01:12.06 | Kevin2 | Hi |
01:48.20 | *** join/#htc-linux ivor (n=ivor@kde/developer/ivor) |
02:00.26 | *** join/#htc-linux psokolovsky (n=psokolov@ip.85.202.124.214.dyn.sub-9.broadband.voliacable.com) |
04:09.02 | *** join/#htc-linux rob_w (n=bob@p85.212.167.244.tisdip.tiscali.de) |
06:53.48 | *** join/#htc-linux rob_w|mis (n=rob_w@p549BBCCD.dip0.t-ipconnect.de) |
08:00.59 | *** join/#htc-linux dullard (n=jim@adsl-static-1-30.uklinux.net) |
08:28.27 | *** join/#htc-linux booba (n=booba@AAnnecy-256-1-216-182.w90-9.abo.wanadoo.fr) |
08:56.21 | *** join/#htc-linux rejon (n=rejon@64-121-195-22.c3-0.sfpo-ubr4.sfrn-sfpo.ca.cable.rcn.com) |
08:56.38 | *** part/#htc-linux rejon (n=rejon@64-121-195-22.c3-0.sfpo-ubr4.sfrn-sfpo.ca.cable.rcn.com) |
15:15.40 | *** join/#htc-linux apt (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
15:15.40 | *** topic/#htc-linux is HTC LINUX ! check also -> http://handhelds.org/moin/moin.cgi/HTC_2dPhones | http://wiki.xda-developers.com/index.php?pagename=BlueangelResearch | http://gnulinux.biz/files/ LOGS: at http://ibot.rikers.org/%23htc-linux/ |
16:16.12 | *** join/#htc-linux lamikr (n=chatzill@aragorn.kortex.jyu.fi) |
16:39.02 | *** join/#htc-linux LunohoD (n=alex@e180109115.adsl.alicedsl.de) |
16:56.36 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
17:19.54 | lkcl | i've cut out practically everything - got a little bit more to remove e.g. the ts and udc from arch/arm/mach-pxa/htcsable/Makefile - and i'm still getting unrecognised filesystem on the root partition |
17:23.56 | *** join/#htc-linux pleemans (n=peter@d51A5E421.access.telenet.be) |
17:25.27 | lkcl | it was cos of the .gz extension on initrd26.gz - gnuharet.exe doesn't like fileextensions |
17:25.30 | lkcl | ARSE! |
17:33.35 | psokolovsky | lkcl, mega-lol ;-) |
17:34.16 | psokolovsky | lkcl, just in case, you *should* use quotes around any string params, period. I'll make mailline haret error on otherwise. |
17:35.36 | lkcl | nggh |
17:35.41 | lkcl | oh yeh. |
17:35.44 | lkcl | arse |
17:37.07 | lkcl | hmm, mmc has been mounted read-only... |
17:38.54 | lkcl | off we go again... |
17:39.16 | lkcl | mmc loaded read-only... hmmm... |
17:40.28 | lkcl | hmm... anyone know how to make devfs the default? |
17:40.55 | lkcl | /dev/mmc/blk0/part2 doesn't exist |
17:44.24 | lkcl | hmm, /dev was missing from the initrd /etc/fstab |
17:45.58 | lkcl | hmm i think i gonna have to have a mknod in the /linuxrc for /dev/mmcblk0p2. YUK! |
17:47.03 | psokolovsky | devfs is dead |
17:47.12 | psokolovsky | just create static in /dev |
17:47.20 | psokolovsky | in initrd image |
17:47.36 | psokolovsky | otherwise, udev is the answer |
17:48.02 | lkcl | ah much joy. |
17:48.19 | lkcl | udev in an initrd. just what the debian guys didn't order. hurrah! |
17:48.47 | psokolovsky | I use static entry in /dev for hack initrd ;-) |
17:49.05 | lkcl | debian is the odd-one-out: gentoo don't use initrd, deadrat and suse use udev/uclibc, debian use some weirdness... |
17:49.13 | lkcl | which causes much aggro with se/linux... |
17:49.46 | lkcl | gotta stop this mmc card from being mounted auto-read-only, next... |
17:51.27 | lkcl | wtf mmcblk0 being recognised as a read-only device??? |
17:53.01 | psokolovsky | so, it recongizes some signal as such ;-) |
17:53.21 | lkcl | hmmm.... |
17:53.26 | lkcl | bastards. |
17:53.54 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-62-245-160-34.dynamic.mnet-online.de) |
17:54.49 | lkcl | but only a little bit because it's a nice table |
17:55.42 | psokolovsky | why worry such? it's all solvable ;-) |
17:56.28 | psokolovsky | iirc, r/o bit comes from asic3 sdio controller, but cr2 already mentioned there're issues with that on htc models |
17:56.40 | lkcl | yeh he did |
17:56.47 | psokolovsky | find twhat to use instead, let's add handler for that |
17:57.43 | lkcl | oh i ain't worrying |
17:58.17 | *** join/#htc-linux asylumed (n=insanity@196.211.28.91) |
17:58.42 | psokolovsky | maybe the bit is just inverted, lol? |
17:59.02 | psokolovsky | asylumed, Hi! just committed uour udc pathc |
17:59.06 | psokolovsky | patch |
17:59.38 | asylumed | :) |
17:59.44 | asylumed | awesome... |
17:59.55 | psokolovsky | what is resume issue exactly? |
18:00.15 | asylumed | it dies in __wake_up_common... |
18:00.21 | asylumed | killing various processes |
18:00.27 | asylumed | none in particular... |
18:00.48 | psokolovsky | any dmesg dump? |
18:00.49 | asylumed | (things die in __wake_up_common for various reasons it seems... |
18:00.51 | asylumed | ) |
18:01.20 | asylumed | shall i mail them to you? |
18:01.28 | asylumed | (or kernel discuss?) |
18:01.31 | psokolovsky | asylumed, yeah, feel free |
18:01.32 | asylumed | i'm stumped... |
18:01.33 | lkcl | ah... yeh good point... _let's_ try just... ignoring that bit :) |
18:01.46 | lkcl | several other bits have been inverted, so far, already... |
18:01.49 | psokolovsky | asylumed, feel free to ML too. |
18:02.25 | psokolovsky | asylumed, do you care about MSCx PXA regs on resume? |
18:02.58 | lkcl | i should really be doing this with nfs not initrd - i can make, rmmod, etc... |
18:03.05 | lkcl | xxxx it can't be bothered :) |
18:03.47 | psokolovsky | lkcl, well, do you have resume issues too? any care for MSC? |
18:04.26 | lkcl | resume? what's MSC? |
18:04.32 | psokolovsky | lkcl, I meant, try to activate lock and mount that ;-) |
18:04.39 | lkcl | resume issues? |
18:04.44 | psokolovsky | lkcl, memory config regs |
18:04.49 | lkcl | sorry dude i have no idea what you mean |
18:04.57 | psokolovsky | lkcl, yes, did you report them for sable? |
18:05.05 | lkcl | absolutely not a clue. |
18:05.19 | psokolovsky | resume works on sable? |
18:05.26 | lkcl | oh - you mean here: http://wiki.xda-developers.com/index.php?pagename=Ipaq6915_Research |
18:05.37 | asylumed | MSCx? |
18:05.53 | lkcl | it's all numbers to me ... |
18:06.04 | lkcl | <PROTECTED> |
18:06.04 | lkcl | <PROTECTED> |
18:06.05 | lkcl | <PROTECTED> |
18:06.11 | psokolovsky | lkcl, asylumed, yeah |
18:06.12 | lkcl | no idea what they are |
18:06.33 | *** join/#htc-linux rejon (n=rejon@adsl-75-5-124-98.dsl.pltn13.sbcglobal.net) |
18:06.37 | psokolovsky | per Linux èùùåäùôâóê çêùåùñùäá èùùåäùôâóê ûðùãäâ åôëó ñôêó ùà åðùûó |
18:07.13 | psokolovsky | sh&t |
18:07.34 | psokolovsky | per Linux bootloader protocol, MSCx should be handled by bootloader |
18:07.53 | psokolovsky | butt they are crucial for system work, including after resume |
18:08.26 | lkcl | ah ha! am getting somewhere - ignoring ro worked... |
18:08.30 | asylumed | :) |
18:08.36 | asylumed | worked for me too :P |
18:10.20 | lkcl | oops... |
18:10.31 | lkcl | exec /sbin/init and it went 'blerk'... |
18:10.46 | lkcl | hey at least i got a familiar login: prompt! |
18:12.53 | psokolovsky | asylumed, lkcl: you're listening about MSCx? how do you handle them? |
18:13.11 | *** join/#htc-linux ljp (n=lpotter@203.94.178.46) |
18:13.23 | asylumed | i don't myself... |
18:13.40 | asylumed | there is some old suspend code that seems to save them i think |
18:13.52 | asylumed | but there's little about them in the resume function... |
18:13.52 | lkcl | haven't got that far, psokolovsky |
18:14.34 | psokolovsky | asylumed, are you sure? until you get the thing settled, you won't have working resume |
18:14.36 | lkcl | like ... _really_ haven't got a clue. am more concerned about devices and drivers at this stage. there are other people more knowledgeable who know about pxa n stuff than i |
18:14.51 | psokolovsky | asylumed, but I don' think your current issues related to that though... |
18:14.59 | lkcl | i'd like... y'know... a keyboard, sound, and to recognise the serial device first :) |
18:15.04 | psokolovsky | lkcl, ok, just fyi |
18:15.20 | lkcl | actually i'll settle for just sound and the gsm serial link |
18:15.32 | psokolovsky | I banged my head for 3 month about that issue with h4000 ;-). |
18:15.55 | asylumed | k |
18:15.56 | lkcl | _why_ do we do this??? |
18:16.09 | psokolovsky | asylumed, make sure you cc: dmesg to me, as MLs doens't work as of now ;-( |
18:16.29 | lkcl | i've worked on/with the himalaya, the blueangel, the universal and now the sable - and i _still_ don't get it |
18:16.35 | lkcl | oh well |
18:16.43 | lkcl | now. where was i |
18:17.01 | asylumed | psokolovsky: sent |
18:17.08 | lkcl | arse. |
18:17.10 | lkcl | initrd. |
18:17.17 | lkcl | anyone got a good initrd i can start from? |
18:17.25 | lkcl | one suited to 2.6.17? |
18:17.37 | psokolovsky | asylumed, got them |
18:17.46 | lkcl | and perhaps an opie/familiar .gz that uses udev? |
18:18.12 | lkcl | cos cr2's universal initrd and opie/familiar are suitable for 2.6.12 - with devfs of course |
18:18.12 | psokolovsky | lkcl, I used old good initrd made by unknown few years ago (was that you? ;-)). Works with any kernel ;-E |
18:18.27 | lkcl | yeh *sigh* it was probably me... |
18:18.32 | psokolovsky | lkcl, use 0.8.4 image closest to your hw |
18:18.47 | lkcl | is it the one with xfbdev and dropbear in it? |
18:19.05 | psokolovsky | lkcl, dropbear, yes, but no X at all |
18:19.36 | psokolovsky | it has "rootme" as root passwd if that uncovers its personality ;-) |
18:28.07 | psokolovsky | asylumed, is blueangel_suspend.c is CVS up to date with what you use? |
18:28.35 | asylumed | yip... using the version in cvs |
18:30.59 | psokolovsky | asylumed, excuse me, but what crap is done there? lotsa memory is overwritten without being preserved, and you expect that to work? |
18:31.13 | psokolovsky | rob_w: any comments? |
18:31.15 | asylumed | psokolovsky: yeah... was comparing that to the other pxa devices... |
18:31.26 | asylumed | psokolovsky: tried commenting some of it out at a time... |
18:31.32 | asylumed | psokolovsky: never resumed... |
18:31.49 | asylumed | psokolovsky: (left the wake up on power button in) |
18:32.30 | psokolovsky | asylumed, you should preserve all the mmeory it barfs to, 48 bytes at memstart at least |
18:32.59 | psokolovsky | asylumed, also, remoev any asic3 stuff from it. asic3_base is what we have to handle any asic3 suspend/resume chores. |
18:33.13 | asylumed | i did :) |
18:33.24 | asylumed | right this moment... |
18:34.28 | psokolovsky | memgpio is likely not needed too |
18:36.25 | asylumed | :/ it froze without it... |
18:37.48 | psokolovsky | asylumed, PXA handling code should preserve all "user-level" register of PXA itself |
18:38.07 | psokolovsky | again, it doens't preserve few "system-level", like MSCx |
18:39.06 | asylumed | any examples? (like a device that does mess with them on resume and suspend?) |
18:39.20 | psokolovsky | asylumed, anyway, maybe that's what another creep-breed of HTC bootloader in blueangle wants to have. But again, you should rpeserve memory area where it is dumped too. |
18:39.34 | psokolovsky | asylumed, h4000 |
18:39.39 | asylumed | k |
18:39.54 | psokolovsky | h4000_pm.c |
18:40.09 | asylumed | yeah... having a look |
18:41.18 | asylumed | eish... where can those values be sourced? can they just be read before suspend? |
18:41.35 | psokolovsky | asylumed, sure |
18:42.18 | psokolovsky | asylumed, again, it's bootloader matter. h4000 bootloader doesn't initialize MSC registers before jumping to "resume routine" |
18:42.50 | psokolovsky | asylumed, some other pocketpc bootloader may initialize. but to be on safe side, better to just set them in init/resume code, voila |
18:43.11 | asylumed | will try that now... |
18:43.31 | asylumed | also... any idea what the code is doing with all those other registers? |
18:44.07 | psokolovsky | asylumed, you see, there're 2 types of pocketpc bootloaders: easy and f^%&ed up |
18:44.41 | psokolovsky | easy ones after resume will just jump back at address stored in predefined address in memory |
18:45.08 | asylumed | and this is where the blueangel comes in it seems :P |
18:45.20 | psokolovsky | f$@#$% up will be calculating checksums on memory, etc, etc, tehn restore some of regsters, and only then jump where needed ;-) |
18:45.49 | psokolovsky | asylumed, h4000 is example of "easy". h3900, hx4700, blueangel of another one ;-) |
18:46.38 | psokolovsky | but I heard from Palm people that PalmOS is where the real fun lives ;-) |
18:46.53 | psokolovsky | they can't have resume support for few models w/o reflashing ;-) |
18:47.14 | asylumed | since it *does* resume... kinda... i'll try saving memory now quickly... then the MSCx's too |
18:47.23 | psokolovsky | asylumed, yep |
18:48.07 | asylumed | well well well |
18:48.17 | asylumed | no crash ! well... still no crash |
18:49.59 | psokolovsky | try few suspend/resumes in row, and count when it hangs ;-) |
18:50.18 | asylumed | number 2... |
18:50.35 | asylumed | number 3 :) |
18:50.55 | asylumed | psokolovsky: thx man... |
18:51.22 | psokolovsky | asylumed, cool, glad it worked out! |
18:51.59 | asylumed | hmm... ok... now i just need to know why you said 48 bytes? |
18:52.57 | asylumed | uint * 10 + ? |
18:53.16 | psokolovsky | static unsigned long calc_csum(u32 *buf) { |
18:53.20 | psokolovsky | for(i = 0 ; i < 48 ; i++) |
18:53.29 | psokolovsky | it counts checksum on 48 bytes |
18:53.51 | psokolovsky | so, naive thought is that it doens't need anything else |
18:54.04 | psokolovsky | it == bootloader |
18:54.11 | asylumed | yeah... |
18:54.16 | psokolovsky | again, just naive ;-) |
18:54.47 | psokolovsky | you should disassemble bootloader yourself to be sure, or leave everything as is for now, and catch rob_w or someone and ask |
18:54.47 | asylumed | but i worry about the other addresses it's writing over like 0xa0000c00 ? |
18:55.19 | asylumed | will you accept the patch that makes thing work for now? |
18:55.38 | asylumed | (after i test it summore) |
18:55.41 | psokolovsky | sure |
19:03.28 | *** part/#htc-linux rejon (n=rejon@adsl-75-5-124-98.dsl.pltn13.sbcglobal.net) |
19:10.37 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
19:24.02 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
19:30.06 | *** join/#htc-linux Kmarc (i=kari@markos.biz) [NETSPLIT VICTIM] |
19:34.40 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
20:05.54 | lkcl | holy cow - you got suspend/resume working on the blueangel? |
20:14.29 | *** join/#htc-linux xery (i=jaahans@dsl-olubrasgw1-fec9de00-173.dhcp.inet.fi) |
20:15.18 | lkcl | anyone know how to unpack a jffs2 partition? |
20:16.29 | lkcl | ok i found this http://familiar.handhelds.org/releases/v0.8.4/install/files/hx4700/opie-image-v0.8.4-ipaq-pxa270.rootfs.tar.bz2 |
20:29.24 | asylumed | lkcl: yeah :) stoked |
20:36.37 | *** join/#htc-linux pH5 (n=ph5@e178204216.adsl.alicedsl.de) |
20:37.48 | lkcl | whoaaaa screen just went white - is that supposed to happen on opie? |
20:38.12 | lkcl | hmmm.... |
20:38.17 | lkcl | allo pH5 |
20:39.47 | pH5 | hey lkcl |
20:42.03 | lkcl | bootstrap-image time, instead.... |
20:42.18 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
20:42.35 | lkcl | i found my other two mini-sdcards... on the damn floor. |
20:47.48 | lkcl | what the _hell_ does the opie / bootstrap stuff _do_??? it just wiped out the screen - even the bootstrap stuff. |
20:48.35 | pH5 | lkcl: kernel console blanking? |
20:49.23 | lkcl | naah - this was more... kinda... fade to white. i've seen it before on himalaya-going-wrong |
20:49.45 | lkcl | you know the one - where it looks like the lcd has been powered off but the backlight is still on |
20:49.52 | lkcl | ok. tired. rest time |
20:50.12 | pH5 | rest well |
20:50.40 | lkcl | back later |
21:11.45 | *** join/#htc-linux rob_w (n=bob@p85.212.133.249.tisdip.tiscali.de) |
22:50.29 | *** join/#htc-linux florian_ (n=fuchs@84-245-181-237.bln.bpool.celox.de) |
23:22.57 | *** part/#htc-linux pH5 (n=ph5@e178204216.adsl.alicedsl.de) |
23:23.12 | psokolovsky | lkcl, please don't commit patches to common drivers without discussing them on kernel-discuss |
23:23.46 | psokolovsky | lkcl, also, please don't commit device-specific hacks to common drivers without first having them accepted upstream |
23:23.52 | psokolovsky | thanks ;-) |
23:27.31 | psokolovsky | lkcl, confirmed to be a-bad-thing, going to revert |
23:29.36 | lkcl | que, que? |
23:29.40 | lkcl | oops. |
23:29.56 | lkcl | i like bad. |
23:30.30 | lkcl | ... you mean... i can't just... like... hack things into submission? |
23:31.14 | psokolovsky | lkcl, yep, we're struggling to make HH.org *cleaner* and start submit stuff upstream, not dirtier |
23:31.15 | lkcl | ... perhaps now would be a good time to point me in the direction of a page describing thingies. conventions-which-really-should-be-followed. |
23:32.37 | psokolovsky | lkcl, the thing is that we kinda have port maintainers which are free to have lead on their port, but for common stuff, we follow the same procedure as main kernel - conservative community review ;-) |
23:32.39 | lkcl | so i don't make mistakes like that |
23:33.02 | lkcl | ok - that's the sort of thing i need to know about |
23:33.21 | lkcl | ... did someone have the foresight to write that kind of thing down? |
23:33.22 | psokolovsky | lkcl, well, the idea is simple - device independent drivers should not have any device-dependent bits. mmc_block is such driver |
23:33.40 | psokolovsky | lkcl, yes, we all ;-) |
23:33.49 | lkcl | hurrah! |
23:34.19 | psokolovsky | the general idea is that we should ask review/suggestion for all stuff on kernel-discuss |
23:35.06 | psokolovsky | lkcl, so, I'd appreciated if you reverted that patch, and let's think how to solve it properly (asic3_mmc.c would be appropriate level to sort this out) |
23:35.15 | lkcl | ack. |
23:36.04 | lkcl | hmm... don't have a ... oh cvs diff -D '1 day ago' will do it. darn been a long time since i used cvs... |
23:36.48 | lkcl | ok done |
23:36.52 | *** join/#htc-linux ivor (n=ivor@kde/developer/ivor) |
23:37.47 | psokolovsky | lkcl, Thanks! |
23:39.38 | psokolovsky | lkcl, you're on kernel-discuss, right? |
23:40.00 | lkcl | no prob |
23:41.53 | lkcl | i used to be on about 100 to 150 _detailed_ emails a day - it cured me of mailing lists, for life. that, and the samba shit. they _still_ haven't apologised, but have acknowledged, privately (not to me) that the things i was talking about were right. |
23:42.21 | psokolovsky | lkcl, sorry for such a questions, just as I told, can't seem to get cr2 on it ;-) |
23:42.25 | lkcl | which made their treatment of me _absolutely_ unnecessary. |
23:43.18 | psokolovsky | so, let's dicuss the issues there once it's live again (and cr2's experience with that issue would help too) |
23:43.28 | lkcl | well... to be honest, some of my best work - and best working practices - have been through wikis and irc. some of the worst, most ineffective and socially completely unacceptable working practices i've seen have been through mailing lists. |
23:43.34 | lkcl | ack, dude. |
23:44.34 | psokolovsky | well, I agree partly, IRC is pretty good for quick work. But it's easy to miss something when using only irc, so maillist have their use too. |
23:44.38 | lkcl | as we're way off topic, anyway, i might as well mention that i've been meaning to write this up for some time (the irc+wiki vs mailing lists thing) |
23:44.51 | lkcl | well, 'missing' something is what the irc logs plus the wiki is for. |
23:45.29 | psokolovsky | lkcl, I'm trying to do some maintence on HH.org wiki. Well, many things I perceive, again, need to be run thru other people first... |
23:45.34 | lkcl | wikis remind you, in an ordered fashion, of what you _should_ have remembered via the mailing lists. you don't get repeat me-too questions as much. |
23:45.44 | lkcl | i understand. |
23:46.29 | lkcl | i mean - heck - even _i_ have a hard time returning to this stuff after each 6-10 month break. wikis are an absolute god-send. |
23:46.32 | lkcl | anyway. |
23:46.49 | lkcl | what am i going to try to break, next, is what i wanna know. |
23:54.56 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-62-245-160-34.dynamic.mnet-online.de) |