr2_rcsr is checked 4 times.
23:10.35cr2_last time with mmu on.
23:11.46cr2_how much is 0xb0000004- 0x07100000
23:11.58Marexwell I started analysing the codeflow in IPL ... it really does weird stuff, I give up :S
23:12.34cr2_this is the last check before warm reset
23:13.11cr2_a8f00004
23:13.24cr2_it's 40f00004
23:13.39cr2_#define PSSR            __REG(0x40F00004)  /* Power Manager Sleep Status Register */
23:13.41cr2_ok.
23:14.19cr2_enteros
23:14.22Marexcr2_, I guess I will stick with palms a little longer ;)
23:14.30cr2_check gplr4
23:14.36cr2_wtf is that ?
23:14.50Marexgplr4 doesnt exist
23:14.55cr2_Marex: i'll check if the mmu is turned off at the end :)
23:15.18cr2_http://www.handhelds.org/moin/moin.cgi/LooxC550Hardware
23:15.25cr2_battery cover says this wiki
23:16.56cr2_25318 is checked.
23:17.42cr2_and compared to 20525555
23:18.16cr2_it's _URR ?
23:18.26cr2_hm.
23:18.30cr2__UUR
23:18.33cr2_RUU
23:18.45cr2_rom update utility ?
23:18.56cr2_should be the linux sig :)
23:20.10cr2_253a4 ULDR
23:21.21cr2_ULDR will be loaded at 0xa0200000
23:21.22cr2_bad.
23:22.22Marexcr2_, I really dont follow anymore :)
23:22.47MarexI ended at [1:09] :(
23:23.22cr2_it may be a neat way to load linux from the SPL, but is not the resume
23:23.44cr2_i'm going through the spl, and it does a lot of weird stuff
23:24.26cr2_ULDR needs the CECE sig of course ;)
23:24.27Marexipl is the same
23:24.38cr2_otherwise it's "invalid"
23:25.03cr2_calcULDRchecksum
23:25.12cr2_not only ECEC
23:25.32cr2_and then jump to 0xa0200000
23:25.52cr2_the tags are at +0x100
23:26.11cr2_the page table is at +0x4000
23:26.13cr2_?
23:26.23*** join/#htc-linux dzo (n=dzo@121.98.128.127)
23:26.35cr2_how long are the tags ?
23:26.58cr2_hi dzo
23:27.26dzohi cr2, i hear youve been looking at the kaiser spl
23:27.46dzofound anything interesting?
23:27.56Marexdzo, n560 spl ;)
23:29.29cr2_Marex: flash IPL success
23:29.38cr2_dzo: yes.
23:29.48cr2_dzo: where is the vogue CPLD located ?
23:29.50Marexcr2_, what? you reflashed n560 ?
23:30.05cr2_Marex: no, just the SPL message :D
23:30.24Marex:-)
23:30.50dzoI'm trying to figure out how audio is routed on my vogue, i think its done by the cpld.
23:31.24cr2_dzo: is it at the same address as on titan ?
23:31.30dzoyes
23:31.41dzoits cpld3
23:32.24cr2_what is cpld3 ?
23:32.25dzo16x16 bits with the top 8 bits of all doing strange things
23:32.44dzojust the htc name for it I guess
23:32.53cr2_dzo: http://wiki.xda-developers.com/index.php?pagename=Kaiser_CPLD ?
23:33.01cr2_where did you pick this name ?
23:33.19dzoyes, similar I think
23:33.23cr2_the top 8bits on htc CPLDs are just random junk
23:33.38cr2_but they are always accessed in 16bit mode
23:33.59dzoOK, that confused me a bit.
23:34.22cr2_you should mask them from the very beginning.
23:34.44cr2_dzo: i think you'd look at the htc 4 param gpio function
23:35.37cr2_gpio(NUM,i,j,k)
23:35.48cr2_NUM is the gpio number
23:36.00cr2_i can be 1 or 2
23:36.08dzowhere is that? in the spl?
23:36.14cr2_j is 0,1,3,4
23:36.16cr2_yes
23:36.25cr2_and k is 0,1
23:36.42cr2_it's some gpio config stuff
23:36.58dzospl might not help for audio, it probably only needs to init sd and usb and stuff
23:37.24cr2_and the keyboard.
23:37.39cr2_spl does not control the sound
23:38.04cr2_and wavedev.dll uses rpc afair.
23:38.05dzohmm, ill see what wavedev.dll does.
23:38.28dzoi can't see any rpc calls on my phone.
23:38.39dzofrom wavedev.
23:38.39cr2_rpc=DMOV
23:38.47cr2_hm, i'll check.
23:38.50dzoit might access adsp directly
23:38.52Marexgnight
23:39.21dzooh i thought dmov was DMA.
23:40.04dzothe android kernel does rpc using shared mem channels.
23:40.28cr2_android is a bit different.
23:41.20dzothere is something in wince called DEX (data exchange i think) that looks like low level comms with A9.
23:42.13cr2_i'll check the wavedev
23:44.01dzook, how do you extract wince dlls, i've been traceing things in haret but really don't understand how dlls are stored and loaded. seems like a horrible mess to me.
23:44.39cr2_from the rom update
23:44.51cr2_have you tried testwm5.exe ?
23:45.01dzono, what does it do?
23:45.05cr2_it comes with the full source
23:45.22dzothanks, i'll try it out.
23:45.23cr2_it loads and dumps the xip dlls
23:45.36cr2_http://hands.com/~lkcl
23:45.53cr2_the binary is a the top
23:46.12dzothanks

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