IRC log for #openezx on 20081104

00:10.25WyrMhappy birthday to myself ;)
00:10.43WyrMaltough here in .br its not my birthday yet.
00:10.46WyrM;)
00:14.59johanbrWyrM: Happy birthday in advance! Sit back, relax and have a caipirinha. :)
00:15.46WyrMjohanbr: thanks! :)
00:16.14WyrMjohanbr: im do prefer a beer, im not much a fan of cachaça.
00:19.03WyrMlol
00:19.28WyrMwikipedia says that in .br we consume 8 litres of cachaça per head per year.
00:19.40johanbrWyrM: Ahh. So have a Brahma. :)
01:04.55*** join/#openezx sabrod_ (n=mac@lns-bzn-37-82-253-42-111.adsl.proxad.net)
01:37.42*** join/#openezx LoneStar (n=xinzhen@218.80.193.59)
01:53.46*** join/#openezx Wyr1 (n=wyrm@201.58.150.250)
02:11.48*** join/#openezx tombhadAC (n=tombhad@pD95DEE91.dip.t-dialin.net) [NETSPLIT VICTIM]
02:40.20*** join/#openezx xLoneStar (n=xinzhen@218.80.193.59)
03:13.57*** join/#openezx johanbr (n=j@blk-7-128-3.eastlink.ca)
03:22.45*** join/#openezx WyrM (n=wyrm@20158139112.user.veloxzone.com.br)
04:34.20*** join/#openezx Wyr1 (n=wyrm@20158130224.user.veloxzone.com.br)
05:13.13WyrMpurl: openezx?
05:13.14purlit has been said that openezx is an effort from volunteers to provide a 100% free software stack for Motorola EZX phones, currently the project is focused on providing a 2.6 kernel with all required hardware support, see more at http://www.openezx.org/
05:13.27WyrMpurl: openezx git?
05:13.28purli guess openezx git is at http://git.openezx.org/
05:13.58WyrMpurl: openezx snapshots?
05:13.58purlopenezx snapshots are at http://people.openezx.org/wyrm/images/
05:14.09WyrMpurl: wyrm?
05:14.10purlfrom memory, wyrm is Daniel Ribeiro, a lazy OpenEZX developer that talks to me instead of doing something useful.
05:14.16WyrM:)
05:15.20WyrMlol, nice bot.
06:01.27*** join/#openezx the_mgt_ (n=ich@dslb-082-083-210-059.pools.arcor-ip.net)
06:14.25tmztWyrM: still trying to update this kernel
06:48.47*** join/#openezx rwhitby (n=rwhitby@nslu2-linux/rwhitby)
07:12.46yangWhen they launched the EZX phones, was Motorola's intention to run some sort of Linux on them or is this simply a reverse hack made by you guys ?
07:13.23yangIf it was Motorola's intention for Linux I wonder why they haven't set up the linux team, or any sort of support
07:56.38*** join/#openezx pleemans (n=toi@116.54-246-81.adsl-static.isp.belgacom.be)
08:10.24*** join/#openezx james___ (n=james@gimel.esc.cam.ac.uk)
09:38.39*** join/#openezx florian_ (n=fuchs@217.146.132.69)
11:09.48*** join/#openezx droebje (n=quirc@92.117.214.47)
11:16.40*** join/#openezx ao2 (n=u@2001:1418:117:0:0:0:0:1)
12:00.40*** join/#openezx gmzhuo (n=gmzhuo@118.113.160.226)
12:09.50*** join/#openezx WyrM (n=wyrm@189.13.101.154)
12:21.14gmzhuoHi,WyrM
12:22.47WyrMhi gmzhuo
12:24.46gmzhuodo you think BP will check codec register on gen2 when phone call?
12:25.01WyrMgmzhuo: no.
12:26.01gmzhuoBut if when I make a call,if I change mixer will cause bp dead,many times
12:26.21WyrMbut I think that it can check the adc during a phone call, to setup radio circuitry depending on battery level.
12:26.40WyrMhm.
12:26.46WyrMgmzhuo: what do you change on mixer?
12:27.06gmzhuochange bp codec to mono or stero
12:27.24WyrMyeah, in this case it is possible.
12:27.50WyrMas if you change the dai to mono or stereo bp will lose the ssp connection to the codec.
12:28.42gmzhuoWhich will send the frame sync,bp or pcap?
12:28.48WyrMpcap
12:28.55WyrMframe and clock
12:30.17gmzhuoWhen phone call record,dose motorola use ssp2 connect to bp and ssp3 connect to pcap?
12:30.27WyrMdon't know.
12:31.25WyrMI don't support network mode currently, and it is needed to use both ssp ports.
12:32.56WyrMand in fact, it don't need both ssp ports.
12:33.11WyrMthe same port may connect to BP and PCAP at the same time.
12:33.18*** join/#openezx yansa (n=yans@89.108.255.71)
12:34.34gmzhuoYou mean ap as a ssp spy to make phone record ?
12:34.37gmzhuo:)
12:34.50WyrMnot as spy.
12:34.57WyrMas a man-in-the-middle.
12:35.14WyrMthe same way that it would work with 2 ports.
12:35.24WyrMbut with 2 slots on network mode.
12:36.00gmzhuoI can't image one ssp as a man in the middle :(
12:36.36WyrMit receives from BP on slot 1, and send to pcap on slot 2.
12:36.57WyrMpcap drives the clock on the port (for both slots)
12:38.01gmzhuothen how about mic?
12:38.09WyrMsame
12:38.19WyrMssp is full dublex.
12:38.25WyrMs/b/p/
12:38.51WyrMit always transmit the same amount of bits that it receives, simultaneously.
12:41.05gmzhuobut ssp has two single,how to recv from from bp and send to pcap,and recv from pcap and send to bp simultaneously?four signal
12:46.11WyrM<slot>: <SSPRXD>, <SSPTXD>
12:46.12WyrMslot1: receive audio frame 1 from BP, do nothing
12:46.12WyrMslot2: receive audio frame 1 from PCAP, send audio frame 1 to PCAP
12:46.12WyrMslot1: receive audio frame 2 from BP, send audio frame 1 to BP
12:46.12WyrMslot2: receive audio frame 2 from PCAP, send audio frame 2 to PCAP
12:54.56WyrMssp network mode allows you up to 8 time slots, so you can use more than one device on the same ssp port.
12:55.15WyrMthis not using chip selects, which is yet another way to share the ssp bus.
12:58.30WyrMahh, and to use network mode some reconfiguration is needed on the BP side.
12:59.03WyrMthere is a big chance that you will see some at command when you start recording a call, to setup the net mode on the bp side.
12:59.28WyrMor the net mode is used since the beginning, who knows. :)
13:12.12gmzhuoI am watching prison break :)
13:12.35WyrMis waiting for heroes.
13:12.40WyrM:)
13:13.08WyrMI still not started season4 of prison break.
13:14.30*** part/#openezx BadCluster (n=badclust@190.50.180.205)
13:45.22*** join/#openezx ade|desk (n=adavey@195.153.131.254)
13:46.13gmzhuoWyrM,do you have more pcap2 document than http://wiki.openezx.org/PCAP2
14:17.46*** join/#openezx bartv (n=magick@demands.xs4all.nl)
15:07.13*** join/#openezx Wyr1 (n=wyrm@189.13.121.158)
15:07.58WyrMgmzhuo: no, we don't have a datasheet or any other document.
15:12.07*** join/#openezx LoneStar (n=xinzhen@114.92.179.21)
15:22.51WyrMao2: ping
15:23.38*** join/#openezx BadCluster (n=badclust@200.26.114.34)
15:27.18*** join/#openezx johanbr (n=j@JBrannlund2.MathStat.Dal.Ca)
15:43.22WyrMwhat should I work on today? a780 camera or a1200 bp?
15:47.23johanbrIsn't today your birthday? Take a day off.
15:47.23stefan_schmidtWyrM: What are you more interested in today?
15:47.48WyrMjohanbr: I took the day off, that's why im asking ;)
15:47.58johanbr:)
15:48.05stefan_schmidtWyrM: Happy birthday!!
15:48.13WyrMstefan_schmidt: Thanks! :)
15:49.28webmindWyrM, congratulations
15:49.31stefan_schmidtWyrM: Then make sure you hack on code you like.
15:49.48stefan_schmidtWyrM: No ugly bug fixing, but working with pleasure today :)
15:52.41WyrMwebmind: Thanks! :)
15:53.07WyrMstefan_schmidt: I will go for a1200 bp, as my a780 batts are all empty
15:53.59stefan_schmidtWyrM: That is of course a good reason :)
15:55.44*** join/#openezx BadCluster (n=badclust@200.26.114.34)
15:55.48*** part/#openezx BadCluster (n=badclust@200.26.114.34)
16:48.45ao2WyrM, happy birthday
16:55.18*** join/#openezx WyrM (n=wyrm@20158183227.user.veloxzone.com.br)
17:00.25WyrMhah!
17:00.27WyrMbingo.
17:00.41ao2WyrM, happy birthday (again), did you want to ask me about adding a780 camera support?
17:00.55WyrMI think I also discovered why gmzhuo needs to change hub.c to make bp work on E6.
17:01.02WyrMtiming is the key.
17:01.07WyrMhi ao2
17:01.09WyrMthanks!
17:01.37WyrMYes, I was going to ask you to send me the userspace software you used to test the camera.
17:01.51WyrMand instructions on how to use it and/or how to fix the colors.
17:04.22WyrMso, how should a fix the bp issue on 2nd gen? a one line change on hub.c? an ugly hack on our files only, to delay usb enumeration? or always using bp code as a module? :)
17:13.02ao2WyrM, I put everything on people
17:22.23WyrMao2: where?
17:23.32ao2I'll put, sorry
17:23.47WyrMahh, ok.
17:23.47WyrM:)
17:33.17ao2WyrM, http://people.openezx.org/ao2/v4l2_capture_example.tar.gz
17:41.57WyrMao2: thanks.
17:43.30WyrMboh is a color format?
17:44.04WyrMor the sensor format is yuv?
17:48.39*** part/#openezx ade|desk (n=adavey@195.153.131.254)
17:49.55ao2WyrM, "boh" is an Italian inrejection for "I really don't know"
17:50.17ao2interjection
17:50.41WyrMpurl: boh?
17:51.36WyrMas it seems the color format is yuv, I don't know what is this boh.
17:54.05ao2it was a silly joke I used when I didn't know where the problem was, so I was making no assumption about the format
17:54.40WyrMahh, so boh is invented by you ;)
17:54.42WyrMlol
17:55.05ao2sure, it's like unknown2rgb but more funny
17:55.35ao2for me at least :|
17:55.36ao2:)
17:56.27l403boh/
17:56.29l403?
17:57.33WyrMpurl: boh is ao2's strange naming for the color format used on mt9m111 camera sensors.
17:57.33purlokay, WyrM
17:59.02l403that explains why the most relevant reesult on wikipedia is the BOH tea
18:05.15ao2WyrM, there is discussion on linux-video about mt9m111 and pxa-camera, the final fix may not be like the one I sent, but for now it is ok for us. The problem is that pxa-camera may be wrong saying it supports YUYV, it looks like the datasheet states UYUV, so if this is confirmed also mt9m111 will be changed (they will add UYUV). But, well, you can safely ignore all this stuff, tho.
18:50.38*** join/#openezx WyrM (n=wyrm@20158183227.user.veloxzone.com.br)
19:33.15*** join/#openezx pleemans (n=toi@d54C2AAB7.access.telenet.be)
19:37.51*** join/#openezx WyrM (n=wyrm@20158151245.user.veloxzone.com.br)
20:46.23ao2a page about gen-blob http://wiki.openezx.org/Dual_boot
22:21.03*** join/#openezx WyrM (n=wyrm@20158170203.user.veloxzone.com.br)
22:23.51WyrMI still can't figure a clean way to delay usb enumeration. :/
22:31.43*** join/#openezx droebje (n=quirc@92.116.179.245)
22:49.47*** join/#openezx l403 (n=l@85.132.159.231)
22:56.03*** join/#openezx l403 (n=l@85.132.159.231)
22:57.39ao2cool, on CNN they are using one of those multi touch screens. I've never seen them on tv before.
23:00.02*** join/#openezx l403 (n=l@85.132.159.231)
23:07.58tmztWyrM: oh, what for?
23:08.33WyrMbecause BP needs to have the handshake done before being enumerated.
23:08.58WyrMotherwise it fails to accept address with -ETIMEDOUT.
23:13.51WyrMoh, I f*cked my topgit tree.
23:14.06WyrMas it seems topgit don't like git reset. :)
23:14.25tmztI think my ezx/current has something extracted on it (like tg update), but my pc is off right now
23:14.49tmztyeah, the SABM thing?
23:15.01WyrMyes
23:16.41tmztI need to try booting with the phone disabled and see if it will come up in AT mode on that usb3, but I've been reading on ezx wiki, and mailing list, and the 27.0170 spec
23:18.02tmztso it's not just when ts0710_mux_usb loads, but when the actual usb is enum'd?
23:18.28WyrMI think its different with ezx than it is with motoq.
23:18.34tmztthe bootloader (on ezx) needs the usb host to be active?
23:18.39tmztyeah
23:18.51tmztto access the BP?
23:18.54WyrMthere is a handshake needed on ezx.
23:18.59tmztthe bootloader (on ezx) needs the usb host to be active to access the BP
23:19.02tmzt?
23:19.08WyrMand you don't seem to need it (or haret already done it to you)
23:19.15WyrMno
23:19.28tmztso you can keep the usb port powered down?
23:19.50WyrMit needs us to don't use the original bootloader flash mode.
23:19.58tmztyeah, the mux is completely different
23:19.58WyrMbecause it puts the baseband on flash mode too.
23:20.24tmztflash mode is on a usb interface?
23:20.35WyrMso, we follow the same code path that the original bootloader takes when booting the ordinary firmware (not the flash mode)
23:20.51WyrMand add another bootloader to the kernel partition on flash.
23:21.27WyrMyes, flash mode is usb.
23:23.16*** join/#openezx l403 (n=l@85.132.159.231)
23:25.50WyrMim cloning my tree again :(
23:26.28tmzthow does 2.4 boot then? something that's not in the wiki?
23:26.49WyrMwhat do you mean?
23:26.56WyrMwe are not talking the same thing here :)
23:27.17tmztI'm asking about ezx not motoq, just wondering what is happening
23:28.01WyrM1. there are 2 code paths on BP firmware that are decided right after boot depending on the level on a gpio.
23:28.12WyrMone code path is the normal baseband firmware.
23:28.24WyrMthe other is the flash mode.
23:28.50WyrMthere is no way to go from the flash mode to the normal baseband, without power-cycling.
23:29.32WyrM2. when you use the bootloader key-combo to enter the bootloader screen, it toggles the level on the bp flash gpio, putting it on the so called flash mode.
23:30.22WyrMit doesn't matter wheter you send a 2.4 motorola kernel or openezx. If you send it to the original bootloader the BP will not work.
23:31.01WyrM3. what we do to trick the BP software is: we never enter the bootloader with the key combo.
23:31.21WyrMinstead, we flash another bootloader to the phone, on the partition that was originally designed to hold the kernel.
23:32.01WyrMthis way, the original bootloader never puts the BP on flash mode, and give control to our bootloader.
23:32.23WyrMour bootloader can receive a kernel via USB, or load it from the MMC card.
23:37.21tmztok, that makes sense
23:37.41tmztso your bootloader initializes the usb host?
23:42.57*** join/#openezx Wyr1 (n=wyrm@189.13.221.88)
23:43.10Wyr1damn, my adsl is so unstable today :/
23:43.16Wyr1where did I stop?
23:44.08tmzt<PROTECTED>
23:44.11tmzt19:29 < tmzt> ok, that makes sense                                              
23:44.14tmzt19:29 < tmzt> so your bootloader initializes the usb host?
23:44.49WyrMno.
23:45.23WyrMit just give control to a kernel, without doing nothing, neither on the usb nor on the handshake pins :)
23:45.38tmztok
23:45.46WyrM(21:34:46) WyrM: 4. the second code path:
23:45.47WyrM(21:35:20) WyrM: the real baseband firmware needs to shake hands with the AP to be activated.
23:45.47WyrM(21:35:53) WyrM: this handshake also starts a watchdog, and enables communication via USB (ipc-usb)
23:45.47WyrM(21:36:29) WyrM: but if we try to enumerate the AP usb port, the device(BP) will timeout unless this handshake is done.
23:45.47WyrM(21:37:20) WyrM: that's why im looking into a way to delay the enumeration until the handshake is ready (variable time, we put a interrupt on a gpio to see when bp is ready).
23:46.27tmztso the watchdog is not related to ts0710 timers?
23:46.31WyrMbefore powering the usb, we need to finish the handshake that is done via gpio levels, this is done by kernel.
23:46.37WyrMyes, it is.
23:47.02WyrMthe watchdog triggers if bp asks for communication and ap don't responds.
23:47.44WyrMso, if you finish the handshake, and don't start talking with BP via the /dev/mux lines it will ask the PCAP to shutdown power.
23:48.47WyrMwhat starts the watchdog is the last handshake step. And what resets it is any communication via mux.
23:49.58tmztbp controls pcap so there is no way to restart it?
23:50.09WyrMboth processors control pcap.
23:50.40WyrMthere are 2 ssp/spi ports on pcap, both processors can read/write pcap's registers.
23:51.38WyrMbut for this particular power off case, any of the 2 processors can abruptelly cut all the phone power simply toggling a gpio.
23:55.27WyrMhum..
23:55.45WyrMif I delay the UP3OCR write, will this delay the enumeration?
23:55.53WyrMlet me try :)
23:56.28*** join/#openezx flo_lap (n=fuchs@g227195137.adsl.alicedsl.de)
23:57.04tmztyou could make ohci-hcd a module? don't know if it will find platform_data that way
23:57.14WyrMtmzt: it will.
23:57.21WyrMand it does work as a module.
23:57.28WyrMbut I want it builtin.
23:57.57tmztI guess if you power up the port later it will reenumerate
23:57.59WyrMafter the handshake is done, I have just some seconds to enumerate the usb port and start talking with bp.
23:58.14tmztoh, so you udelay it
23:58.28tmztthe builtin bootloader starts the handshaking process?
23:58.34WyrMyes.
23:58.34tmztgpios
23:58.50WyrMthe bootloader starts and the kernel finishes it.
23:59.03WyrMand this varies depending on the phone generation :/
23:59.07tmztcould you shut it down from pcap in your bl and then bring it up in the kernel?
23:59.17WyrMshut what down?
23:59.21WyrMBP?
23:59.22tmztjust shut down the bp
23:59.26WyrMno.
23:59.44ao2a first draft about testing audio calls with the current code http://wiki.openezx.org/Call doesn't really work with latest code, I tried so many times...
23:59.44WyrMI can't shut down bp without shuting down AP.

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