00:06.38 | lkcl | one wire at a time.... bringing the house down :) |
00:06.43 | lkcl | mwahahahah |
01:36.21 | *** join/#htc-linux ljp (n=lpotter@203.94.178.46) |
02:03.24 | *** join/#htc-linux LunohoD (n=alex@e180125093.adsl.alicedsl.de) |
03:29.02 | *** join/#htc-linux ellisway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
05:33.18 | *** join/#htc-linux parmaster (i=par@dipole.idlepattern.com) |
05:35.42 | parmaster | !statusofblueangel |
05:56.09 | *** join/#htc-linux rob_w (n=bob@p85.212.150.69.tisdip.tiscali.de) |
06:56.54 | *** join/#htc-linux rob_w (n=rob_w@p549B938F.dip0.t-ipconnect.de) |
07:21.20 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
07:22.40 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
07:40.41 | *** join/#htc-linux RoEn_PC (n=roen@p54A674DF.dip.t-dialin.net) |
08:45.32 | *** join/#htc-linux pH5 (n=ph5@e178203168.adsl.alicedsl.de) |
10:46.09 | *** join/#htc-linux Fossi (n=cmewes@195.177.48.115) |
10:50.46 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
10:52.47 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
10:56.48 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
10:57.57 | *** join/#htc-linux TripleDES (n=sergio@89.129.136.254) |
10:57.59 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
10:58.06 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
11:03.17 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
12:10.34 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
13:00.24 | *** join/#htc-linux psokolovsky_ (n=psokolov@237.usernat.ip.net.ua) |
13:03.56 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
13:37.39 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
14:23.24 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
14:35.53 | drno | cr2: just a short status update, Ive been fidling around with my hw6915 and got the kernel to startup, but unable to find the SDcard (I found out that lkcl had the same problem on the online IRC logs) but Ill try the NFS setup next time. So just to let you know that Im working on it |
14:36.06 | lkcl | drno: hello |
14:36.17 | lkcl | i got a small patch to stop that happening |
14:36.21 | lkcl | gimme 1sec... |
14:36.24 | drno | ah k |
14:37.09 | lkcl | psokolovsky wouldn't let me cvs commit my hack-patches :) |
14:37.24 | drno | the thing is, that he detects the card, and even knows about its partitions, but then he still gives a kernel panic |
14:37.27 | drno | haha |
14:37.45 | drno | isnt that why they invented cvs? ^_^ |
14:38.02 | psokolovsky_ | lkcl, yep, if they affect any common code ;-) send them to mainline instead ;-P |
14:38.05 | lkcl | he he |
14:38.15 | lkcl | *cackle* |
14:39.37 | lkcl | http://hands.com/~lkcl/hp6915/mmc_block.c.patch |
14:39.43 | lkcl | yes i know. |
14:40.07 | lkcl | it's to do with the mmc read-only detection being ignored on mini-sd cards. |
14:40.25 | lkcl | there's no read-only tag so it can't be detected, can it?? :) |
14:41.06 | psokolovsky_ | lkcl, is that *same* issue which was 2 months ago?! you really should drop prophets and other stuff and finish sable in *nice* way ;-) |
14:41.12 | lkcl | unfortunately, the mmc hardware on the sable it's set permanently to read-only |
14:41.16 | drno | true :) |
14:41.28 | lkcl | oh, alll riiight theen |
14:41.59 | drno | thanks lkcl Ill try that patch in a minute |
14:43.28 | lkcl | psoko: i'll get onto it. |
14:44.18 | Fossi | nah, you should drop everything else and help us get the prophet running ;p |
14:45.32 | psokolovsky_ | lkcl, I believe, we pretty well discussed, that non-RO patch should be applied to asic3_mmc, not to anything else, and in such manner, so it was possible to specify that feature via platform_data. so, why can't you do just that? ;-) |
14:45.43 | lkcl | yeh i'm going to. |
14:46.52 | lkcl | i need _some_ sort of callback... |
14:47.41 | lkcl | this is going to involve a lot of files.... |
14:48.13 | lkcl | it's a chain, like that pxa serial port configure function... |
14:51.57 | psokolovsky_ | lkcl, not many, just 2 (plus a header), maybe. and yep, doing something right is a bit harder then just do. just in case, it's in my todo too, I just wait when I'll be ready to tackle h3900 asic3_mmc support |
14:54.58 | lkcl | psokolovsky: i was going to trace it through all the way to a hermes-specific file, htchermes/hermes_mmc.c - or just a setup function which is called from htchermes/hermes.c |
14:55.54 | lkcl | mmc_block.c mmc_blk_readonly needs to call a function to test read-only status. |
14:56.31 | lkcl | and that goes through to the host-specific data struct, so the struct mmc_host needs to contain a callback function. |
14:56.50 | lkcl | then, asic3_mmc.c and all other *_mmc.c need to set up that function. |
14:57.05 | lkcl | _then_ the devices (e.g. hermes) need to set up _that_ function. |
14:57.49 | lkcl | ok. mmc_host_ops. |
14:58.39 | lkcl | huh. |
14:58.39 | BabelO | lkcl: ok :) need to do the same for blueangel |
14:58.40 | lkcl | there's a get_ro function already. |
14:58.40 | lkcl | oh, yes - i patched that already, i remember now. ha ha |
14:59.21 | BabelO | i read a kernel discuss about this already |
14:59.55 | lkcl | yep. |
15:00.06 | lkcl | ok that's host ops, not card ops. |
15:00.08 | lkcl | argh. |
15:00.12 | cr2 | lkcl: wtf that has to do with hermes ? |
15:00.20 | cr2 | hi guys |
15:00.25 | BabelO | hi cr2 |
15:00.29 | lkcl | cr2: sd card read-only status bit is inverted / ignored. |
15:00.40 | cr2 | on sable |
15:00.41 | cr2 | not hermes. |
15:00.42 | lkcl | mmc_block_readonly() always returns ... |
15:00.44 | lkcl | oops :) |
15:00.54 | lkcl | oh yeh, you're right ha ha |
15:01.43 | lkcl | ah _ha_. |
15:02.02 | lkcl | mmc_get_ro comes from the host fn... |
15:02.14 | lkcl | and is set through to the card simply as status info... |
15:02.30 | lkcl | ok i think i got it. |
15:06.58 | cr2 | on ba it's more or less easy to override, because the ro pin is on asic3. |
15:08.50 | drno | lkcl: ok, thx for the patch, it continus to boot now from the SD card |
15:08.54 | lkcl | hurrah. |
15:09.23 | lkcl | drno: you'll be delighted to know that you're the third person (that i know of) who is trying linux on their hw6915. |
15:09.31 | drno | haha |
15:10.02 | drno | well, Im left handed and the microsoft OS on it isnt really build for left handers |
15:10.03 | lkcl | are you a linux baby, then? cos we could really use the help. suspend needs updating |
15:10.43 | lkcl | to be in line with what's in the other 2.6.19 devices that are up-to-date |
15:10.44 | drno | well, lets say it like this: I know how to install the average linux distro, and even could do a LFS |
15:10.49 | lkcl | cool. |
15:10.56 | drno | but building drivers, I never tried that |
15:11.27 | lkcl | aww, our first userrr :) |
15:11.36 | lkcl | how cute |
15:11.37 | drno | but, Im willing to learn and help you guys |
15:11.40 | drno | ^_^ |
15:11.44 | lkcl | teehee |
15:11.45 | *** join/#htc-linux BOGDAN-S (n=BOGDAN-S@ACA264E5.ipt.aol.com) |
15:12.00 | drno | Id be ideal for trying to break the software |
15:13.04 | lkcl | actually - tell you what: would you be willing to set up an opie/familiar or gpe environment and edit some files etc? |
15:13.06 | drno | atleast Im trying to install gentoo on my PDA |
15:13.14 | lkcl | oooo, scarey :) |
15:13.31 | lkcl | you're aware that it's a little bit more involved than that, right? |
15:13.41 | drno | well I could try that if I get another SD card |
15:13.59 | lkcl | the distros (e.g. opie/familar / gpe) are specifically targetted at PDAs, whereas gentoo certainly isn't. |
15:14.13 | lkcl | and not even familiar or gpe are targetted at mobile phones |
15:14.21 | drno | haha, yes, I do realize that writing the software is ALOT more complicated then just installing it ;) |
15:14.46 | lkcl | but with openmoko from FIC and with opie II from trolltech that's all about to change, early next year |
15:15.15 | lkcl | anyway. i gotta go write this patch so that you can download the latest stuff without having to have that hack of mine, ok? |
15:15.18 | drno | I know that they are designed for PDA's, but I still would like to run gentoo on it in the end |
15:15.26 | drno | ok |
15:15.40 | lkcl | well, it's a laudable aim, but the only significant advantage of gentoo is portage. |
15:15.50 | lkcl | and http://openembedded.org has bitbake |
15:16.01 | lkcl | which is actually a hell of a lot more sophisticated than portage |
15:16.13 | iggy | drno: you have certainly been to #gentoo-embedded |
15:16.22 | iggy | if not, I'd suggest it |
15:16.35 | lkcl | oh, there's a gentoo-embedded? cool - i take it all back. |
15:16.55 | cr2 | we have a working and advanced config for universal |
15:16.55 | drno | yes, indeed, the whole compiling for specific hardware gets kinda lost on PDA's for the consumer, there isnt really much hardware to excange |
15:17.09 | iggy | I've actually run a version of gentoo on my old ipaq and on my apache |
15:17.12 | cr2 | there are minor patches needed to make sable happy. |
15:17.33 | cr2 | the problem is to do it in a portable way. |
15:17.57 | drno | yes, there is, but its sort of out of date and the wiki's need alot of work so you dont have to read 5 pages to summarise it into on e useable page, but they will get there :) |
15:18.27 | drno | iggi, is that on freenode as well? |
15:19.25 | *** join/#htc-linux pH5 (n=ph5@e178203168.adsl.alicedsl.de) |
15:23.03 | lkcl | allo pH5 |
15:23.17 | pH5 | hey lkcl |
15:23.32 | lkcl | what's the prototype in C for typecasting something to a function? |
15:23.43 | lkcl | i have this: |
15:23.45 | *** join/#htc-linux pH5 (n=ph5@e178203168.adsl.alicedsl.de) |
15:23.46 | lkcl | #define MMC_CARD_ALWAYS_WRITEABLE ((int)(platform_device*))(-1) |
15:23.51 | lkcl | #define MMC_CARD_ALWAYS_WRITEABLE ((int)(platform_device*))(-1) |
15:23.54 | pH5 | whoops, wrong key |
15:23.55 | lkcl | but it doesn't look right. |
15:23.59 | lkcl | well done pH5 :) |
15:24.41 | lkcl | this is the function pointer in the struct: |
15:24.43 | lkcl | int (*card_is_read_only)(platform_device *dev); |
15:25.03 | elysion | what's the status of the hermes port at the moment, and where can i find the working sources of that project? |
15:25.23 | lkcl | hermes, hermes... we just started on it yesterday, elysion. |
15:25.38 | pH5 | elysion: cvs.handhelds.org, linux/kernel26 repo |
15:25.39 | elysion | i thought that if i go through them i might understand a bit more of the stuff |
15:25.57 | lkcl | http://wiki.xda-developers.com/index.php?pagename=Hermes_Linux |
15:25.57 | elysion | pH5: is that somekind of a general kernel source, or hermes specific? |
15:26.38 | lkcl | cr2 committed a basic kernel configuration to hh.org cvs, yesterday. i spent yesterday evening trying to get it to boot. |
15:27.12 | elysion | oh, so it's that early state |
15:27.30 | elysion | i just tested the haret app yesterday |
15:27.55 | lkcl | yep. |
15:28.21 | lkcl | also there's a version of haret around which has gpio polling, which i'm using to fill out the HermesGPIO page, as-and-when. |
15:28.30 | elysion | i tought that if i try to understand what you have done this far, i might get an understanding of the stuff that will be done in the future |
15:28.31 | lkcl | i found the vibra GPIO pin yesterday, and i am _so_ proud. |
15:28.32 | lkcl | ha ha |
15:28.47 | lkcl | elysion: that sounds very zen, dude :) |
15:29.45 | lkcl | well, you're in luck, cos it's not a lot so far he he. i am having haret-booting-difficulties which are reminiscent of my hermes-booting-difficulties, until cr2 pointed out that i need to specify a boot option mem=63M and then it worked. |
15:30.06 | lkcl | oops. |
15:30.22 | lkcl | make that sable-booting-difficulties |
15:49.33 | lkcl | erm.. why is the asic3_mmc platform_device not done as an asic3 platform device?? |
15:49.47 | lkcl | e.g. like this? |
15:49.49 | lkcl | static struct platform_device *htcsable_asic3_devices[] __initdata = { |
15:49.50 | lkcl | <PROTECTED> |
15:49.50 | lkcl | <PROTECTED> |
15:50.29 | cr2 | historical grounds. |
15:50.42 | lkcl | oops. |
15:50.50 | lkcl | because it needs to change. |
15:53.08 | lkcl | damn. |
15:57.18 | cr2 | lkcl: write an asic3_mmc2 driver in this way, test it that it works and post to k-d. |
15:57.43 | lkcl | arg. |
15:58.27 | lkcl | you know what? i _really_ like the approach that independent vendors take, to create wince os. |
15:58.56 | lkcl | hack-it-together with hexadecimal values spewed forth across different devices |
15:59.14 | lkcl | and never the code from another platform shall meet. |
15:59.24 | cr2 | :) |
15:59.42 | lkcl | maintaining so many different devices is a complete pain. |
15:59.48 | cr2 | this driver needs a modification anyway. |
16:00.00 | cr2 | yes, and then kill old asic3_mmc |
16:00.12 | lkcl | *grumbles*. |
16:00.16 | cr2 | the driver should be capable of sdio. |
16:00.35 | cr2 | but i have seen only some basic sdio init support patches. |
16:00.53 | lkcl | ok, i'm going to hack it into asic3_platform_device, at psoko's suggestion. |
16:01.07 | lkcl | next to tmio_mmc_hwconfig. |
16:01.15 | lkcl | really bad, really dire, but i don't care. |
16:01.39 | lkcl | i'm back to holland on the 7th jan and i want to get as much done as i can before then. |
16:02.37 | cr2 | ok. |
16:02.38 | lkcl | btw cr2 you don't _have_ to boot off the sdcard, you could always copy haret and the kernel onto the device's memory or flash or the extended_rom |
16:02.58 | cr2 | of course. you are put into ram anyway. |
16:03.06 | cr2 | s/you/they/ |
16:03.11 | lkcl | it's possible to unlock extended_rom btw - pof's program works (you have to run it twice) though |
16:03.15 | cr2 | i'm crazy today. |
16:03.47 | cr2 | basic boot first :) |
16:05.30 | lkcl | yes. haven't got that far. tried several times different things, yesterday. |
16:05.37 | lkcl | mem=64M didn't do it. |
16:06.15 | lkcl | compiling up haret.exe from Kevin2 (got the oegcc) and enabling the s3c dma disabling didn't do it |
16:07.41 | cr2 | i suspect usb host |
16:07.49 | lkcl | ahh. |
16:07.56 | cr2 | need to look for the ohci base reg. |
16:08.08 | cr2 | and the infamous 16bit counter. |
16:08.32 | lkcl | i think i disabled that... |
16:08.34 | cr2 | it was fcking everything up on jornada820 and universal. |
16:08.46 | cr2 | it's cpu -> MSM link. |
16:08.56 | cr2 | usb host, not UDC. |
16:09.22 | lkcl | oh you mean if you have the thing plugged in _at all_ when haret boots? |
16:09.23 | lkcl | well |
16:09.30 | cr2 | i think the setup is quiet similar to the universal. |
16:09.42 | cr2 | s/quiet/quite/ |
16:09.47 | cr2 | too much work today. |
16:10.03 | lkcl | lsusb hung until i removed the usb plug. |
16:10.06 | lkcl | burble :) |
16:10.53 | *** join/#htc-linux jjazz (n=jjazz@cpe-24-90-14-159.nyc.res.rr.com) |
16:11.58 | lkcl | ok if i remove g_ether then lsusb works |
16:12.31 | lkcl | ok - actually: if i boot with the usb cable removed it is happy when i put it back. |
16:13.07 | BabelO | is there some change in usbnet support ? my suse linux detect BA as a Microsoft USB corp |
16:13.10 | BabelO | eurk !! |
16:13.30 | lkcl | but i've put 'vibra on' into start_kernel() and it doesn't switch on. |
16:14.34 | lkcl | and that's _really_ early. |
16:15.45 | lkcl | probably too early, due to the irqs not being initialised... |
16:16.33 | lkcl | ok trying to move it to after irq init... |
16:17.24 | cr2 | BabelO: there are 3 usb ids on universal. |
16:17.27 | lkcl | hey btw i managed to get the micro-sd-card to ping across the room over 1 metre yesterday. |
16:17.42 | lkcl | steady, Babel0... |
16:17.44 | cr2 | lkcl: why do you need irqs ? |
16:17.59 | lkcl | local_irq_save() and restore() in the s3c gpio set function :) |
16:18.01 | lkcl | brb |
16:18.09 | cr2 | if the mmu is not turned on , just write to the reg. |
16:32.04 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
16:33.16 | lkcl | cr2: ack. |
16:33.26 | *** join/#htc-linux JTRipper86 (n=jtripper@p54B1F283.dip.t-dialin.net) |
16:36.44 | Kevin2 | Hi |
16:38.22 | lkcl | allo Kevin2 |
16:38.46 | lkcl | i managed to compile haret.exe yesterday. got caught out by that gdb thing, but.. |
16:39.32 | lkcl | linboot() is being called. |
16:39.52 | lkcl | but it's splat after that. |
16:40.00 | Kevin2 | Do you get the colors on the screen? |
16:40.13 | lkcl | <PROTECTED> |
16:40.13 | lkcl | <PROTECTED> |
16:40.14 | lkcl | yep. |
16:40.22 | lkcl | just confirmed by adding an extra black line |
16:40.53 | Kevin2 | So, either your kernel is toast or you don't have all DMA disabled. |
16:41.17 | lkcl | whoopsiee |
16:42.41 | lkcl | huh. preHardwareShutdown is called but HardwareShutdown is not. |
16:42.43 | lkcl | errr... |
16:44.33 | Kevin2 | What makes you think that? |
16:45.16 | lkcl | ok got it. case-sensitive. duh |
16:45.18 | Kevin2 | Try adding: memPhysWrite(0x49000004, 1); to MachineS3::hardwareShutdown |
16:45.59 | Kevin2 | Wow, Intel docs are light years advanced over Samsung docs. |
16:46.24 | lkcl | oops... |
16:46.55 | lkcl | what does that do - reset udc? |
16:47.18 | Kevin2 | It should reset UHC. |
16:47.29 | Kevin2 | It is similar to what arch-pxa27x.cpp does. |
16:48.04 | lkcl | ack. |
16:48.09 | lkcl | oh btw what's this mean? |
16:48.12 | lkcl | preload=424@002A3000/32e30000 sj=002A3000 stack=002A1000/32e2e000 data=002A2000/32e2f000 exec=32e300ec^M |
16:48.56 | lkcl | why is it 32e00000 instead of 0x3000000 like i specified ramaddr 0x30000000 in the default.txt? |
16:50.01 | Kevin2 | That is where the preloader is copied to. (The Haret code that runs once the MMU is off.) It can't be early in memory, because the preloader needs to copy the kernel/initrd to the top of ram. |
16:50.52 | lkcl | nope still splat. |
16:50.53 | lkcl | ok. |
16:51.03 | lkcl | hmm |
16:51.43 | cr2 | Kevin2: not only intel docs. when i see the gpio mess on samsung it hurts my eyes. |
16:52.30 | cr2 | and the docs errata list is really amazing. |
16:53.03 | lkcl | i think that processor is a bit... early days, yet |
16:55.53 | Kevin2 | Well, I've been in this situation of a kernel silently failing. I did a couple of different things - first I enhanced haret to crc the kernel and report on any crc differences - this will help see if you're not disabling dma correctly. The second thing is to pepper the kernel init sequence with writes to the framebuffer - this will allow you to see where in the kernel it dies. |
16:57.02 | cr2 | Kevin2: it's better to check the usb host first. |
16:57.07 | cr2 | easier. |
16:57.24 | Kevin2 | lkcl did that --- memPhysWrite(0x49000004, 1); |
16:58.45 | Kevin2 | Oops - I'm an idiot - it should read -- memPhysWrite(0x49000008, 1); |
16:58.52 | lkcl | ha ha |
16:59.03 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
16:59.14 | lkcl | 1sec... |
17:00.33 | lkcl | nope. |
17:02.28 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-75-133.dynamic.mnet-online.de) |
17:16.13 | lkcl | hmm i get exactly the same splat with the g500 |
17:16.29 | lkcl | mind you i have a g500+ not a g500 |
17:17.06 | *** join/#htc-linux rob_w (n=bob@p85.212.150.69.tisdip.tiscali.de) |
17:21.10 | lkcl | ok i'm just getting grumpy, here :) |
17:50.07 | *** join/#htc-linux skodde (n=skodde@85-18-136-66.fastres.net) |
17:55.47 | lkcl | ok - polling of gpio isn't quick enough to catch a 2-instruction gpio change. |
17:55.51 | lkcl | hmmmm... |
17:57.43 | cr2 | why is it so fast ? |
18:09.47 | lkcl | i dunno. |
18:10.02 | elysion | lkcl: which device are you working on? |
18:10.02 | lkcl | but it's GPIO 89, set and cleared within about 3 instructions. |
18:10.09 | *** join/#htc-linux FossiFilm (n=Fossi@e176098071.adsl.alicedsl.de) |
18:10.27 | lkcl | elysion: well i keep flitting between several different devices, as i get depressed about not being able to solve a particular issue :) |
18:10.40 | elysion | :) |
18:10.41 | lkcl | the triumph of hope over experience, ha ha |
18:10.52 | elysion | lkcl: but at the moment? |
18:10.56 | lkcl | it stops me going insane. |
18:11.03 | lkcl | que? oh - blueangel. sound. again. |
18:11.08 | elysion | ok |
18:11.22 | lkcl | after trying for a week. |
18:12.12 | *** join/#htc-linux asylumed (n=insanity@196.211.28.91) |
18:12.17 | lkcl | Notes: the reset time should be at least 1 microsecond... |
18:12.39 | lkcl | uda1380_4.pdf data sheet, section 8.10 |
18:13.32 | lkcl | ok - let's try that... |
18:16.43 | cr2 | lkcl: 89 ? |
18:16.52 | cr2 | in 3 instructions ? |
18:16.53 | lkcl | cr2: yep. lemme just double-check... |
18:17.12 | lkcl | yes 89. |
18:17.15 | lkcl | http://wiki.xda-developers.com/index.php?pagename=BlueangelGPIO |
18:17.16 | cr2 | where ? |
18:17.19 | lkcl | it was already marked in there. |
18:17.21 | lkcl | 1sec... |
18:17.43 | lkcl | .text:015B9D58 loc_15B9D58 |
18:17.53 | cr2 | ok. |
18:18.34 | cr2 | .text:015B9D58 LDR R0, =PXAGPIO |
18:18.38 | cr2 | .text:015B9D5C CMP R1, #0 |
18:18.42 | cr2 | .text:015B9D60 LDR R1, [R0] |
18:18.43 | cr2 | .text:015B9D64 MOV R0, #0x2000000 |
18:18.43 | cr2 | .text:015B9D68 ADDNE R2, R1, #0x20 |
18:18.43 | cr2 | .text:015B9D6C STRNE R0, [R2] |
18:18.43 | cr2 | .text:015B9D70 ADDEQ R2, R1, #0x2C |
18:18.44 | cr2 | .text:015B9D74 STREQ R0, [R2] |
18:19.02 | lkcl | yep - that's it. |
18:19.07 | cr2 | if (r1 != 0) set, if == 0 , clear . |
18:19.16 | lkcl | oh oops :) |
18:19.18 | cr2 | why 3 instructions ? |
18:19.21 | lkcl | ha ha |
18:19.31 | lkcl | sillee meee :) |
18:19.34 | cr2 | :) |
18:19.43 | lkcl | bollocks. |
18:19.52 | lkcl | _where_ is this damn codec reset gpio/asic3gpio???? |
18:21.23 | Kevin2 | lkcl: Can you try -- http://www.handhelds.org/~koconnor/haret/haret-20061218.exe |
18:21.30 | lkcl | Kevin2: ok. 1sec... |
18:21.38 | Kevin2 | Add the line "set kernelcrc 1" to your boot script. |
18:22.02 | Kevin2 | This will cause the bootup to crc your kernel before and after the preloader. It may help detect a dma issue. |
18:23.27 | lkcl | cr2: we've got B14, GPIO20 and GPIO89 left. |
18:23.39 | lkcl | ok - gotta just get the other sdcard out... |
18:28.08 | lkcl | ok got a nice yellow line |
18:28.16 | lkcl | now a black one |
18:28.40 | cr2 | lkcl: which one had the same "reset" number as universal ? |
18:28.46 | Kevin2 | There was some delay between yellow and black? |
18:28.50 | lkcl | Kevin2: yep |
18:29.01 | Kevin2 | Your kernel/initrd crc'd correctly. You have a kernel issue I'd say. |
18:29.24 | lkcl | cr2: none of them. you mentioned 0x800 was the codec_reset - there wasn't one. well, there was a fall-back option... |
18:29.30 | Kevin2 | As one final check, you might want to look at the logs, and compare the CRC in the log to the the output of "cksum" for your kernel/initrd. |
18:29.41 | lkcl | Kevin2: argh. well, that's not a surprise, because g500 doesn't load either. |
18:29.41 | cr2 | ok. |
18:30.16 | cr2 | Kevin2: is it crc32 ? |
18:30.31 | Kevin2 | cr2: Yes. It will match cksum output. |
18:30.36 | cr2 | ok. |
18:31.09 | cr2 | food time. |
18:31.16 | lkcl | ok no there's no 0x800. oops. |
18:31.17 | lkcl | ok dude. |
18:43.47 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
18:51.51 | *** join/#htc-linux Hawk|- (n=Hawk@p57A547FC.dip0.t-ipconnect.de) |
18:57.07 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
19:24.55 | elysion | hey, umm... i read the sc3244 users manual, and i would like to watch some memory addresses. How can i for example watch the adc control register at 0x5800000 with haret? |
19:26.08 | elysion | i tried it with watch p2v(0x5800000), but that didn't do anything (might be that there was nothing going on in that address) |
19:26.38 | elysion | and with watch 0x5800000 the pda just shuts down the haret :-/ |
19:27.11 | elysion | woops... i mean addwatch and then watch x |
19:29.59 | psokolovsky_ | cr2, lkcl: so, any special compiler needed to build kernel for s3? |
19:30.17 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
19:31.16 | Kevin2 | elysion: You should have seen a message at the start with the current value of that register. If you didn't see any updates after that, then the register probably didn't change. |
19:31.40 | elysion | Kevin2: got that, and no change after that... humm |
19:31.59 | elysion | so is the p2v function reliable? |
19:32.28 | Kevin2 | Should be. Try: print "%08x" P2V(0x5800000) |
19:32.37 | elysion | i mean is it okay to use it, or should i find the physical to virtual address table from somewhere and use those addresses instead? |
19:33.16 | Kevin2 | elysion: It should work - haret is just looking up the tables for you. |
19:33.22 | elysion | ok |
19:33.56 | cr2 | psokolovsky_: i'll switch verbose on while compiling. |
19:35.40 | cr2 | -D__LINUX_ARM_ARCH__=4 -march=armv4t -mtune=arm9tdmi |
19:36.40 | cr2 | this is ugly as sin |
19:36.44 | cr2 | arch/arm/mach-s3c2410/s3c2410-pm.c:36:27: error: asm/arch/g500.h: Datei oder Verzeichnis nicht gefunden |
19:36.44 | cr2 | arch/arm/mach-s3c2410/s3c2410-pm.c: In function ‘s3c2410_pm_prepare’: |
19:36.44 | cr2 | arch/arm/mach-s3c2410/s3c2410-pm.c:69: error: ‘G500_SUSPEND_RESUMEAT’ undeclared (first use in this function) |
19:37.51 | psokolovsky_ | cr2, do you use gcc3.4.1? it understands all those -mtune's? |
19:38.02 | cr2 | 4.1.1 |
19:38.22 | cr2 | but i guess mtune= is not really necessary. |
19:38.48 | cr2 | i doubt gcc takes it really serious. |
19:39.03 | psokolovsky_ | cr2, ok, I have such suspicion too ;-) armv4 should be enough |
19:39.03 | cr2 | maybe some inline asm will be activated in the kernel. |
19:39.43 | cr2 | but it's what is found in the Makefile |
19:41.21 | cr2 | lkcl: the CVS is broken again. can you just keep the g500-specific stuff somewhere local ? |
19:51.53 | *** join/#htc-linux pH5 (n=ph5@e178203168.adsl.alicedsl.de) |
20:28.13 | *** join/#htc-linux psokolovsky_ (n=psokolov@237.usernat.ip.net.ua) |
20:30.21 | *** join/#htc-linux psokolovsky__ (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
20:38.03 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
20:52.28 | cr2 | goxboxlive: found new bugs ? |
20:56.34 | goxboxlive | hi cr2: No, i havent had any time today. But i have to say that i am impressed of the accuracy of the battery status :-) |
20:57.15 | goxboxlive | I was waiting for Opie II today, but i havnet seen any released yet. |
20:58.09 | cr2 | we have made jokes with BabelOued yesterday about your battery status :) |
20:58.21 | goxboxlive | :-) |
20:58.25 | *** join/#htc-linux ellisway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
20:58.38 | cr2 | it should have switched off your wifi earlier. |
20:58.51 | cr2 | wince does it at 10% |
20:59.17 | cr2 | imho it's not good for the battery to be discharged to 3.5V |
20:59.26 | goxboxlive | Well it didnt, I dont see the point of switching it off. You get alot of warnings, i will disabel the shutoff anyway |
20:59.33 | goxboxlive | realy |
21:00.08 | goxboxlive | I thought that was good to completely discharge it now and then |
21:02.58 | lkcl | psokolovsky: argh, i hope not |
21:03.42 | lkcl | cr2: i've got to cvs commit a file for g500 - but you should cvs update again to avoid that |
21:03.55 | lkcl | cr2: i fixed itlready |
21:04.40 | lkcl | psokolovsky: i have gcc 4.1.1-19 and gcc 3.4.6-4 |
21:04.44 | lkcl | i'll try 3.4-6... |
21:06.43 | cr2 | lkcl: i'll try to boot on hermes. which haret should i use ? |
21:06.58 | lkcl | the latest one from Kevin |
21:07.20 | lkcl | i got less than a minute to set something going then i'll be back in a bit. |
21:08.28 | cr2 | default.txt ? |
21:08.31 | lkcl | <Kevin2> lkcl: Can you try -- http://www.handhelds.org/~koconnor/haret/haret-20061218.exe |
21:08.35 | elysion | if i wan't to set the value of the gpb data register at 0x56000014 to value 0x0000004e, would something like this work: VFW P2V(0x56000014) 1 0x0000004e? |
21:08.56 | lkcl | log "bootlog.txt" |
21:08.56 | lkcl | set ramaddr 0x30000000 |
21:08.56 | lkcl | set ramsize 0x04000000 |
21:08.56 | lkcl | set KERNEL "zImage" |
21:08.56 | lkcl | set CMDLINE "mem=64M" |
21:08.57 | lkcl | set MTYPE 1213 |
21:08.59 | lkcl | set bootspd 0 |
21:09.01 | lkcl | set kernelcrc 1 |
21:09.03 | lkcl | bootlinux |
21:09.05 | lkcl | is what i have so far |
21:09.08 | lkcl | ok gotta go back later |
21:09.09 | cr2 | ok. thanks. |
21:09.09 | psokolovsky__ | lkcl, I'm moslty just interested, don't spend too much time on that now ;-) |
21:10.10 | cr2 | elysion: whhy not 'pfw 0x56000014 1 0x0000004e' ? |
21:10.16 | elysion | well... :D |
21:10.30 | elysion | but should that work the way i expected? |
21:10.47 | cr2 | it should. |
21:13.57 | elysion | wow, neat :) |
21:14.05 | *** join/#htc-linux rob__w (n=bob@p85.212.153.187.tisdip.tiscali.de) |
21:14.11 | elysion | found the register that controls the backlight of the screen :) |
21:15.08 | cr2 | pwm? |
21:15.22 | elysion | no, just the on off thing |
21:15.25 | cr2 | for brightness. |
21:15.33 | cr2 | ok. |
21:16.12 | elysion | not such a big discovery for you propably, but now i can be sure that i understand at least something of this cpu+haret+telnet thing :) |
21:16.54 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
21:17.07 | cr2 | :) |
21:17.36 | cr2 | you can try to check the "alt" gpio settings. |
21:17.45 | elysion | ? |
21:18.34 | cr2 | some gpio pins have alternate functions, depending on the config bit. |
21:18.57 | cr2 | my microSD is not recognized by hermes... |
21:21.15 | elysion | :( |
21:22.52 | cr2 | hmm. |
21:23.06 | cr2 | both are 256M, a780 read both. |
21:23.16 | cr2 | hermes only one. |
21:24.54 | goxboxlive | cr2: What kind of phones and pda's do you have now? |
21:25.29 | goxboxlive | BA, Hermes, Sable, UNI, A780 Have i forgot any? |
21:28.27 | cr2 | no. |
21:29.07 | cr2 | ok. does not boot for me too. |
21:38.58 | *** join/#htc-linux psokolovsky_ (n=psokolov@237.usernat.ip.net.ua) |
21:40.50 | cr2 | s3c24xx_init_clocks(16934000); |
21:40.51 | cr2 | i hate magic numbers. |
21:41.18 | elysion | what's that supposed to do :o |
21:43.34 | *** join/#htc-linux lamikr_ (n=chatzill@aragorn.kortex.jyu.fi) |
21:46.50 | cr2 | touchscreen does not work anymore. |
21:47.37 | cr2 | ah. it's locked. |
21:48.49 | cr2 | hmm. black stripe, and then it's dead. |
21:53.11 | Kevin2 | cr2: Hermes? |
21:54.19 | cr2 | yes. |
21:56.00 | cr2 | my hermes wifi is not really great. |
21:56.15 | *** join/#htc-linux Ro_En_PC (n=roen@p54A64FB2.dip.t-dialin.net) |
21:56.29 | cr2 | can't connect even if i put the access point before the phone. |
21:56.51 | Kevin2 | Black stripe from Haret? That means it relocated the kernel and is about to jump into it. |
21:57.31 | Kevin2 | What I've done in the past is pepper the kernel bootup code with writes to the framebuffer. |
21:57.56 | cr2 | i'll try. |
22:00.05 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
22:10.57 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
22:13.27 | *** join/#htc-linux jjazz (n=jjazz@cpe-24-90-14-159.nyc.res.rr.com) |
22:45.41 | Kevin2 | Hrmm. I'd like to define the lcd startup/shutdown routine on apache -- should i use the pxafb pxafb_lcd_power callback, or create an lcd_properties device? |
22:48.20 | pH5 | Kevin2: I have the feeling that both don't work properly |
22:48.47 | pH5 | pxafb already has some serialization, but the wince powerup/down routines have two parts. |
22:49.34 | pH5 | both methods do the job (in the end the lcd is powered down), but what I see on the screen during powerup/down doesn't tell me that I'm doing things in the right order. |
22:49.47 | Kevin2 | pH5: The magician code seems to use lcd_properties.. Why did you go that route? |
22:50.21 | Kevin2 | Or is it all commented out? |
22:50.26 | pH5 | copy&paste. today I rewrote it to use pxafb |
22:51.12 | Kevin2 | Heh, can I copy and paste yours? |
22:51.22 | pH5 | wait, I'll upload |
22:55.48 | pH5 | Kevin2: cvs up |
23:01.53 | lkcl | elysion: every gpio discovered is a victory. |
23:02.05 | elysion | :) |
23:02.12 | lkcl | we need to find every single one. fill it in on the wiki page, dude. |
23:02.35 | elysion | i actually found two already :) |
23:02.45 | lkcl | yaay, well done! |
23:02.47 | elysion | :D |
23:03.16 | Kevin2 | elysion: Which haret version are you using? |
23:05.19 | Kevin2 | pH5: Thanks. |
23:05.37 | elysion | it's the 20061218 version at http://handhelds.org/~koconnor/haret/ |
23:08.04 | elysion | lkcl: i don't quite get the num column on the hermes gpio wiki page :-/ |
23:08.20 | lkcl | bit numbers and hexadecimal |
23:08.36 | lkcl | 1<< number == 0xN00 whatever. |
23:08.41 | lkcl | it's for convenience. |
23:08.58 | lkcl | when looking at the page you don't have to do the conversion in your head all the time. |
23:09.19 | elysion | ok |
23:10.02 | Kevin2 | Cool, thanks. |
23:10.07 | elysion | so how would i add there that the lsb of gpb is the backlight toggle? |
23:11.54 | lkcl | number 0, then 1 left shifted by 0 is 1, so 0 and 0x0001 |
23:12.15 | lkcl | put this: 0 | 0x0001 | backlight (1=on) |
23:12.28 | lkcl | if in fact that's true. |
23:12.34 | elysion | oh, ok, now i get it :) |
23:12.42 | lkcl | if it's 'clear' to be on, put 0=on instead. |
23:13.00 | elysion | hmm... can't remember anymore :-/ |
23:13.09 | elysion | i'll have to check that later |
23:13.14 | elysion | have to go to bed soon |
23:13.56 | lkcl | k dude |
23:16.01 | elysion | so is it commonly known that the keyboard is connected to spi channel 0 on hermes? |
23:31.11 | elysion | hope i got the stuff right in the wiki |