| 00:25.11 | Speedy2 | re |
| 01:24.04 | *** join/#htc-linux dullard (n=jim@adsl-static-1-30.uklinux.net) |
| 04:00.38 | lkcl | had to get some sleep - too tiring. |
| 04:00.47 | lkcl | more on the asm->c tomorrow |
| 04:04.11 | *** join/#htc-linux toi (n=pleemans@d5152D3B4.access.telenet.be) |
| 04:09.06 | Speedy2 | re |
| 04:47.27 | *** join/#htc-linux psokolovsky_ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
| 04:48.46 | *** join/#htc-linux psokolovsky (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) [NETSPLIT VICTIM] |
| 04:58.00 | *** join/#htc-linux ljp (n=lpotter@203.94.178.46) |
| 05:31.56 | Speedy2 | ljp: Yo man, long time no see |
| 06:08.52 | *** join/#htc-linux WizMaui (n=WizMaui@62.112.90.179) |
| 06:38.20 | *** join/#htc-linux scatt293 (i=scatterp@packet.filtered.ca) |
| 06:56.56 | *** join/#htc-linux toi (n=pleemans@d5152D3B4.access.telenet.be) |
| 07:09.25 | *** join/#htc-linux rob_w (n=bob@p85.212.195.201.tisdip.tiscali.de) |
| 07:27.07 | ljp | hey speedy |
| 07:33.37 | Speedy2 | How's life in tha .AU? |
| 07:45.42 | *** join/#htc-linux alex-begin (n=345rt@84.232.212.62) |
| 07:56.22 | *** join/#htc-linux lilliput (n=Touriste@82-47-214-98.stb.ubr11.brad.blueyonder.co.uk) |
| 08:00.47 | *** join/#htc-linux dullard (n=jim@adsl-static-1-30.uklinux.net) |
| 08:04.41 | *** join/#htc-linux Fossi (n=cmewes@195.177.48.115) |
| 08:05.42 | ljp | good. getting warmer +* |
| 08:05.48 | ljp | :) |
| 09:56.24 | *** join/#htc-linux TripleDES (n=sergio@84.77.161.251) |
| 10:00.15 | *** join/#htc-linux booba (n=booba@AAnnecy-256-1-79-119.w90-10.abo.wanadoo.fr) |
| 10:12.54 | *** join/#htc-linux psokolovsky_ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
| 10:48.54 | *** join/#htc-linux pH5 (n=ph5@e178201129.adsl.alicedsl.de) |
| 11:07.52 | *** join/#htc-linux rob_w (n=bob@p85.212.195.201.tisdip.tiscali.de) |
| 11:08.58 | rob_w | hey all |
| 11:15.18 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
| 11:17.46 | cr2 | hi rob_w |
| 11:18.47 | cr2 | rob_w: i have found a bug in the blueangel_audio |
| 11:21.18 | pH5 | hi all |
| 11:22.37 | pH5 | is it considered to be the right way(tm) to turn on the gsm chipset in the pxa serial startup routine (and have the gsm chipset shutdown as soon as the serial device is closed) |
| 11:23.09 | cr2 | hi PH5 |
| 11:23.13 | cr2 | i think yes. |
| 11:23.34 | cr2 | if the device is closed, nobody can answer the call/get sms anyway. |
| 11:23.51 | cr2 | or you need a hotplug driver to start such a program. |
| 11:23.53 | pH5 | we need some special-casing for the suspend/resume case then, because all serial uarts are shut down on suspend |
| 11:24.16 | pH5 | but we don't want to shut down the gsm chip just because of that |
| 11:24.26 | cr2 | ah, ok- |
| 11:25.31 | cr2 | maybe this hotplug driver is needed anyway. |
| 11:26.47 | pH5 | tricky... |
| 11:27.45 | cr2 | you get a gpio phone event, start the program (hmm, text mode?) it opens the port. |
| 11:32.54 | pH5 | I'm not sure, perhaps it is better to leave that to userspace. Some kind of daemon talking to the ttySx. |
| 11:35.57 | cr2 | pH5: like pcmcia. |
| 11:41.40 | rob_w | cr2, good work |
| 11:41.57 | rob_w | i was away for some days ..and i need to catch up a bit |
| 11:41.59 | rob_w | bbl |
| 11:49.19 | *** join/#htc-linux doepiedoe (n=doepiedo@84-53-96-51.fiber.unet.nl) |
| 11:49.45 | doepiedoe | hey cr2 |
| 11:49.51 | doepiedoe | are you here? ;) |
| 13:07.19 | lkcl | dwmw2 on #mtd has a g3 spec from m-systems |
| 13:08.04 | lkcl | ph5: http://www.handhelds.org/hypermail/kernel-discuss/19/1961.html |
| 13:08.22 | doepiedoe | lkcl are you working on universal kernel too? or blueangel |
| 13:08.48 | lkcl | eventually yes, when i get my devices back from holland |
| 13:10.08 | doepiedoe | ahh |
| 13:10.13 | doepiedoe | ben je ook nederlander? ;) |
| 13:11.20 | doepiedoe | i was just thinking, my SD slot is defect.... |
| 13:11.33 | doepiedoe | the extended rom has dedicated flash am i right? |
| 13:12.04 | doepiedoe | if i'm right, could it be possible to make a small rootfs image, boot it, format the flash and put opie on the flash so i can still run linux on it? |
| 13:23.22 | cr2 | doepiedoe: there is 1 g3 DoC and it's partitioned in a funny m-sys way. |
| 13:24.27 | cr2 | lkcl: where do we get an (open) g3 spec ? the version from m-sys website is not enough as you know. |
| 13:24.41 | doepiedoe | hmmmm |
| 13:24.56 | doepiedoe | ahhh thats the g3 thing you talked about yesterday ;) |
| 13:28.42 | cr2 | yes. |
| 13:28.42 | doepiedoe | for some reason it crashed at mmc_scan |
| 13:28.42 | cr2 | boot with the ramdisk |
| 13:28.42 | doepiedoe | yeah then too |
| 13:28.42 | cr2 | and load the asic3_mmc driver manually. |
| 13:28.42 | doepiedoe | got 1 kernel that gets over mmc_scan (2.6.16) but doen't recognize ramdisk |
| 13:28.43 | cr2 | post the crsh message |
| 13:28.43 | doepiedoe | crsh? |
| 13:28.43 | cr2 | carsh |
| 13:28.43 | cr2 | :) |
| 13:28.43 | doepiedoe | ;) |
| 13:28.43 | cr2 | crash |
| 13:28.43 | doepiedoe | only get mmc_scan |
| 13:28.43 | doepiedoe | nothing else |
| 13:28.44 | doepiedoe | does respond on clamshell stuff |
| 13:28.44 | cr2 | crash == oops |
| 13:28.44 | doepiedoe | so it does recognize open close rotate.... |
| 13:28.44 | doepiedoe | sorry ;) |
| 13:28.44 | cr2 | does it hang ? |
| 13:28.45 | doepiedoe | there's no oops |
| 13:28.45 | doepiedoe | somelike |
| 13:28.45 | cr2 | clamshell has nothing to do with mmc |
| 13:28.45 | doepiedoe | it does hang at mmc_scan |
| 13:28.46 | cr2 | ok. |
| 13:28.46 | doepiedoe | yeah but device doesn't hang |
| 13:28.46 | doepiedoe | so only kernel hangs |
| 13:28.46 | cr2 | cat /proc/interrupts |
| 13:28.46 | doepiedoe | i cant |
| 13:28.46 | cr2 | does the clamshell counter change ? |
| 13:28.46 | cr2 | why ? |
| 13:28.47 | doepiedoe | it doesnt get into rootfs.... |
| 13:28.47 | doepiedoe | just hangs at kernel mmc_scan |
| 13:28.47 | cr2 | mount the ramdisk. |
| 13:28.47 | doepiedoe | thing it's compiled in kernel |
| 13:28.47 | doepiedoe | think. |
| 13:28.47 | doepiedoe | sd slot is defect.... |
| 13:28.47 | cr2 | then don't dow it. take the .12 kernel+ramdisk as documented in xda-dev wiki. |
| 13:28.47 | doepiedoe | yes i did |
| 13:28.47 | doepiedoe | ;) |
| 13:28.47 | doepiedoe | but that kernel tries to load the mmc_scan |
| 13:28.47 | cr2 | and it can't mount the ramdisk ? |
| 13:28.48 | doepiedoe | nope |
| 13:28.49 | cr2 | recompile the kernel |
| 13:28.51 | doepiedoe | yeah thats the only way i think |
| 13:28.55 | cr2 | ok. |
| 13:29.13 | doepiedoe | is there a precompiled toolchain available? |
| 13:29.23 | cr2 | 3.4.1 on hh.org |
| 13:29.26 | doepiedoe | ok |
| 13:30.22 | doepiedoe | and cvs on the xanadux project? or hh.org too |
| 13:31.09 | doepiedoe | got ir |
| 13:31.11 | doepiedoe | got it |
| 13:31.13 | doepiedoe | sorry |
| 13:31.14 | doepiedoe | ;) |
| 13:34.29 | lkcl | cr2: there isn't one, yet - but dwmw2 has one. so he could at least confirm that we have things correct. |
| 13:35.13 | doepiedoe | lkcl what isn't? |
| 13:35.18 | lkcl | g3 spec |
| 13:35.22 | cr2 | hmm. "intellectual property" ;-) |
| 13:35.33 | lkcl | neither of which two words is correct. |
| 13:35.49 | cr2 | lkcl: that's why it is in "" :) |
| 13:36.35 | cr2 | are you working on the g3.c file ? |
| 13:37.00 | cr2 | i suggest to replace the #3E access by a DELAY() macro. |
| 13:37.38 | cr2 | then there is a strange 1byte store of 0x1000. are you sure ? |
| 13:37.46 | cr2 | i didn't look at the asm. |
| 13:48.32 | doepiedoe | hey can't you use a driver from an other ipaq? or is this just a stupid suggestion |
| 13:48.46 | doepiedoe | e.g. hx2000 http://www.handhelds.org/moin/moin.cgi/HpIpaqHx2000 |
| 13:48.55 | doepiedoe | Processor: Intel PXA270 at 312 Mhz |
| 13:49.10 | doepiedoe | it hasnt the driver too |
| 13:49.11 | doepiedoe | ;) |
| 13:49.29 | doepiedoe | hx2000: No DiskonChip G3 driver exists in Linux but the specifications are available to write an mtd/jffs2 driver. Status of any FTL unknown for compatibility/reading of wince data. |
| 14:03.14 | lkcl | no i got to go to thingy. chinese doctor. |
| 14:04.42 | *** join/#htc-linux ekkobelli (n=michi@p54A3FF3C.dip.t-dialin.net) |
| 14:05.38 | *** part/#htc-linux ekkobelli (n=michi@p54A3FF3C.dip.t-dialin.net) |
| 14:06.13 | *** join/#htc-linux JTRipper86 (n=jtripper@p54B1CC53.dip.t-dialin.net) |
| 14:21.51 | *** join/#htc-linux alex-begin (n=345rt@84.232.212.62) |
| 14:37.52 | *** join/#htc-linux rob__w (n=bob@p85.212.193.180.tisdip.tiscali.de) |
| 14:50.54 | doepiedoe | hey lkcl, if i add ramdisk support to the kernel it should support ramdisk support true?, got an error that /dev/ram0 could not be found..... |
| 14:52.47 | doepiedoe | error: could not mount root, tried reiserfs ext3 ext2 cramfs minix msdos vfat ntfs romfs jfs |
| 15:15.45 | *** join/#htc-linux asylumed (n=insanity@196.211.28.91) |
| 15:20.51 | *** part/#htc-linux asylumed (n=insanity@196.211.28.91) |
| 15:20.53 | *** join/#htc-linux asylumed (n=insanity@196.211.28.91) |
| 16:45.51 | *** join/#htc-linux iggy_ (n=iggy@gentoo/developer/iggy) |
| 17:00.27 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
| 17:17.29 | goxboxlive | 'mount -t usbfs none /proc/bus/usb' |
| 17:27.14 | *** join/#htc-linux pleemans (n=peter@d51A5E421.access.telenet.be) |
| 17:32.47 | cr2 | goxboxlive: do you have 'lsusb' ? |
| 17:34.57 | goxboxlive | i dont kbnow cr2, i am trying to compile the usb support now in ti the kernel |
| 17:44.05 | goxboxlive | s/kbnow/know/ |
| 17:51.29 | goxboxlive | cr2: I have now booted linux and yes, i have 'lsusb' |
| 17:52.49 | cr2 | lsusb |
| 17:54.25 | goxboxlive | nothing |
| 17:56.06 | goxboxlive | root@htcuniversal:~$ lsusb -t |
| 17:56.06 | goxboxlive | cannot open /proc/bus/usb/devices, No such file or directory (2) |
| 17:59.12 | cr2 | do you have usb host ? |
| 17:59.23 | goxboxlive | how do i check? |
| 17:59.31 | cr2 | in .config |
| 18:00.40 | goxboxlive | root@htcuniversal:~$ ls -la /proc/bus/ |
| 18:00.40 | goxboxlive | dr-xr-xr-x 4 root root 0 Nov 14 23:10 . |
| 18:00.40 | goxboxlive | dr-xr-xr-x 55 root root 0 Nov 21 1996 .. |
| 18:00.41 | goxboxlive | dr-xr-xr-x 2 root root 0 Nov 14 23:16 input |
| 18:00.41 | goxboxlive | drwxr-xr-x 2 root root 0 Nov 14 23:10 usb |
| 18:01.04 | doepiedoe | wow, is USB host working? ;) |
| 18:03.06 | cr2 | OHCI HCD yes... |
| 18:03.13 | doepiedoe | nice |
| 18:03.14 | doepiedoe | ;) |
| 18:03.22 | goxboxlive | OHCI HCD support (USB_OHCI_HCD) is enabled |
| 18:03.28 | cr2 | doepiedoe: it's not what you think |
| 18:03.32 | doepiedoe | oh |
| 18:03.34 | doepiedoe | only for gsm? |
| 18:03.49 | goxboxlive | wich else should be enabled to have usb host support? |
| 18:05.33 | cr2 | USB serial. generic. |
| 18:05.42 | cr2 | doepiedoe: it's what we know. |
| 18:06.44 | cr2 | usb driver for GSM and CDMA modems ? wtf is that ? |
| 18:07.17 | goxboxlive | no idea |
| 18:07.38 | cr2 | pcmcia + usb host. a bit misleading, but worth looking at. |
| 18:07.57 | cr2 | let's start with generic. |
| 18:09.57 | cr2 | the GSM driver name is 'option.c'. great. |
| 18:11.10 | cr2 | <PROTECTED> |
| 18:11.10 | cr2 | <PROTECTED> |
| 18:11.10 | cr2 | <PROTECTED> |
| 18:13.46 | cr2 | ok. it's a hack on top of usbserial |
| 18:14.15 | doepiedoe | ahhh |
| 18:14.16 | doepiedoe | ;) |
| 18:14.34 | doepiedoe | now it's clear to me ;) |
| 18:15.20 | cr2 | we have ohci on pxa270 |
| 18:15.22 | goxboxlive | cr2: so for now, only this one: (USB_SERIAL_GENERIC) |
| 18:15.27 | cr2 | yes. |
| 18:15.33 | *** join/#htc-linux rejon (n=rejon@adsl-75-5-124-98.dsl.pltn13.sbcglobal.net) |
| 18:15.45 | goxboxlive | why dont this one ? (USB_G_SERIAL) |
| 18:15.45 | cr2 | we need our own driver it seems. |
| 18:16.13 | cr2 | because "we" are the host, not the phone. |
| 18:16.19 | goxboxlive | ok |
| 18:16.39 | *** part/#htc-linux rejon (n=rejon@adsl-75-5-124-98.dsl.pltn13.sbcglobal.net) |
| 18:20.03 | goxboxlive | Everytime i have compiled a kernel my laptop is getting so slow. Why is that? |
| 18:22.09 | psokolovsky_ | Hi! |
| 18:22.10 | cr2 | cpu and memory bandwidth, slow disk. |
| 18:22.27 | cr2 | psokolovsky_: hi |
| 18:22.52 | psokolovsky_ | goxboxlive, had a chance to build OPIE from OE.dev lately? any changes noticed? something not working? ;-) |
| 18:25.17 | cr2 | psokolovsky_: dynamic irq alloc question here. can the "generic" code cope with the "irq holes" efficiently ? |
| 18:26.45 | psokolovsky_ | cr2, what "generic" code you mean? |
| 18:26.57 | cr2 | non-hh specific |
| 18:27.16 | cr2 | imagine i have 2 cpus |
| 18:27.24 | lkcl | allo darlins |
| 18:27.28 | cr2 | one has 20 gpio irqs |
| 18:27.37 | cr2 | another has 200 |
| 18:27.41 | psokolovsky_ | cr2, standard API doesn't allow to "allocate" IRQ afaik. you request specific IRQ, voila. |
| 18:27.53 | cr2 | we put a static 200 base for both. |
| 18:28.10 | cr2 | is there a performance hit from 180 irq hole ? |
| 18:28.10 | psokolovsky_ | cr2, and as for pluggable CPUs, I have totally zero knowlegde how they work, lol ;-) |
| 18:28.51 | psokolovsky_ | cr2, allocation happens once anyway, right? so after it is done, no perf hits obviosly |
| 18:28.53 | cr2 | psokolovsky_: not everything that the mainline kernel does is smart. it has its roots in x86, and this shit shows. |
| 18:29.15 | lkcl | allo psoko. |
| 18:29.18 | cr2 | if there is a linear search of something... |
| 18:29.57 | lkcl | psoko: who were those really nice people on h3600 who did m-sys mill+ flash? i want to send them a thank you |
| 18:30.09 | psokolovsky_ | cr2, many things mainline does are plain dumb, up to the level that I like to waste my time and point them that, lol ;-) |
| 18:30.35 | psokolovsky_ | cr2, *our* (hh.org) does linear search for inital allocation. again, it happens once |
| 18:30.36 | cr2 | well, if there is no performance penalty, why don't we set the irq base at 10000 ? |
| 18:30.40 | cr2 | ok, i see :D |
| 18:31.04 | cr2 | lol |
| 18:31.37 | psokolovsky_ | cr2, ah, what is involved here is wasted memory, of course |
| 18:32.09 | psokolovsky_ | cr2, array of irq descriptors is statically allocated |
| 18:32.10 | cr2 | i've looked how SA1101 "soc" handles irqs, and it was going through the bits to find the irq line. |
| 18:32.22 | cr2 | memory is not a problem, imho. |
| 18:32.50 | cr2 | i have concerned only about the runtime performance. |
| 18:33.34 | lkcl | cr2: what you reckon that the problems we have with audio (and me with acx) are something to do with asic3 not everything being switched on or something? |
| 18:33.39 | lkcl | you said something about clocks? |
| 18:33.44 | goxboxlive | psokolovsky_ : I havent tried for a couple of days, and it worked. But i havent had any problems with it since i added the invisibility line you added in the qte-common.inc. |
| 18:33.57 | psokolovsky_ | cr2, we can have 512 IRQ, and set ASIC2 base at 256, for example. should work for any cpu, and memory waste should be decent ;-) |
| 18:34.32 | cr2 | agreed. but what is the highest CPU IRQ number now ? |
| 18:34.32 | psokolovsky_ | goxboxlive, ok |
| 18:34.41 | cr2 | lkcl: it's hima/ba specific. |
| 18:34.48 | lkcl | ah. |
| 18:35.03 | psokolovsky_ | cr2, dunno, but I guess, with 256 we should have some reserve for a year or two, no? ;-) |
| 18:35.13 | cr2 | yes, i think so. |
| 18:35.25 | cr2 | some omap must have 180 or so. |
| 18:36.03 | cr2 | btw, there are older ipaqs with two asics ? what about them ? |
| 18:36.19 | cr2 | do they have 2 SoCs ? |
| 18:36.38 | doepiedoe | i'm getting crazy! |
| 18:36.47 | cr2 | or blueangel. |
| 18:36.52 | doepiedoe | each time i disable mmc support kernel doesn't boot |
| 18:36.56 | cr2 | it has asic3 and tsc2200 |
| 18:37.12 | doepiedoe | is mmc support needed for boot of kernel on Universal? |
| 18:37.19 | cr2 | the second is not a SoC now, but should be probably a SoC. |
| 18:37.28 | psokolovsky_ | cr2, asic3 has ~70 irqs, right, so can have something at 256+128, for example, or do even more packaged assignment |
| 18:37.45 | cr2 | or asic3 + TSC2046 on hx47000. |
| 18:38.08 | cr2 | ok, let's hope there are no devices with 3 SoCs :) |
| 18:39.24 | cr2 | doepiedoe: impossible. |
| 18:39.49 | doepiedoe | ok... |
| 18:39.49 | cr2 | lkcl: let's cleanup the g3.c |
| 18:40.26 | lkcl | doepiedo: what do you mean 'doesn't boot'? does it say 'kernel panic cannot find root fs' and have you still left the line in root=/dev/mmcblk0p2 in your default.txt? |
| 18:40.59 | lkcl | did you leave boot_mmc=y in default.txt? |
| 18:41.03 | lkcl | etc. etc. |
| 18:41.05 | cr2 | lkcl: when you have said "chinese doctor", it reminded me about an old worm "hacked by chinese". |
| 18:41.10 | lkcl | it should be quite straightforward |
| 18:41.16 | cr2 | lkcl: boot_mmc is a ba hack. |
| 18:41.20 | lkcl | oh MAN the ... |
| 18:42.38 | lkcl | i'm undergoing the chinese equivalent of physiotherapy for my shoulder (my knee is better) |
| 18:42.50 | lkcl | and jeeeeezus it friggin hurts |
| 18:42.54 | *** join/#htc-linux scatterp (i=scatterp@packet.filtered.ca) |
| 18:43.24 | lkcl | he puts the needles about 40mm into my shoulder and hand to reach the points where it hurts. |
| 18:43.45 | lkcl | then digs around a bit to make it hurt some more. |
| 18:44.01 | cr2 | ;) |
| 18:44.12 | lkcl | ARGH! :) |
| 18:44.37 | lkcl | i swear people would run away if they were in the shop and heard me going ouch |
| 18:44.48 | cr2 | :) |
| 18:45.02 | goxboxlive | <6>usbcore: registered new driver usbserial_generic |
| 18:45.02 | goxboxlive | <6>drivers/usb/serial/usb-serial.c: USB Serial Driver core |
| 18:45.25 | goxboxlive | root@htcuniversal:~$ lsusb -t |
| 18:45.25 | goxboxlive | cannot open /proc/bus/usb/devices, No such file or directory (2) |
| 18:45.27 | cr2 | goxboxlive: you need the right parameters |
| 18:45.40 | goxboxlive | ok, and they are |
| 18:45.42 | cr2 | for modprobe usbserial |
| 18:45.50 | cr2 | modinfo usbserial |
| 18:46.46 | goxboxlive | i dont have usbserial as a module, it is compiled in to the kernel |
| 18:47.02 | *** join/#htc-linux BabelOued (n=Fabrice@82.238.28.28) |
| 18:47.03 | cr2 | insmod usbserial vendor=0x#### product=0x#### |
| 18:47.26 | cr2 | you should, because this BB is "vendor-specific". |
| 18:47.54 | BabelOued | hi |
| 18:47.55 | cr2 | it must be in the xda-dev "Linux on Universal" thread. |
| 18:48.28 | cr2 | BabelOued: hi. i want to correct a bug in xanadux CVS .12 kernel. can you then test it ? |
| 18:48.41 | doepiedoe | cr2, if i want to have ramdisk support for the rd ramdisk on gnulinux.biz, would Initial RAM filesystem and RAM disk (initramfs/initrd) support be enough then? |
| 18:48.48 | cr2 | yes. |
| 18:49.10 | doepiedoe | so i don't need ram disk support? |
| 18:49.39 | cr2 | you need ramdisk+initrd options. |
| 18:49.48 | doepiedoe | alrighty ;) |
| 18:50.09 | doepiedoe | but should the amount of ramdisks still be 16? |
| 18:51.04 | cr2 | yes, it does not matter. |
| 18:51.10 | cr2 | goxboxlive: http://forum.xda-developers.com/showthread.php?p=903987&highlight=usbserial#post903987 |
| 18:52.00 | BabelOued | cr2: I've only 2.6.18-hh0 tree on my computer now :( What corrected ? |
| 18:52.00 | lkcl | cr2: the trigger-type for ASIC GPIO C is different when the acx is switched on. |
| 18:52.18 | cr2 | After |
| 18:52.18 | cr2 | modprobe usbserial vendor=0x05c6 product=0x3197 |
| 18:52.18 | cr2 | root@blueangel:~# cat /proc/bus/usb/devices |
| 18:52.20 | goxboxlive | ok i recompile then |
| 18:52.33 | cr2 | lkcl: ? |
| 18:52.51 | cr2 | BabelOued: do you have working audio in .18-hh0 ?? |
| 18:54.26 | BabelOued | cr2: not for moment, i've have LCD, keys and mayve battery on evening, i've nice patch from asylumed on tsc2200, it's not finished but works |
| 18:55.27 | BabelOued | cr2: camera link from this web pages does not work http://wiki.xda-developers.com/index.php?pagename=HimalayaCamera |
| 18:56.22 | BabelOued | cr2: i want to take this one instead and extract adcm driver http://ftp.arm.linux.org.uk/pub/armlinux/people/xscale/mainstone/02-25-2005/src/kernel/patch-2.6.9-intc1 |
| 18:56.35 | cr2 | it was a patch. |
| 18:56.50 | cr2 | isn't it in some other kernel tree ? |
| 18:57.44 | cr2 | sure, if there is no better version, take it from this one. |
| 18:58.12 | cr2 | btw, where is the tsc2200 patch ? |
| 18:58.41 | asylumed | cr2: eish... are you sure you want it? |
| 18:58.50 | BabelOued | hi asylumed |
| 18:58.57 | asylumed | BabelOued: hey |
| 19:00.03 | BabelOued | asylumed: i make a test with battery this evening :) if i can compile |
| 19:00.15 | goxboxlive | cr2: |
| 19:00.17 | goxboxlive | root@htcuniversal:~$ modprobe usbserial vendor=0x05c6 product=0x3197 |
| 19:00.18 | goxboxlive | root@htcuniversal:~$ cat /proc/bus/usb/devices |
| 19:00.18 | goxboxlive | cat: /proc/bus/usb/devices: No such file or directory |
| 19:00.18 | goxboxlive | root@htcuniversal:~$ |
| 19:00.48 | goxboxlive | root@htcuniversal:~$ modinfo usbserial |
| 19:00.48 | goxboxlive | filename: /lib/modules/2.6.18-hh0/kernel/drivers/usb/serial/usbserial.ko |
| 19:00.49 | goxboxlive | author: Greg Kroah-Hartman, greg@kroah.com, http://www.kroah.com/linux/ |
| 19:00.49 | goxboxlive | description: USB Serial Driver core |
| 19:00.49 | goxboxlive | license: GPL |
| 19:00.49 | goxboxlive | vermagic: 2.6.18-hh0 mod_unload ARMv5 gcc-4.1 |
| 19:00.51 | goxboxlive | depends: |
| 19:00.53 | goxboxlive | parm: product:User specified USB idProduct (ushort) |
| 19:00.55 | goxboxlive | parm: vendor:User specified USB idVendor (ushort) |
| 19:00.57 | goxboxlive | parm: debug:Debug enabled or not (bool) |
| 19:01.19 | asylumed | BabelOued: kewl... i'm having issues getting the tsc2200 adc to work properly alas... if you get a proper reading let me know :) |
| 19:01.41 | goxboxlive | dmesg gives me: |
| 19:01.42 | cr2 | asylumed: is it a SoC driver now ? |
| 19:01.45 | goxboxlive | <6>usbcore: registered new driver usbserial |
| 19:01.45 | goxboxlive | <6>drivers/usb/serial/usb-serial.c: USB Serial support registered for generic |
| 19:01.45 | goxboxlive | <6>usbcore: registered new driver usbserial_generic |
| 19:01.45 | goxboxlive | <6>drivers/usb/serial/usb-serial.c: USB Serial Driver core |
| 19:01.45 | goxboxlive | root@htcuniversal:~$ |
| 19:02.10 | asylumed | cr2: yip yip... |
| 19:02.20 | cr2 | asylumed: great. |
| 19:02.22 | asylumed | cr2: tsc2200 and tsc2200_keys... |
| 19:02.31 | cr2 | battery ? |
| 19:02.50 | asylumed | well... it's on the todo list |
| 19:02.59 | cr2 | i've seen the OWMINT in nk.exe and i'm confused now. |
| 19:03.26 | cr2 | wince shows only very simple data, too. not the kind of data you get over owm. |
| 19:03.52 | cr2 | you have both batteries ? |
| 19:04.10 | cr2 | goxboxlive: what does lsusb say ? |
| 19:04.46 | goxboxlive | cr2: root@htcuniversal:~$ lsusb -t |
| 19:04.46 | goxboxlive | cannot open /proc/bus/usb/devices, No such file or directory (2) |
| 19:04.48 | cr2 | goxboxlive: you can also add debug=1 |
| 19:04.57 | cr2 | mount usbfs ? |
| 19:05.07 | goxboxlive | i'll try |
| 19:05.32 | cr2 | goxboxlive: it was working in .12 |
| 19:05.56 | goxboxlive | it's not working now |
| 19:06.04 | goxboxlive | where to add debug=1 |
| 19:07.25 | doepiedoe | modprobe maby? |
| 19:07.34 | goxboxlive | ok, thx |
| 19:07.34 | doepiedoe | maybe |
| 19:08.11 | goxboxlive | cr2: FATAL: Module vendor=0x05c6 not found. |
| 19:08.29 | doepiedoe | vendor is always 0x00 format i think |
| 19:08.47 | cr2 | lsusb must work independed of usbserial |
| 19:08.47 | lkcl | you criticising my rubbish tsc2200 code, asylumed? |
| 19:08.54 | doepiedoe | it's not |
| 19:09.08 | asylumed | lkcl: hehe, no... |
| 19:09.35 | asylumed | lkcl: i'm using all of the code you developed... mutated now though :) |
| 19:09.51 | cr2 | lkcl: read Documentation/spi/pxa2xx how to write a proper NSSP driver :D |
| 19:10.03 | lkcl | well you should! |
| 19:10.16 | lkcl | good god, you mean there's documentation?? |
| 19:10.28 | asylumed | cr2: isn't the generic spi overkill for this case? |
| 19:10.28 | cr2 | have a look |
| 19:10.31 | lkcl | bloody hell. burn it! |
| 19:10.31 | goxboxlive | it might be taht i am missing something in the kernel |
| 19:10.42 | cr2 | LOL. |
| 19:10.59 | cr2 | asylumed: it is. but's it a great practice :) |
| 19:11.39 | cr2 | asylumed: then you will not have the mess we have with the tsc2046 touchscreen driver ;-) |
| 19:11.47 | asylumed | cr2: doh... yeah i do need practice... but gonna use a two callbacks for now until it all works at least |
| 19:12.09 | cr2 | it's not generic, because the generic pxa2xx-spi does not support 24bit spi transfer. |
| 19:12.20 | cr2 | asylumed: of course. |
| 19:13.01 | asylumed | cr2: one last thing though... you said i should update the backlight driver... |
| 19:13.02 | asylumed | cr2: which macro? |
| 19:13.03 | *** join/#htc-linux g3gg0__ (n=g3gg0@ppp-82-135-73-127.dynamic.mnet-online.de) |
| 19:13.06 | cr2 | but you have 8bit (afaik, or was it 16), so it should not be a problem in the future. |
| 19:13.12 | doepiedoe | backlight_turn_off |
| 19:13.32 | cr2 | asylumed: 1sec |
| 19:13.38 | lkcl | cr2: it looks like i might need to change a few more ASIC params ... AH BOLLOCKS the asic3gpio d 'alt' function is set. |
| 19:15.08 | cr2 | asylumed: look here arch/arm/mach-pxa/h3900/h3900_bl.c |
| 19:15.39 | cr2 | asylumed: but it's for asic2. you should use the asic3 macros defined in ipaq-asic3. |
| 19:15.49 | cr2 | i'll tell you the initial values. |
| 19:16.21 | cr2 | you shoule setup _both_ pwm0 and pwm1 channels. |
| 19:17.01 | asylumed | k |
| 19:17.39 | cr2 | hmm. i have only the handwritten notes. |
| 19:17.46 | goxboxlive | cr2: did you get the last message? Module vendor=0x05c6 not found. |
| 19:18.42 | asylumed | cr2: aaah right... cos i'm relying on the lingering setup from wince for some values... |
| 19:18.44 | cr2 | goxboxlive: i need the full boot dmesg output. is your OHCI a module ? |
| 19:19.01 | doepiedoe | lol |
| 19:19.06 | doepiedoe | tell him to pm you ;) |
| 19:19.17 | cr2 | pastebin |
| 19:19.20 | doepiedoe | ah |
| 19:19.21 | doepiedoe | ;) |
| 19:19.54 | cr2 | asylumed: so -- |
| 19:19.58 | cr2 | pwm1 -> alt |
| 19:20.02 | cr2 | pwm0 -> alt |
| 19:20.07 | goxboxlive | cr2: no it's not a module. |
| 19:20.21 | cr2 | pwm1 & fff0 | b |
| 19:20.24 | cr2 | pwm1_cdex |
| 19:20.34 | cr2 | pwm0 & fff0 | 4 |
| 19:20.42 | cr2 | pwm0_cdex |
| 19:21.25 | cr2 | sleep 0x3e8 -> usleep 1000 |
| 19:21.44 | cr2 | pwm1+2 = 0x40 |
| 19:21.52 | cr2 | pwm0+2 = 0x200 |
| 19:22.05 | cr2 | pwm1 | 0x10 |
| 19:22.35 | cr2 | pwm0 | 0x10 |
| 19:22.35 | cr2 | that's all |
| 19:23.17 | cr2 | +2 is PeriodTime |
| 19:23.30 | asylumed | k |
| 19:23.45 | cr2 | b and 4 -> PWM_TIMEBASE_VALUE(x) ((x)&0xf) |
| 19:23.59 | cr2 | 0x10 -> PWM_TIMEBASE_ENABLE |
| 19:24.50 | cr2 | then your +4 DutyTime control is ok. |
| 19:25.34 | cr2 | goxboxlive: ok, it's not a problem. |
| 19:25.44 | cr2 | or should not be a problem. |
| 19:25.46 | goxboxlive | noe it is a module |
| 19:25.56 | cr2 | then load it. |
| 19:26.12 | goxboxlive | i have also enabled debus message |
| 19:26.26 | cr2 | ok. will produce a lot of messages. |
| 19:28.44 | goxboxlive | i'll use pastebin |
| 19:31.30 | goxboxlive | <7>2005 April 22 USB 1.1 'Open' Host Controller (OHCI) Driver (pxa27x)<7>block sizes: ed 64 td 64 |
| 19:31.30 | goxboxlive | <7>In ohci_hcd_pxa27x_drv_probe<6>usbcore: registered new driver usbserial |
| 19:31.30 | goxboxlive | <6>drivers/usb/serial/usb-serial.c: USB Serial support registered for generic |
| 19:31.30 | goxboxlive | <6>usbcore: registered new driver usbserial_generic |
| 19:31.30 | goxboxlive | <6>drivers/usb/serial/usb-serial.c: USB Serial Driver core |
| 19:32.11 | cr2 | lsusb ? |
| 19:33.08 | goxboxlive | root@htcuniversal:~$ ls -la /proc/bus/usb/ |
| 19:33.08 | goxboxlive | drwxr-xr-x 2 root root 0 Nov 14 23:11 . |
| 19:33.08 | goxboxlive | dr-xr-xr-x 4 root root 0 Nov 14 23:11 .. |
| 19:33.08 | goxboxlive | root@htcuniversal:~$ lsusb |
| 19:33.09 | goxboxlive | root@htcuniversal:~$ lsusb -t |
| 19:33.09 | goxboxlive | cannot open /proc/bus/usb/devices, No such file or directory (2) |
| 19:33.54 | cr2 | mount usbfs ? |
| 19:34.04 | goxboxlive | yes it's mounted |
| 19:34.14 | cr2 | umount and mount again |
| 19:34.40 | goxboxlive | but it doesnt show up in 'df' |
| 19:35.17 | cr2 | cat 7proc/mounts |
| 19:35.20 | cr2 | cat /proc/mounts |
| 19:35.58 | goxboxlive | root@htcuniversal:~$ cat /proc/mounts |
| 19:35.58 | goxboxlive | rootfs / rootfs rw 0 0 |
| 19:35.58 | goxboxlive | /dev/root / ext3 rw,data=ordered 0 0 |
| 19:35.58 | goxboxlive | proc /proc proc rw 0 0 |
| 19:35.58 | goxboxlive | sysfs /sys sysfs rw 0 0 |
| 19:36.00 | goxboxlive | /dev/root /dev/.static/dev ext3 rw,data=ordered 0 0 |
| 19:36.02 | goxboxlive | tmpfs /dev tmpfs rw 0 0 |
| 19:36.04 | goxboxlive | /dev/mmcblk0p3 /media/mmcblk0p3 ext3 rw,data=ordered 0 0 |
| 19:36.06 | goxboxlive | /dev/mmcblk0p4 /media/mmcblk0p4 ext3 rw,sync,data=ordered 0 0 |
| 19:36.08 | goxboxlive | /dev/mmcblk0p1 /media/mmcblk0p1 vfat rw,sync,fmask=0022,dmask=0022,codepage=cp850,iocharset=iso8859-15 0 0 |
| 19:36.10 | goxboxlive | devpts /dev/pts devpts rw 0 0 |
| 19:36.12 | goxboxlive | tmpfs /var tmpfs rw 0 0 |
| 19:36.14 | goxboxlive | tmpfs /media/ram tmpfs rw 0 0 |
| 19:36.16 | goxboxlive | usbfs /proc/bus/usb usbfs rw 0 0 |
| 19:36.33 | cr2 | looks good. |
| 19:37.02 | cr2 | but your ohci should spew a lot of messages about found and created busses , etc. |
| 19:37.38 | goxboxlive | it doesnt |
| 19:37.58 | cr2 | then we need more complex setup in .18 |
| 19:38.14 | goxboxlive | ok |
| 19:38.24 | cr2 | platform_device ohci... |
| 19:38.26 | goxboxlive | ok |
| 19:38.41 | goxboxlive | It said that it was an unkown vendor |
| 19:38.59 | goxboxlive | maybe i have forgoten something in the .config |
| 19:40.35 | goxboxlive | cr2: is it working on the hx4700? |
| 19:43.00 | cr2 | hx4700 does not have usb host |
| 19:43.17 | cr2 | better, does not need. |
| 19:44.51 | goxboxlive | :-) |
| 19:45.17 | cr2 | asus a730 |
| 19:45.32 | goxboxlive | cr2 why does my bandsel in cu show me that it is connected to UMTS? |
| 19:45.34 | *** join/#htc-linux Fossi (n=Fossi@e176098063.adsl.alicedsl.de) |
| 19:47.36 | cr2 | how do you init the phone ? |
| 19:47.49 | goxboxlive | what do you mean? start cu? |
| 19:48.27 | goxboxlive | 'cu -l /dev/ttyS0 -s 115200' |
| 19:48.36 | cr2 | no. |
| 19:48.49 | cr2 | you should send ATZ;and so on after the setup. |
| 19:48.56 | cr2 | after the connect. |
| 19:49.08 | cr2 | magician has ohci. for bluetooth. |
| 19:49.45 | goxboxlive | 'at+cfun=1' and at+cops=0 |
| 19:49.52 | goxboxlive | then i am up and go |
| 19:49.58 | doepiedoe | at+&t |
| 19:50.21 | doepiedoe | doesn't it show UPS instead of UMTS? :) |
| 19:50.26 | goxboxlive | i have disabled the pin code |
| 19:50.26 | goxboxlive | i'll try |
| 19:51.38 | cr2 | then you use some hardcoded BB values |
| 19:52.11 | goxboxlive | uucp@htcuniversal:/home/root$ cu -l /dev/ttyS0 -s 115200 |
| 19:52.12 | goxboxlive | Connected. |
| 19:52.12 | goxboxlive | AT-Command Interpreter ready |
| 19:52.12 | goxboxlive | +BANDSEL: 2,1 |
| 19:52.12 | goxboxlive | 0 |
| 19:52.32 | goxboxlive | and since it tells me "2" then it is UMTS |
| 19:53.41 | cr2 | yes. |
| 19:55.00 | goxboxlive | doesnt that say that UMTS is working? |
| 19:55.37 | cr2 | don't know. you should add AT+CREG=2 |
| 19:55.49 | cr2 | and the CSQ/HTCCSQ |
| 19:56.05 | cr2 | and watch the signal level, LAC data. |
| 19:56.35 | cr2 | i have made some changes to the touchscreen driver and would like to commit them. |
| 19:56.45 | goxboxlive | +CREG: 2,5 |
| 19:56.56 | goxboxlive | ok |
| 19:56.57 | cr2 | can you recompile and test the kernel ? |
| 19:57.10 | doepiedoe | cr2, can this changes cause the problem of linux not booting? |
| 19:57.11 | cr2 | ohci needs some additional code. |
| 19:57.17 | goxboxlive | yes, |
| 19:57.32 | cr2 | doepiedoe: no, they are mainly cosmetic. |
| 19:58.06 | doepiedoe | alright :) |
| 19:58.12 | goxboxlive | Do you know if there is any command to exit 'cu' ? I allways use killall cu |
| 19:58.26 | *** join/#htc-linux lamikr (n=chatzill@aragorn.kortex.jyu.fi) |
| 19:58.50 | doepiedoe | ctrl+]? |
| 19:59.05 | cr2 | ok. done. |
| 19:59.10 | doepiedoe | or ctrl+c |
| 19:59.24 | cr2 | ask goxbxolive |
| 20:09.05 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
| 20:17.58 | goxboxlive | cr2: I am booting now. What do you want me to test? |
| 20:20.54 | lkcl | hey is it worthwhile me putting in polling/print-out of the alt status etc. of the asic3? |
| 20:21.00 | Speedy2 | Hey all. |
| 20:21.06 | lkcl | hiii speedy2 |
| 20:21.34 | lkcl | into gnuharet? |
| 20:21.59 | lkcl | or, cr2, shall i do more asm -> c |
| 20:22.08 | cr2 | goxboxlive: touchscreen |
| 20:22.19 | cr2 | lkcl: ALT is already there ? |
| 20:22.28 | lkcl | not in the wasic3gpio it isn't |
| 20:22.34 | goxboxlive | it works |
| 20:23.33 | cr2 | lkcl: you have masked it |
| 20:23.46 | cr2 | goxboxlive: great. |
| 20:24.05 | cr2 | lkcl: let's fix the asm->c bugs |
| 20:24.11 | lkcl | cr2: okie |
| 20:24.53 | cr2 | <PROTECTED> |
| 20:24.53 | cr2 | i have found one more register. |
| 20:24.53 | cr2 | already in wiki. +1042 |
| 20:28.58 | BabelOued | cr2: can i read battery value under wince with haret ? |
| 20:30.59 | lkcl | ack |
| 20:31.00 | cr2 | BabelOued: dump power ? no. 'powermon' |
| 20:32.12 | BabelOued | cr2: ok thanks.i take latest gnuharet |
| 20:42.46 | doepiedoe | DiskOnChip Millenium+ (aka G2) works on h4000 |
| 20:42.54 | doepiedoe | so this G3 is just the next generation? |
| 20:44.52 | lkcl | just setting up some macros for this stuff, cr2. |
| 20:45.06 | cr2 | ok. |
| 20:45.36 | *** part/#htc-linux WizMaui (n=WizMaui@62.112.90.179) |
| 20:46.01 | alex-begin | seen the new HTC P4350 |
| 20:46.02 | alex-begin | :)) |
| 20:48.08 | lkcl | EOR R8, R8, R8 |
| 20:48.09 | lkcl | wtf??? |
| 20:48.22 | lkcl | why not just put R8 = 0 ?? |
| 20:50.21 | BabelOued | cr2: powermon see only one battery power |
| 20:51.13 | cr2 | both. |
| 20:51.30 | cr2 | lkcl: it's the compiler. |
| 20:51.33 | BabelOued | i've this line |
| 20:51.33 | lkcl | joy |
| 20:51.38 | BabelOued | 494824 8 81 % 4030 0 0 0 1 3 -1 -1 100 1 100 1 -1 -1 0 0 1397 4 |
| 20:51.45 | cr2 | 1397 |
| 20:52.01 | cr2 | is the second battery. |
| 20:52.36 | cr2 | looks like 2 ADCs, and nothing else. some hack job for the 81%. |
| 20:53.01 | cr2 | i wonder how they generate the colour bars in wince ;) |
| 20:53.11 | BabelOued | ok i've downloaded datasheet and look into it |
| 20:53.30 | cr2 | lkcl said there were bugs there. |
| 20:54.07 | lkcl | bugs in where, dude? |
| 20:54.15 | cr2 | goxboxlive: http://www.earth.li/pipermail/e3-hacking/2006-September/000601.html |
| 20:54.28 | cr2 | lkcl: tsc2200 datasheet TI |
| 20:54.33 | lkcl | oh yes. there are. |
| 20:55.02 | lkcl | if you look at the data-exchange, the bastards send less bytes than the number that they read |
| 20:55.52 | lkcl | what the little fuckers at TI told me was 'oh, if you KNEW how SSP works (you ignorant little shit) then you would KNOW that you have to send an extra word 0x0000 to 'pad out' the data to the right length' |
| 20:56.25 | cr2 | :) |
| 20:56.31 | lkcl | cr2: well-remembered. that was nearly 3 years ago |
| 20:56.48 | cr2 | yeah, even i remember it :) |
| 20:56.55 | goxboxlive | cr2: Are you able to get it in to our kernel? |
| 20:57.25 | cr2 | we need to find out how to calculate the clock on our non-omap machine. |
| 20:58.41 | cr2 | i have written an e-mail to the ovcamchip maintainer, but did not get any reply. |
| 20:58.42 | cr2 | imho, we should try to put it into ovcamchip |
| 20:58.42 | cr2 | and the ov9650 too. |
| 20:58.42 | cr2 | btw, there is a capture program too. |
| 20:58.53 | cr2 | so it'l possible to test it. |
| 20:58.57 | lkcl | <PROTECTED> |
| 20:58.57 | lkcl | what that mean, where R12=0xa0000000 |
| 20:58.58 | lkcl | ? |
| 20:59.16 | cr2 | combined with the i2c logs it should be easy to fix the driver. |
| 20:59.33 | cr2 | copy R5 to RAM. |
| 20:59.38 | Speedy2 | I think xor rx,rx,rx is the fastest way to clear out a register in ARM |
| 21:00.03 | cr2 | lkcl: not only on arm. |
| 21:00.04 | lkcl | does it also mean 'advance r12 by 2'? |
| 21:00.11 | lkcl | cool |
| 21:00.25 | lkcl | because it's in a loop - i think i found the area which reads from memory |
| 21:00.27 | cr2 | lkcl: it's (post?) increment |
| 21:00.32 | lkcl | that figures. |
| 21:00.42 | goxboxlive | ok, let me now when it's ready, i'll try it out. |
| 21:00.45 | cr2 | you should keep in mind what this code does. |
| 21:01.00 | lkcl | i think i found the area which does the actual copy from the read-register - 0x40 |
| 21:01.03 | cr2 | there is the SPL in flash. |
| 21:01.34 | cr2 | 512byte blocks (0x200) interleaved (two copies) |
| 21:01.49 | cr2 | you should copy 256K from flash to RAM |
| 21:02.11 | cr2 | is two blocks are not the same, there is a bug, i don't know what they do. |
| 21:02.24 | cr2 | maybe read the first, if it fails, read the second. |
| 21:02.43 | lkcl | all very odd. |
| 21:03.21 | cr2 | i think mickeysoft does the same on xbox. |
| 21:03.51 | cr2 | hmm. maybe this 0x40 is also good for writes... |
| 21:04.10 | cr2 | yes, and the SPL is located on a "binary partition". |
| 21:04.41 | cr2 | it has "BIPO" BInary Partition something signature. |
| 21:04.49 | lkcl | i think... i think... there must be some sort of 16-byte checksum on every 512 bytes. or something |
| 21:05.00 | cr2 | it seems they go through the flash and seek for it. |
| 21:05.22 | cr2 | there is ECC builtin in the chip ASIC afaik. |
| 21:05.22 | lkcl | and if something goes wrong, they go back and do the partition again. |
| 21:05.25 | lkcl | oooo |
| 21:05.47 | cr2 | if something goes wrong, you get a special interrupt status. |
| 21:05.55 | cr2 | ECC error has its own number. |
| 21:07.29 | goxboxlive | cr2: later in the email, there where one who had ported the cam driver to .17 |
| 21:07.33 | *** join/#htc-linux florian_ (n=fuchs@84.245.168.106) |
| 21:07.46 | cr2 | *((char*)(R1 + 0x1E) = 0x1000; /* unknown reg */ |
| 21:07.53 | cr2 | lkcl: this is strange. |
| 21:08.04 | cr2 | goxboxlive: .17 ? |
| 21:08.37 | florian | re |
| 21:09.00 | goxboxlive | cr2: yes |
| 21:09.30 | goxboxlive | http://www.earth.li/pipermail/e3-hacking/2006-September/000608.html |
| 21:10.23 | goxboxlive | http://www.earth.li/pipermail/e3-hacking/2006-September/000612.html |
| 21:10.46 | cr2 | yes, it's a minor patch and it's omap-specific. |
| 21:13.56 | lkcl | cr2: yes it is :) it's STRB R1, [R1, #1E] |
| 21:15.14 | lkcl | yep - STRB R1, [R1,#0x1E] |
| 21:15.15 | cr2 | weird. but then it's empty ?? |
| 21:15.23 | lkcl | gotta love it. |
| 21:15.55 | cr2 | is #1E used at another place ? |
| 21:18.38 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
| 21:20.14 | cr2 | yes, once more. it looks like a macro. |
| 21:20.27 | cr2 | probably a bug :) |
| 21:21.45 | lkcl | another quick update |
| 21:21.47 | lkcl | yes it is |
| 21:22.49 | cr2 | lkcl: they use 0 in your g4 :) |
| 21:22.55 | lkcl | just got to the bottom |
| 21:23.01 | lkcl | hooray, i am so happy. |
| 21:23.13 | cr2 | MOV R2, #0 |
| 21:23.20 | cr2 | STRB R2, [R1,#0x1E] |
| 21:23.28 | cr2 | STRB R0, [R1,#0x3E] |
| 21:23.46 | cr2 | can you post the file ? |
| 21:23.52 | lkcl | same place. |
| 21:24.00 | lkcl | http://hands.com/~lkcl/hp6915/g3.c |
| 21:24.04 | cr2 | ok. |
| 21:24.49 | cr2 | SETB(0xA, 0x5); /* another device id select? */ |
| 21:24.53 | cr2 | it is device select |
| 21:26.19 | cr2 | g3_delay(0x39); |
| 21:26.31 | cr2 | it's probably not what you meant. |
| 21:26.47 | lkcl | ok |
| 21:27.12 | lkcl | 'another device id select' was when you had a bug in the wiki page |
| 21:28.28 | cr2 | yes. |
| 21:28.38 | cr2 | SETW(0x1A, 0x1042); /* device select */ |
| 21:28.55 | cr2 | it's read address here. |
| 21:34.41 | cr2 | they look for BIPO by reading from 0x3a. |
| 21:36.22 | lkcl | ack |
| 21:36.52 | cr2 | hmm. but copy to RAM from 0x40 |
| 21:37.39 | cr2 | R0 = 0x210 |
| 21:37.57 | cr2 | there are 10 extra bytes per block ? |
| 21:38.01 | cr2 | for ECC ? |
| 21:40.34 | cr2 | uuh. g4 has some extra regs. #0x50 |
| 21:41.15 | lkcl | <PROTECTED> |
| 21:41.17 | lkcl | que?? |
| 21:42.27 | lkcl | arg i removed something by mistake |
| 21:44.27 | cr2 | hmm. why R5 = GETW(0x40); ? |
| 21:44.41 | lkcl | 1sec... |
| 21:45.25 | cr2 | on g4 it's 0x3C, and here it should be 0x38+2 = 0x3a ? |
| 21:46.02 | cr2 | ok, g4 is more complicated. |
| 21:46.37 | lkcl | because R7 = R1 + 0x38 |
| 21:47.00 | lkcl | then, it was actually STRH R5, [R7, #2] |
| 21:47.15 | cr2 | yes, 0x38+2=0x3a |
| 21:47.18 | lkcl | so i just turned it back into sensible code. i hope. |
| 21:47.21 | lkcl | oops! |
| 21:47.32 | cr2 | :) |
| 21:48.04 | cr2 | i was surprised because there is a direct write into 0x40 |
| 21:48.13 | lkcl | ok that's it |
| 21:48.46 | lkcl | that's the lot. |
| 21:48.53 | cr2 | g4 has a bit more. |
| 21:49.11 | cr2 | mov r6, |
| 21:49.20 | cr2 | there are a lot of such strings. |
| 21:49.59 | cr2 | the copy is from 0x800 area. it's logical. |
| 21:50.15 | cr2 | not we need to understand where is the page selector. |
| 21:50.26 | cr2 | 0x36 ? |
| 21:50.51 | cr2 | because of that: |
| 21:50.54 | cr2 | big_wr8_R8_to_0x36_wr16_0x800_to_0x1a |
| 21:52.37 | cr2 | ok. now all reads go from 0x3a. undocumented register:) thank you msys ;) |
| 21:53.58 | cr2 | now why 0x210 for 0x200 bytes. ECC ? i have read something about data layout on the flash in some msys docs. |
| 21:54.58 | cr2 | lkcl: SETW(0x1A, 0x1042); /* device select */ |
| 21:55.02 | lkcl | 1sec... |
| 21:55.02 | cr2 | this is address select |
| 21:56.06 | cr2 | it seems 0x3a read register also autoincrements. |
| 21:56.34 | cr2 | <PROTECTED> |
| 21:56.38 | cr2 | <PROTECTED> |
| 21:56.38 | cr2 | <PROTECTED> |
| 21:56.38 | cr2 | <PROTECTED> |
| 21:56.39 | cr2 | <PROTECTED> |
| 21:56.39 | cr2 | <PROTECTED> |
| 22:04.02 | cr2 | lkcl: http://www.pi-sync.net/html/kp_pi.html |
| 22:06.18 | lkcl | sorry am back |
| 22:06.36 | lkcl | o god. really? |
| 22:07.00 | cr2 | they say it does not really work. |
| 22:07.10 | lkcl | what's that about, dude? |
| 22:07.53 | lkcl | what, kphone doesn't work? |
| 22:08.26 | cr2 | on E3 device that has the same small camera like universal. |
| 22:09.08 | cr2 | i suggest you dump all addresses from the bottom of g3, and use the full address in macros. (not the offset). |
| 22:09.26 | cr2 | it will be more readable. |
| 22:09.49 | cr2 | still, looks great :) |
| 22:10.13 | cr2 | i wonder if we can read (write?) the whole flash in this way. |
| 22:12.57 | *** join/#htc-linux booba (n=booba@lns-bzn-47f-81-56-252-135.adsl.proxad.net) |
| 22:16.37 | cr2 | hmm. and this code does not touch the dma registers. |
| 22:16.47 | doepiedoe | hey guys, you made this code by anazing the asm of original bootloader? |
| 22:17.01 | doepiedoe | analyzeing |
| 22:17.03 | doepiedoe | ;) |
| 22:17.07 | doepiedoe | analyzing |
| 22:17.53 | cr2 | yes. |
| 22:22.25 | cr2 | lkcl: should i clean the code ? |
| 22:26.51 | doepiedoe | wow thats really nice;) |
| 22:35.26 | lkcl | sure cr2 go for it |
| 22:36.41 | lkcl | doepiedoe: it's one thing i'm happy to do if propietary companies are going to be fucking stupid. |
| 22:36.41 | cr2 | ok. |
| 22:37.18 | lkcl | god i'm gonna love to be able to see the look on the face of that bozo from global locate if we do the same thing with their gps chip. |
| 22:37.32 | cr2 | LOL!!!!!!!!!! |
| 22:37.42 | doepiedoe | lol |
| 22:37.42 | doepiedoe | :D |
| 22:37.45 | doepiedoe | thats awefull :P |
| 22:38.12 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
| 22:38.19 | lkcl | he actually told me 'no there's no need for you to reverse-engineer anything because we're going to provide a [proprietary] program for you to run on linux' |
| 22:38.35 | doepiedoe | ;) |
| 22:39.24 | cr2 | usually that means that there is not much information there to hide. |
| 22:39.40 | doepiedoe | i don't get it |
| 22:39.53 | lkcl | ok. i gotta try setting ALT functions on asic3 gpio |
| 22:40.08 | cr2 | lkcl: what for ? |
| 22:40.12 | lkcl | cr2: i don't think he had much of a clue on how things worked in order to be able to say that |
| 22:40.15 | lkcl | acx100 |
| 22:40.19 | doepiedoe | if company's would know how much more money they can make if they just release linux drivers for their products.... |
| 22:40.45 | doepiedoe | but hey, where's the challenge then ;) |
| 22:40.48 | lkcl | when i checked a PD of the asic3 memory area, some of the ALT bits were set |
| 22:41.08 | lkcl | oh no, there's nooo money in linux: it's all free, right? |
| 22:41.17 | doepiedoe | yeah ;) |
| 22:41.34 | doepiedoe | but if a company releases the source for linux..... |
| 22:41.36 | lkcl | no, it's to do with their investors, doepiedoe - 'intellectual property' and all that |
| 22:41.48 | doepiedoe | yeah i know |
| 22:41.50 | lkcl | then they would 'lose their intellectual property' and the investors would get upset. |
| 22:41.55 | doepiedoe | but investors should know better ;) |
| 22:42.04 | lkcl | they obviously haven't looked at zope's business model |
| 22:42.10 | lkcl | anyway. |
| 22:42.11 | lkcl | enough. |
| 22:42.14 | doepiedoe | yes |
| 22:42.14 | doepiedoe | ;) |
| 22:55.56 | *** join/#htc-linux psokolovsky_ (n=psokolov@237.usernat.ip.net.ua) |
| 22:57.19 | cr2 | lkcl: cleaned. ~/g3.c |
| 22:57.28 | lkcl | ack |
| 22:58.52 | lkcl | ok i make it more c-like |
| 22:58.55 | lkcl | loops n stuff |
| 22:59.08 | lkcl | anyone want to risk running this? :) |
| 22:59.20 | doepiedoe | me |
| 22:59.21 | doepiedoe | ;) |
| 22:59.23 | cr2 | on blueangel :) |
| 22:59.26 | doepiedoe | ow |
| 22:59.27 | doepiedoe | ;) |
| 22:59.28 | lkcl | ha ha. |
| 22:59.35 | doepiedoe | too bad, don't have a BA :D |
| 22:59.35 | cr2 | it's read. |
| 22:59.40 | cr2 | not write. |
| 22:59.42 | lkcl | actually - i have a BA - one with a broken screen. |
| 22:59.42 | doepiedoe | till it writes |
| 22:59.43 | doepiedoe | ;) |
| 22:59.50 | doepiedoe | me2 |
| 22:59.51 | doepiedoe | :) |
| 22:59.57 | cr2 | we don#t know yet how to write. |
| 23:00.05 | lkcl | exactly what is the difference between read and write? |
| 23:00.05 | doepiedoe | it's the same lik read |
| 23:00.11 | doepiedoe | replace read with write |
| 23:00.11 | doepiedoe | ;) |
| 23:00.16 | cr2 | i guess :) |
| 23:00.25 | lkcl | can we ha ha dopiedoe make it do writes by mistake? |
| 23:00.29 | lkcl | what's the chances? |
| 23:00.39 | doepiedoe | 50:50 |
| 23:00.40 | doepiedoe | :D |
| 23:00.45 | cr2 | the parameters may be different. |
| 23:00.49 | doepiedoe | it does, or it does not ;) |
| 23:01.02 | cr2 | there are some magic numbers written to magic places. |
| 23:01.11 | doepiedoe | and it creates a magic crash? |
| 23:01.18 | cr2 | it is possible to run 'wi' with itsutils! |
| 23:01.35 | doepiedoe | itsutils from itsme? |
| 23:01.40 | cr2 | then we may even see the dma. |
| 23:01.47 | doepiedoe | did you ever heard of devmem2? |
| 23:02.03 | cr2 | yes. reuse the functions itsutils.dll |
| 23:02.26 | cr2 | ttfs* |
| 23:02.34 | cr2 | for raw reads. |
| 23:02.40 | doepiedoe | ftp://ftp.buici.com/pub/arm/devmem/devmem2.c |
| 23:02.52 | doepiedoe | For situations where system software needs to access hardware resources and there isn't time or a practical justification for writing a driver, the /dev/mem<//tt> give user-mode applications access to the physical address space of the machine. |
| 23:03.00 | cr2 | it's for RAM. |
| 23:03.13 | doepiedoe | http://wiki.buici.com/wiki/Using_/dev/mem |
| 23:03.15 | doepiedoe | no |
| 23:03.16 | cr2 | yes, you can read IPL (0x0-0x800) with it in linux. |
| 23:03.34 | cr2 | but not the rest. |
| 23:03.44 | doepiedoe | IPL==bootloader? |
| 23:03.52 | cr2 | 1stage bootloader |
| 23:03.54 | doepiedoe | ok |
| 23:03.56 | cr2 | SPL second stage |
| 23:04.14 | doepiedoe | i read somewhere that bootloader gets overwrittes by ce at the last 10% |
| 23:04.16 | doepiedoe | is this true? |
| 23:04.28 | cr2 | no. |
| 23:04.51 | doepiedoe | ;) then it should be a broodje aap verhaal :) |
| 23:05.12 | cr2 | wince trashes the bootloader in RAM by putting te OHCI at 0x14000 |
| 23:05.16 | cr2 | on universal |
| 23:05.29 | doepiedoe | aha |
| 23:07.58 | cr2 | lkcl: we can modify g3.c to allocate the 256K buffer and operate on /dev/mem, and save the buffer. |
| 23:07.58 | lkcl | yep. |
| 23:07.59 | lkcl | o dis is gonna be fun... |
| 23:07.59 | cr2 | if you'll find your SPL in the buffer, then it works. |
| 23:08.06 | lkcl | remember hw6915 has g4 |
| 23:08.16 | cr2 | yes. |
| 23:08.29 | cr2 | but you have the file, and the experience now :) |
| 23:08.40 | cr2 | most of the regisetrs are the same. |
| 23:08.54 | cr2 | at least the documented are the same. |
| 23:09.21 | cr2 | in ~/paq69 |
| 23:10.29 | lkcl | o right |
| 23:10.32 | lkcl | ha ha v funny |
| 23:12.54 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
| 23:17.45 | cr2 | there are 2 "read" regs 0x3c and 0x50. magic numbers are different. the structure is the same. |
| 23:18.22 | lkcl | ok. |
| 23:19.06 | lkcl | can you take a look at loc_37C and tell me how i'm supposed to turn that into c? i don't get it |
| 23:19.32 | lkcl | i think it's a for loop |
| 23:20.20 | lkcl | nuts. i'm missing a } |
| 23:21.58 | lkcl | no i'm not. |
| 23:23.10 | cr2 | spaghetty programming at its finest :) |
| 23:25.15 | lkcl | *sigh* |
| 23:40.55 | *** join/#htc-linux lamikr (n=chatzill@aragorn.kortex.jyu.fi) |
| 23:42.36 | lkcl | LSL << 31 moving all bits but bit 0 out the way, to test bit 0 equal to zero. clever |
| 23:44.52 | doepiedoe | hey guys |
| 23:44.57 | doepiedoe | i'm gonna sleep :) |
| 23:45.38 | *** part/#htc-linux doepiedoe (n=doepiedo@84-53-96-51.fiber.unet.nl) |