| 00:10.44 | *** join/#openezx jonwil (n=jonwil@220-245-202-116-wa-pppoe.tpgi.com.au) |
| 00:10.59 | jonwil | hi |
| 00:21.39 | blkhawk | lo |
| 00:21.42 | blkhawk | and gnight |
| 03:14.14 | *** join/#openezx jaebird (n=jae@53-89.netblk-69-41-89.coolaccess.net) |
| 04:20.57 | *** join/#openezx jaebird (n=jae@53-89.netblk-69-41-89.coolaccess.net) |
| 04:59.21 | *** join/#openezx mwester (n=mwester@nslu2-linux/mwester) |
| 05:58.07 | *** part/#openezx ade|desk (n=ADAVEY@213-232-83-137.dsl.prodigynet.co.uk) |
| 05:59.03 | *** join/#openezx _dion_ (n=dionua@xen.inhex.net) |
| 07:21.42 | *** join/#openezx ade|desk (n=ADAVEY@194.200.143.249) |
| 09:10.19 | Hopscotch | guten morgen |
| 10:25.05 | ade|desk | elllo |
| 12:11.37 | *** join/#openezx florian_kc (n=fuchs@dns.kernelconcepts.de) |
| 12:12.02 | florian_kc | hi all |
| 12:16.52 | stefan_schmidt | hi florian_kc |
| 12:33.15 | *** join/#openezx [g2] (n=g2@nslu2-linux/g2) |
| 12:42.11 | *** join/#openezx _dion_ (n=dionua@xen.inhex.net) |
| 13:04.20 | jonwil | hi all |
| 13:17.15 | stefan_schmidt | hi [g2], _dion_, mickeyl, jonwil |
| 13:17.28 | [g2] | morning stefan_schmidt ! |
| 13:18.25 | [g2] | stefan_schmidt can you talk about the usb->serial console driver at all ? |
| 13:18.53 | [g2] | I thought I saw something about using the usb port as a console yesterday |
| 13:19.18 | _dion_ | stefan_schmidt: |
| 13:20.45 | stefan_schmidt | [g2]: Yes, no driver needed. Just reconfigured pcap2. |
| 13:21.24 | [g2] | did you have to tweak the kernel to support that console on usb ? |
| 13:21.24 | jonwil | so when do you play further with this phone? |
| 13:22.01 | jonwil | hopefully we will soon be able to make a phonecall on an EZX phone using nothing but self-built code on the AP side :) |
| 13:25.32 | stefan_schmidt | [g2]: Yes, you need to change a function call. That's all. Later we should be able to change this from userspace. |
| 13:25.57 | stefan_schmidt | jonwil: My plan is to have a first phone call this year. :) |
| 13:26.12 | stefan_schmidt | jonwil: I have a bet with alphaone about that. :) |
| 13:27.53 | *** join/#openezx blkhawk[work] (n=blkhawk@IP-213157024141.dialin.heagmedianet.de) |
| 13:27.54 | blkhawk[work] | lo |
| 13:27.54 | stefan_schmidt | wb blkhawk[work] |
| 13:28.47 | jonwil | I think we can do be4tter than just "make a phone call" |
| 13:29.04 | jonwil | I think once we get as far as "make a phonecall" we can go further |
| 13:29.18 | jonwil | and be able to make and recieve phonecalls using only selfbuilt code on the AP side |
| 13:29.27 | jonwil | just like the normal firmware |
| 13:29.55 | jonwil | so you can use the selfbuilt software on the phone and still be able to use it as a functioning phone |
| 13:30.28 | jonwil | if we could figure out how to get our own kernel onto an A1200 or E2, we could probobly do some good things there too |
| 13:30.39 | jonwil | but as far as I know so far no-one has done that yet |
| 13:36.24 | stefan_schmidt | jonwil: It is still some work before the first phone call. |
| 13:37.05 | jonwil | Assuming we can get mux_cli going, that should (in theory) result in a stable phone that doesnt shut down anymore |
| 13:37.23 | stefan_schmidt | jonwil: And I don't start to write to much userspace stuff, instead of some test scripts, as I wait for openmoko here. |
| 13:37.30 | stefan_schmidt | jonwil: right |
| 13:37.37 | jonwil | is openmoko available though |
| 13:37.39 | jonwil | > |
| 13:37.40 | jonwil | ? |
| 13:37.51 | stefan_schmidt | But handling the complete AT commands is way more work. |
| 13:37.55 | stefan_schmidt | jonwil: january |
| 13:39.20 | jonwil | I know enough about AT commands as they apply to my own L6 to know that only maybe half a dozen AT commands (all of them specified in the GSM AT command spec) are required to make and recieve phonecalls |
| 13:39.33 | jonwil | most of the others are for fancy stuff like phonebooks and SMS and stuff |
| 13:40.01 | jonwil | of course, if the phone sends any AT commands to the BP on startup, thats a different thing alltogether. |
| 13:40.09 | jonwil | But if it does, we can just send exactly what they send |
| 13:43.11 | jonwil | Other than mux_cli, what other pieces of hardware dont work in 2.6? |
| 13:43.52 | stefan_schmidt | jonwil: audio, bluetooth, powermanagement, GPS, |
| 13:44.05 | stefan_schmidt | jonwil: And all other drivers need more fixing. |
| 13:44.47 | jonwil | does this "make a phonecall" bet actually mean that you can talk also or just dial number and have the other phone ring? :) |
| 13:45.05 | alphaone | jonwil: A complete communication |
| 13:45.07 | alphaone | :-) |
| 13:45.10 | jonwil | cool |
| 13:45.16 | stefan_schmidt | alphaone: wow, that's new to me. :) |
| 13:45.18 | stefan_schmidt | ;) |
| 13:45.33 | alphaone | stefan_schmidt: Got to earn your club mate :-) |
| 13:45.35 | jonwil | I am going to side with alphaone on this :) |
| 13:45.42 | stefan_schmidt | alphaone: Keep in mind neolynx is actually working on the also driver. :) |
| 13:46.04 | stefan_schmidt | s/also/alsa/ |
| 13:47.02 | alphaone | jonwil: We just haven't settled who gets how many club mate if I make the phone call. |
| 13:47.10 | stefan_schmidt | jonwil: Short term we will write some code to test phone call and sms, perhaps with an existing gsm lib. Long term plan is to use the openmoko framework- |
| 13:47.19 | alphaone | The original bet was if stefan can call me before the CCC |
| 13:47.37 | stefan_schmidt | alphaone: Can I get my phone back? ;) |
| 13:47.54 | alphaone | You wish :-) |
| 13:48.12 | jonwil | all you really need is working mux_cli, possibly working audio (if audio for phone call goes through ezx AP side), a function to send AT commands (and if needed data) to the right /dev/mux ports |
| 13:48.19 | jonwil | and from there writing a phonedial app isnt hard |
| 13:48.46 | stefan_schmidt | jonwil: ./dial $number is fine for first tests. :) |
| 13:48.59 | alphaone | jonwil: I totally aggree and it is looking much better now than it was last weekend. |
| 13:49.31 | alphaone | s/gg/g/ |
| 13:50.13 | jonwil | I saw somewhere that some people are trying to cleanroom libezxtapi |
| 13:50.35 | jonwil | i.e. have someone disassemble libezxtapi (and if needed tapisrv) and reverse engineer it |
| 13:50.39 | alphaone | Where did you see that? |
| 13:50.45 | jonwil | I cant remember exactly |
| 13:51.02 | stefan_schmidt | jonwil: You mean proberts efforts |
| 13:51.09 | jonwil | probobly |
| 13:51.11 | jonwil | yeah |
| 13:57.37 | jonwil | I can read x86 and even (to some extent) 68000, 6502 and 65816 ASM |
| 13:57.51 | jonwil | but ARM assembler is greek to me |
| 13:58.44 | alphaone | Learn a new dialect today :-) |
| 13:58.45 | stefan_schmidt | jonwil: For the magic it should be enough to sniff the traffic. |
| 13:59.00 | jonwil | yep |
| 14:00.20 | jonwil | Can we safetly flash a 2.6 kernel to the phone (i.e. safetly as in being able to flash other stuff on top of it later) yet? |
| 14:03.43 | _dion_ | i never do this but afaik yes |
| 14:04.59 | jonwil | if I was going to do hacking on EZX phone what I would do next is (in no particular order): 1.Get MUX_CLI working on a flashed-to-phone 2.6 kernel |
| 14:05.25 | jonwil | i.e. get it working enough that the it can talk to the right USB device (i.e. the BP) |
| 14:05.40 | jonwil | and so that any kernel items necessary to keep the phone from shutting down |
| 14:06.54 | jonwil | 2.Log the /dev/mux devices (especially the ones labled "AT commands") during a bootup of a phone with a normal motorola userland in place and also during making and recieving calls |
| 14:07.21 | stefan_schmidt | exactly our plan. When we have fixed mtd support in 2.6 we can reflash a new kernel from our own system. |
| 14:07.29 | jonwil | 3.Once we know how to send AT commands to the BP side, send AT+CLAC (list all commands) to each of the relavent /dev/mux devices and see what it returns back |
| 14:07.31 | alphaone | tapisrv is at least partly responsible for keeping the phone alive |
| 14:07.57 | jonwil | well items 2 and 3 will help a lot with understanding whats going on |
| 14:08.31 | alphaone | jonwil: Point 3 could be done on a 2.4 kernel |
| 14:08.34 | jonwil | I dont know if linux has such functionality but if there is a way to log all the devices that tapisrv opens, reads from, writes to and sends IOCTLs to, that would help too |
| 14:08.39 | jonwil | yeah, thats what I mean |
| 14:08.45 | jonwil | doing 2 and 3 on the 2.4 kernel |
| 14:08.53 | alphaone | jeak |
| 14:08.56 | alphaone | s/k/h |
| 14:09.09 | jonwil | that wat we know exactly what the phone sends to the BP when it starts up |
| 14:09.19 | jonwil | and what it sends while it is alive |
| 14:09.27 | jonwil | and what it sends when a phonecall is made and recieved |
| 14:10.05 | _dion_ | i download it.. but i forget, where |
| 14:34.26 | *** join/#openezx agd5f (n=DeucherA@65.242.105.130) |
| 14:45.46 | *** join/#openezx ao2 (n=u@2001:1418:117:0:0:0:0:1) |
| 14:51.54 | nh_work | this is great guys. :) |
| 14:52.31 | stefan_schmidt | hi agd5f, ao2, nh_work |
| 14:52.43 | nh_work | stefan: heya, sounds like you are making good progress! |
| 14:53.24 | stefan_schmidt | nh_work: Yes |
| 14:54.05 | nh_work | stefan: do you know much about flashing regarding sbf and smg files? |
| 14:55.28 | stefan_schmidt | nh_work: no |
| 14:56.39 | nh_work | stefan: ok. There's a post on motomodders discussing it. It sounds like the smg files contain various carmfs file systems and other mtd data when the phone gets flashed (at least for the E2). |
| 14:57.45 | nh_work | I read earlier that /dev/mtd is a 1-to-1 mapping of what's in flash, so I might be able to figure out how to create a backup image from my phone, and then delete or modify the acpolicy.cfg file. |
| 14:57.58 | nh_work | er what's in the flash image |
| 14:58.04 | nh_work | not in flash memory. :) |
| 15:05.59 | ao2 | hi stefan_schmidt, any news different from yesterday? |
| 15:06.47 | stefan_schmidt | ao2: No. Just sleep. :) |
| 15:07.03 | ao2 | a good news anyway :) |
| 15:08.01 | stefan_schmidt | heh |
| 15:08.32 | ao2 | playing with the 2.6 botted via usb we found (again) that it works when using the touchscreen but than it shut down if used remotely, because no event come from ts and powering off the lcd means powering off the whole device given the current pm handling |
| 15:10.18 | stefan_schmidt | hmm, IIRC I had something like this myself. Perhaps problems with going into powersave and non working pwermanagement |
| 15:10.27 | ao2 | so neolynx wondered if the shutdown problem you had with flashed kernels can be more a similar problem (hence on AP side) than not a watchdog from BP |
| 15:11.48 | alphaone | ao2: No since we also had the problem with a self compiled 2.4 kernel |
| 15:11.50 | stefan_schmidt | ao2: I don't think so. |
| 15:12.01 | stefan_schmidt | ah, alphaone is right here. |
| 15:12.30 | alphaone | It seems that tapisrv is involved in the keep-alive process |
| 15:12.57 | ao2 | ah, right, the 2.4 is supposed to do pm "correctly" |
| 15:13.47 | alphaone | And it does. I have one kernel which works - unless you kill tapisrv |
| 15:14.07 | stefan_schmidt | ao2: Let us say it _does_ pm. Not sure if it does it the right way. :) |
| 15:14.27 | alphaone | heh :-) |
| 15:17.07 | ao2 | prawling tapisrv ops with strace may be of some help, I saw once a strace version compiled for ezx |
| 15:18.28 | alphaone | ao2: I will try to strace the tapisrv tomorrow |
| 15:19.51 | ao2 | alphaone, the 2.4 kernel you compiled is 2.4.20 from opensource.motorola.com, isn't it? |
| 15:20.47 | alphaone | Yes, with the mux_cli module motorola released compiled into the kernel |
| 15:21.12 | ao2 | and which toolchain you used to build it? ezx-crosstool ? |
| 15:22.37 | alphaone | cu |
| 15:24.28 | alphaone | ao2: Yes, ezx-crosstool. You can find the patch and config at http://thebe.totalueberwachung.de/~alphaone/openezx/ |
| 15:25.07 | ao2 | good, I'll try to get my 2.4 kernel, too. Thanks. |
| 15:29.57 | alphaone | np |
| 16:09.07 | *** join/#openezx blkhawk (i=blkhawk@pD9E06137.dip.t-dialin.net) |
| 16:09.09 | blkhawk | lo |
| 16:38.07 | *** join/#openezx kenws (n=ken@dslb-088-064-129-127.pools.arcor-ip.net) |
| 16:38.18 | *** join/#openezx TimRiker (n=timr@216.49.181.128) |
| 16:45.35 | *** join/#openezx florian_ (n=fuchs@87.193.44.59) |
| 16:55.27 | *** join/#openezx florian__ (n=fuchs@87.193.4.46) |
| 17:11.04 | *** join/#openezx florian_ (n=fuchs@87-193-29-90.dpool.celox.de) |
| 17:55.54 | *** join/#openezx _dion_ (n=dionua@xen.inhex.net) |
| 18:21.03 | *** join/#openezx vmaster_ (i=vmaster@p549B4C58.dip.t-dialin.net) |
| 18:53.33 | *** join/#openezx mickeyl (n=nnnnnmic@linuxtogo.org) |
| 19:03.23 | *** join/#openezx proberts (n=proberts@cpe-76-167-78-52.socal.res.rr.com) |
| 19:03.28 | proberts | hi |
| 19:07.08 | proberts | nh_work: Hey :) I think the disable module thing may not have worked because it was too late in the boot-up process. I'm going over the disassembly of ag_getflexbit, which is called in rc2 and looks at two files to determine if security should be enabled or disabled. I'm trying to figure out which values we can try to change in the flex files to disable security. |
| 19:08.02 | nh_work | proberts: ah! What are the two files that it looks at? |
| 19:09.06 | proberts | /usr/setup/ezx_flexbit.cfg and /user/securesetup/ezx_security.cfg |
| 19:10.22 | proberts | There's /etc/rc.d/rc2.d/S10acgetflex.sh that just calls /etc/ac_getflexbit (which would be really easy to hack if we had +w access to) |
| 19:11.31 | proberts | /etc/ac_getflexbit uses /usr/lib/ezx/libezxflex.so to access the files. I'm trying to figure out what value it asks for (it's part of either a struct or >4 args passed to the lib routine) |
| 19:12.47 | proberts | The lib routine returns 1 or 0, and 1==enable, 0==disable. The app then passes that value to "SWI 0x900FD" (system call), which I'm trying to find right now, but I'm guessing that's the ac_* routine I was calling from my module. |
| 19:14.07 | nh_work | proberts: What is it that's preventing us from calling it after the bootup process is complete? |
| 19:16.28 | proberts | I think the setting may be propagated to other parts of the boot process (and can't go backwards). |
| 19:18.08 | proberts | The selinux securty stuff for the filesystem is (somewhere) stuffed into the inodes. |
| 19:18.33 | proberts | I'm still not exactly sure why though. |
| 19:21.16 | proberts | Their motoac stuff may also not be called in the module, since it uses 'fopen' to open the acpolicy.cfg file, and fopen (normally) isn't available to kernel modules. It may be that the "acpolicy.cfg" file is somehow "applied" or already baked into the system. |
| 19:24.20 | proberts | ...gotta reboot my laptop back into windows - later :) |
| 19:38.09 | *** join/#openezx nevarrie_ (n=jgrant@CPE-24-163-233-209.kc.res.rr.com) |
| 21:01.10 | *** join/#openezx |dion| (n=dionua@80.84.181.50) |
| 21:02.11 | *** join/#openezx _kenws (n=ken@dslb-088-064-128-178.pools.arcor-ip.net) |
| 22:04.54 | *** join/#openezx neolynx (n=u@167-123-204-62-pool.dsl.fcom.ch) |
| 22:36.35 | *** join/#openezx jonwil (n=jonwil@220-245-202-116-wa-pppoe.tpgi.com.au) |
| 22:37.15 | jonwil | hi |
| 22:39.46 | neolynx | hello |
| 23:24.10 | *** part/#openezx agd5f (n=DeucherA@65.242.105.130) |