00:00.29 | cr2 | asylumed: also after resume ? :) |
00:00.40 | asylumed | hmm... dunno yet... |
00:01.29 | asylumed | cr2: but... issues... the ds1wm driver hard codes rising edge interrupt... while ba needs falling edge... :/ |
00:01.43 | asylumed | cr2: will mail kernel discuss tomorrow after some sleep... |
00:01.49 | cr2 | btw, wince report only the voltage. how do you decode the datastream ? |
00:01.53 | cr2 | ok. |
00:02.49 | pH5 | cr2: look at hwmon/battery/h2200_battery.c and w1/slave/w1_ds2760.c. I suspect it works nearly the same way on all ds2760-using devices. |
00:03.24 | pH5 | good night |
00:03.25 | cr2 | pH5: i don't see ds2760 on ba |
00:03.29 | lkcl | night pH5 |
00:03.30 | cr2 | good night. |
00:04.19 | pH5 | cr2: it's there, look at battery.dll or wait until asylumed posts his findings ;) |
00:04.20 | *** part/#htc-linux pH5 (n=ph5@e178195199.adsl.alicedsl.de) |
00:57.47 | *** join/#htc-linux apt (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
00:57.47 | *** topic/#htc-linux is HTC Linux Channel: Find logs at http://apt.rikers.org/%23htc-linux/ | please check http://handhelds.org/moin/moin.cgi/HTC_2dPhones | http://wiki.xda-developers.com/index.php?pagename=Xanadux |
01:16.49 | *** join/#htc-linux apt_ (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) |
01:16.49 | *** topic/#htc-linux is HTC Linux Channel: Find logs at http://apt.rikers.org/%23htc-linux/ | please check http://handhelds.org/moin/moin.cgi/HTC_2dPhones | http://wiki.xda-developers.com/index.php?pagename=Xanadux |
01:46.01 | lkcl | http://www.pharscape.org/index.php?option=content&task=view&id=28%22 |
01:53.45 | BabelO | nice howto lkcl :) |
02:08.58 | *** join/#htc-linux LunohoD_ (n=alex@e180070144.adsl.alicedsl.de) |
02:39.23 | *** join/#htc-linux g3gg0__ (n=g3gg0@ppp-82-135-86-240.dynamic.mnet-online.de) |
03:16.45 | *** join/#htc-linux rob_w_T01 (n=bob@p85.212.151.97.tisdip.tiscali.de) |
03:24.39 | Kevin2 | FYI - I've updated http://www.handhelds.org/moin/moin.cgi/ApachePhoneTrace with more tracing tips. I've also uploaded a newer haret binary with slightly better poll reporting. http://www.handhelds.org/~koconnor/haret/haret-20061216.exe |
03:26.05 | *** join/#htc-linux Paynters (i=Paynters@d207-216-231-218.bchsia.telus.net) |
03:42.17 | Paynters | hi |
03:42.35 | Paynters | can anyone recommend a kernel for an htc apache? |
04:16.22 | Kevin2 | Paynters: We use hh.org cvs. The CVS tree is up to date for apache. |
04:16.52 | Paynters | oooo thanks |
04:16.54 | lkcl | Kevin2: i've added CDEX watching for blueangel to haret-gnu |
04:16.58 | Kevin2 | :pserver:anoncvs:anoncvs@anoncvs.handhelds.org:/cvs linux/kernel26 |
04:17.14 | lkcl | and some descriptions to the ASIC3 GPIO banks A to C |
04:17.19 | Kevin2 | lkcl: What is CDEX watching? |
04:17.32 | lkcl | well it's like ASIC3 watching only it's CDEX :) |
04:18.01 | lkcl | asic3 has some clock enables - 16 of them. |
04:18.09 | lkcl | something like that. |
04:18.47 | lkcl | that chip is just... _so_ significant, it really is. |
04:19.05 | lkcl | it's only been three years working out how it's used and how it works. |
04:19.39 | lkcl | i'm just compiling up pierrox's eten g500 kernel. |
04:27.09 | lkcl | um.... why don't i have an htc magician??? this is a pxa272 device, uda1380, gsm/gprs etc. etc. ermm.... |
04:29.06 | lkcl | ok, well _that's_ sorted out, thanks to ebay. |
04:29.07 | lkcl | duh. |
05:02.04 | *** join/#htc-linux Paynters_ (n=Paynters@d207-81-246-67.bchsia.telus.net) |
05:02.32 | *** join/#htc-linux Paynters__ (n=Paynters@d207-81-246-67.bchsia.telus.net) |
06:14.12 | *** join/#htc-linux iggy (n=iggy@gentoo/developer/iggy) [NETSPLIT VICTIM] |
06:14.12 | *** join/#htc-linux lpotter (n=ljp@c210-49-127-23.rochd1.qld.optusnet.com.au) |
06:14.12 | *** join/#htc-linux drno (n=drno@eki.student.utwente.nl) [NETSPLIT VICTIM] |
06:14.15 | *** join/#htc-linux eway (n=ellis@host-87-74-241-174.bulldogdsl.com) |
07:09.41 | *** join/#htc-linux ralith_ (n=ralith@soggy202.drizzle.com) |
08:11.31 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
08:34.12 | *** join/#htc-linux rmoravcik (n=rmoravci@icm7-orange.orange.sk) |
08:43.43 | *** join/#htc-linux pH5 (n=ph5@e178227177.adsl.alicedsl.de) |
08:58.46 | *** join/#htc-linux pwr (n=pwr@86.121.146.153) |
09:11.17 | *** join/#htc-linux veni (n=veni@LSt-Amand-152-32-13-217.w82-127.abo.wanadoo.fr) |
09:11.21 | veni | hi |
09:16.13 | pH5 | hi veni |
09:17.57 | veni | =) |
09:43.44 | *** join/#htc-linux RoEn_PC (n=roen@p54A64964.dip.t-dialin.net) |
09:46.04 | *** join/#htc-linux pwr (n=pwr@86.121.147.64) |
09:50.34 | goxboxlive | good morning |
09:53.50 | *** join/#htc-linux BabelO (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
10:00.36 | pH5 | good morning goxboxlive |
10:04.42 | BabelO | hi all |
10:05.56 | pH5 | hi BabelO |
10:08.15 | goxboxlive | pH5: Do you know how gpe backlight control is handeled. It adjust my backlight down to value "1" and that is not off. I want it to "0" wich is off. (in the /sys/class/.../brightness) |
10:10.40 | pH5 | goxboxlive: I think somebody who doesn't have a transflective screen complained that the backlight turns off completely when pulling the control down to the bottom |
10:11.47 | goxboxlive | hmm ok, but isent that the point? Shouldnt the backlight turn completly off when the control is in the bottom? Do you know where in the source i can change this? wich file? |
10:15.54 | pH5 | goxboxlive: yes, for our devices this would be the correct way. but for devices without a transflective screen, the display just turns black without a backlight and you have no way to increase the brightness again because you won't find the on-screen control :) |
10:16.06 | pH5 | the package is minilite |
10:16.52 | goxboxlive | ok |
10:16.54 | goxboxlive | thx |
10:16.59 | pH5 | in sysclass_set_level(int level): |
10:17.03 | pH5 | <PROTECTED> |
10:17.07 | pH5 | <PROTECTED> |
10:17.23 | goxboxlive | I'll change it localy |
10:19.33 | pH5 | there already is some model detection in main() |
10:19.45 | pH5 | case H36XX: |
10:19.47 | pH5 | <PROTECTED> |
10:19.51 | pH5 | <PROTECTED> |
10:19.52 | pH5 | etc. |
10:21.02 | pH5 | we could expand that with our models, remove the check in sysclass_set_level and use gtk_vscale_new_with_range (1, 255, 1) for devices without transflective/reflective screens |
10:22.41 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
10:24.14 | goxboxlive | pH5: Do you have write access to gpe source? |
10:24.25 | goxboxlive | good morning psokolovsky |
10:24.50 | goxboxlive | pH5: Or do we have to ask florian to do it? |
10:26.23 | pH5 | goxboxlive: we should ask florian. I have only ever contributed one patch to GPE, so I didn't bother to ask for write access. |
10:26.25 | pH5 | hi psokolovsky |
10:27.07 | goxboxlive | ok |
11:20.25 | *** join/#htc-linux rmoravcik (n=rmoravci@icm1-orange.orange.sk) |
11:27.59 | *** join/#htc-linux asylumed (n=insanity@196.211.116.2) |
11:45.23 | cr2 | hi all |
11:45.38 | BabelO | hi cr2 |
11:46.00 | cr2 | asylumed: if the irq handling is pushed down to the ds1wm, why do you enable the irq ? |
11:46.04 | cr2 | +irq = asic3_irq_base(&blueangel_asic3.dev) + BLUEANGEL_OWM_IRQ; |
11:46.04 | cr2 | <PROTECTED> |
11:46.04 | cr2 | +enable_irq(irq); |
11:46.05 | cr2 | +blueangel_ds1wm_resources[1].start = irq; |
11:46.18 | cr2 | hi BabelO |
11:46.47 | asylumed | cr2: hmmm that may be an artifact of trying everything to get it to work... |
11:47.58 | asylumed | cr2: will test now... |
11:48.06 | asylumed | thx |
11:50.03 | cr2 | the rest looks good :) |
11:50.15 | cr2 | should i apply it ? |
11:50.15 | asylumed | cr2: seems it's needed |
11:50.30 | BabelO | yes cr2 it works fine for me too :) |
11:50.35 | cr2 | it may be needed, but it does not look right. |
11:50.47 | asylumed | cr2: i don't think the ds1wm enables the irq... |
11:51.06 | cr2 | but it does set_irq_type ? |
11:51.11 | asylumed | cr2: yip... |
11:51.25 | asylumed | cr2: i agree... i don't like the look of it either... |
11:51.32 | *** join/#htc-linux Hawk|- (n=Hawk@p57A52852.dip0.t-ipconnect.de) |
11:52.07 | cr2 | you have set it up, and the driver should handle the irq itself. |
11:55.13 | asylumed | cr2: so, should it be changed to enable the irq? (ds1wm i mean?) |
11:55.37 | pH5 | asylumed: are asic3 irqs disabled by default? |
11:55.47 | asylumed | pH5: seems so... |
11:56.43 | cr2 | asylumed: gpiod10 changed ? i need to look into wiki. |
11:57.07 | asylumed | hmmm, though asic3_keys does not enable_irq's |
11:57.22 | pH5 | asylumed: that's why I was wondering |
11:57.27 | cr2 | ok. looks like a bug in the header. |
11:58.27 | cr2 | GPIOD6 0x0040 Kbd controller reset O |
11:58.30 | cr2 | is missing. |
11:58.56 | asylumed | cr2: damn... this lack of sleep is getting to me... |
11:59.32 | cr2 | :) |
12:00.01 | asylumed | cr2: was just thinking about how complete it looked :P seems i can't count to 16... |
12:00.40 | cr2 | well, a lot of things are missing there. some may be incorrect. |
12:01.26 | asylumed | cr2: GPIOD i mean... |
12:02.20 | asylumed | cr2: where does it say GPIOD10 ? |
12:03.41 | cr2 | in the old header. |
12:03.57 | cr2 | do you want to update blueangel_defconfig ? |
12:05.18 | asylumed | cr2: k, will send you one soon... |
12:05.35 | cr2 | udc is disabled |
12:06.28 | cr2 | arch/arm/mach-pxa/htcblueangel/blueangel_ds1wm.c:47: error: unknown field ‘falling_edge’ specified in initializer |
12:07.15 | asylumed | cr2: is the ds1wm patch applied? |
12:07.36 | cr2 | you did not post it ? |
12:07.44 | asylumed | *sigh* |
12:08.42 | cr2 | i'll commit the rest. |
12:09.02 | asylumed | cr2: done... |
12:11.35 | asylumed | was there ever any solution for the missing find_bus problem? i see the conversation just died in kernel-discuss... and i now i have a similar problem as with the h2200 |
12:14.31 | pH5 | asylumed: psokolovsky re-added find-bus |
12:14.41 | pH5 | the alternative is to export w1_bus_type |
12:14.58 | cr2 | when 10 socs will use it, it will make a difference. |
12:14.59 | asylumed | pH5: thx |
12:15.06 | pH5 | asylumed: like http://en.pastebin.ca/282241 |
12:15.26 | pH5 | I don't like the whole mechanism, actually. |
12:16.13 | pH5 | The battery driver shouldn't scan the bus. It should be the bus that calls the battery driver's probe function or something like this. |
12:17.18 | pH5 | I have added my magician battery driver to cvs. with the ba there will be three very similar battery drivers, so we could start thinking about how to abstract things into a common driver. |
12:17.42 | psokolovsky | Hi! |
12:17.49 | asylumed | psokolovsky: hi |
12:17.55 | BabelO | pH5: you have have two battery in magician ? |
12:17.59 | BabelO | hi psokolovsky |
12:18.08 | psokolovsky | pH5, exactly, why did you add it, instead of refactoring common driver? ;-) |
12:18.11 | rmoravcik | psokolovsky: hi |
12:18.43 | psokolovsky | rmoravcik, how's your ds1wm code? is it in good shape to submit it? |
12:19.22 | pH5 | psokolovsky: as always, time. oh well. |
12:19.42 | rmoravcik | psokolovsky: testing, i have little problem with wrong clock rate, but it is working |
12:19.49 | pH5 | BabelO: there is a backup battery, but for that I can only read the voltage with tsc2046 AUX_IN |
12:20.28 | psokolovsky | rmoravcik, nice! |
12:20.43 | BabelO | pH5: yes same in BA, i modify bluangel_battery.c to have Main and Backup in /sys/class tree |
12:20.45 | cr2 | isn't IRQT_* a compatibility hack ? |
12:21.10 | cr2 | the g500 defconfig does not compile for me. |
12:21.33 | goxboxlive | root@htcuniversal:/etc/bluetooth$ hcitool cc rfcomm1 |
12:21.33 | goxboxlive | Can't create connection: Input/output error |
12:21.34 | goxboxlive | root@htcuniversal:/etc/bluetooth$ |
12:22.12 | goxboxlive | I just cant connect any BT devices to my universal. The only one i have managed to cennect is my bluetooth mouse, but i missed the mouse service. |
12:22.13 | pH5 | cr2: oh, you're right. IRQ_TYPE_* are the correct ones.. |
12:25.10 | cr2 | rmoravcik: i'd like to have the rx3715 defconfig |
12:27.51 | rmoravcik | cr2: http://morgan.fbi.cz/tmp/config |
12:28.11 | psokolovsky | asylumed, so, tell me truth - how asic3, aic2 and ds1wm all fit together? ;-) |
12:28.41 | psokolovsky | rmoravcik, what about committing it to CVS? |
12:28.43 | asylumed | psokolovsky: magic :) |
12:29.12 | psokolovsky | asylumed, well, let's put that magic to order. where's actually ds1wm block contained? |
12:29.24 | rmoravcik | psokolovsky: good idea, but first want have udc, touchscreen and backlight in cvs |
12:30.05 | asylumed | psokolovsky: physical address 0x11000000 |
12:30.17 | psokolovsky | rmoravcik, well, I always thought that patches you submitted are complete ;-). so, please post them - I'm suure your patches are in better shape than g500's |
12:30.24 | asylumed | psokolovsky: just after the asic3 gpio memory... |
12:30.35 | psokolovsky | asylumed, what chip is that? |
12:30.46 | asylumed | psokolovsky: apparently as HTC_AIC2 |
12:30.51 | rmoravcik | psokolovsky: did you read my mail http://www.handhelds.org/hypermail/kernel-discuss/current/2495.html |
12:30.58 | cr2 | asylumed: imho, it's better to force the ds1wm users to specify the irq type explicitly. |
12:31.03 | psokolovsky | asylumed, what you mean "asic3 gpio memory"? |
12:31.22 | psokolovsky | rmoravcik, yep. So, those are not your patches, but from somebody else? |
12:31.53 | asylumed | psokolovsky: http://wiki.xda-developers.com/index.php?pagename=BlueangelVirtualMemoryMapping |
12:31.56 | cr2 | psokolovsky: maxim has distributed the VHDL for w1, so everybody and his friend just use it in their custom asics. |
12:32.19 | rmoravcik | psokolovsky: me and pierrox are using rtp's patches from h1940 |
12:32.22 | psokolovsky | cr2, that's old news ;-). |
12:32.36 | asylumed | psokolovsky: where it says BT/wifi led, etc... |
12:33.14 | cr2 | so htc asic3/aic2/aic3 have it too :) |
12:33.15 | psokolovsky | rmoravcik, can you please post another mail, with original links for those patches, so I can specify them in commit msg, and we know where to get newer versions? |
12:33.43 | psokolovsky | cr2, asylumed: specifically, for BA, what chip has ds1wm used for battery? |
12:33.55 | rmoravcik | psokolovsky: http://lechat.rtp-net.org/ipaq/patches/ |
12:34.03 | psokolovsky | rmoravcik, thanks |
12:34.22 | asylumed | psokolovsky: a ds2760 |
12:34.38 | psokolovsky | asylumed, wrong answer ;-) |
12:34.38 | lkcl | allo darlins |
12:34.54 | cr2 | rmoravcik: compiling. |
12:34.56 | psokolovsky | cr2, what chip has it? |
12:35.11 | pH5 | psokolovsky: the ds1wm is in the AIC2 chip, as in the magician (AIC3) |
12:35.12 | cr2 | chip ? |
12:35.28 | lkcl | pH5: why aren't your patches in the hh.org cvs tree, dude? |
12:35.34 | pH5 | which ones? |
12:35.39 | psokolovsky | pH5, asylumed: cool. how close are AIC2 and AIC3? |
12:35.47 | pH5 | severely |
12:35.49 | cr2 | psokolovsky: there are some custom HTC CPLDs |
12:36.03 | lkcl | oh, all of them :) |
12:36.26 | pH5 | lkcl: but they are |
12:36.39 | lkcl | ehh? oh, ok. cool! |
12:36.48 | psokolovsky | cr2, pH5: I understand. What I don't understand is why we don't have aic2_base, aic3_base or just common driver is they are the same ;-) |
12:37.08 | pH5 | psokolovsky: patience. we will. |
12:37.48 | cr2 | rmoravcik: compilation failed. |
12:37.51 | cr2 | <PROTECTED> |
12:37.52 | cr2 | drivers/input/touchscreen/s3c2410_ts.c: In function ‘touch_timer_fire’: |
12:37.52 | cr2 | drivers/input/touchscreen/s3c2410_ts.c:228: error: called object ‘3’ is not a function |
12:37.52 | cr2 | drivers/input/touchscreen/s3c2410_ts.c:228: warning: statement with no effect |
12:37.52 | cr2 | drivers/input/touchscreen/s3c2410_ts.c:237: error: called object ‘3’ is not a function |
12:38.14 | cr2 | the s3c2410 looks a bit fcked up now. |
12:38.28 | psokolovsky | pH5, asylumed: Well, maybe it's good idea to separate BA's ds1wm change to aic2_base right now? that's what your blueangel_ds1wm.c does right now, AFAICT |
12:38.54 | pH5 | psokolovsky: which battery driver do you use on h4000? |
12:39.09 | cr2 | psokolovsky: it sets the asic3 irq, and it is device-specific code. |
12:39.12 | psokolovsky | pH5, don't ask ;-). so far, ads7846 ;-) |
12:39.54 | psokolovsky | cr2, so, that's why I ask: how asic3, asic2 and ds1wm fit together |
12:40.10 | pH5 | the set/get register functions should go into a common aic2/3_base, as well as the led handling |
12:40.19 | pH5 | ds1wm is already abstracted sufficiently |
12:40.22 | cr2 | asylumed: i don't see ds276x in the cracked BA battery |
12:40.54 | pH5 | cr2: are there other chips that are compatible on the w1 command level? |
12:40.56 | cr2 | psokolovsky: it's the same ds1wm core + some other external setup, like clocks. |
12:40.57 | psokolovsky | cr2, what I'm sure, is that "static struct clk ds1wm_clk" likely doesn't belong to BA code, but to aic2/asic3 |
12:41.12 | pH5 | psokolovsky: in this case, it belongs into asic3 |
12:41.15 | psokolovsky | cr2, we ahndle clocks on soc level, no? |
12:41.19 | asylumed | cr2: i've successfully read data from it though... or something that behaves like it? family 30? |
12:41.30 | cr2 | pH5: the htc alpine CPLD. maybe the hermes, but i need to test it first. |
12:41.34 | psokolovsky | pH5, gotcha, so when we'll have it there? |
12:42.17 | cr2 | asylumed: i need to look in the logs, it was some microcontroller. |
12:42.20 | pH5 | asic3 clock handling needs a major revamp, I'm a bit afraid of touching this |
12:43.02 | pH5 | cr2: the commands that wince sends are exactly similar to what I have in the ds2760 spec sheet, from command bytes to the register positions. |
12:43.32 | psokolovsky | pH5, ok, so now we'll have it smeared across all machine-specific drivers? ;-) |
12:44.24 | psokolovsky | well, I hope rmoravcik posts his version (which is known to be vanilla asic3), so I can do that refactor. |
12:44.30 | pH5 | psokolovsky: I don't feel guilty, magician doesn't have asic3 ;) |
12:45.17 | psokolovsky | pH5, yep, that's why I poke asylumed for that ;-) |
12:45.43 | pH5 | actually, I do. hx4700 does have it, after all. |
12:46.11 | pH5 | but I too hoped to piggyback on rmoravcik's changes :-P |
12:46.28 | psokolovsky | ;-) ok, so let's wait for him ;-) |
12:48.12 | cr2 | pH5: on blueangel? |
12:48.59 | pH5 | cr2: more context please, I don't understand. |
12:49.03 | cr2 | wince does not report temperature and current on BA, and i don't see the physical ds276x in the battery. |
12:50.06 | pH5 | ah. yes. given that the battery.dll that asylumed sent me is indeed from a blueangel rom, the register reads are the same as on the magician. |
12:50.18 | cr2 | pH5: the asic3 clocks. i suggest to add the new api, and post the deprecation mail for those who write to the CDEX directly. then remove it as an exported symbol later. |
12:53.19 | lkcl | psokolovsky: i'm compiling a response to your message on kernel-discuss. pierrox is a newbie hacker whom i'm encouraging to get involved. he's pointed out that whilst there are a lot of s3c24xx ports around, very few of them are merged into a common cvs repository _anywhere_. |
12:53.35 | pH5 | sounds reasonable |
12:54.17 | psokolovsky | lkcl, I agree, but still please be a bit conservative a bit about merging newbies' patches ;-) |
12:54.53 | psokolovsky | lkcl, rmoravcik posted above a link for upstream of many those patches, and we should use latest version of them (and update regularly) |
12:55.01 | lkcl | he's pretty damn good: he's got the g500 nearly to the point where it's a useable linux PDA |
12:55.12 | lkcl | psoko: ok. |
12:55.53 | psokolovsky | lkcl, yep, cool. but try to encourage *him* to cleanup hist stuff ;-). I'm very glad he conquered MMC problem, etc.! |
12:56.04 | lkcl | rmoravcik: allo |
12:56.43 | cr2 | ok. the g500 and rx3715 do _not_ compile from the current CVS. for different reasons, btw. |
12:56.43 | *** join/#htc-linux rmoravcik (n=rmoravci@icm4-orange.orange.sk) |
12:57.09 | lkcl | o gawd, found rm's stuff, psoko... |
12:57.16 | *** join/#htc-linux rmoravcik (n=rmoravci@icm4-orange.orange.sk) |
12:57.32 | *** join/#htc-linux rmoravcik (n=rmoravci@icm4-orange.orange.sk) |
12:58.44 | rmoravcik | cr2: did you apply patches from my mail http://www.handhelds.org/hypermail/kernel-discuss/current/2495.html? |
12:59.05 | cr2 | rmoravcik: no :) |
12:59.05 | rmoravcik | i have little problem with my internet |
13:00.55 | cr2 | imho, the code can be suboptimal, not the latest, etc, but we need the CVS in a state which compiles and runs. |
13:10.49 | cr2 | rmoravcik: what was the patch against that: |
13:10.53 | cr2 | arch/arm/mach-s3c2410/built-in.o: In function `s3c2410_pm_drvinit': |
13:10.53 | cr2 | mach-rx3715.c:(.init.text+0xba4): undefined reference to `s3c2410_sysclass' |
13:11.20 | rmoravcik | cr2: making patch for s3c24xx |
13:11.52 | rmoravcik | cr2: want fix all problems |
13:12.50 | rmoravcik | cr2: can you try this patch http://lists.arm.linux.org.uk/pipermail/linux-arm-kernel/attachments/20061205/606be8d3/s3c24xx-cpu-ifdefs.bin |
13:13.39 | lkcl | rmoravcik: i'm happy to test s3c24xx stuff for the g500 |
13:14.52 | lkcl | rmoravcik: that patch is already in the LKCL_ETEN_G500_MERGE branch, and it makes the g500 stuff compile cleanly. |
13:15.11 | cr2 | ok, i'll apply it. |
13:16.22 | cr2 | compiles. |
13:16.26 | cr2 | WARNING: arch/arm/mach-s3c2410/rx3715_bt.o - Section mismatch: reference to .init.text:rx3715_bluetooth_probe from .data between '$d' (at offset 0x0) and 'dev_attr_power_control' |
13:17.02 | *** join/#htc-linux pH5_ (n=ph5@e178227177.adsl.alicedsl.de) |
13:17.10 | rmoravcik | cr2: i will took to it, you can disable rx3715_bluetooth support |
13:17.17 | cr2 | imho, time to commit it to CVS. |
13:17.43 | lkcl | cr2: yep please do cos then i can remove it from the merge stuff i'm doing. |
13:17.57 | cr2 | running diff. |
13:18.10 | lkcl | there's another one - with the Kconfig - which you really do have to apply as well. |
13:18.25 | cr2 | otherwise we will be always in such half-assed state. |
13:18.27 | *** join/#htc-linux rob_w (n=bob@p85.212.151.97.tisdip.tiscali.de) |
13:19.01 | lkcl | cr2: tell me about it.... |
13:19.29 | lkcl | s3c24xx really needs some love... |
13:20.58 | lkcl | rmoravcik: do you want me to apply your udc and touchscreen patches to s3c, including the rx3715? |
13:21.13 | rmoravcik | lkcl: yes please |
13:21.16 | lkcl | ok. |
13:21.22 | cr2 | done. |
13:21.36 | cr2 | lkcl: do cvs up first. |
13:21.40 | lkcl | cr2: ok. |
13:21.41 | cr2 | and then rediff. |
13:21.52 | lkcl | ok. |
13:21.55 | cr2 | i can compile rx3715 now. |
13:22.05 | lkcl | cool! |
13:22.09 | cr2 | rmoravcik: can you test it and tell me if it works ? |
13:22.34 | lkcl | i'll let you know about eten g500 in a mo. wife's taken the phone away with her so i'll double-check it works later, plus at least pierrox can try it. |
13:23.06 | rmoravcik | cr2: i will test it |
13:24.10 | lkcl | cr2: you added the s3c touchscreen as well, i notice. |
13:24.30 | cr2 | i've added a lot ;) |
13:24.40 | cr2 | even hermes :) |
13:24.48 | lkcl | hooray! |
13:25.18 | cr2 | but it's broekn |
13:25.51 | lkcl | o well. i can look at it |
13:25.55 | cr2 | HTC Hermes aka TyTN |
13:25.57 | cr2 | :) |
13:27.05 | psokolovsky | cr2, where're from those patches you committed? |
13:27.25 | psokolovsky | cr2, and how htcuniversal_ak4641.c relates to rx3715? |
13:27.36 | lkcl | cr2: yaay! |
13:29.34 | cr2 | psokolovsky: it was a minor change, forgot about it. edited out the #if 0 #endif block. |
13:30.05 | cr2 | psokolovsky: all the patches that rmoravicik needs for the rx3715 to work. |
13:30.13 | psokolovsky | sure, it's all just one sloppy commit ;-). but nevermind ;-) |
13:30.31 | psokolovsky | cr2, could you include upstream info in commit msg? |
13:31.21 | cr2 | can i edit commit messages aposteriori ? |
13:31.51 | cr2 | http://www.handhelds.org/hypermail/kernel-discuss/current/2495.html |
13:32.01 | cr2 | http://lists.arm.linux.org.uk/pipermail/linux-arm-kernel/attachments/20061205/606be8d3/s3c24xx-cpu-ifdefs.bin |
13:32.15 | cr2 | these two patchsets. |
13:33.20 | psokolovsky | nope ;-) but again, rmoravcik said those patches are from http://lechat.rtp-net.org/ipaq/patches/ . and we really should make sure we have latest versions, and know from where to update. |
13:34.45 | cr2 | psokolovsky: i think it should be left to rmoravcik right now. i'll take care of it when i'll have hermes up and running. |
13:35.36 | psokolovsky | cr2, I guess you're right - that's you guys who have s3 stuff ;-) |
13:35.38 | cr2 | at least the core + LCD |
13:37.34 | rmoravcik | cr2: backlight cleanup and support for rx3715 is in k-d |
13:38.09 | lkcl | rmoravcik: you want to take a look at pierrox's touchscreen driver and see if his driver works for you? |
13:38.30 | cr2 | rmoravcik: a link ? and the most important, does the current kernel boot ? |
13:38.44 | lkcl | rmoravcik: he told me that he tried the rx3715 touchscreen driver and it was _really_ bad, really jittery. so he had to cut/paste it to create a separate one for the g500 |
13:40.10 | rmoravcik | lkcl: i think i have the same problem with original driver, i will test pierrox alternative version |
13:40.48 | lkcl | ok great! you're more experienced at this than pierrox: would you mind tidying things up / merging to one driver, if it works? |
13:41.12 | rmoravcik | lkcl: yes |
13:41.15 | lkcl | ta dude |
13:41.43 | lkcl | i'll ask pierrox if he can sort out the backlight code, like posokolovsky asked. |
13:42.04 | lkcl | cr2: you committed rmoravcik's backlight driver for rx3715, right? |
13:42.13 | rmoravcik | lkcl: my code is the same |
13:42.20 | lkcl | oh. |
13:42.22 | lkcl | huh. |
13:43.05 | cr2 | lkcl: create the g500 subdirectory and do whatever you like there :) |
13:43.28 | lkcl | i'm not keen on being so evil quite just yet, cr2 |
13:43.37 | cr2 | lol. |
13:44.17 | lkcl | cr2: where's the hermes support in what you've committed, then, dude? |
13:44.21 | cr2 | i want to see the penguin on all 3 devices. touchscreen can wait. |
13:44.51 | cr2 | arch/arm/mach-s3c2410/htchermes |
13:45.20 | lkcl | huh. not there. |
13:45.40 | lkcl | maybe i need to cvs update duh.. |
13:46.25 | lkcl | ahh here we go, yep. |
13:46.36 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
13:47.24 | lkcl | rmoravcik: you got backlight working, right, on 3715? |
13:47.36 | rmoravcik | lkcl: yes |
13:48.04 | lkcl | huh. can't find it in hh.org cvs - where should i be looking? |
13:48.43 | lkcl | ok duh just received your email with a patch in it. |
13:49.03 | rmoravcik | cr2: compiling kernel |
13:49.18 | goxboxlive | scripts/kconfig/mconf arch/arm/Kconfig |
13:49.19 | goxboxlive | arch/arm/mach-s3c2410/Kconfig:124: can't open file "arch/arm/mach-s3c2410/htchermes/Kconfig" |
13:49.19 | goxboxlive | make[1]: *** [menuconfig] Fejl 1 |
13:49.19 | goxboxlive | make: *** [menuconfig] Fejl 2 |
13:50.24 | psokolovsky | Anyone worked with rtc in Linux? |
13:50.41 | cr2 | goxboxlive: cvs up |
13:51.02 | goxboxlive | i've just done it |
13:52.41 | goxboxlive | There is no subdiarctory named htchermes in mach-s3c2410 |
13:52.53 | cr2 | goxboxlive: wait a bit, probably there is some delay. |
13:52.59 | goxboxlive | ok |
13:53.03 | cr2 | is it empty ? |
13:53.13 | goxboxlive | there is no directory |
13:53.35 | pH5 | cr2: I suspect that is the typical anoncvs delay. |
13:54.14 | goxboxlive | I guess it's not ready for users yet |
13:54.18 | cr2 | http://handhelds.org/cgi-bin/cvsweb.cgi/linux/kernel26/arch/arm/mach-s3c2410/htchermes/ |
13:54.18 | rmoravcik | cr2: :) |
13:54.33 | goxboxlive | the cvs web is faster than anoncvs |
13:54.37 | goxboxlive | dont ask me why |
13:54.45 | cr2 | hmm. 30min ? |
13:54.56 | lkcl | sync delay stuff, depends on how anoncvs is updated. |
13:54.57 | cr2 | rmoravcik: it work ? |
13:55.18 | rmoravcik | cr2: compilation without errors |
13:55.25 | cr2 | ok. |
13:57.10 | lkcl | cr2, rmoravcik: i'm just about to cvs commit your 03-lcd.patch, for backlight |
13:57.29 | lkcl | if i can resolve cvs.handhelds.org via dns, that is... |
13:57.51 | rmoravcik | cr2: i will prepare update for s3c2410_buttons today, pierrox use it, i don't use it, i'm using gpio_keys |
13:58.30 | cr2 | lkcl: too late :) |
13:58.34 | lkcl | teehee |
13:59.36 | cr2 | rmoravik: do a cvs up for 03-lcd.patch |
13:59.36 | rmoravcik | cr2: thanks |
14:00.29 | cr2 | as long as you don't touch the mainline|pxa code nobody is affected afaik. |
14:00.57 | rmoravcik | cr2: by the way, i have success with ds1wm and asic3 |
14:01.29 | goxboxlive | I checked out a new kernel tree and the htchermes filder where included now. I have had simular problems problem when new folders have been added. |
14:02.18 | cr2 | rmoravcik: post the patch to k-d |
14:03.56 | cr2 | hx4700 uses it too. |
14:05.32 | rmoravcik | cr2: it is development version, i will look to it evening, it start working when i set inten_ias bit in int_en register in ds1wm |
14:07.53 | lkcl | cr2: hermes compile: arm-linux-gnu-ld: no machine record defined |
14:08.17 | cr2 | lkcl: i've said it's broken. |
14:08.21 | lkcl | oops :) |
14:08.33 | cr2 | i'm adapting the LCD driver now. |
14:09.01 | cr2 | the htchermes.o is not compiled. |
14:09.43 | *** join/#htc-linux elysion (n=elysion@k203c.lanhovi.tontut.fi) |
14:13.34 | lkcl | ok |
14:15.56 | cr2 | rmoravcik: i get a warning |
14:15.59 | cr2 | <PROTECTED> |
14:15.59 | cr2 | arch/arm/mach-s3c2410/s3c244x.c: In function ‘s3c244x_init_clocks’: |
14:15.59 | cr2 | arch/arm/mach-s3c2410/s3c244x.c:124: warning: implicit declaration of function ‘s3c2410_baseclk_add’ |
14:18.13 | cr2 | added obj-$(CONFIG_MACH_HTCHERMES) += htchermes/ |
14:18.15 | cr2 | to Makefile |
14:18.28 | cr2 | got a lot of errors :) |
14:18.47 | lkcl | wheeeee |
14:20.57 | cr2 | lkcl: update |
14:21.00 | lkcl | ok |
14:21.49 | pof | i see something big happening here :) |
14:22.09 | cr2 | fixed. |
14:22.18 | cr2 | kernel compiled :9 |
14:22.26 | cr2 | pof: ready to blow your hermes ? |
14:23.19 | pof | cr2: yah :D |
14:23.59 | cr2 | lkcl: update again. it compiles now. maybe something should be tuned in defconfig, though. i've taken the rx3715 defconfig. |
14:24.05 | cr2 | pof: wait a bit :) |
14:24.38 | cr2 | and we don't have asic3 & other things may be improperly setup. |
14:25.30 | lkcl | ok. |
14:25.49 | cr2 | removed asic3 from config. |
14:26.02 | cr2 | the kernel still compiles. |
14:26.28 | cr2 | # CONFIG_MACH_RX3715 is not set |
14:26.29 | cr2 | CONFIG_MACH_HTCHERMES=y |
14:26.29 | cr2 | CONFIG_HTCHERMES_LCD=y |
14:26.29 | cr2 | # CONFIG_MACH_OTOM is not set |
14:26.54 | cr2 | big bashing can begin! |
14:28.37 | lkcl | rx3715 defconfig isn't in hh.org cvs, cr2 |
14:29.08 | cr2 | added both, sorry. |
14:29.16 | cr2 | cvs up |
14:29.27 | lkcl | ok. |
14:31.09 | goxboxlive | cr2: The rtc-sa1100 works now. |
14:31.11 | lkcl | food. i _really_ gotta eat. hands getting cold. |
14:31.18 | lkcl | compile... brb... |
14:31.41 | cr2 | goxboxlive: ok. |
14:31.55 | goxboxlive | what is it good for? |
14:32.26 | cr2 | wall clock. |
14:32.29 | *** join/#htc-linux rmoravcik (n=rmoravci@icm3-orange.orange.sk) |
14:32.42 | goxboxlive | ok |
14:32.43 | cr2 | hard interrupts. |
14:33.01 | *** join/#htc-linux rmoravcik (n=rmoravci@icm3-orange.orange.sk) |
14:33.01 | cr2 | if you want realtime stuff. |
14:34.57 | lkcl | don't tell Babel0 i'm having porridge, ribena, glace cherries and barbeque-flavoured pringles. for breakfast. at 2:30pm. |
14:35.17 | BabelO | :) ah ah |
14:40.07 | cr2 | pof: i can upload the compiled zimage somewhere. |
14:40.17 | lkcl | zzz compiling on 1ghz machine.... zzzz... |
14:40.38 | cr2 | my microSD card has a lot of useful things ;) |
14:41.44 | cr2 | and goxboxlive knows, i prefer that others destroy their devices :) |
14:42.22 | lkcl | mwahahaha |
14:42.44 | cr2 | well, a lot of haret hacking coming. |
14:43.20 | cr2 | rmoravcik: where do i get a good s3c24xx CPU manual ? |
14:45.06 | lkcl | cr2: your caution and paranoia is touching |
14:45.09 | lkcl | really. |
14:45.35 | cr2 | :) |
14:46.29 | cr2 | my microSD is from a780, and has the maps. |
14:46.39 | lkcl | ahh. |
14:46.56 | lkcl | you got backups, right??? |
14:47.02 | lkcl | honestly... |
14:47.27 | cr2 | if it will be screwed, i need to find a windows machine to get them from CDs. |
14:47.34 | cr2 | of course not :) |
14:47.47 | cr2 | i have raid1 for my home, though. |
14:48.06 | *** join/#htc-linux rmoravci1 (n=rmoravci@icm8-orange.orange.sk) |
14:48.09 | cr2 | after the 180+180+250GB diskes failed |
14:48.27 | cr2 | taking some useful data with them ;) |
14:48.31 | *** join/#htc-linux rmoravci1 (n=rmoravci@icm8-orange.orange.sk) |
14:48.33 | goxboxlive | cr2: I noticed something. When i played a mp3 file while running gomunicator i was able to both hear music form the headphones and hear the voicecall form the lcd speaker. |
14:48.44 | pof | cr2: yes please (sory I was away) |
14:49.35 | cr2 | goxboxlive: music on the lcd speaker ? |
14:49.39 | cr2 | pof: where ? |
14:49.46 | lkcl | goboxlive: that's too funny |
14:50.15 | pof | ftp://xdaupload:xda@ftp.xda-developers.com/Hermes/ |
14:50.18 | cr2 | lkcl: with the right mixer driver you can mix them :) |
14:50.28 | cr2 | pof: ok. |
14:50.33 | lkcl | or, as it seems, if you are exploring the chip and make a mistake, even |
14:50.34 | pof | create a folder named linux if you want |
14:50.58 | goxboxlive | cr2: No, music in the headphones and voice (during a call) from the lcd speaker. It was at same time. |
14:51.24 | lkcl | ok. default.txt.... |
14:51.50 | cr2 | goxboxlive: i guess the mixer should switch to mono out, so you'll get the mic -> phone. |
14:52.16 | cr2 | and the aux_in, for phone -> speakers/BT/headset. |
14:53.05 | cr2 | have i told you that the current i2c/hx4700-ak4641.c mixer is too dumb ;) |
14:53.46 | lkcl | splat - exited. oops. |
14:53.55 | cr2 | pof: uploaded. |
14:54.17 | pof | thanks, going to test now! :D :D |
14:54.18 | cr2 | lkcl: which aret are you using ? |
14:54.26 | lkcl | s3c dedicated one. |
14:54.44 | lkcl | the one that's suitable for eten g500 |
14:54.47 | cr2 | ok. |
14:54.55 | goxboxlive | what device is pof using? |
14:55.00 | cr2 | hermes |
14:55.10 | goxboxlive | ok cool |
14:55.25 | cr2 | he has universal too, but likes hermes better :) |
14:55.35 | lkcl | what's the mtype again? |
14:55.38 | goxboxlive | :-) |
14:55.39 | cr2 | not surprising, considering the weight. |
14:55.59 | cr2 | but hermes lcd is really shitty, compared to universal :) |
14:56.06 | pof | i like the univ screen for browsing |
14:56.11 | pof | at vga resolution |
14:56.15 | goxboxlive | hmm seems like there is only me now who not have a hermes. |
14:56.28 | cr2 | i'd like to see somebody who dislikes universal LCD :D |
14:56.28 | pof | but hermes is my day-to-day phone :) |
14:56.30 | lkcl | everything's a compromise / step down from the universal, cr2 |
14:57.05 | cr2 | lkcl: we will see how it works out on openmoko. |
14:57.10 | lkcl | what's the mtype for hermes? |
14:57.17 | cr2 | hmm. |
14:57.43 | cr2 | MACH_HTCHERMES HTCHERMES 1213 |
14:57.52 | cr2 | dumb old haret. |
14:58.08 | cr2 | can't autodetect the machine. |
14:58.12 | lkcl | ok. |
14:58.55 | lkcl | splat again. |
14:58.58 | lkcl | hmmm... |
14:59.33 | cr2 | lkcl: does it crash/hang of exit ? |
14:59.42 | cr2 | s/ of/ or/ |
14:59.48 | lkcl | exits |
14:59.56 | lkcl | but haret is still running. |
14:59.58 | cr2 | looks like a haret bug. |
15:00.01 | lkcl | and leaves wince in an unstable state |
15:00.24 | cr2 | try Kevin's haret, but i'm unsure here. |
15:00.37 | lkcl | well, anything will do. |
15:00.44 | lkcl | where's his stuff, again? |
15:01.08 | cr2 | <PROTECTED> |
15:01.10 | lkcl | ta. |
15:02.17 | lkcl | allocated memory will overwrite itself - oops! |
15:02.29 | lkcl | at least it said 'found machine HTC hermes' |
15:02.51 | lkcl | there's a limit on initrd size, isn't there... |
15:02.59 | lkcl | on haret. |
15:03.02 | cr2 | don't use initrd. |
15:03.05 | lkcl | ack. |
15:03.06 | cr2 | yes. 4MB |
15:03.18 | cr2 | need to ask Kevin why. |
15:03.37 | cr2 | maybe he allocates 1 page for preloaded. |
15:04.51 | lkcl | nope - same error. |
15:05.12 | cr2 | 1200004 2006-12-17 15:53 zImage |
15:05.17 | lkcl | time for you to make a backup of that mini-sdcard, dude |
15:05.28 | cr2 | 1.2MB is not much. |
15:05.33 | cr2 | :) |
15:05.46 | *** join/#htc-linux g3gg0_ (n=g3gg0@ppp-82-135-64-106.dynamic.mnet-online.de) |
15:06.01 | lkcl | <PROTECTED> |
15:06.45 | pH5 | hehe. 1090668 2006-12-17 15:48 arch/arm/boot/zImage |
15:07.14 | lkcl | you showing off, pH5? |
15:07.26 | cr2 | hmm. hermes has 100% battery after 3 days. |
15:07.36 | pH5 | lkcl: yeah, "mine is smaller than yours" |
15:07.48 | lkcl | give it a couple of seconds, cr2... |
15:08.05 | lkcl | wince shows you the previous battery status and _then_ updates it... |
15:08.20 | cr2 | yeah, with linux |
15:08.28 | lkcl | ok. suggestions... |
15:08.57 | lkcl | merge in s3c244x support into haret-gnu? |
15:08.57 | lkcl | so we have some source code to start from? |
15:09.17 | cr2 | remove unused devices from htchermes.c |
15:09.23 | lkcl | ok. |
15:09.51 | cr2 | we should be able to boot with g500 haret. |
15:10.39 | lkcl | gpiokeys |
15:10.53 | lkcl | this is what pierrox uses: |
15:10.56 | lkcl | set ramaddr 0x30000000 |
15:10.57 | lkcl | set ramsize 0x04000000 |
15:10.57 | lkcl | set kernel "zImage" |
15:10.57 | lkcl | set mtype 1178 |
15:10.57 | lkcl | set bootspd 0 |
15:10.57 | lkcl | set cmdline "root=254:2 init=/sbin/init" |
15:10.59 | lkcl | bootlinux |
15:11.08 | lkcl | but i deliberately didn't put the two first lines |
15:11.21 | lkcl | ramaddr and ramsize. |
15:11.22 | cr2 | why? |
15:11.25 | lkcl | any idea what they're for? |
15:11.29 | lkcl | cos i don't know what they are! |
15:11.40 | cr2 | base RAM addr |
15:11.51 | cr2 | and the 64MB RAM size |
15:12.17 | lkcl | ahh |
15:12.21 | lkcl | ok |
15:12.41 | pierrox | hi all |
15:12.49 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HermesMemoryMap |
15:12.52 | lkcl | yep - that did it. |
15:13.00 | lkcl | hiya pierrox, we been having fun |
15:13.02 | pierrox | i see you are speaking about g500 :-) |
15:13.12 | lkcl | i have the haret penguin, at least. |
15:13.24 | cr2 | btw, you smash the bootloader. may be unhealthy, if you have usb host. |
15:13.28 | pierrox | lkcl: cool, and more ? |
15:13.36 | cr2 | waaaaah!!!!!!!!!!!!!!!!!!! |
15:14.11 | lkcl | i haven't a thingy-penguin. linux penguin, though. |
15:14.13 | cr2 | pierrox: coming soon . |
15:14.29 | lkcl | you need to do some cvs updating, pierrox |
15:14.43 | lkcl | could you test what's been cvs committed? |
15:15.06 | pierrox | yes i'll do |
15:15.17 | Kevin2 | Hi. |
15:15.32 | pierrox | i guess i arrive a bit too late : i have some new patches (broken in pieces) |
15:15.49 | cr2 | Kevin2: we ae trying to boot on hermes. |
15:15.56 | Kevin2 | Latest haret can load files >1MB -- I fixed that earlier. |
15:16.19 | Kevin2 | s/1MB/4MB/ |
15:16.22 | lkcl | pierrox: he he. well, it's fine - go for it. |
15:16.44 | cr2 | pierrox: make a diff to the current CVS |
15:16.44 | lkcl | guys my wife and baby are back so i'm going to bounce with baby for a bit. |
15:17.34 | lkcl | Kevin2: i got a 'allocated memory will overwrite itself' error on bootlinux with hermes. |
15:17.37 | Kevin2 | lkcl: Can you post the exact haret messages you get when it fails with "allocated memory will overwrite itself". |
15:18.04 | Kevin2 | Heh - Yeah, if you can forward all the info from the telnet session it will help me fix it. |
15:18.05 | lkcl | that's it, Kevin2 - that's all i get. i'm booting from the screen not via telnet |
15:18.08 | lkcl | brb.... |
15:18.10 | lkcl | will do - later. |
15:18.12 | lkcl | gotta go |
15:18.21 | Kevin2 | lkcl: Sure, you can use "log somefile.txt" too. |
15:18.21 | pierrox | lkcl: bye |
15:19.29 | cr2 | Kevin2: isn't the zImage relocatable by itself ? |
15:19.34 | pierrox | cr2: i am currently doing a cvs co, i will do the diff soon |
15:19.43 | cr2 | pierrox: ok. |
15:20.15 | pierrox | cr2: what has been put into cvs for the s3c24xx ? |
15:20.33 | cr2 | pierrox: a lot ;) |
15:20.43 | Kevin2 | cr2: I added a bunch more checks than original haret. I think original haret would have just crashed silently. For some reason on hermes wince is returning memory from the tip of memory, and when we try to relocate it to the tip, it will overwrite itself. |
15:21.31 | cr2 | Kevin2: ok, but imho it is possible to put zimage anywhere in RAM ? |
15:21.53 | pierrox | cr2: super, i may begin to work with the cvs tree so |
15:22.00 | cr2 | so it will decompress itself to the destination. |
15:22.08 | Kevin2 | cr2: I tried that, but it didn't work. Zimage is relocatable, but tags and initrd are not. |
15:22.18 | cr2 | i don't know the limits, though. |
15:22.20 | cr2 | ok. |
15:22.41 | cr2 | non-relocatable tags is silly ;-) |
15:22.49 | Kevin2 | One easy workaround to the "overwrite itself" issue is to try to boot without initrd. |
15:24.30 | cr2 | Kevin2: there was a wince flag to allocate at the max addr. m$ says it is deprecated, but afaik it is not. |
15:24.48 | cr2 | hmm. maybe it's impossible with allocphysmem. |
15:29.27 | Kevin2 | cr2: In order to make the boot code more portable, I stopped using allocphysmem. I was wondering if the "overwrite itself" message would appear - I think the way to fix it is to just allocate some additional pages so that we don't get that problem. |
15:31.15 | cr2 | Kevin2: do you use LockPages ? and do you still use "malloc" instaed of VirtualAlloc() ? |
15:31.36 | cr2 | the "original" haret code is not very portable too. |
15:32.18 | cr2 | the really portable code is in "VMALLOC" in haret-gnu |
15:32.18 | Kevin2 | cr2: No - I don't use LockPages or VirtualAlloc. |
15:32.52 | Kevin2 | I couldn't get LockPages to work - so instead, I just touch all the pages - that seems to work for me. |
15:33.02 | cr2 | ? why ? |
15:33.16 | cr2 | does vmalloc work for you ? |
15:33.33 | Kevin2 | What's wrong with just malloc? |
15:34.27 | cr2 | "officially" lockpages wants a pointer allocated by vmalloc. malloc calls LocalAlloc internally. |
15:36.29 | Kevin2 | I don't think I need LockPages - Just dirtying the memory will force wince to keep it locked down (where else can it put it?). |
15:39.04 | cr2 | sounds reasonable. but did not work for me... |
15:40.08 | cr2 | not really locked, it may be dropped 1ms after dirtying. |
15:40.45 | cr2 | with lockpages you know that the page tables are marked appropriately. |
15:42.00 | cr2 | if the machine is low on memory. like with 16MB or so, with 8MB available for alloc. |
15:42.10 | Kevin2 | cr2: I don't understand - where would WinCE move the pages to? It didn't it work for you when you were trying what? |
15:43.02 | cr2 | hmm. |
15:43.44 | cr2 | the calls to virt_to_phys failed for some reason for me at that time. |
15:44.42 | cr2 | i called malloc, copied the kernel there, but got '-1' for physical addresses on some pages. |
15:45.04 | cr2 | spent really a lot of time to debug these things. |
15:45.20 | Kevin2 | cr2: Yes - I've seen that. You must dirty the pages first, otherwise wince doesn't allocate them at all. |
15:45.37 | cr2 | maybe it's ok now after your changes in memory.c |
15:45.50 | cr2 | copying from disk to RAM does not dirty them ? |
15:46.06 | Kevin2 | I allocate with calloc to ensure they're dirtied. I haven't had a failure yet, and I do test for that scenario. |
15:46.17 | cr2 | ok. |
15:46.34 | Kevin2 | cr2: Hrmm.. It should.. |
15:50.18 | pierrox | what is the hh.ork kernel cvs version ? is it 2.6.18 ? |
15:50.31 | cr2 | .19-hhX |
15:51.24 | cr2 | VERSION = 2 |
15:51.24 | cr2 | PATCHLEVEL = 6 |
15:51.24 | cr2 | SUBLEVEL = 19 |
15:51.24 | cr2 | EXTRAVERSION = -hh6 |
15:51.48 | pierrox | cr2: ok the irq handler prototype changed |
15:52.19 | cr2 | pt_regs ? |
15:52.23 | pierrox | cr2: compilation fail on s3c2440mci : regs-sdi.h is not up todate |
15:52.29 | pierrox | cr2: yes |
15:52.37 | cr2 | pt_regs is dropped. |
15:52.57 | pierrox | cr2: my new patch has been updated for this, not the one in cvs |
15:53.03 | cr2 | which .config are you using ? |
15:53.16 | pierrox | cr2: i tried the g500_defconfig |
15:53.20 | cr2 | ok. |
15:54.02 | cr2 | i committed rx3715 and htchermes. |
15:54.21 | cr2 | hmm. need to update my tree. |
15:54.27 | pierrox | cr2: can i do some modifications, do a diff between the two, and send my diff for re-incorporation in cvs ? |
15:55.13 | cr2 | yes. |
15:56.33 | pierrox | cr2: ok, i do it, compile, run and test and i send you another diff |
15:56.38 | cr2 | ok. |
15:56.42 | cr2 | <PROTECTED> |
15:56.42 | cr2 | arch/arm/mach-s3c2410/mach-g500.c: In function ‘g500_machine_init’: |
15:56.42 | cr2 | arch/arm/mach-s3c2410/mach-g500.c:365: warning: implicit declaration of function ‘s3c_irq_wake’ |
15:57.36 | pierrox | cr2: seems old too, i have updated this too |
15:58.54 | pierrox | cr2: try http://www.pierrox.net/G500/20061217/patch-linux-2.6.19-g500-phe.tar.gz : the full patch is now broken into smaller ones |
16:01.35 | cr2 | ok. |
16:01.47 | lkcl | allo darlins |
16:01.55 | lkcl | i'm sort-of back |
16:02.22 | pierrox | re lkcl |
16:02.35 | pierrox | lkcl: i have read your mail about k-d |
16:02.55 | pierrox | lkcl: i did not subscribe to it yet |
16:04.33 | lkcl | you can see the archives.. |
16:05.56 | cr2 | pierrox: the convenience patch is of course unacceptable :) |
16:06.29 | cr2 | btw, where can i get a good GPIO reference manual ? |
16:06.31 | pierrox | cr2: i know :-) |
16:07.03 | pierrox | cr2: hence its name |
16:07.07 | rmoravci1 | cr2: datasheet |
16:07.23 | cr2 | maybe it's better to put the alt ts in g500 dir ? |
16:07.36 | cr2 | rmoravci1: where ? |
16:07.41 | lkcl | huh. odd. Kevin2: adding a couple of lines about ramaddr and ramsize and it booted. |
16:07.47 | lkcl | ok i got lines and a crash |
16:07.55 | lkcl | blue red cyan black |
16:07.57 | rmoravci1 | cr2: s3c2442? |
16:07.58 | lkcl | splat. |
16:08.07 | cr2 | rmoravci1: yes. |
16:08.09 | lkcl | and 'go go go' onscreen |
16:08.45 | rmoravci1 | cr2: i download it from samsung.com |
16:08.48 | pierrox | cr2: i think the ts driver could be far simpler without timers. i just miss how to detect stylus is up |
16:09.11 | cr2 | pierrox: you don't have a gpio for it ? |
16:09.29 | pierrox | cr2: no, it is in adc registers |
16:09.52 | lkcl | pierrox: that's just plain weird. |
16:09.59 | cr2 | rmoravci1: can you send me a link ? |
16:10.13 | pierrox | lkcl: i am not sure to understand samsung docs |
16:10.15 | lkcl | i wonder if the ati has some extra functionality like the asic3? |
16:10.27 | cr2 | pierrox: strange. i'm already thinking about tsc2046. |
16:10.39 | cr2 | lkcl: there is a CPLD ;) |
16:10.41 | rmoravci1 | cr2: try this http://www.samsung.com/Products/Semiconductor/MobileSoC/ApplicationProcessor/ARM9Series/SC32442/SC32442A43_Users_Manual_Rev1.1.pdf |
16:10.46 | lkcl | cr2: o dear |
16:10.50 | Kevin2 | lkcl: blue/red/cyan/black means it successfully loaded it to memory and jumped into the kernel. |
16:10.57 | lkcl | Kevin2: ack. ok. |
16:11.05 | rmoravci1 | cr2: or http://www.samsung.com/Products/Semiconductor/MobileSoC/ApplicationProcessor/ARM9Series/SC32442/SC32442A45_Users_Manual_Rev11.pdf |
16:11.20 | cr2 | rmoravci1: thanks. |
16:11.23 | lkcl | boot KERNEL=zImage INITRD=initrd |
16:11.23 | lkcl | Opening file zImage |
16:11.23 | lkcl | Opening file initrd |
16:11.23 | lkcl | Failed to load file \Storage Card\initrd |
16:11.23 | lkcl | boot MTYPE=1213 CMDLINE='init=/linuxrc' |
16:11.23 | lkcl | Allocated 298 pages (tags=00181000/32d9a000 kernel=00182000/32d9b000 initrd=002A7000/32ec0000 index=002A7000/32ec0000) |
16:11.27 | lkcl | Video buffer at phys=10800000 |
16:11.29 | lkcl | preload=412@002AA000/32f3a000 sj=002AA000 stack=002A8000/32ec1000 data=002A9000/32f38000 exec=32f3a0ec |
16:11.31 | lkcl | Reading 1199744 bytes... |
16:11.33 | lkcl | Read complete |
16:11.35 | lkcl | Launching to physical address 32f3a00c |
16:11.37 | lkcl | Trampoline setup (tram=132@00025000/1e025000/30c8a000) |
16:11.38 | *** join/#htc-linux rmoravcik (n=rmoravci@icm8-orange.orange.sk) |
16:11.39 | lkcl | MMU setup: mmu=A0250000/30850000 |
16:11.40 | lkcl | Video buffer at virt=4e800000 |
16:11.45 | lkcl | Go Go Go... |
16:12.09 | cr2 | lkcl: do you stop the DMA and usb host ? |
16:12.15 | cr2 | i guess no. |
16:12.38 | lkcl | usb host - no. lsusb locks up when i put the thingy back in. usb. |
16:12.40 | Kevin2 | lkcl: What is S3's start of memory? |
16:12.53 | lkcl | Kevin2: que? |
16:13.11 | lkcl | clueless, Kevin2 |
16:13.25 | Kevin2 | What did you set ramaddr and ramsize to? |
16:13.48 | lkcl | set ramaddr 0x30000000 |
16:13.48 | lkcl | set ramsize 0x04000000 |
16:13.54 | lkcl | absolutely no idea if it's correct :) |
16:14.45 | cr2 | http://wiki.xda-developers.com/index.php?pagename=HermesMemoryMap |
16:15.17 | pierrox | lkcl: ramsize should be easy to set... if you don't trust wince |
16:15.39 | lkcl | he he i have no idea what i'm doing |
16:16.37 | pierrox | lkcl: so it is perhaps time to sleep so :-) |
16:16.50 | cr2 | lkcl: do you set mem= ? |
16:17.47 | Kevin2 | cr2: So should start of me be 0x30000000 or 0x30600000? |
16:17.52 | Kevin2 | s/me/mem/ |
16:17.59 | cr2 | lkcl: add mem=64MB |
16:18.04 | lkcl | cr2: ok |
16:18.09 | pierrox | Kevin2: i suppose 0x30000000 |
16:18.14 | cr2 | Kevin2: 0x30000000 of course |
16:18.16 | lkcl | wheeeeee |
16:18.25 | Kevin2 | It won't matter if you set ramsize 0x04000000 |
16:18.27 | cr2 | Kevin2: what do you do with DMA ? |
16:18.41 | cr2 | mem=64M |
16:18.42 | Kevin2 | cr2: Nothing. |
16:18.43 | cr2 | actually |
16:18.58 | cr2 | Kevin2: i could have guessed :) |
16:19.16 | lkcl | oops. |
16:19.34 | Kevin2 | There is code in haret to disable dma on s3cxx processors, but the s3xxx haret had it disabled. |
16:19.35 | cr2 | what is the full mem= syntax ? |
16:20.02 | cr2 | mem=0x30000000@64M ? |
16:20.44 | Kevin2 | cr2: You really need to set those parameters via haret. The early decompression code can't see the cmdline, so it needs that info via the tags structure. |
16:21.18 | Kevin2 | s/can't see/doesn't use/ |
16:22.01 | cr2 | ok. but what do we have in tags now ? |
16:22.31 | Kevin2 | (11:13:47) lkcl: set ramaddr 0x30000000 |
16:22.31 | Kevin2 | (11:13:48) lkcl: set ramsize 0x04000000 |
16:23.00 | pierrox | pierrox: these are needed for the g500 and the old haret version |
16:23.35 | cr2 | ok. |
16:26.07 | pierrox | make mrproper fails in hh kernel tree : no fastfpe ? |
16:26.14 | Kevin2 | I'll add a default ramaddr to next version of haret. |
16:30.07 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
16:33.20 | *** join/#htc-linux JTRipper86 (n=jtripper@p54B1D86A.dip.t-dialin.net) |
16:34.29 | Kevin2 | Where is "g500 haret" source? |
16:35.38 | pierrox | Kevin2: on http://www.handhelds.org/moin/moin.cgi/EtenG500Downloads, there is the original old haret cvs snapshot and a patch against it |
16:41.34 | pierrox | cr2 lkcl, here is a diff against hh.org cvs tree : http://www.pierrox.net/G500/20061217/linux-hh-cvs-g500.diff |
16:41.42 | pierrox | it would be fine if you could add also : |
16:41.50 | pierrox | http://www.pierrox.net/G500/20061217/patch-linux-2.6.19-g500-phe/005-remove-excessive-printk.patch |
16:41.52 | pierrox | and |
16:42.00 | pierrox | http://www.pierrox.net/G500/20061217/patch-linux-2.6.19-g500-phe/006-pm.patch for suspend resume |
16:42.32 | Kevin2 | pierrox: Were you able to boot linux with that haret? |
16:42.51 | pierrox | Kevin2: yes, i use only this one |
16:47.29 | Kevin2 | pierrox: Any reason you use 8MB to start of initrd? |
16:47.56 | lkcl | cr2 can u patch? i got baby |
16:48.14 | pierrox | Kevin2: i noticed that my kernel was getting really big, above 4MB uncompressed |
16:48.49 | Kevin2 | pierrox: Mainline now uses 5MB - do you think that will be okay? |
16:49.43 | pierrox | Kevin2: yes, now i removed things, the 8MB limit was excessive |
16:50.34 | pierrox | Kevin2: and the g500 can now use modules, since it has a sd card to load them from somewhere :-) |
16:51.33 | Kevin2 | Okay - there is some weird stuff in the DMA shutdown code - you can't call Output at this phase of shutdown (nor can you call memPhysMap, but that is less fatal). |
16:53.05 | Kevin2 | pierrox: can you see if http://www.handhelds.org/~koconnor/haret/haret-20061216.exe works for you? You'll need to run "set ramsize 0x30000000". |
16:53.14 | pierrox | Kevin2: ok |
16:53.47 | pierrox | Kevin2: no need for ramaddr ? |
16:54.43 | Kevin2 | Oops - I meant set ramaddr. |
16:54.57 | pierrox | Kevin2: yes, and me ramsize, lol ! |
16:55.39 | pierrox | Kevin2: that works !!! |
16:56.23 | pierrox | Kevin2: it said "found machine: Default" |
16:56.34 | Kevin2 | pierrox: What machine do you have? |
16:56.57 | pierrox | Kevin2: a s3c2440 eten g500 |
16:57.24 | Kevin2 | Heh - duh. :-) |
16:57.53 | Kevin2 | pierrox: Can you create a file "earlyharetlog.txt" in the same directory as haret.exe, and then rerun haret.exe? |
16:58.02 | lkcl | agh, i not sure what to do. |
16:58.13 | Kevin2 | It should create a file "haretlog.txt" in the same directory. If you mail me that file, I'll add a G500 Machine type to haret. |
16:59.11 | pierrox | Kevin2: ok |
17:02.00 | pierrox | Kevin2: here it is : http://www.pierrox.net/G500/20061217/haretlog.txt |
17:02.22 | Kevin2 | 403 - Forbidden |
17:03.12 | pierrox | Kevin2: oh, sorry, stupid umask :-( corrected |
17:04.15 | *** join/#htc-linux pwr (n=pwr@86.121.148.184) |
17:07.07 | Kevin2 | pierrox: What do you call the machine? "G500" or "ETEN G500"? |
17:08.21 | Kevin2 | Are there any special steps you need to do? For example - overriding default memory size? |
17:08.23 | pierrox | Kevin2: let's say G500, that is enough for me |
17:09.13 | pierrox | Kevin2: in the old version haret was discovering 128MB instead of 64MB, because something has change between wm2003 and 2005. this may not be needed anymore now |
17:09.33 | pierrox | Kevin2: since haret has been improved a lot |
17:11.59 | Kevin2 | pierrox: can you try http://www.handhelds.org/~koconnor/haret/haret-20061217.exe |
17:12.28 | Kevin2 | You shouldn't need to set ramaddr. I didn't hardcode ramsize - let me know if it needs it. |
17:13.59 | pierrox | Kevin2: it doesn't seem to work, boot failed on initrd decompressing |
17:14.34 | Kevin2 | Did you set ramsize? |
17:14.58 | pierrox | Kevin2: i removed ramsize and keep ramaddr |
17:15.18 | Kevin2 | You shouldn't need ramaddr any more. I'm not sure about ramsize. |
17:15.52 | Kevin2 | It did say "Detected machine g500" right? |
17:16.35 | pierrox | Kevin2: ok for "Detected machine g500" |
17:17.40 | pierrox | Kevin2: without ramaddr it fails with "Allocated memory will overwrite itself" |
17:21.39 | Kevin2 | pierrox: Okay - can you download haret-20061217.exe again and try to boot. Don't set ramaddr, mtype, or ramsize - then post haretlog.txt". |
17:27.24 | *** join/#htc-linux RoEn_PC (n=roen@p54A64964.dip.t-dialin.net) |
17:27.43 | pierrox | Kevin2: looks almost good : ramaddr is ok, mtype too, but ramsize seems bad (kernel failed to uncompress initrd) |
17:29.48 | Kevin2 | pierrox: What does haretlog.txt say the wince memory size was? |
17:30.40 | pierrox | Kevin2: 0x08000000 : 128MB, this is the nand size |
17:31.49 | Kevin2 | pierrox: Okay - can you download haret-20061217.exe again and try to boot again? This version overrides ramsize on g500. |
17:34.14 | pierrox | Kevin2: that's ok this time :-) |
17:34.26 | pierrox | Kevin2: good work ! |
17:39.12 | Kevin2 | pierrox: Okay - it is all checked into CVS now. If you get a chance, you may want to check out CVS and have a look at the "src/mach/arch-s3.cpp" file. I have pulled in the dma shutdown code, but I have not activated it. |
17:40.19 | pierrox | Kevin2: ok |
17:53.41 | lkcl | ok am back again |
17:54.33 | pierrox | re lkcl |
17:54.41 | lkcl | Kevin2: doesn't work on hermes. yet. |
17:54.48 | lkcl | allo pierrox. |
17:54.57 | lkcl | did cr2 commit those patches while i was away? |
17:55.37 | lkcl | Kevin2: http://hands.com/~lkcl/hermes/haretlog.txt |
17:55.42 | pierrox | lkcl: i did not have news from cr2 since some time, so i suppose no |
17:55.48 | lkcl | ok i'll do it in a mo |
17:57.48 | Kevin2 | lkcl: Grab haret cvs from hh.org cvs, and modify src/mach/arch-s3.cpp so that it tries to disable dma. |
17:57.52 | pierrox | lkcl: i saw that there are still irq handlers in drivers/usb/gadget/s3c2410_udc.c and drivers/input/keyboard/s3c2410_buttons.c with old proto |
17:58.16 | lkcl | Kevin2: does haret compile with gnu tools? |
17:58.22 | lkcl | pierrox: ok. oops. |
17:58.26 | lkcl | patch time... |
17:59.03 | Kevin2 | lkcl: Yes - follow instructions in INSTALL file. |
17:59.12 | lkcl | ok. |
18:01.36 | lkcl | o fer goodness sake :) |
18:05.46 | lkcl | Kevin2: has anyone tried arm-wince-pe-gcc? |
18:06.50 | pierrox | lkcl: i tried with haret but i had some problems |
18:06.57 | lkcl | oh |
18:08.07 | lkcl | pierrox: of your patch regs-sdi.h already apparently includes some of your changes. |
18:08.46 | pierrox | lkcl: ? i did the diff against the cvs hh tree |
18:08.50 | lkcl | :) |
18:09.01 | pierrox | lkcl: did you commit this file ? |
18:09.14 | lkcl | anoncvs is delayed behind the main cvs |
18:09.43 | pierrox | lkcl: oh, this won't help, really cvs is a pain |
18:09.56 | lkcl | he he it is if you only have anon cvs. |
18:10.35 | pierrox | lkcl: this is the case, so i'll wait a bit, have you an idea of the delay ? |
18:11.12 | lkcl | it depends. if someone doesn't want their server overloaded with constant copying, they can make it 30 mins. |
18:11.14 | lkcl | but it depends. |
18:11.50 | pierrox | ok |
18:12.34 | lkcl | so if you ask on kd@hh.org for cvs access then at least you can stay up-to-date. and promise to review things with psokolovsky and others before doing any cvs commits. |
18:12.44 | *** join/#htc-linux pwr (n=pwr@86.121.147.76) |
18:14.38 | pierrox | lkcl: i just did a cvs update and some files have been patched :-) |
18:14.46 | lkcl | hooray. well, then that's good then. |
18:15.38 | lkcl | oh god, what a frickin mess the s3c24xx code is. |
18:16.10 | lkcl | pierrox: can you at least put #ifdef CONFIG_MACH_G500 around that pm section if (machine_is_g500())... |
18:17.15 | pierrox | lkcl: i thought machine_is_g500 was enough, since i suppose it should be discarded by the compiler seeing that the condition is always false |
18:17.29 | lkcl | no, that means that the test is carried out on every single machine. |
18:18.12 | pierrox | lkcl: so i wonder what this definition is good for |
18:18.27 | lkcl | it's also very bad practice, because if inside that test you had e.g. access to a specific device, then it would mean that everyone else has to link with that device driver code. |
18:18.53 | Kevin2 | lkcl: You mean arm-wince-pe-gcc from the videolan toolset? If so, no, it wont work. |
18:19.40 | pierrox | lkcl: the argument about linking is true, i will add this #ifdef |
18:21.11 | lkcl | Kevin2: argh :) |
18:22.57 | *** join/#htc-linux asylumed (n=insanity@196.211.116.2) |
18:23.37 | lkcl | pierrox: any idea what this lot is? |
18:23.40 | lkcl | drivers/usb/gadget/s3c2410_udc.c: In function 'usb_gadget_register_driver': |
18:23.40 | lkcl | drivers/usb/gadget/s3c2410_udc.c:1547: warning: ignoring return value of 'device_add', declared with attribute warn_unused_result |
18:23.40 | lkcl | drivers/usb/gadget/s3c2410_udc.c: In function 's3c2410_udc_probe': |
18:23.40 | lkcl | drivers/usb/gadget/s3c2410_udc.c:1737: warning: passing argument 2 of 'request_irq' from incompatible pointer type |
18:23.41 | lkcl | drivers/usb/gadget/s3c2410_udc.c:1752: warning: passing argument 2 of 'request_irq' from incompatible pointer type |
18:23.44 | lkcl | drivers/usb/gadget/s3c2410_udc.c:1769: warning: ignoring return value of 'device_create_file', declared with attribute warn_unused_result |
18:24.17 | pierrox | lkcl: the request_irq is because irq handlers proto has change, just remove the last argument in the handler proto |
18:24.21 | lkcl | ack. |
18:27.24 | pierrox | lkcl: http://www.pierrox.net/G500/20061217/pm.patch for the #ifdef CONFIG_MACH_G500 |
18:27.27 | lkcl | ok |
18:28.25 | lkcl | ahh it's an unsigned int to irq_handler_t |
18:29.07 | pierrox | lkcl: the third argument has been removed |
18:29.34 | lkcl | arg ok. |
18:30.25 | lkcl | ideally the entire codebase for s3c24xx needs a rewrite to update so that things like the pm.c stuff don't happen. |
18:32.58 | pierrox | lkcl: what do you mean about the s3c pm ? |
18:33.18 | lkcl | ok there needs to be a subdirectory arch/arm/mach-s3c2410/g500 |
18:33.31 | lkcl | and then a specific file g500-pm.c in there |
18:33.33 | lkcl | etc. etc. |
18:33.55 | pierrox | lkcl: this could be shared with h1940 and rx3715 |
18:34.08 | lkcl | the initialisation you've created? hmmm... |
18:34.48 | pierrox | lkcl: which initialisation ? |
18:34.53 | *** join/#htc-linux pwr (n=pwr@86.121.147.76) |
18:35.00 | lkcl | sorry, pierrox - i'm not making much sense. |
18:35.12 | lkcl | the power management initialisation - that memcpy. |
18:35.26 | lkcl | ok, then that's probably fine as it is. |
18:35.45 | pierrox | lkcl: yes, that may seem a bit weird, but it avoid to do a "reserve_bootmem_node" |
18:37.02 | pierrox | lkcl: in the h1940 and rx3715 there is a call to reserve_bootmem_node because the bootloader jumps in this memory space after resume |
18:37.38 | pierrox | lkcl: but for an unknown reason (yet), reserve_bootmem_node produces oops on my g500 later |
18:40.31 | pierrox | lkcl: see http://trinity.fluff.org/ben/linux-26/2619/2618-rc6-h1940-pm.patch |
18:42.30 | lkcl | oops |
18:43.27 | pierrox | lkcl: null pointer dereference, not good at all, some memory mess somewhere |
18:43.31 | lkcl | oops. |
18:44.51 | *** join/#htc-linux FossiFoo (n=Fossi@e176098065.adsl.alicedsl.de) |
18:49.47 | pierrox | lkcl: there is more than s3c2410-pm.c in the patch, the second pm.patch was just for the #ifdef CONFIG_MACH_G500, but there are other things in the base patch (sleep.S, g500.h and pm.h) |
18:50.02 | lkcl | ok. what the trinity one? |
18:50.32 | lkcl | no, not that one. 006- something |
18:50.35 | pierrox | http://www.pierrox.net/G500/20061217/patch-linux-2.6.19-g500-phe/006-pm.patch |
18:50.42 | pierrox | yes |
18:51.59 | lkcl | pierrox: arch/arm/mach-s3c2410/s3c2410-pm.c:35:27: error: asm/arch/g500.h: No such file or directory |
18:52.05 | lkcl | where's the g500.h file? |
18:52.16 | pierrox | yes, it is in http://www.pierrox.net/G500/20061217/patch-linux-2.6.19-g500-phe/006-pm.patch |
18:52.33 | lkcl | and that should be #ifdef'd too. |
18:53.01 | pierrox | lkcl: that should even probably not exist |
18:53.01 | lkcl | that's exactly why you use #ifdef. i'm compiling up hermes not g500 |
18:53.24 | pierrox | lkcl: i understand |
18:53.37 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
18:53.59 | lkcl | ok gotta feed baby. again |
18:54.18 | pierrox | :-) |
19:01.41 | pierrox | lkcl: you have been a bit quick with the removal of g500.h, why not #ifdef CONFIG_MACH_G500 ? |
19:11.13 | lkcl | hehe |
19:11.47 | lkcl | ok, it's back |
19:11.56 | *** join/#htc-linux goxboxlive (n=goxboxli@9.80-202-160.nextgentel.com) |
19:11.57 | lkcl | i thought you said it didn't matter! |
19:13.05 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
19:13.39 | *** join/#htc-linux rmoravcik (n=rmoravci@pc-3s0zt5w2e4y0vzmhnrzq3a21zqajzfw.users.student.utc.sk) |
19:14.00 | pierrox | lkcl: i said that it *should* not be here :-) |
19:26.02 | *** join/#htc-linux asylume1 (n=insanity@196.211.116.2) |
19:50.37 | *** join/#htc-linux beg (n=chatzill@80.92.96.55) |
19:56.00 | goxboxlive | Whats the the BA status in .19? Is it working? |
20:04.37 | BabelO | goxboxlive: yes ba is working, missing sound driver and acx |
20:05.32 | goxboxlive | BabelO: I c, what about gomunicator, are you able to do any calls or sms? |
20:06.06 | BabelO | goxboxlive: last time i ve tried in .12 sms and call works |
20:07.08 | BabelO | but now, we ( asylumed and me) have some problem with gpe, specially xfbatiw server |
20:08.11 | goxboxlive | i c |
20:08.43 | BabelO | goxboxlive: but we need prefix and postfix before at command |
20:09.06 | goxboxlive | Ok, They have becomed so cheap now. |
20:09.47 | BabelO | maybe doing a daemon or a userspace module to do thaht externaly |
20:10.18 | goxboxlive | ok |
20:10.36 | BabelO | i thinking about bought a HTC trinity |
20:11.03 | goxboxlive | wich one is that, have to check. |
20:11.42 | BabelO | i think it is same as hermes without keyboard |
20:13.02 | goxboxlive | ok, looks good. |
20:13.41 | goxboxlive | But is is expencive. |
20:14.13 | BabelO | oh i have Orange for Operator and enougth points |
20:15.34 | goxboxlive | ok, we dont have such in Norway. but what is the diffrents of the hermes deluxe and standard release? |
20:16.02 | goxboxlive | It is over €100 in diffrents |
20:16.53 | goxboxlive | Ohh i c. It is provided maps for the gps |
20:18.02 | BabelO | goxboxlive: do you know what is HSDPA ready. ? |
20:18.15 | *** join/#htc-linux asbjorn (n=asbjorn@2.81-167-1.customer.lyse.net) |
20:20.10 | goxboxlive | BabelO no |
20:21.23 | lkcl | an htc magician is only $EUR 200, new, guys. |
20:21.25 | pH5 | BabelO: "high-speed downlink packet access", see http://en.wikipedia.org/wiki/HSDPA |
20:24.36 | BabelO | pH5: thanks |
20:25.15 | BabelO | pH5, lkcl how goes sound with uda1380 ? |
20:26.21 | *** join/#htc-linux TripleDES (n=sergio@89.129.175.52) |
20:26.27 | BabelO | pH5: it is nice to monitor current and temperature on battery now :) |
20:26.47 | pH5 | BabelO: indeed! |
20:27.12 | pH5 | it is really about time. took us far too long to get proper battery drivers for ba/magician. |
20:28.21 | pH5 | lkcl: besides the fact that mine seem to run stable at 624MHz so far, the magician is a bit outdated now. |
21:00.34 | *** join/#htc-linux goxboxliv1 (n=root@9.80-202-160.nextgentel.com) |
21:30.23 | lkcl | Babel0: i've been doing eten g500 patches today. |
21:30.38 | BabelO | lkcl: i see i see :) |
21:30.46 | BabelO | is it a good phone ? |
21:31.00 | BabelO | battery life ? |
21:31.17 | lkcl | pH5: well, that doesn't really matter - i just have to have allll of them |
21:31.30 | lkcl | so i bought one on ebay yesterday errr this morning |
21:31.38 | *** join/#htc-linux goxboxlive|opie (n=goxboxli@9.80-202-160.nextgentel.com) |
21:32.26 | goxboxlive | goxboxlive|opie: test |
21:32.39 | BabelO | :) test goxboxlive |
21:32.39 | pH5 | lkcl: great! :) |
21:33.05 | goxboxlive | I am so lonely so i have to chat with my self :-) |
21:33.15 | goxboxlive|opie | goxboxlive test |
21:37.05 | *** join/#htc-linux rob_w_T01 (n=bob@p85.212.151.97.tisdip.tiscali.de) |
21:37.24 | pH5 | BabelO: I didn't touch sound again, still waiting for lrg's changes to ASoC |
21:57.24 | cr2 | hmm. adding samsung sump stuff to haret will take some time. |
21:59.58 | goxboxlive | Ihave used acx for over an hour now. Havent had any break. Also installed a lot of stuff. looks good. |
22:07.24 | cr2 | goxboxlive: cyclic suspend/resume as psokolovsky suggests ? |
22:08.00 | goxboxlive | What did he meant? |
22:10.27 | lkcl | allo again cr2 |
22:11.04 | goxboxlive | :-) |
22:11.06 | cr2 | hi lkcl |
22:11.11 | lkcl | he he there you go |
22:11.20 | cr2 | goxboxlive: do we have a list of bugs ? |
22:11.29 | cr2 | keyboard resume is no. 1 |
22:11.54 | lkcl | well, in between feeding baby and watching telly i've been trying to get thingy working. hermes. |
22:11.56 | goxboxlive | yes i have it in my head :-) |
22:13.24 | goxboxlive | cr2: keypad below lcd screen no.2 |
22:13.49 | cr2 | yes, switches and the night sensor. |
22:13.53 | goxboxlive | cr2: Sound so we can use it as a phone. no. 2,5 |
22:14.02 | cr2 | sound mixer. |
22:14.19 | goxboxlive | cr2: leds (keyboardled and flashlight led) |
22:14.51 | cr2 | asic3_leds and other leds. |
22:14.57 | goxboxlive | cr2: Camera no rush, the other are more importent, but it would have been cool thow. |
22:15.17 | cr2 | pxa_camera + ov9650 + ov6650 |
22:15.43 | cr2 | brf6150 custom bts. |
22:15.50 | goxboxlive | ok |
22:15.53 | cr2 | for headset setup. |
22:16.00 | goxboxlive | yes |
22:16.15 | cr2 | goxboxlive: btw, we don't need this TI_Init.... |
22:16.47 | cr2 | only one command from it, to switch to 921600, but it can be done with hcitool in init.d script. |
22:17.19 | cr2 | for the rest we should either upload the native bts script, or (better) create a full new one. |
22:17.20 | goxboxlive | ok, good |
22:17.28 | goxboxlive | right |
22:17.52 | cr2 | because there are some init custom ti bt commands, that are not in the script. |
22:19.05 | goxboxlive | i c |
22:20.22 | goxboxlive | I'll make a list at hh.org bugs tab. ok? |
22:20.48 | cr2 | ok. |
22:23.55 | lkcl | cr2: any clues/hints on this hermes startup thing? |
22:25.54 | cr2 | lkcl: what can we disable there ? |
22:26.14 | lkcl | i've removed several devices |
22:26.29 | lkcl | i wonder if i can do the same trick with vibra to find out what's going on |
22:26.30 | cr2 | we need to write "dump " haret support. the cpu is really awkward compared ti pxa, but it's doable. |
22:26.38 | cr2 | comment them out. |
22:26.47 | cr2 | have you found vibra gpio ? |
22:27.00 | lkcl | i'll take a look in 1 min.... back in 2mins... |
22:28.44 | *** join/#htc-linux psokolovsky (n=psokolov@237.usernat.ip.net.ua) |
22:43.38 | goxboxlive | cr2: todo updated: http://www.handhelds.org/moin/moin.cgi/UniversalBugs?action=show remind me if there anything i have forgot. |
22:46.01 | cr2 | it is event2 (together with the volume slider, just something is broken in asic3_keys driver). |
22:47.13 | cr2 | <PROTECTED> |
22:48.10 | goxboxlive | cr2: Ok, i c. But, before the new keybaord driver it was an own event together with the arrows on the keyboard. They where linked parallel |
22:48.32 | cr2 | vibration is led. |
22:48.54 | goxboxlive | ok, this is good, then the list isent that long. |
22:48.58 | cr2 | ok, it was, but now it is not :) |
22:49.03 | goxboxlive | I'll update |
22:49.30 | cr2 | <PROTECTED> |
22:53.07 | goxboxlive | I have one more, fix the htcuniversal_acx so it is not crashing the system when unloading the module. |
22:53.25 | cr2 | yes. |
22:55.01 | goxboxlive | I have had uni running now for two hours and 30 min with acx running. |
22:56.28 | goxboxlive | The Opie battery applet is saying that i have 34 min left |
22:59.30 | cr2 | maybe :) |
22:59.40 | goxboxlive | -) |
23:00.11 | BabelO | cr2: non idea for ba wifi ? |
23:00.34 | *** join/#htc-linux goxboxlive|opie (n=goxboxli@9.80-202-160.nextgentel.com) |
23:01.12 | goxboxlive|opie | root@htcuniversal:~$ uptime 00:00:49 up 2:35, |
23:02.44 | goxboxlive|opie | cr2 there are something i miss on the keyboard, and that is pipe. |
23:03.41 | cr2 | goxboxlive: in opie ? |
23:04.08 | goxboxlive|opie | i have tried to add it in the keymap file, but it fails loading. I tried the combination fn+1 |
23:04.14 | goxboxlive|opie | no not i opie |
23:04.17 | cr2 | BabelO: there should be a pcmcia setup. |
23:04.39 | cr2 | goxboxlive|opie: then it should work. |
23:05.05 | goxboxlive|opie | is the name of it >pipe< |
23:06.36 | cr2 | my console has crashed, can't check ;) |
23:07.17 | goxboxlive|opie | You have tried to resume ) |
23:09.32 | BabelO | cr2: one thing i can ' t understand is why i can change irq and firmware always load ? |
23:11.17 | cr2 | BabelO: then it does not need irq. irq is gpio17, you don't need to change it. |
23:12.13 | goxboxlive|opie | i dont understand why wlan is so stable today. I have even installed KonqE over wlan now. |
23:12.54 | cr2 | goxboxlive|opie: good weather :) |
23:14.14 | goxboxlive|opie | yes in opie, yes its cold today ) |
23:15.33 | goxboxlive|opie | ohh system message. The battery is runnnning very low. Well if i get disconnected you probably know why )) |
23:16.26 | cr2 | :) coonect the usb cable. |
23:19.13 | goxboxlive|opie | no i have to go to bed now. Tomorrow is work again. But next week i have some time off again. )) |
23:20.04 | goxboxlive|opie | 9 min left. I have to see how exact the battery applet is |
23:21.02 | cr2 | is... |
23:22.50 | goxboxlive|opie | BT, IRDA, and wlan activated. 4% and 6 min )) |
23:23.37 | goxboxlive|opie | and 100% light |
23:24.42 | cr2 | does IRDA work ? |
23:24.45 | goxboxlive|opie | uptime 3 hours. I think that is good with wlan on |
23:25.11 | goxboxlive|opie | i dont know, because irda is not working on my laptop. |
23:25.27 | BabelO | goxboxlive|opie: and with cpufrq you can save some time :) |
23:25.46 | goxboxlive|opie | i can check it tomorrow with my wifes dell x50v |
23:26.10 | goxboxlive|opie | babelO i have tried but it speeds up mu uni to 624MHz |
23:26.26 | goxboxlive|opie | and i am affraid of frying it |
23:26.51 | goxboxlive|opie | 1% and 2 min left |
23:27.06 | BabelO | now tehre is a special governor for mobile device :) |
23:27.14 | cr2 | goxboxlive|opie: already run haret on dell x50v ? :) |
23:31.26 | BabelO | cr2: gauge meter for uni battery life seem to be ok :) |
23:33.37 | cr2 | true. |
23:33.56 | cr2 | it shuts off wifi in wince at 10%. |
23:34.48 | BabelO | cpufreq governor is "conservative" |
23:42.10 | lkcl | cr2: s3c GPIO B4 is vibrate on hermes |
23:50.52 | BabelO | cr2: ok irq gpio 17 is requested only after a ifconfig wlan0 up |