00:00.09 | miknix | I always press the run button instead of listen on haret |
00:01.21 | cr2_ | 2c->3f are also lcd related ? |
00:03.17 | cr2_ | 2,0,0x30 and 2,0,0xc0 are queried over i2c . also lcd related |
00:04.13 | cr2_ | cpld(5,4) and gpio0x13. storage. |
00:04.22 | cr2_ | miknix: what is storage on wiza ? |
00:04.32 | miknix | SD card |
00:04.46 | infernix | where can I learn more about GPIOs? |
00:05.55 | cr2_ | miknix: ok. so bad news for you. |
00:06.03 | miknix | : \ |
00:06.04 | cr2_ | miknix: ^^ |
00:06.19 | miknix | cr2_, why? |
00:06.22 | *** join/#htc-linux surgex0 (n=surge@pool-71-186-234-216.bflony.fios.verizon.net) |
00:06.23 | cr2_ | gpio0x13 and cpld(5,4) are sd power |
00:06.40 | miknix | cr2_, hum... |
00:07.02 | miknix | so it is the cpld that is messing around with hw on boot |
00:07.35 | BabelO | cr2_: do i have sd power ? ;) |
00:07.55 | cr2_ | miknix: so you need the cpld driver to power the SD card |
00:07.59 | cr2_ | BabelO: check wiki |
00:08.12 | cr2_ | infernix: ? |
00:08.51 | miknix | cr2_, that is the main problem we are currently facing.. we have to rely on the state hw is left on wm5 |
00:08.58 | infernix | i'm trying to find out how to use haret to find out what the gpios are for kaiser so I can add to the wiki |
00:09.25 | miknix | cr2_, if gsm is left up, some obscure pin mux is enabled and usb power is disabled (for example) |
00:10.28 | cr2_ | miknix: i've added the usb pins, it's easy to fix |
00:10.45 | cr2_ | miknix: but to powerup the SD you need the i2c cpld driver. |
00:12.16 | cr2_ | miknix: so these weird wince hacks you use are only doing exactly that thing. |
00:12.49 | cr2_ | the obscure pin mux settings can be checked and dcoumented too. |
00:12.56 | miknix | lol |
00:13.47 | miknix | HTC has joined the Open Handheld Allience |
00:13.53 | miknix | s/Allie/Allia/ |
00:14.09 | miknix | where is the openness? |
00:15.24 | miknix | cr2_, seriously: http://www.openhandsetalliance.com/oha_members.html |
00:15.50 | cr2_ | miknix: 7200A probably. see "android" |
00:16.53 | *** join/#htc-linux BabelO (n=Fabrice@unaffiliated/babelo) |
00:17.06 | *** join/#htc-linux diogene31_ (n=rj@mur31-2-82-243-122-54.fbx.proxad.net) |
00:17.18 | infernix | cr2_: for instance, on kaiser the jog wheel generates in1: in1-2(194)=0 in1-3(195)=0, in1: in1-3(195)=1, and in3: in3-22(278)=1 when pressed + in3: in3-22(278)=0 when depressed. |
00:17.33 | infernix | how do I interpret this data so it can go into the wiki (http://wiki.xda-developers.com/index.php?pagename=Kaiser_GPIO)? |
00:17.42 | cr2_ | miknix: http://wiki.xda-developers.com/index.php?pagename=WizardGPIO |
00:18.10 | cr2_ | miknix: you'd probably create the WizardCPLD page, so we can put the cpld(M,N) pairs there. |
00:18.28 | miknix | cr2_, oh. that was nice |
00:18.50 | infernix | i'm eager to learn but have no idea where to start |
00:19.57 | cr2_ | infernix: i don't remember the kaiser gpio nubering |
00:20.30 | cr2_ | infernix: i think the matrix keyboard pins for 7x00 are documented |
00:21.09 | cr2_ | miknix: but these pairs have different meaning for i2c cpld than for artemis et al. |
00:24.03 | cr2_ | miknix: gpio0x21 is the usb_en |
00:24.25 | cr2_ | isn't it the same on artemis ? |
00:28.55 | cr2_ | miknix: the keyleds are on the i2c cpld too. |
00:34.33 | miknix | BabelO, omap_spi_100k.c:258 if (tx != NULL) // shouldnt be *tx ? |
00:38.46 | miknix | BabelO, no. forget it. but that comparison doesnt make sense for me. tx is always incremented, how it could be null? |
00:46.20 | bigdissaved | Is anyone free for a sec? I have a question about the HTC Apache.. |
00:46.46 | BabelO | bigdissaved: ask |
00:47.49 | bigdissaved | my phone met its demise a week ago. The touchscreen got a spider crack, and, now, useless. I have stripped it, pulling all the heat shilding off the chips. |
00:48.20 | bigdissaved | Where would U like the pics sent of each chip, as indicated on the todo list for the HTC apache page? |
00:48.34 | miknix | BabelO, http://pastebin.com/d70044be0 |
00:48.57 | bigdissaved | ok, I have spent a bit to much atm, will p |
00:49.28 | bigdissaved | srry, stupid slow conn, wrong chan |
00:50.30 | BabelO | bigdissaved: maybe Kevin2 or cr2_can like to have pics, else try to put it on the xda forum |
00:50.57 | bigdissaved | ok, I have spent a bit to much atm,ok |
00:51.02 | bigdissaved | will do. |
00:54.44 | miknix | BabelO, I get it now why you check the NULL under the loop :) |
00:55.09 | BabelO | miknix: :) |
00:56.09 | BabelO | miknix: i go to bed now, i ll check tomorrow why my egpio does not work and why i did not get irq |
00:56.14 | BabelO | good night |
00:56.19 | miknix | BabelO, good night |
00:59.40 | bigdissaved | brb, supper |
01:29.29 | *** join/#htc-linux the_sys0p (i=the_sys0@gateway/tor/x-ccba2801210015e6) |
01:48.28 | dcordes_ | hola |
01:50.45 | Marex | cr2_, bah |
01:50.54 | Marex | cr2_, maybe palmos's bootloader corrupts memory |
01:51.31 | Marex | cr2_, if you still have the asm file I sent you, check 0x258 |
01:52.27 | dcordes_ | hi Marex |
01:53.42 | Marex | dcordes_, hello |
01:53.57 | dcordes_ | Marex: working on artemis? |
01:54.05 | Marex | palm lifedrive |
01:55.14 | dcordes_ | nice device. looks like you have access to plenty of handhelds |
01:56.33 | Marex | yea, but I had to exchange something for it ... |
01:56.34 | dcordes_ | Marex: is lifedrive htc manufactured? |
01:56.52 | dcordes_ | Marex: your soul? |
01:57.00 | Marex | probably yes, not really sure, but most handhelds are HTC made |
01:57.22 | Marex | dcordes_, something like that ... yea |
01:57.30 | dcordes_ | most? I knew plenty models are, many ipaqs among them. but most.. |
01:57.58 | Marex | well ipaqs, htcs itself, palm treo series |
01:58.08 | Marex | palmtc was made by Asus |
01:58.14 | Marex | palmtt3 was Asus too |
01:59.32 | dcordes_ | ok |
02:00.05 | dcordes_ | I'm really impressed about how active you are with all those different devices |
02:00.29 | dcordes_ | resepect even it's not much worth from a person of my level of knowledge |
02:02.06 | Marex | dcordes_, there are many other people that should have your respect (cr2 for his endless patience, aquadran, pH5 etc) ;-) |
02:04.59 | dcordes_ | no question they all do have it |
02:06.56 | dcordes_ | my utmost pardon in case that was not visible before. I hoe in future I will be able to be more productive |
02:07.10 | dcordes_ | hope |
02:07.38 | *** join/#htc-linux surgex0 (i=surge@pool-71-186-234-216.bflony.fios.verizon.net) |
02:07.42 | Marex | well I cant speak for cr2, but I dont care ;) |
02:08.27 | Marex | in my case I dont have anything else than handheld hacking and school so I have to stick to it |
02:10.56 | dcordes_ | Marex: wow you still going to school? |
02:12.00 | Marex | I started uni this year |
02:12.45 | dcordes_ | congrats I guess you study IT stuff? |
02:14.17 | Marex | yea, something like that |
02:14.33 | Marex | http://www.mff.cuni.cz/toISO-8859-2.en/ |
02:16.00 | dcordes_ | Marex: which of those? Information Services? |
02:16.53 | Marex | no idea ... |
02:17.03 | Marex | I study "programming" there |
02:24.28 | dcordes_ | Marex: who's aquadran? |
02:24.44 | dcordes_ | btw |
02:24.53 | Marex | one guy from hh.org |
02:25.08 | dcordes_ | never heard about him. what does he do? |
02:25.29 | Marex | mypal a716 |
02:26.17 | dcordes_ | I tried to get some asus from my brother for researching. but he is not interested in what i do |
02:26.24 | dcordes_ | works for asus |
02:27.25 | dcordes_ | anyway timefor some sleep. I'm more productive at daytime |
02:27.28 | dcordes_ | night Marex |
02:28.51 | Marex | gn |
02:29.08 | Marex | 4.30am here |
03:49.09 | *** join/#htc-linux ltxda (n=ltxda@c-98-196-125-205.hsd1.tx.comcast.net) |
07:03.23 | *** join/#htc-linux LunohoD_ (n=alex@e180070102.adsl.alicedsl.de) |
07:52.00 | *** join/#htc-linux rmoravcik (n=rmoravci@adsl-dyn8.78-98-50.t-com.sk) |
08:20.23 | dcordes_ | hi |
08:29.27 | *** join/#htc-linux goxboxlive (n=goxboxli@208.84-48-176.nextgentel.com) |
08:44.10 | dcordes_ | cr2_: can you tell me something about "80-VA736-2 Rev C" it is in the gpio_hw.h. Is that a reference to part of the datasheets? |
08:54.50 | swetland | it's a qualcomm document number for one of the databooks for the 7K family |
08:56.11 | dcordes_ | ok I probably though it was available documentation on the gpio. |
08:56.35 | dcordes_ | I'm looking at what is required to modify in order to make sdc give more than timeouts |
08:57.57 | swetland | need to get the gpio mux configured correctly and enable the appropriate clocks. in our world we use proc_comm to ask the A9 to do that, since it's the master as far as that stuff is concerned |
09:00.32 | dcordes_ | swetland: I have seen the comment concerning that situation in the code. So I can realize sd also with those shadowed gpios? |
09:01.01 | dcordes_ | and concerning the mux: is that already available in the code? I have read a comment somewhere which says it is not |
09:01.49 | swetland | the mux/config stuff is owned by the A9 -- we don't have any code that directly interacts with it |
09:02.53 | dcordes_ | long story short I can't make SD working w/o a9 side access? |
09:02.54 | swetland | if the memory protection stuff is configured "correctly" the A11 cannot even access the registers |
09:03.23 | swetland | depends a lot on how your particular amss build on whatever device is setup |
09:03.44 | dcordes_ | sdc1, sdc2 |
09:03.56 | swetland | and the stuff prior to amss 6.x (what we're working with) is a complete black box to me |
09:04.38 | dcordes_ | oh you mean the board by whatever device? that is kaiser |
09:04.59 | swetland | well specifically what version of the baseband software is running on it |
09:05.16 | swetland | and how it's configured/customized by the oem |
09:05.32 | dcordes_ | I would have to ask cr2_ |
09:06.00 | dcordes_ | or can I look that up in the wince bootloader? It has that rom version line |
09:06.05 | dcordes_ | radio rom |
09:06.14 | swetland | for devices that are already out there, that'll be some version prior to 6.x, and I honestly don't know how much (if any) of the shared memory interfaces are compatible |
09:06.53 | dcordes_ | I know it runs some pistachio microkernel |
09:07.00 | dcordes_ | and thumbcode? or anything |
09:08.41 | dcordes_ | swetland: do you think SanMehat could help with that? |
09:10.01 | swetland | nobody on our side deals with older radio stuff. we're working on what will ship in future products (and working with qualcomm on integration with it), not trying to adapt existing stuff that was not designed to support linux |
09:10.53 | dcordes_ | that sounds much like we will sooner or later have to access arm9 and run our oun stuff |
09:11.38 | swetland | our world ends at the shared memory interface to the arm9. it is not something we have source access to, etc. |
09:13.04 | dcordes_ | swetland: so having information on the shared memory region would in this case sufice to control the gpio registers? |
09:14.01 | swetland | yeah -- the proc_comm interface (it's in the sources) provides calls for voltage regulator, clock, and gpio configuration/mux control |
09:14.02 | dcordes_ | maybe we can obtain that from the pistachio code |
09:14.36 | dcordes_ | swetland: and that communication done by proc_comm is specific to what runs on your particular arm9? |
09:15.10 | swetland | a lot of it is new with the new version of amss we're working against |
09:15.35 | swetland | presumably there's equivalent features somewhere in the old stuff |
09:17.59 | dcordes_ | swetland: is amss=arm9 software?# |
09:18.42 | swetland | yes. |
09:18.46 | swetland | apparently it stands for Advanced Mobile Subscriber Software |
09:18.56 | dcordes_ | that's neat :) |
09:18.58 | swetland | according to google search |
09:31.12 | dcordes_ | swetland: how could I just try if the current gpio mux setting would work to give sd some clock? |
09:34.54 | dcordes_ | I mean if it works for 7500A it should somehow work for 7200 |
09:37.11 | dcordes_ | swetland: can you tell me where in the code I can look up how you tell proc_comm to provide voltage/clock for sdcc or other peripherals so I can get an idea of how that is done? |
09:49.04 | *** join/#htc-linux AlGe (n=alge@chello080109231226.4.uni-klu.teleweb.at) |
09:57.19 | dcordes_ | hi AlGe |
10:01.18 | AlGe | hi |
10:41.35 | *** join/#htc-linux rob_w|laptop (n=rob@X1d92.x.pppool.de) |
10:53.22 | *** join/#htc-linux Marex (n=marex@85-132-216-250-eth3-gwfm10-user.802.cz) |
11:25.17 | *** join/#htc-linux patoche (n=patoche@lns-bzn-56-82-255-206-113.adsl.proxad.net) |
11:28.33 | dcordes_ | cr2_: are you around? |
12:33.55 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
12:35.42 | cr2_ | yes |
12:36.41 | BabelO | hi |
12:36.45 | cr2_ | dcordes_: it seems that the MMU protection on kaiser is configured more "correctly" than on all 7500 devices |
12:37.08 | cr2_ | dcordes_: because we don't see the SD and MDP registers from wince |
12:37.11 | cr2_ | hi BabelO |
12:38.52 | cr2_ | dcordes_: but nothing is lost. the SPL accesses all this hardware directly, so if we will get a chance to overwrite it the NAND, we will take complete control of the hardware |
12:44.39 | *** join/#htc-linux Banji (n=ontv@87-196-69-187.net.novis.pt) |
12:45.08 | dcordes_ | cr2_: what is spl again? |
12:46.07 | Banji | hi everybody, i have qtek9090 how i install linux???yesterday i´m try to install butneed the familiar login and i try other version and i can but the keypad of qtek9090 is ever on they write alone |
12:46.32 | Banji | how i installone correct version? |
12:46.36 | Banji | ??????? |
12:46.58 | Banji | how i install one correct version? |
12:48.00 | dcordes_ | ah that bootloader right |
12:48.21 | Banji | yes the bootloader right |
12:48.59 | Banji | but in one version stop and they appear in the screen "familiar login" ando don´t know the login |
12:49.11 | Banji | and the other version the bootloader |
12:49.18 | Banji | the go to the linux painel |
12:49.56 | Banji | but every keys on, and then my qtek9090 write alone |
12:50.38 | Banji | soory for my english, but i don´t write good english |
12:51.03 | dcordes_ | no te apures |
12:52.04 | Banji | do you can help me? |
12:52.18 | Banji | or one person to help me |
12:52.44 | dcordes_ | Banji: I don't understand your problem. you mix a lot of things up at a time. |
12:53.24 | Banji | my problem is:i can´t install the linux in my pocket pc qtek9090 |
12:53.37 | dcordes_ | cr2_: what would we need to overwrite in the NAND? amss? spl? |
12:54.23 | dcordes_ | Banji: ahhh :) HTC Blueangel you mean |
12:54.30 | Banji | yesterday i try to install , but in my first tentative is ak me for "familiar login" and i don´t know the login |
12:55.16 | dcordes_ | Banji: try root and no password |
12:55.27 | dcordes_ | Banji: where did you download your image? |
12:55.35 | Banji | and in my second tentative, i can bootloader to finish, i can to entry in linux, but the my keys are every on, and then my pocket pc write alone |
12:56.15 | dcordes_ | Banji: that is a known bug. No fix for that yet. Try not to press too many buttons. |
12:56.41 | Banji | [dcordes_]: i don´t press any button |
12:56.50 | Banji | is press only |
12:57.32 | Banji | do you know one website where explain step by step one installation for the qtek9090 |
12:57.38 | dcordes_ | yes |
12:58.04 | Banji | what the website? |
12:58.07 | dcordes_ | Banji: try using the name blueangel rather than qtek9090 |
12:58.26 | dcordes_ | Banji: http://handhelds.org/moin/moin.cgi/BlueAngelHowtoInstallLinux |
12:59.00 | BabelO | cr2_: egpio have no effect :( |
12:59.15 | Banji | this site is the my first tentative |
12:59.24 | dcordes_ | hey BabelO :) can you give Banji a link to your latest qtopia image? |
12:59.41 | dcordes_ | I lost my bookmarks |
12:59.58 | Banji | and this tentative is ask the "familiar login" |
13:00.10 | Banji | and i dont know the familai login |
13:00.42 | BabelO | i already told yesturday to take http://linuxtogo.org/~htcpxa/htcblueangel/Qtopia/ both file here the .exe and .bz2 to copy this file on root of the sd card fat partiotion and start .exe from wince |
13:00.45 | dcordes_ | Banji: you are using a very very old image. There is new linux |
13:01.26 | infernix | dcordes_: im completely lost in all this gpio stuff |
13:02.04 | infernix | i lack a good deal of knowledge to do anything useful in kernel space it seems |
13:02.10 | dcordes_ | infernix: so am I |
13:02.42 | Banji | is this my second tentative and my keyboard are every ON, and is write only, i don´t press anything buton and is write only |
13:02.55 | dcordes_ | Banji: do what BabelO says. |
13:03.03 | Banji | yes |
13:03.13 | dcordes_ | did he say something about familiar? |
13:03.19 | Banji | in this moment i load |
13:03.27 | Banji | is can bootloader |
13:03.50 | Banji | the screen appear no appointmanets today |
13:03.54 | dcordes_ | good. no login needed there. only put both files on sd main directory and start exe |
13:04.06 | Banji | [dcordes_]: i can loader |
13:04.12 | dcordes_ | good |
13:04.41 | cr2_ | dcordes_: http://wiki.xda-developers.com/index.php?pagename=WizardCPLD |
13:04.44 | Banji | [dcordes_]: but know my keyboard of the my qtek9090 is turn on, and they write nly |
13:05.55 | dcordes_ | cr2_: what's up with this? |
13:06.13 | Banji | ohhh loll i can know |
13:06.15 | Banji | :) |
13:06.17 | Banji | thanks |
13:06.18 | Banji | :P |
13:06.31 | dcordes_ | that's great :) |
13:06.39 | dcordes_ | have fun with it |
13:06.40 | Banji | i have to wait, to go started |
13:06.48 | Banji | and i dont wait |
13:07.06 | Banji | but know i wait and pocket pc is on :) |
13:07.25 | Banji | thanks for all help they need me thanks a lot |
13:07.26 | Banji | :) |
13:07.33 | Banji | thanks dcordes_ |
13:07.41 | dcordes_ | Banji: thank BabelO. He mad the image |
13:08.09 | Banji | thankd BabelO |
13:09.10 | dcordes_ | cr2_: is the kaiser cpld unit also controlled by amss/arm9? |
13:09.58 | dcordes_ | I don't quite get the role of the cpld. All I know is it provides extra gpios |
13:10.58 | dcordes_ | if I could just fire some of GPIOB0 0x0001 sd power |
13:11.55 | dcordes_ | cr2_: It is not in there but if sd power is on the cpld, it will also provide the clock setting right? We already have sd clock |
13:12.20 | cr2_ | dcordes_: it's the way to world domination |
13:12.34 | dcordes_ | ! |
13:12.46 | cr2_ | dcordes_: no, the kaiser cpld is mmaped and can be accessed freely |
13:12.51 | bigdissaved | I see good news! |
13:12.51 | bigdissaved | http://www.linuxdevices.com/news/NS4897933970.html |
13:12.59 | bigdissaved | its a short artical |
13:13.12 | cr2_ | dcordes_: what we really need is the same style documentation for kaiser. also for gpio setup. |
13:13.34 | cr2_ | bigdissaved: May 06, 2005 |
13:13.37 | dcordes_ | cr2_: the same style of what docu? |
13:14.14 | bigdissaved | cr2_, My ATC Apache bit the dust, and I have taken it all apart with all the heat shields and stuff. If I can, U want pics of all the chips? |
13:14.18 | cr2_ | dcordes_: there is a 4 parameter gpio config function |
13:14.19 | bigdissaved | ir a list of all the chips? |
13:14.39 | bigdissaved | s/ir/or/ |
13:14.45 | dcordes_ | cr2_: am I right with the statement that sd clock is turned on by the cpld and we already do this? (I see the "mmc0: n Hz" kernel message) |
13:15.12 | cr2_ | bigdissaved: i think apache is well supported. the only missing part is the CDMA modem, and nobody wanted to hack on that. |
13:15.42 | bigdissaved | ok, I saw on its poage on the todo list is better pics of it inside.... |
13:15.51 | cr2_ | dcordes_: CPLD only has some extra gpios. |
13:15.52 | dcordes_ | where n > 0 |
13:16.33 | dcordes_ | cr2_: yes but am I right with the clock thing? |
13:16.47 | cr2_ | dcordes_: htc does not go into such layering schema like the androids, they control the clock registers directly from wince. |
13:17.13 | cr2_ | dcordes_: so they saved us the rpc hassle |
13:18.45 | dcordes_ | cr2_: so that's why |
13:18.57 | dcordes_ | ..sd is already clocked |
13:19.29 | dcordes_ | what about the cpld now? we need documentation you say? |
13:19.49 | cr2_ | dcordes_: see the 'sdX related' clock registers in wiki ? |
13:20.38 | dcordes_ | the one I pasted above? |
13:20.54 | cr2_ | a0 - ac |
13:21.09 | cr2_ | in MSM_CLK_CTL |
13:21.11 | cr2_ | http://wiki.xda-developers.com/index.php?pagename=KaiserMemoryMap |
13:21.39 | dcordes_ | yea |
13:21.50 | cr2_ | can you trace this area while inserting the SD card ? |
13:22.02 | dcordes_ | in an instance |
13:24.34 | dcordes_ | ok console fired up |
13:27.50 | dcordes_ | ah just noticed my comment there "exception" |
13:28.05 | dcordes_ | but that's not in the range we talk about .. |
13:28.35 | cr2_ | you should see other clocks too. |
13:28.46 | cr2_ | for example the backlight (pwm?) clock |
13:29.23 | *** join/#htc-linux CVirus (n=GoD@196.205.193.110) |
13:30.26 | dcordes_ | cr2_: are we talking about addlist mmutrace; wirq n? |
13:30.48 | dcordes_ | HaRET(1)# addlist mmutrace 0xa86000a0 0xac |
13:30.53 | dcordes_ | 000.000 IRQS IRQ=00000100: IRQ8=1 |
13:30.57 | dcordes_ | 005.030 IRQS IRQ: IRQ6(6)=1 |
13:31.03 | dcordes_ | 008.530 IRQS IRQ: IRQ6(6)=1 |
13:33.25 | cr2_ | btw, do you see that the MDP are is not mapped by ARM9 ? |
13:34.05 | dcordes_ | cr2_: no I don't know how to determine which cpu does what |
13:35.00 | dcordes_ | cr2_: did I do the trace right? |
13:35.05 | cr2_ | no |
13:35.17 | cr2_ | you forgot the p2v on the address |
13:35.23 | dcordes_ | ?? |
13:35.35 | cr2_ | mmutrace works on virtual addresses |
13:35.46 | cr2_ | 0xa86000a0 is physical address |
13:36.09 | cr2_ | should be p2v(0xa86000a0) |
13:36.36 | dcordes_ | offset is ok? |
13:36.48 | cr2_ | and if you start at 0xa0 you'd use 0xc for the "size" |
13:37.06 | dcordes_ | ok |
13:37.10 | cr2_ | addlist mmutrace p2v(0xa8600000) 0xac |
13:37.17 | cr2_ | then you'll see all clock regs |
13:37.25 | dcordes_ | let's roll |
13:37.58 | dcordes_ | 00: Mapping a8600000(@00700000) accesses to e1100000 (tbl 00700402) |
13:37.58 | dcordes_ | 01: Mapping 92e00000(@a8600000) accesses to e1200000 (tbl a8602402) |
13:38.27 | cr2_ | b2e is ok. |
13:38.32 | dcordes_ | that's it and of course the sd detect irq |
13:38.45 | dcordes_ | what's b2e? |
13:39.00 | cr2_ | 92e00000 = b2e00000 |
13:39.20 | cr2_ | so oyu don't see any writes to that area. |
13:40.00 | dcordes_ | which means we can't see sd clock r/w from wince? |
13:40.15 | cr2_ | with the current config no. |
13:40.19 | cr2_ | it seems so. |
13:40.24 | dcordes_ | I thought it is freely accesible |
13:40.40 | dcordes_ | which is why we don't need rpc routing and stuf |
13:40.46 | cr2_ | that's why i always talk about better documenting the gpio access |
13:41.05 | cr2_ | yes, in the SPL it is accessed directly |
13:41.29 | dcordes_ | which is before the good working mmu protection is turned on |
13:41.31 | dcordes_ | ? |
13:41.46 | cr2_ | but wince kernel may turn some magic bits off on boot, so you are out of luck. |
13:42.09 | dcordes_ | can't we trace from the spl :D |
13:42.35 | cr2_ | and if these bits are no "reversable", they the only chance is to reflash the spl. |
13:43.00 | dcordes_ | how can we try if they are? |
13:43.06 | cr2_ | ok, but wince has some mechanism to tell ARM9 it wants the clock being turned on |
13:43.14 | dcordes_ | first how can we identify this mechanism |
13:43.22 | cr2_ | on android 7200A it's their rpc |
13:43.41 | dcordes_ | cr2_: don't I already have clock on the sd?? |
13:43.54 | cr2_ | swetland says it's different on 7200, but it must be similar to it. |
13:45.37 | Banji | hi again dcordes_ i put my sim card in my pocket pc and the linux is bloqued "crash" is normal? |
13:45.57 | *** join/#htc-linux exco (n=excogita@e181090047.adsl.alicedsl.de) |
13:46.09 | dcordes_ | Banji: yes reboot |
13:48.46 | Banji | i cant reboot because the button of the power is crah |
13:48.49 | Banji | crash |
13:49.06 | Banji | i can take off the battery |
13:49.10 | Banji | ? |
13:49.11 | dcordes_ | rip it off |
13:49.44 | excogitation | Magician anyone? |
13:49.47 | Banji | ok i rip it off |
13:49.57 | Banji | and now loader the linux again? |
13:51.24 | dcordes_ | cr2_: maybe wince disassemblies can help us to identify those "magic bits" which prevent us from accesing registers? |
13:52.40 | Banji | and now loader the linux again? |
13:52.46 | dcordes_ | yes |
13:53.03 | Banji | ok i´m loader the linux again |
13:53.14 | Banji | and now he don´t crash? |
13:53.21 | dcordes_ | no idea |
13:53.22 | dcordes_ | try |
13:53.39 | Banji | ok |
13:55.15 | cr2_ | dcordes_: this list shows which areas are dual, and which are not -> http://wiki.xda-developers.com/index.php?pagename=KaiserMemoryMap |
13:55.54 | cr2_ | the first column shows arm9 map, the second column the arm11 map |
13:56.01 | dcordes_ | :) |
13:56.43 | *** join/#htc-linux CyberScript32_ (n=ontv@87-196-69-187.net.novis.pt) |
13:56.45 | cr2_ | SHARED_MEM A9/A11 is not mapped by arm9 btw, so it may be wrong for kaiser |
13:56.57 | CyberScript32_ | dcorders he don´t crash, but if i don´t touch in the pocket pc for one minute he crash |
13:57.20 | banji_dk1 | dcorders he don´t crash, but if i don´t touch in the pocket pc for one minute he crash |
13:57.48 | dcordes_ | cr2_: in case that is mapped differently by our amss? |
13:58.00 | banji_dk1 | if i touch in pocket pc he don´t never crash |
13:58.19 | banji_dk1 | if i don´t touch with one minute is crash |
13:58.22 | dcordes_ | banji_dk1: does it turn white or black? |
13:58.59 | banji_dk1 | the light is on |
13:59.22 | banji_dk1 | the light of screen is on |
14:00.20 | dcordes_ | banji_dk1: go to settings, power management and turn off sleep, display dim. |
14:00.54 | banji_dk1 | in the linux mode or in the windows mode? |
14:01.00 | dcordes_ | in the linux mode |
14:01.08 | dcordes_ | aka in qtopia |
14:01.39 | banji_dk1 | [ImCoKeMaN]: but if i turn off sleep, the pocket pc is ever turn on |
14:02.11 | banji_dk1 | but if i turn off sleep, the pocket pc is ever turn on |
14:02.20 | dcordes_ | banji_dk1: open a terminal in linux mode and type "cat /proc/cpuinfo" |
14:02.42 | banji_dk1 | when i put this text? |
14:02.51 | dcordes_ | banji_dk1: in linux mode |
14:03.01 | banji_dk1 | ok but where? |
14:03.08 | dcordes_ | go to applications. there is qterminal |
14:03.12 | banji_dk1 | ok |
14:03.26 | banji_dk1 | i´m loading the linx mode |
14:05.18 | dcordes_ | cr2_: why is a9 not in charge of MSM_GPIO1/2. looking at my conversation with swetland, it sounds like arm9 controls the gpio |
14:07.04 | banji_dk1 | i need to restart he crash wait a moment ok |
14:09.49 | cr2_ | dcordes_: these areas are "virtual gpios" . swetland said before that the "real" gpios are controlled by arm9. but we don't know where they are. yet |
14:11.19 | dcordes_ | cr2_: you've got the code |
14:13.31 | banji_dk1 | dcordes_what do you know? |
14:14.04 | banji_dk1 | of the cpuinfo |
14:14.14 | banji_dk1 | .... |
14:14.22 | *** join/#htc-linux diogene31_ (n=rj@mur31-2-82-243-122-54.fbx.proxad.net) |
14:14.53 | banji_dk1 | processor:xscale-pxa255 rev |
14:15.11 | banji_dk1 | bogomips:397.28 |
14:15.16 | dcordes_ | rev |
14:15.26 | dcordes_ | boardid |
14:15.49 | banji_dk1 | 6(v5l) |
14:16.10 | banji_dk1 | dont have boardid |
14:16.51 | banji_dk1 | processor:xscale-pxa255 rev 6(v5l) |
14:17.11 | banji_dk1 | hardware: htc blueangel |
14:17.19 | banji_dk1 | revision:0005 |
14:17.26 | banji_dk1 | serial:00000000000000000 |
14:19.02 | dcordes_ | banji_dk1: "revision:0005" you are lucky. you have the good "boardid 5". that means your htc blueangel supports sleep. |
14:19.23 | banji_dk1 | he are crash again :s and know i don´t know why?:s |
14:19.28 | dcordes_ | cr2_: can't we simply look up where the real gpios are in the pistachio code? |
14:19.46 | dcordes_ | banji_dk1: I told you to turn off display dim. |
14:19.50 | dcordes_ | it does not crash. |
14:20.00 | banji_dk1 | i turn off display dim |
14:20.06 | dcordes_ | if it really crashs I don't know sorry |
14:20.07 | banji_dk1 | but her are crash again |
14:20.31 | banji_dk1 | he is normal to crash? |
14:20.34 | dcordes_ | no |
14:20.37 | dcordes_ | works fine here. |
14:21.08 | banji_dk1 | i have badexperience with everything with linux :( |
14:21.14 | banji_dk1 | i like linux |
14:21.27 | banji_dk1 | but i have bad experiences of this |
14:21.31 | dcordes_ | get it on your computer. easier to learn. |
14:21.49 | banji_dk1 | i have a laptop and i can´t install linux |
14:22.11 | dcordes_ | look for help in distro wiki, forum or irc channel |
14:22.14 | banji_dk1 | the ubuntu,the opensource,the fedora |
14:22.33 | banji_dk1 | don´t work in my laptop , they crash when the kernel loading |
14:22.46 | banji_dk1 | i put linux in virtual machine |
14:22.49 | banji_dk1 | :s |
14:23.02 | dcordes_ | uncool |
14:23.03 | banji_dk1 | in my pocket pc the linux install but crash |
14:23.27 | banji_dk1 | in my machines linux never don´t work, i don´t choose machines |
14:23.44 | dcordes_ | banji_dk1: without SIM it does not crash? |
14:23.46 | banji_dk1 | or i never choose machines like linux |
14:23.53 | dcordes_ | :) |
14:24.19 | banji_dk1 | yes without sim does not crash |
14:24.50 | dcordes_ | what happens when it does crash? |
14:25.55 | banji_dk1 | they simply crash i go to the contacts or everything and don´t work, the clock stop, the lights are turn on, the keyboard light if i press one key turn on and off |
14:26.15 | banji_dk1 | pocket pc is stop to responding for everything |
14:26.20 | banji_dk1 | simply crash |
14:26.35 | dcordes_ | ok maybe BabelO knows something |
14:27.11 | banji_dk1 | and i need to rip off batery to restart pocket pc |
14:28.12 | banji_dk1 | BabelO:do you knowm what happens? |
14:29.04 | cr2_ | dcordes_: which lcd panel you have on kaiser ? |
14:29.31 | dcordes_ | cr2_: I can't recall. toppoly? |
14:29.58 | dcordes_ | if there is only one kaiser lcd it is toppoly |
14:30.20 | cr2_ | ok |
14:30.22 | *** join/#htc-linux lamikr (n=lamikr@dyn60-31.dsl.spy.dnainternet.fi) |
14:31.20 | banji_dk1 | why linux don´t like me :( |
14:31.20 | banji_dk1 | i like the linux |
14:31.20 | cr2_ | dcordes_: do the mddi_tx and mddi_rx work ? |
14:31.20 | cr2_ | lamikr: hi |
14:31.47 | lamikr | cr2_ Hi! |
14:32.52 | cr2_ | lamikr: how are you doing ? didn't hear from you for a long time |
14:34.31 | dcordes_ | cr2_: isn't mddi the lcd processor? Display works |
14:35.13 | lamikr | cr2_ I am fine, and started playing again :-) I invested for new AMD 780G motherboard and x86-64 cpu and I have now build out the first images for a long time. Needed to make some fixes to old build scripts to get things build with new version of everything... |
14:35.15 | cr2_ | dcordes_: get_id, and panel setup. after wince, and after resume |
14:35.37 | cr2_ | lamikr: ok. |
14:36.33 | dcordes_ | cr2_: sorry did not understand |
14:36.58 | lamikr | cr2_ I am actually now mostly interested of investigating the olpc project... I already made the default build and tested it with qemu. I am interesred in getting it running on out phones :-) |
14:37.21 | cr2_ | lamikr: i thought you have an olpc :) |
14:37.56 | cr2_ | dcordes_: we need full hw support |
14:38.26 | lamikr | cr2_ h6300 and the xda II are almost like that one :-) |
14:40.09 | dcordes_ | cr2_: any action required? |
14:40.23 | cr2_ | lamikr: i guess it's time to check the hima kernel again. |
14:42.11 | lamikr | cr2_ I can try to do it by myself in the beginning once I get the toolchain for XDA ready. I will bother you if I get stuck with the kernel. |
14:45.00 | cr2_ | lamikr: get the toolchain from qtopia |
14:47.42 | cr2_ | lamikr: Cross Toolchain here http://www.qtopia.net/modules/devices/openmoko.php |
14:47.53 | dcordes_ | cr2_: can we get usb going in the current situation? |
14:47.55 | banji_dk1 | dcordes do you know some peron to help me |
14:48.01 | banji_dk1 | person* |
14:48.07 | dcordes_ | banji_dk1: as I said maybe BabelO |
14:48.50 | banji_dk1 | but he dont talk |
14:53.39 | cr2_ | dcordes_: it seems to me that it's easier to make wifi work than usb ;) |
14:54.54 | dcordes_ | cr2_: how's that? you see sd registers in pistachio? |
14:56.22 | cr2_ | dcordes_: do you have htc-egpio driver working on kaiser ? |
14:57.52 | dcordes_ | cr2_: no where can I get it? |
14:58.40 | cr2_ | for 2.6.25 ? |
14:58.51 | dcordes_ | ah in linux.. |
14:59.12 | cr2_ | i'm slowly getting tired of all the kernel version mess that we have ;) |
14:59.43 | dcordes_ | android is 24 now |
15:00.12 | cr2_ | dcordes_: you are not going to move forward with SD without sd power gpio. the same as for wizard. |
15:00.13 | dcordes_ | where can I grab the source? |
15:00.22 | cr2_ | dcordes_: and this gpio is on CPLD. |
15:00.50 | cr2_ | the working source is in hh.org |
15:01.16 | cr2_ | but pH5 did something for 2.6.24+ but i don't know its status, or where to get it in finished form |
15:01.58 | dcordes_ | cr2_: is that a module? |
15:03.28 | cr2_ | yes |
15:03.53 | cr2_ | BabelO added it to the omap tree. you may ask him |
15:04.42 | dcordes_ | cr2_: I looked in web cvs but don't know where to search |
15:05.34 | cr2_ | drivers/mfd |
15:05.45 | dcordes_ | ./drivers/mfd/htc-egpio.c |
15:06.00 | dcordes_ | have ba tree local |
15:07.14 | dcordes_ | cr2_: of course I have not the bit of an idea how to compile that against msm git x_x |
15:08.38 | cr2_ | dcordes_: does halibut have a cpld ? vogue ? |
15:11.20 | dcordes_ | cr2_: good question |
15:11.38 | dcordes_ | don't know where I would look that up. I'll google |
15:12.53 | *** join/#htc-linux patoche (n=patoche@lns-bzn-56-82-255-206-113.adsl.proxad.net) |
15:13.28 | dcordes_ | cr2_: you asked dzo in the xda wiki and he didn't knpw |
15:13.32 | dcordes_ | s/wiki/forums/ |
15:14.06 | dcordes_ | http://forum.xda-developers.com/showthread.php?p=2085452 |
15:18.13 | dcordes_ | swetland: do you guys have a cpld on halibut? |
15:23.43 | cr2_ | dcordes_: hm. |
15:28.31 | dcordes_ | cr2_: you know halibut manuf? |
15:30.47 | cr2_ | it's an android development board |
15:30.58 | cr2_ | a big desktop box := |
15:31.16 | dcordes_ | I can't find any specs |
15:37.31 | cr2_ | dcordes_: i've created a big mess here http://wiki.xda-developers.com/index.php?pagename=KaiserMemoryMap |
15:38.10 | cr2_ | dcordes_: by adding 1 more column. the +0x register offsets are moved left now ;) |
15:39.41 | dcordes_ | cr2_: I'm confused. what's virt now? |
15:43.42 | cr2_ | spl mapping |
15:44.06 | dcordes_ | ok |
15:44.08 | cr2_ | we have 3 different mappings for physical address spaces |
15:53.19 | dcordes_ | cr2_: are the registers you wrote down all spl does access? |
15:55.28 | dcordes_ | BabelO: can you give me the source of htc-egpios module? |
15:55.58 | BabelO | dcordes_: driver/mfd/htc-egpio.c in hh kernel |
15:57.36 | dcordes_ | BabelO: I just copy that to msm git? |
15:58.59 | BabelO | dcordes_: yes, then add it to Kconfig and Makefile |
15:59.04 | dcordes_ | ok |
15:59.15 | BabelO | , after that you need to add the definition into board file |
15:59.45 | dcordes_ | BabelO: which board in hh uses egpio? (so I can look up) |
16:01.02 | BabelO | dcordes_: looxc550 |
16:01.33 | dcordes_ | thx |
16:03.02 | dcordes_ | BabelO: hm where in hh Makefile can I find egpio entry? msm already had mfd folder |
16:04.07 | BabelO | dcordes_: in drivers/mfd |
16:05.05 | Kevin2 | Hi |
16:05.22 | BabelO | hi Kevin2 |
16:06.10 | Kevin2 | Hi BabelO. |
16:07.07 | Kevin2 | Seems like lots of work on msm chips recently. |
16:07.43 | dcordes_ | BabelO: shit board init looks too complicated |
16:08.23 | dcordes_ | Kevin2: hello. did you have any idea on msm unreliable big ramdisks? |
16:08.25 | Kevin2 | bigdissaved: That's great! Can you put the pics on the web somewhere and send a link? |
16:08.30 | *** join/#htc-linux exco (n=excogita@e181080214.adsl.alicedsl.de) |
16:09.13 | Kevin2 | dcordes_: No. Is it always with large ramdisks? (No ramdisk boots reliably and small ones do too?) |
16:09.53 | dcordes_ | Kevin2: I can say with a smaller initrd, it works much more frequently |
16:10.20 | dcordes_ | if not always if < 10M |
16:10.26 | Kevin2 | Could be a dma issue. |
16:10.46 | Kevin2 | You're booting with ramdisk greater than ten megs? |
16:11.04 | dcordes_ | I have a funny ritual that helps sometimes: I copy the ramdisk from SD to NAND overwriting the old one before booting. |
16:11.41 | dcordes_ | Kevin2: I have an angstrom-console image eg, 12M that works almost always |
16:11.49 | dcordes_ | an X11 image with 20 didn't boot one time |
16:12.04 | dcordes_ | I should at least see the bootmessages in case the images i broken in a way |
16:12.11 | dcordes_ | but I think there is some memory problem |
16:12.54 | cr2_ | Kevin2: the htc-egpio is going to be used on omap850 and msm7x00 |
16:13.30 | Kevin2 | dcordes_: Even a large image sometimes boots? (So, it seems random as to when it does / does not boot?) |
16:13.33 | BabelO | cr2_: but no effect on artemis :( cr2_ i look at cpld_77 function |
16:13.50 | Kevin2 | cr2_: Cool. Have you seen pH5's changes to it? He's been trying to get it into mainline. |
16:14.14 | cr2_ | Kevin2: yes, but i don't know the outcome |
16:14.47 | cr2_ | Kevin2: btw, where is the hermes spi keyboard in the linuxtogo git ? |
16:14.57 | dcordes_ | can somebody help me fiddle the htc-egpio init in this board.c http://rafb.net/p/vUF0gt27.html ? |
16:17.57 | Kevin2 | cr2_: arch/arm/mach-s3c2442/htchermes/htchermes.c and drivers/input/keyboard/htc-spi-kbd.c |
16:18.47 | Kevin2 | dcordes_: Do you have a serial console on your machine? Do you rely solely on framebuffer for debug messages? |
16:21.04 | dcordes_ | Kevin2: I think there is a serial port on board |
16:21.08 | cr2_ | Kevin2: ok, i'll check |
16:21.26 | cr2_ | dcordes_: there are 3 serial ports, but are they connected ?? |
16:21.27 | dcordes_ | Kevin2: maybe I could solder one cable on it |
16:21.47 | dcordes_ | cr2_: wait I think laf0rge posted something about that in his blog |
16:21.56 | cr2_ | dcordes_: link ? |
16:22.00 | dcordes_ | looking |
16:22.20 | dcordes_ | he could give us his pcb photos |
16:22.34 | cr2_ | he gave me the pics already |
16:22.43 | dcordes_ | http://laforge.gnumonks.org/weblog/ |
16:23.09 | dcordes_ | cr2_: did he gave them to you under special proprietary hack liscense or can you pass me them on? |
16:24.10 | dcordes_ | I've already managed to get into the bootloader, which apparently offers a standard USB serial emulation that can be accessed even from a Linux PC. |
16:24.20 | BabelO | Kevin2:what do i need on egpio ? i think i miss something when i put it for omap |
16:24.26 | dcordes_ | I think that is nothing new to you. though he was mantioning a 'real' serial interface |
16:24.30 | cr2_ | dcordes_: 150MB i think |
16:24.37 | BabelO | Kevin2: i never see egpio set: reg %d = 0x%04x\n message |
16:24.45 | *** join/#htc-linux goxboxlive (n=goxboxli@208.84-48-176.nextgentel.com) |
16:24.47 | cr2_ | dcordes_: i think these are really hires pictures |
16:25.31 | dcordes_ | can you upload them somewhere? I'm really interested in how that beast looks from inside |
16:25.46 | Kevin2 | BabelO: I don't know - what does your code look like? |
16:26.20 | Kevin2 | dcordes_: Where is the source for the kernel you are using? |
16:27.24 | BabelO | Kevin2: i just use your htc_egpio driver, add the gpio_desc to gpio.c in omap arch and then do this declaration according to cr2_ http://pastebin.com/m2bc27f5c |
16:27.26 | dcordes_ | Kevin2: http://git.android.com/?p=kernel.git;a=summary;h=android-msm |
16:28.09 | dcordes_ | Kevin2: kaiser vogue friends only boot with this patch http://baliniak.pl/android/kernel.diff there are several revisions of the patching. this is one which I use currently |
16:28.55 | dcordes_ | cr2_: how can we determine whether or not there is a serial console connected? |
16:31.35 | cr2_ | dcordes_: where are the kaiser links there ? |
16:31.36 | dcordes_ | is going out, lets his kaiser get some fresh air |
16:31.38 | dcordes_ | bbl |
16:31.57 | banji_dk1 | hi everybody |
16:32.05 | *** join/#htc-linux AlGe (n=alge@chello080109231226.4.uni-klu.teleweb.at) |
16:32.29 | banji_dk1 | BabelO hi have a problem, my pocket pc with the sim card is crash |
16:32.33 | Kevin2 | BabelO: I hope you're adding &htcartemis_egpio to your main platform device list somewhere later in the file. |
16:32.40 | banji_dk1 | did you konw what happen? |
16:33.42 | BabelO | Kevin2: yes yes ;) i did not forgot that i see this message EGPIO phys=%08x virt=%p\n |
16:33.58 | BabelO | banji_dk1: maybe a bad radio rom |
16:34.15 | BabelO | banji_dk1: i can not help you with that |
16:34.27 | banji_dk1 | ok thanks :s |
16:35.02 | BabelO | banji_dk1: i use the 1.06.02 version |
16:35.26 | banji_dk1 | ok thanks |
16:35.37 | cr2_ | Kevin2: no htc-spi-keyboard.c here http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=tree;f=drivers/input/keyboard;h=5a8b0a2abbf34fb0889e4eaedc3222a277abc671;hb=HEAD |
16:35.44 | banji_dk1 | your pocket pc is a qtek9090 |
16:36.06 | BabelO | banji_dk1: yes spvm2000 which is a qtek9090 too |
16:36.15 | banji_dk1 | [BabelO]: ok |
16:36.31 | banji_dk1 | where i can download this version? |
16:37.05 | BabelO | banji_dk1: look at xda-developers website |
16:37.27 | BabelO | banji_dk1: http://wiki.xda-developers.com/index.php?pagename=HTC_Blueangel |
16:37.45 | banji_dk1 | [BabelO]: ok very thanks |
16:38.02 | Kevin2 | cr2_: I did all my changes on the "work" branch - see http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=blob;f=drivers/input/keyboard/htc-spi-kbd.c;h=cdcdf9b0867b8f295d98c584006fd542c5573eca;hb=refs/heads/work |
16:40.19 | cr2_ | Kevin2: ok :) does it mean that apache and hermes will run with 2.6.25 kernel ? |
16:41.39 | Kevin2 | cr2_: I haven't tried |
16:42.43 | *** join/#htc-linux mistadman (n=mistadma@adsl-6-113-151.msy.bellsouth.net) |
16:43.49 | mistadman | Does anyone have any experiance with using haRET with the Athena? |
16:44.18 | cr2_ | mistadman: it works, and was always working. |
16:44.57 | mistadman | When I click "Listen for network connection" it just hangs. I am using ver 0.5.1. |
16:45.07 | mistadman | I believe this is the latest version... |
16:45.54 | Kevin2 | What do you mean "it just hangs"? Does the whole machine lockup? |
16:46.32 | mistadman | Nope, the spinning hourglass(?) just keeps spinning... |
16:46.39 | *** join/#htc-linux miknix (n=miknix@bl4-81-16.dsl.telepac.pt) |
16:46.49 | cr2_ | mistadman: it's the right behaviour |
16:46.55 | mistadman | Hmmm |
16:47.12 | mistadman | Let me trying connecting again... brb |
16:47.17 | Kevin2 | mistadman: You need to run haretconsole on your main machine and connect to the phone. |
16:47.51 | cr2_ | Kevin2: 2 questions about haret now |
16:48.06 | mistadman | Does the fact that I am using Ubuntu matters? |
16:48.13 | cr2_ | mistadman: no |
16:48.32 | mistadman | ok, brb and thanks! |
16:48.45 | Kevin2 | cr2_: Okay. |
16:49.05 | Kevin2 | cr2_: BTW, I was looking through the patch you sent to haret@handhelds.org. Where did the info in the header files come from? |
16:49.10 | cr2_ | Kevin2: i've sent a patch some weeks ago |
16:49.53 | cr2_ | hmm. probably from the wince SDK, but i can grep if you need an exact reference |
16:50.23 | Kevin2 | I don't think we're supposed to redistribute sdk stuff. |
16:50.41 | cr2_ | and maybe from msdn, i can't remember now. |
16:50.58 | cr2_ | sdk headers are not included in cegcc ? |
16:51.16 | cr2_ | sdk and platform builder are 2 different things. |
16:51.40 | Kevin2 | cr2_: As far as I know, everything in cegcc is from msdn. |
16:52.06 | cr2_ | ok, i'll recheck the headers. |
16:52.36 | cr2_ | btw, not all numeric constants are described in msdn, so they must come from the SDK. |
16:52.52 | banji_dk1 | [BabelO]: but my pocket pc crash in linux mode no in windows mode |
16:53.00 | banji_dk1 | the rom is in linux ? |
16:53.44 | cr2_ | Kevin2: ok, another question. what information should be added to 'wi' so it can trace the memory accesses for mmutrace ? |
16:53.59 | BabelO | banji_dk1: it is not os depend, separate rom only for gsm radio |
16:54.21 | banji_dk1 | [BabelO]: ah ok |
16:54.56 | Kevin2 | cr2_: For stuff not in msdn (like constants), pedro from cegcc suggested doing a google code search. |
16:55.51 | Kevin2 | cr2_: I don't understand your question about mmutrace. Doesn't mmutrace do that already? |
16:57.04 | cr2_ | Kevin2: hmm. i'd check |
17:00.39 | banji_dk1 | [BabelO]: but my pockt pc only cras in linux mode |
17:01.31 | mistadman | Hey cr2, THANKS! |
17:02.45 | mistadman | It worked. My problem was that I was trying to connect via USB RNDIS. I forgot about using WIFI on the Athena |
17:04.43 | *** join/#htc-linux Banji (n=ontv@87-196-69-187.net.novis.pt) |
17:07.06 | *** join/#htc-linux banji_dk1 (n=ontv@87-196-69-187.net.novis.pt) |
17:08.46 | *** join/#htc-linux Banji (n=ontv@87-196-69-187.net.novis.pt) |
17:23.39 | *** join/#htc-linux patoche (n=root@lns-bzn-56-82-255-206-113.adsl.proxad.net) |
17:27.21 | cr2_ | BabelO: i'm wondering why do the omap people always program in machine code ;) |
17:27.24 | cr2_ | + omap_writel((omap_readl(0xfffe1080) & 0xf07ffdfe) | 0x08000000, |
17:27.24 | cr2_ | + 0xfffe1080); |
17:27.47 | BabelO | cr2_: because normal write does not work |
17:27.53 | BabelO | and i cannot explain why |
17:28.04 | BabelO | cr2_: still looking at egpio |
17:28.09 | cr2_ | i mean the constants. |
17:28.17 | Kevin2 | (12:55:58 PM) Kevin2: cr2_: I don't understand your question about mmutrace. Doesn't mmutrace do that already? |
17:28.19 | BabelO | cr2_: ah ;) |
17:28.41 | BabelO | cr2_: i copy the htcapache definition of egpio |
17:28.51 | BabelO | cr2_: it is more complete than on looks |
17:28.52 | cr2_ | where is the omap_writeX defined ? |
17:29.22 | cr2_ | loox is the most simple irq-less no.init setup. |
17:30.10 | cr2_ | Kevin2: yes, i'm booting on the atlas3 now, but it's a bit tricky. |
17:30.32 | BabelO | cr2_: #define omap_writew(v,a) (*(volatile unsigned short *)IO_ADDRESS(a) = (v)) in io.h |
17:31.36 | Kevin2 | cr2_: Okay. I'm just confused on what you were asking. |
17:40.10 | cr2_ | BabelO: and IO_ADDRESS ? |
17:41.03 | cr2_ | Kevin2: i was not sure if you need some knowledge about the interrupt controller details. |
17:41.09 | BabelO | cr2_: #define IO_ADDRESS(pa) ((pa) - IO_OFFSET) and #define IO_OFFSET 0x01000000 |
17:41.33 | cr2_ | BabelO: then you need to adjust the .base_address respectively. |
17:41.48 | BabelO | cr2_: ? for egpio ? |
17:41.51 | cr2_ | yes |
17:42.13 | cr2_ | so omap_writew is equivalent to a direct write. |
17:43.22 | cr2_ | BabelO: because htc-egpio simply uses the pointer arithmetic. |
17:43.31 | BabelO | cr2_: sorry, but i not understand how i adjust the egpio, it is too much low level ;) |
17:44.02 | cr2_ | hmm- physical address ? |
17:44.12 | cr2_ | ok, let's think physical first. |
17:44.30 | cr2_ | your base address ix 0x04000400 |
17:44.39 | BabelO | cr2_: full line is #define IO_OFFSET 0x01000000 /* Virtual IO = 0xfefb0000 */ |
17:44.51 | BabelO | cr2_: ok that is theegpio base |
17:45.24 | cr2_ | but for the direct access you need (base_address-0x01000000) |
17:45.27 | jeanseb | Hi |
17:45.46 | jeanseb | BabelO, is there news about blue angel ? |
17:45.46 | cr2_ | BabelO: the problem is that you need to adjust the virtual address here. |
17:46.12 | cr2_ | BabelO: does htc-egpio ioremap() the base address ? |
17:46.35 | BabelO | cr2_: i check |
17:46.50 | BabelO | jeanseb: hi, no news, i m working on my artemis |
17:47.12 | jeanseb | ok |
17:47.14 | cr2_ | BabelO: it seems we need an .io_offset parameter to htc-egpio struct |
17:47.18 | Kevin2 | cr2_: The only thing you need to use 'wi' is an arm machine with AllocPhysMem and the correct cache flush. (Of course, you wont get any useful irq info unless you also populate the IRQS variable.) We setup the cache flush in arch-centrality, so you shouldn't have an issue. There is nothing arch specific to mmutrace, so it should work whenever 'wi' works. |
17:47.40 | cr2_ | Kevin2: ok, testing it now. |
17:47.42 | BabelO | cr2_: yes ei->addrBase = (u16 *)ioremap_nocache(res->start, res->end - res->start); |
17:49.30 | cr2_ | BabelO: ok, i'm a bit confused now. |
17:50.34 | cr2_ | check where are the physical and where are the virtual addresses, and where do you need to subtract 0x01000000 |
17:50.39 | BabelO | cr2_: me too, i don't understand, it seem with 0x4000400 we have the correct adress, because like you say it clear all |
17:51.10 | cr2_ | BabelO: only DMA works on physical addresses directly |
17:51.26 | BabelO | cr2_: http://pastebin.com/m5a03254d |
17:51.43 | dcordes_ | back |
17:51.44 | cr2_ | therefore you always need a virtual mapping. i.e. ioremap |
17:52.30 | dcordes_ | cr2_: can't you take from the photos if there is a serial port? |
17:53.14 | cr2_ | dcordes_: of course not. there are some internal connector pins of test pads. |
17:53.45 | cr2_ | dcordes_: where have you seen the serial data in the laf0rge blog ? |
17:53.56 | cr2_ | BabelO: how is the ioremap done on omap ? |
17:54.54 | dcordes_ | cr2_: I pasted the line above. but that is only for the bootloader. (usb>serial" |
17:55.09 | BabelO | cr2_: how can i check ?the ioremap function is common to arm/mm |
17:55.19 | dcordes_ | 18:24 < dcordes_> I've already managed to get into the bootloader, which apparently offers a standard USB serial emulation that can be accessed even from a Linux PC. |
17:55.49 | cr2_ | dcordes_: usb serial != serial |
17:56.12 | dcordes_ | yes I see it's useless since it is for the bootloader |
17:56.21 | cr2_ | BabelO: hmm. wait, i'd like to finish with the atlas3. |
17:56.30 | dcordes_ | I just thought the blog entry had something about real console |
17:57.05 | cr2_ | dcordes_: htc used usb, 3 serial ports, and a special debug board for development. |
17:57.07 | BabelO | cr2_: ok, i wait, i look more to egpio sourceq |
17:57.13 | dcordes_ | cr2_: in order to get egpio going, all me tree lacks is init in board.c. can you help me? |
17:57.29 | cr2_ | dcordes_: but it does not mean that your kaiser has all these things soldered and connected. |
17:58.01 | cr2_ | dcordes_: look how it's done on looxc550.c |
17:58.18 | dcordes_ | yes I did but that looks very specific for the loox with registers and all |
17:58.30 | dcordes_ | have both board.c in front of me |
17:58.52 | cr2_ | dcordes_: but it's exactly the same setup. |
17:59.06 | cr2_ | only the bit purpose differs and the base address |
17:59.23 | *** join/#htc-linux p3t3r__ (n=p3t3r__@wlan.net-you.de) |
18:00.45 | dcordes_ | bit purpose? |
18:01.33 | cr2_ | Kevin2: Handled 522 irq, 7352 abort, 641 prefetch, 0 lost, 0 errors |
18:01.45 | cr2_ | Kevin2: but nothing on the screen. |
18:02.02 | cr2_ | dcordes_: which bit means what. |
18:02.18 | cr2_ | dcordes_: which bit in which bank. |
18:02.48 | *** part/#htc-linux exco (n=excogita@e181080214.adsl.alicedsl.de) |
18:07.25 | cr2_ | this one looks better :) |
18:07.27 | cr2_ | 000086: mmutrace 03a2179c: e5933468(ldr) ab090468 00000080 (00000000) |
18:08.12 | cr2_ | hmm. it seems i have a bug in my list. |
18:08.32 | *** join/#htc-linux surgex0 (i=surge@pool-71-186-234-216.bflony.fios.verizon.net) |
18:09.02 | cr2_ | of maybe not. |
18:09.06 | cr2_ | HaRET(10)# addr2mod 0x03a2179c |
18:09.06 | cr2_ | Address 03a2179c not process specific |
18:09.06 | cr2_ | <PROTECTED> |
18:11.42 | *** join/#htc-linux zwox_ (i=zwox@par69-7-82-243-234-94.fbx.proxad.net) |
18:11.53 | dcordes_ | cr2_: can you look? http://rafb.net/p/LP15rg42.html |
18:12.21 | dcordes_ | infernix: ping |
18:12.50 | infernix | pong |
18:13.11 | dcordes_ | infernix: did you get SD running now w/ Raven_coda's patch? (I wanna do a clean rebase) |
18:13.16 | infernix | yah |
18:13.20 | dcordes_ | in order to sync with you guys |
18:13.22 | infernix | irq errors and all |
18:13.28 | dcordes_ | what did you do from clear pull+patch? |
18:13.57 | cr2_ | dcordes_: PXA_CS4_PHYS ??? |
18:14.08 | infernix | msm_comm thingie |
18:14.13 | cr2_ | put the kaiser CPLD base address here. |
18:14.16 | dcordes_ | I just copied the part |
18:14.27 | dcordes_ | ok |
18:15.01 | cr2_ | N560_EGPIO_BASE |
18:15.03 | infernix | so checkout android-msm, patch it with ravens patch, change msm_comm |
18:15.05 | dcordes_ | memory map looks so ill now :D |
18:15.13 | cr2_ | this should be defined somewhere. |
18:15.21 | infernix | and optionally you can edit drivers/mmc/msm-sdcc.c and turn on some more verbosity |
18:15.41 | cr2_ | .nrRegs = 7 |
18:15.47 | cr2_ | how much does kaiser have ? |
18:15.50 | dcordes_ | cr2_: why should N560 be defined in my tree? |
18:15.59 | dcordes_ | cr2_: cpld banks? |
18:16.11 | cr2_ | yes |
18:16.37 | cr2_ | <PROTECTED> |
18:16.53 | cr2_ | you need to define where your cpld gpios start |
18:16.55 | dcordes_ | cr2_: 8 banks |
18:17.06 | cr2_ | i have no idea how it is done on msm |
18:17.38 | cr2_ | for 8 banks you need .nrRegs = 8, |
18:17.42 | dcordes_ | cr2_: I have no cpld stuffs in the tree |
18:17.49 | dcordes_ | yes did the 8 regs line |
18:17.58 | cr2_ | .end = PXA_CS4_PHYS + 0x2*8, /* 7 regs */ |
18:18.02 | cr2_ | and 9 here. |
18:18.14 | cr2_ | hm. probably. |
18:18.33 | cr2_ | maybe just my copypaste bug :) |
18:18.47 | cr2_ | i think i've copied it from athena |
18:18.55 | dcordes_ | what? |
18:19.07 | cr2_ | this line |
18:19.58 | dcordes_ | I changed .end = PXA_CS4_PHYS + 0x2*9, /* 8 regs */ ; .nrRegs = 8, so far |
18:20.19 | cr2_ | ok. |
18:20.25 | cr2_ | does not hurt anybody. |
18:21.45 | dcordes_ | " .gpio_base = 0x98," ? |
18:22.15 | cr2_ | msm7200 has 105 gpios |
18:22.33 | cr2_ | but there may be some offset, i don't know. |
18:23.03 | cr2_ | i'm not sure how the generic gpio api works on msm |
18:24.41 | cr2_ | gpio_interrupt0, then gpio4. looks good. |
18:25.13 | dcordes_ | cr2_: can't follow |
18:26.38 | cr2_ | i do 3 things at once. |
18:26.40 | Kevin2 | cr2_: Right. You wont see any irq reports during "wi" if you don't have anything in the IRQS variable. |
18:27.07 | Kevin2 | The IRQS variable is just like the GPIOS variable - it just gets looked at during each irq. |
18:27.08 | cr2_ | Kevin2: ok, now i picked the TS gpio. |
18:27.37 | dcordes_ | infernix: :) did you do any further x11 attempts? |
18:27.45 | infernix | yeah |
18:27.53 | infernix | Xfbdev works fine |
18:28.07 | infernix | but I didnt get a working mouse cursor |
18:28.08 | dcordes_ | infernix: what's image size? could you email? |
18:28.09 | cr2_ | Kevin2: nonmod address. |
18:28.12 | cr2_ | HaRET(19)# addr2mod 0x8c6beedc |
18:28.12 | cr2_ | Address 8c6beedc not process specific |
18:28.14 | dcordes_ | might help with mouse |
18:28.31 | infernix | and x11-image boots but still gives a black screen |
18:28.43 | dcordes_ | doesn't boot here |
18:28.46 | infernix | so i played with the Xserver script |
18:28.48 | infernix | yeah i stripped it |
18:28.56 | cr2_ | Kevin2: it's probably inside the kernel -> 021626: mmutrace 8c6beedc: e5931000(ldr) ab09008c 00000010 (00000000) |
18:28.57 | infernix | i suspect that, when the initrd gets unpacked, there's not enough ram |
18:29.04 | infernix | and that makes larger images fail |
18:29.06 | dcordes_ | good point |
18:29.22 | infernix | you can have a 1GB file with all zeroes packed in a 4mb initrd |
18:29.26 | infernix | but it'll fail completely when unpacked |
18:29.39 | dcordes_ | yea |
18:29.49 | infernix | whereas a 30mb initrd with mostly non-compressable binaries will just work |
18:29.57 | infernix | since it could only be 35mb uncompressed |
18:30.02 | dcordes_ | got it |
18:30.28 | dcordes_ | infernix: you can also strip from OE level by removing tasks from the metadata in packages/images/ |
18:30.28 | infernix | so i ripped out ipkg info and a lot of terminfo from the x11-image, and it'll boot. but no working X at bootup |
18:30.47 | infernix | nah, nothing much to strip from an x11-image |
18:30.55 | infernix | in OE that is |
18:31.27 | dcordes_ | infernix: gpe-config is in x11-image. that is huge. |
18:31.27 | infernix | well you need it :) |
18:31.29 | infernix | you need a display manager |
18:31.30 | dcordes_ | what for? |
18:31.43 | dcordes_ | gpe-config are useless guis eg user manager ;) |
18:32.09 | infernix | oh |
18:32.28 | dcordes_ | but might also include ts calibration but that might be in the gpe-base metapackage |
18:32.33 | infernix | either way, i manged to start Xfbdev just fine |
18:32.41 | dcordes_ | how did you do it? |
18:32.57 | infernix | boot stripped x11 image, rdinit=/bin/sh |
18:33.05 | infernix | then just /usr/bin/Xfvdev -videoTest |
18:33.22 | dcordes_ | so you fall into console after boot? |
18:33.32 | infernix | yup |
18:33.45 | dcordes_ | did you try to play a video :) |
18:33.53 | *** part/#htc-linux p3t3r__ (n=p3t3r__@wlan.net-you.de) |
18:34.06 | infernix | Xfbdev doesnt give you any window manager |
18:34.08 | infernix | so you can't do shit |
18:34.20 | dcordes_ | mplayer with fbdev? (would that work?) |
18:34.42 | infernix | in theory. but where do you leave the video file? |
18:35.02 | dcordes_ | small 2mb into rootfs |
18:35.16 | dcordes_ | wanted to try that all day. will do with modified x11 |
18:35.27 | cr2_ | sdmlc. |
18:35.29 | infernix | alright, i'll try to build an x11 image with mplayer |
18:35.47 | infernix | but i'd much rather have sd working. i tried to figure out more about gpios but its way over my head |
18:36.09 | dcordes_ | echo "ANGSTROM_EXTRA_INSTALL += "mplayer"" >> conf/local.conf |
18:36.41 | dcordes_ | just for motivation :) |
18:37.36 | infernix | a sec while i reconfigure my nfs |
18:40.18 | *** join/#htc-linux patoche (n=pat@lns-bzn-56-82-255-206-113.adsl.proxad.net) |
18:44.17 | dcordes_ | rebasing tree so I can cleanly put egpio stuff in |
18:49.40 | dcordes_ | infernix: can you put Raven_coda patch unextracted to your server? |
18:49.59 | infernix | sure, a sec |
18:50.06 | dcordes_ | it's always shity to point people to a winzip file ina forum when they ask. |
18:50.20 | dcordes_ | thx |
18:50.35 | dcordes_ | Raven_coda: any news? |
18:50.36 | infernix | its up there |
18:50.40 | infernix | dx.infernix.net/kaiser/ |
18:53.46 | dcordes_ | lol my OE is screwed again |
18:54.11 | dcordes_ | have to do things in the kernel anyway now and XorA has the kaiser metadata |
18:55.18 | BabelO | cr2_: :( |
18:56.50 | dcordes_ | infernix: we have board-htckaiser.c but board-kaiser.keypad.c |
18:57.09 | *** join/#htc-linux lamikr (n=lamikr@dyn60-31.dsl.spy.dnainternet.fi) |
18:58.22 | dcordes_ | infernix: shouldn't that be board-htckaiser-keypad.c? |
18:58.42 | infernix | probably |
19:00.00 | cr2_ | Kevin2: base=03990000 size=00007000 |
19:00.12 | cr2_ | HaRET(2)# addr2mod 0x039a3328 |
19:00.12 | cr2_ | Address 039a3328 not process specific |
19:00.29 | cr2_ | hmm. that's true. |
19:02.40 | dcordes_ | cr2_: does egpio depend on anything btw? only have the htc-egpio.c |
19:19.14 | dcordes_ | infernix: we have to do some fancy stuff before we can access sd power gpio |
19:19.24 | dcordes_ | which is provided by kaiser CPLD |
19:22.12 | *** join/#htc-linux ltxda (n=ltxda@unaffiliated/ltxda) |
19:23.44 | *** join/#htc-linux patoche (n=pat@lns-bzn-56-82-255-206-113.adsl.proxad.net) |
19:28.04 | BabelO | cr2_: long time i did not work with spi stuff ;) i forgot to push dummy data to tx buffer to get the previous rx data :) |
19:30.00 | *** join/#htc-linux lamikr_ (n=lamikr@dyn60-31.dsl.spy.dnainternet.fi) |
19:46.31 | *** join/#htc-linux AlGe (n=alge@chello080109231226.4.uni-klu.teleweb.at) |
19:47.20 | BabelO | Kevin2: still around ? |
19:48.41 | *** join/#htc-linux rmoravcik (n=rmoravci@ip-89-102-255-171.karneval.cz) |
20:18.14 | *** join/#htc-linux hollo_ (n=hollo@3e6b025d.rev.stofanet.dk) |
20:21.48 | *** join/#htc-linux pikapika (n=pikapika@mar75-8-88-164-227-147.fbx.proxad.net) |
20:21.59 | pikapika | hi |
20:32.46 | *** join/#htc-linux patoche_ (n=pat@lns-bzn-52-82-65-104-219.adsl.proxad.net) |
20:33.56 | BabelO | hi pikapika |
20:34.34 | BabelO | cr2_: did you remember if when you send D300 do tsc2046 i reply 0 if no touch on screen ? |
20:35.53 | *** join/#htc-linux dcordes (n=lukas@f049175238.adsl.alicedsl.de) |
20:52.19 | cr2_ | BabelO: but you should send it only after the touch irq |
20:52.55 | BabelO | cr2_: that is by big misunderstand |
20:53.43 | BabelO | cr2_: i don't understand why irq is not detected by driver.... |
20:54.07 | BabelO | cr2_: my spi driver have no irq at all |
20:55.39 | BabelO | cr2_: i think it is normal :) |
20:57.02 | BabelO | cr2_: have to debug the egpio stuff too, all peripheral need it |
21:02.05 | BabelO | cr2_: http://pastebin.com/m40e3823b this one seem correct now, on startup it get like a touch |
21:04.26 | *** join/#htc-linux zwox__ (i=zwox@par69-7-82-243-234-94.fbx.proxad.net) |
21:41.48 | *** join/#htc-linux CVirus (n=GoD@196.205.193.110) |
22:02.18 | BabelO | cr2_: that is normal egpio does not work on omap :) |
22:02.36 | BabelO | i found why now, it use gpiodev2 need to add some line |
22:14.46 | BabelO | ahhhh |
22:18.03 | *** join/#htc-linux Marex (n=marex@vasut.kolej.mff.cuni.cz) |
22:26.54 | *** join/#htc-linux zwox (i=zwox@par69-7-82-243-234-94.fbx.proxad.net) |
22:34.25 | BabelO | cr2_: yesssssss, it works now ;) |
22:35.08 | BabelO | cr2_: it was the call to gpiodev2, a .h contain code lines, i does not like that, it is in omap tree |
22:35.38 | cr2_ | hehe |
22:35.41 | cr2_ | :) |
22:36.08 | cr2_ | what is so special in gpiodev2.h ? |
22:36.15 | cr2_ | is it some hh.org hack ? |
22:37.04 | BabelO | cr2_: yes hh hack, each gpio_get_value have a if (gpio < GPIO_BASE_INCREMENT) .... else gpiodev2_get_value(gpio) |
22:37.34 | BabelO | cr2_: same for others function |
22:37.56 | cr2_ | ok. |
22:38.08 | cr2_ | so msm will also need it for htc-egpio |
22:38.08 | BabelO | cr2_: to be sure, i start the egpio stuff to GPIO_BASE_INCREMENT * 2 |
22:38.40 | BabelO | cr2_: i miss all this thing because was looking at a .c file, in omap it is in include/arch/asm/gpio.h |
22:38.42 | BabelO | so bad |
22:40.14 | cr2_ | BabelO: so does your kernel work on artemis now ? |
22:40.23 | cr2_ | an on wiazrd ? |
22:40.50 | BabelO | cr2_: my kernel works on artemis sure :) now i need to fix the spi stuff, |
22:41.11 | Pryan | wenas |
22:41.14 | BabelO | cr2_: on wizard, they use the mvista git tree |
22:41.19 | cr2_ | you need to solve this IO_OFFSET problem |
22:41.33 | BabelO | cr2_: why ? |
22:41.44 | cr2_ | i don't see any big differences between these 2. only for the cpld |
22:42.03 | cr2_ | does htc-egpio work on omap ? |
22:42.19 | cr2_ | without omap_writeX ? |
22:42.46 | cr2_ | this omap_* and #ifdef OMAP is the ultimate evil ;) |
22:43.26 | BabelO | cr2_: yes the driver works now, i set it to vibra, and vibra run |
22:43.38 | BabelO | cr2_: standard htc-gpio driver ;) |
22:43.39 | cr2_ | ok. |
22:43.57 | cr2_ | yes, you mmap the area that you really need. ok |
22:44.17 | cr2_ | so you need to sort out the LCD power now :) |
22:44.31 | cr2_ | btw, does your usb puen is setup properly ? |
22:44.36 | cr2_ | and the uarts. |
22:44.52 | BabelO | cr2_: yes it seem i miss two lcd power from cpld, because when cpld is up lcd goes black ;) |
22:45.10 | BabelO | cr2_: i think for uart i found something |
22:45.22 | cr2_ | if toppoly then lcdspi is needed :) |
22:45.31 | cr2_ | you need to setup the gpios properly. |
22:45.35 | BabelO | cr2_: :( |
22:46.07 | cr2_ | hm. and it's bit different than on others. |
22:46.23 | BabelO | cr2_: why different ? |
22:46.28 | cr2_ | maybe we will create a custom driver out of the Kevin2's BB driver |
22:46.43 | cr2_ | it's more simple on artemis |
22:47.29 | BabelO | cr2_: now i understand how spi works ;) my electronics studies was far away |
22:47.41 | BabelO | i just miss the interrupt handling for touchscreen |
22:47.53 | cr2_ | yes, but for the lcd we need a bitband sync driver. |
22:48.22 | BabelO | cr2_: yes i see Kevin2 have made a bitbang driver |
22:49.18 | cr2_ | he called it pseudo-i2c |
22:49.37 | cr2_ | we need something more spi like. |
22:50.00 | cr2_ | to account for the byteorder, and with the _cs=0 send/recv _cs=1 |
22:50.12 | cr2_ | and the multibyte transfers. |
22:50.18 | cr2_ | at least on the n560 ;) |
22:50.24 | cr2_ | and reads. |
22:50.35 | cr2_ | n560 is really special compared to the others. |
22:51.12 | cr2_ | since it's a one-shot thing at boot/resume, maybe the "real" spi driver is not really worth it. |
22:52.46 | BabelO | cr2_: why do you flags cpld_6_2 and cpld_6_1 to gps and mode ? in .lst it is related to lcd it seem |
22:52.58 | BabelO | cr2_: 8C01A26C |
22:53.48 | cr2_ | i'll check |
22:55.49 | BabelO | cr2_: ok i see why you say artemis lcdspi is more easy, it is a bit to bit write ;) |
22:57.37 | cr2_ | BabelO: it's the same on n560 |
22:57.51 | cr2_ | but with much more options |
22:58.32 | cr2_ | BabelO: yes, but what aboout gps |
22:59.23 | cr2_ | BabelO: 6 is not 7 |
22:59.59 | cr2_ | gps should be 7,0x4 |
23:00.14 | cr2_ | do you have an addr for gps power ? |
23:01.14 | BabelO | cr2_: gps power on http://pastebin.com/m841ffbc |
23:01.17 | cr2_ | hmm |
23:01.55 | cr2_ | 8c01328c |
23:02.11 | BabelO | cpld |
23:02.23 | cr2_ | 6,4,1 too |
23:02.49 | cr2_ | no. |
23:02.55 | cr2_ | lcd is 6,2 and 6,1 |
23:03.25 | cr2_ | these are bitmasks |
23:03.34 | cr2_ | i'll edit your entry in wiki |
23:03.39 | BabelO | ok |
23:05.04 | cr2_ | strh 6 may mean you have gps power on already |
23:05.32 | cr2_ | 2 is ok. |
23:05.51 | cr2_ | and i guess all these LCD gpios need an LCD type check. |
23:06.10 | cr2_ | maybe it's better to check the ddi* |
23:06.30 | cr2_ | on n560 the spl has support for 6 LCD panels or something like that. |
23:06.38 | cr2_ | but ddi has only 1. |
23:06.42 | *** join/#htc-linux MHazem (n=GoD@196.205.193.110) |
23:07.24 | cr2_ | BabelO: your vibrator entry is wrong ? |
23:08.07 | BabelO | cr2_: i use your column |
23:08.27 | cr2_ | ok. |
23:08.28 | BabelO | cr2_: 4_3 |
23:08.40 | BabelO | cr2_: yes so mine is wrong |
23:08.51 | cr2_ | i'll remove the MSB |
23:09.09 | cr2_ | i guess it's not used on all htc cplds |
23:09.16 | BabelO | ok |
23:09.58 | cr2_ | you have many stray LCD "gpios" there |
23:11.03 | BabelO | cr2_: i try to set 6_1 and 6_2 to see if lcd goes back |
23:11.19 | cr2_ | done |
23:11.22 | cr2_ | ok. |
23:11.46 | cr2_ | hmm. it seems i need to create a sed script for the 0xfff* names |
23:11.53 | cr2_ | like i do on pxa. |
23:12.24 | cr2_ | then we can finally to the right thing with the IO* setup. |
23:13.44 | BabelO | cr2_: yes 6_2 and 6_1 are lcd power, it works like a charm |
23:13.57 | cr2_ | lol |
23:14.05 | cr2_ | what about the other lcd gpios |
23:14.41 | BabelO | cr2_: it is like you do mechanics on a car, each time you finish you have some remaining piece |
23:14.51 | cr2_ | lol |
23:15.21 | BabelO | let me check gps power ;) |
23:15.39 | cr2_ | BabelO: this is from the wizard kernel. i really hate it. |
23:15.42 | cr2_ | <PROTECTED> |
23:15.42 | cr2_ | <PROTECTED> |
23:15.42 | cr2_ | <PROTECTED> |
23:15.42 | cr2_ | <PROTECTED> |
23:16.00 | cr2_ | can it be done in a sane way ? |
23:16.05 | BabelO | cr2_: yes i know, i use mux setup for artemis |
23:16.37 | BabelO | cr2_: there is a mux.c file, just need to set the config |
23:16.39 | cr2_ | because it seems htc randomly hashed the gpios ;) |
23:16.47 | cr2_ | ok. |
23:17.09 | BabelO | cr2_: so gps power is 6_4 |
23:17.15 | cr2_ | yes. |
23:17.18 | cr2_ | 4 as in mask |
23:18.55 | cr2_ | but you need a proper uart mux setup too. |
23:19.10 | BabelO | cr2_: already done PCC and others |
23:20.21 | cr2_ | ok. |
23:20.34 | cr2_ | BabelO: is it possible to use these without omap_ ? |
23:20.36 | cr2_ | omap_set_gpio_direction |
23:21.00 | BabelO | cr2_: yes now, i understand where to add the code to wrap it |
23:21.05 | cr2_ | i'll have a similar battle against asic3_ soon :) |
23:21.29 | cr2_ | then we can finally have sane portable drivers. |
23:21.31 | BabelO | cr2_: for set and get it is already done |
23:21.38 | BabelO | yes agree |
23:21.45 | cr2_ | direction should be portable too. |
23:21.56 | cr2_ | _alt is a problem right now. |
23:22.26 | BabelO | cr2_: htc-egpio is a godd driver ;) |
23:22.32 | cr2_ | yes. |
23:23.01 | BabelO | cr2_: btw, why i have some gpio related to cpld ( in cpu gpios ) ? |
23:23.12 | cr2_ | which one ? |
23:23.43 | cr2_ | lcd related / Touchscreen ? |
23:24.42 | BabelO | cr2_: for example gpio 84 |
23:25.42 | cr2_ | hex ? |
23:25.55 | cr2_ | 0x54 |
23:26.02 | BabelO | cr2_: no decimal or 0x54 hax |
23:26.15 | cr2_ | <PROTECTED> |
23:26.37 | BabelO | uh warning, don't touch ;) |
23:26.48 | cr2_ | look for 'cpld prog' |
23:27.02 | cr2_ | something for the virus writers ;) |
23:28.04 | cr2_ | BabelO: did you add the SD gpio ? |
23:28.20 | BabelO | cr2_: no i don't found any of them |
23:28.32 | cr2_ | ok, it's not on the cpld. |
23:29.06 | cr2_ | you need to do something with the sdio/wifi mux. |
23:29.15 | BabelO | cr2_: i think i just miss some power gpio, because driver already dial with it |
23:29.27 | cr2_ | i guess 54mbit acx100 is going to be the next "great thing" |
23:30.03 | cr2_ | how do you set the 0x79 and 0x8d ? |
23:30.07 | BabelO | yes i already set a3,79,8d to zero like we say yesturday |
23:30.12 | cr2_ | i think it must be in the logs. |
23:30.16 | cr2_ | ok. |
23:30.22 | Kevin2 | BabelO: Still looking for me? |
23:30.26 | cr2_ | add these comments into the source. |
23:30.41 | cr2_ | because they are difficult to grep in the logs. |
23:30.42 | BabelO | Kevin2: no, it is ok now ;) i like your htc-egpio driver |
23:30.55 | BabelO | Kevin2: once i found what was missing on omap tree |
23:31.34 | Kevin2 | Okay. :-) |
23:31.39 | cr2_ | Kevin2: we were thinking of adapting your BB driver for spi-type |
23:32.01 | cr2_ | that we need for the lcd setup on at least 3 phones now. |
23:32.27 | Kevin2 | Cool. It would be nice to know what chip the bitbanging connects to. Maybe that way we could figure out how to write reusable code. |
23:33.27 | cr2_ | Kevin2: it seems to me that on many toppoly panels we have real spi. |
23:33.59 | cr2_ | moko1 uses a real spi port for it, but htc consistently does gpio bitbang. |
23:34.37 | cr2_ | Kevin2: i use the term 'spi' because of the following. |
23:34.54 | cr2_ | it uses _cs=0 before send/receive and _cs=1 after it. |
23:35.15 | cr2_ | the transfers may be multibyte. and i think MSB first |
23:35.40 | cr2_ | send and receive are done by using different gpios. |
23:35.48 | cr2_ | so it's certainly not i2c. |
23:36.30 | *** join/#htc-linux marmotta (n=skodde@unaffiliated/skodde) |
23:38.17 | BabelO | cr2_: what is BIC R3, R3, #0x60000000 |
23:38.28 | Kevin2 | The link on hermes and apache uses only two gpios. One works like a clock (always output and just oscillates between 0 and 1), the other like data in/data out. Which is very similar to i2c. |
23:38.41 | cr2_ | clear bits 0x20000000 and 0x40000000 |
23:38.53 | BabelO | cr2_: ok |
23:38.58 | Kevin2 | I guess, theoretically, one could think of it as 3-wire spi without the chip select line. I'd have to read up on spi to be sure. |
23:39.09 | cr2_ | Kevin2: yes, that's why i think these are different drivers. |
23:39.24 | cr2_ | i2c is 2 wire. SDA and SCL |
23:39.42 | Kevin2 | Sure, but it would still be nice to know what chip we're bit-banging against. Maybe there are some docs somewhere. :-) |
23:40.14 | cr2_ | SDA is (may be) used for transmit and receive, i.e. it's not full duplex. |
23:40.36 | cr2_ | afair this is the difference to spi without cs. |
23:40.41 | Kevin2 | BTW, I have a small awk script that can parse mmutrace output and turn it into more readable output for the bitbanging stuff. |
23:41.08 | cr2_ | neat. add it to haretconsole dir. |
23:41.35 | Kevin2 | cr2_: The script is too much a hack. I'll pastebin it though. |
23:41.38 | cr2_ | what we are missing now is the i2c and spi protocol analyzer/parser. |
23:41.42 | cr2_ | ok. |
23:42.14 | cr2_ | i doubt one can write the i2c cam control driver without such analyzer. |
23:42.16 | Kevin2 | http://pastebin.ca/999807 |
23:42.25 | cr2_ | but it's certainly a big and complex task. |
23:43.42 | cr2_ | thanks. |
23:47.43 | cr2_ | BabelO: i don't see any other SD gpios. only these 2. |
23:50.14 | cr2_ | BabelO: a3 -megasim power ? |
23:54.38 | cr2_ | good night. |
23:57.41 | BabelO | cr2_: ok |
23:57.44 | BabelO | good night |
23:58.08 | Kevin2 | cr2_: Good night. |