**** BEGIN LOGGING AT Thu Oct 29 02:59:57 2009 Oct 29 03:58:22 hi there Oct 29 03:59:08 I'm using SHR Sept 06 and I still have the ERROR: Cannto satisfy the following dependencies....message http://pastebin.com/m260951af Oct 29 03:59:39 it doesn't be resolved yet? Oct 29 04:09:13 if you know the answer, please response here, i'll look in the log tomorrow Oct 29 05:35:24 freesmartphone.org: 03mickey 07specs * r3e7aa345d72b 10/ (3 files in 3 dirs): org.freesmartphone.GSM.SMS: add API for computing how many SMSes necessary to send a text Oct 29 06:04:32 freesmartphone.org: 03mickey 07cornucopia * r3d2471890b87 10/fsogsmd/ (9 files in 3 dirs): fsogsmd: sending arbitrary long SMSes works now transparently Oct 29 06:19:54 has anyone gotten the cyanogen mod on theyre openmoko handset yet Oct 29 06:20:17 their* Oct 29 07:22:49 morning Oct 29 08:34:29 hello, iam new in the openmoko project, and i try to run an application on it. i compile the sample application with toolchain and that works great. but when i start the application on the om "there where no output but exit code 2". can anybody help me? sorry for my english :) Oct 29 09:48:21 freesmartphone.org: 03mickey 07cornucopia * r2451a3c97cb8 10/fsogsmd/src/ (3 files in 2 dirs): fsogsmd: show properties in sms w/ type DELIVER Oct 29 10:05:50 hi Oct 29 10:06:48 I'm trying to mount my uSD-card with ssh over the freerunner because I don't have a cardreader Oct 29 10:07:33 sycoso, great and that works? Oct 29 10:07:37 you mean sshfs? Oct 29 10:07:41 I'm not 100% sure dropbear does that Oct 29 10:07:48 but it always says "remote host has disconnected" Oct 29 10:08:17 actually - it should simply be scp Oct 29 10:08:25 so it should - I'd have thought Oct 29 10:09:28 scp is only for copying Oct 29 10:09:52 I mean - the same interfaces as scp Oct 29 10:10:05 can't the FR be put into device mode? Oct 29 10:10:18 how? Oct 29 10:10:29 that would be perfect! Oct 29 10:12:18 Kamping_Kaiser: at least for me usb storage gadget is not reliable Oct 29 10:12:49 lindi-: ah, shame (I've never used it, just seen the option) Oct 29 10:19:40 in which distribution do you have the option, Kamping_Kaiser? Oct 29 10:28:23 since there are no more shr-u updates lately is it ok (compatible) to change to the shr-u repo from mrmoku which seems to got updates lately? Oct 29 10:29:54 sycoso: SHR iirc Oct 29 10:31:05 thx :) Oct 29 10:31:17 I'm just flashing mine :D Oct 29 10:35:42 does someony know why there are no more shr-u updates for about the last month or so? Oct 29 10:36:33 someony=someone :) Oct 29 10:38:11 what was the ideal video format for gta01, mpeg1 320x240? Oct 29 10:43:12 Fox_M|afk: developers are working on new branch of SHR (merge) Oct 29 10:45:10 what does that actually mean? Oct 29 10:45:57 have patience Oct 29 10:46:01 will be better :] Oct 29 10:47:52 Psi: something like that, yes. Though actually in tests, many codecs got about the same fps Oct 29 10:48:12 Psi: err - 240*320 Oct 29 10:48:18 Psi: rotating is a big task Oct 29 10:57:58 if the display is set to the 240x320 is it not alreay rotated? Oct 29 10:58:04 *already Oct 29 10:59:14 no. Oct 29 10:59:25 Err - yes Oct 29 10:59:32 I mean - you need to prerotate the video really Oct 29 11:04:15 so if i got a normal 320x240 video and played it on gta01 that had the lcd set to 240x320 it would try and scale 320 to fit 240 and give a tiny picture Oct 29 11:06:28 scaling is also heavy on the CPU Oct 29 11:06:50 No, it will rotate it fine if you ask it to - mplayer - it's just expensive, and halves or worse the frame rate Oct 29 11:08:50 k Oct 29 11:09:06 im just not getting why it needs rotating Oct 29 11:09:16 I'm assuming you want it fullscreen Oct 29 11:09:30 the hardware cannot do rotation. Oct 29 11:09:31 ya Oct 29 11:09:36 It's just a sideways framebuffer Oct 29 11:10:02 so if you want to display it rotated 90 degrees something - be it X or mplayer - has to do that rotation. Oct 29 11:10:37 how does the desktop manager display rotated when in 240x320 res? Oct 29 11:11:12 That's just X Oct 29 11:11:27 X draws stuff rotated Oct 29 11:12:46 ok Oct 29 11:15:55 what sound format do you recommend? Oct 29 11:16:17 mp3 worked fine. However. Oct 29 11:16:49 The 'normal' mplayer used IIRC 30% CPU for 128k audio - the integer 'mad' I think library one used more like 12% I think. Oct 29 11:41:31 hooray TDK Oct 29 11:41:40 ~lart subaru Oct 29 11:41:40 * apt drops a truckload of VAXen on subaru Oct 29 11:44:25 :O Oct 29 11:53:22 DocScrutinizer: (n810 gps bug) fyi the ticket is about nokia SW department saying "we've checked all our stuff and we are sure it's bug-free. If you suspect a hardware bug, go ask someone else (e.g. nokia customer care). We're not providing you any additional debugging information so you can pinpoint the issue." Oct 29 11:54:36 that's harsh for sure. OTOH I even somewhat understand that POV Oct 29 11:56:25 It seems the driver contains proprietary code of chip manuf (like gllin) which they simply can't disclose. And to me it seems it's usually working quite good Oct 29 11:58:04 so if *some* few customers complain about gps performance, it's probably a hw problem (at least that's what nokia sw dept comes to think), and they impossibly can give further instructions sw-related to do hw-debugging Oct 29 11:58:40 DocScrutinizer: yes but they could easily add several debug printf's to the driver so (quite knowledgeable) people would be able to diagnose the problem they could reproduce. Oct 29 11:59:08 DocScrutinizer: I understand the 810 is a sw gps Oct 29 11:59:13 of course they *could* have reacted more helpful and polite, but hey that's a sw-dept that's really not used to cope with FOSS Oct 29 11:59:27 it uses approximately the same amount of horsepower as the hammerhead on gta01 for hte 'driver'. Oct 29 11:59:59 SpeedEvil: [2009-10-29 12:56:24] It seems the driver contains proprietary code of chip manuf (like gllin) Oct 29 12:00:06 DocScrutinizer: claiming their software is bug-free is utterly unprofessional, imho Oct 29 12:00:20 oh Oct 29 12:02:18 They also made proprietary some other parts without any pressure from TI or whomever, e.g. their agps helper and related soft. Oct 29 12:02:51 thx Kamping_Kaiser, it works now!... with sshfs xD Oct 29 12:03:00 :) Oct 29 12:03:34 sure Oct 29 12:07:08 DocScrutinizer: so all in all after reading this bugreport i was quite disappointed by their immoral attitude. Talking like that to people who clearly know GPS in depth and who're willing to help to make their product better is not just impolite, it's part of being an asshole lifestyle. Oct 29 12:07:23 Well - yes and no. Oct 29 12:07:31 hm, not exactly clear statement Oct 29 12:07:36 The team that did the GPS integration has probably moved on. Oct 29 12:07:38 s/being/leading/ ? Oct 29 12:07:53 And bringing someone up to speed on that, and getting you an answer might cost actual money Oct 29 12:09:35 SpeedEvil: then they should've admitted that instead of saying "we checked everything, it's bugfree; and we don't care a tiny bit about possible HW bugs you must be experiencing because our software is bug-free and we checked on reference hardware that everything works". Oct 29 12:10:03 yeah Oct 29 12:10:31 'we checked it's bug-free' = 'I glanced at the bug tracker, which is empty' Oct 29 12:10:39 at best Oct 29 12:11:18 SpeedEvil: if you have a spare 5 min, you might want to look at the actual ticket i'm talking about to have an unbiased opinion. Oct 29 12:12:13 5 min ? LOL Oct 29 12:12:27 that ticket has some 150 comments Oct 29 12:12:59 and quite some of them rather lengthy Oct 29 12:13:18 DocScrutinizer: sure Oct 29 12:13:43 so chances are not even Nokia cared to read each of them thoroughly Oct 29 12:14:16 DocScrutinizer-8: well, i admit i spent at least 30min reading almost all of them. :) Oct 29 12:17:17 that's a really awesome reading speed still. 30min / (150posts * avrg 15lines/post) o.O Oct 29 12:20:56 and the posts I remember best are those of the maemo devels to nokia, asking to somewhat open up the code Oct 29 12:21:40 and the guy asking "how shall I handle and proceed with this ticket now?" Oct 29 12:39:04 PaulFertser: gps on N810 works absolutely flawlessly here, with AGPS beta. So what kind of printf() would you think could help inside closed source code, for people encountering problems while using absolutely same hw? Oct 29 12:39:46 s/ hw?/ sw?/ Oct 29 12:39:47 DocScrutinizer-8 meant: PaulFertser: gps on N810 works absolutely flawlessly here, with AGPS beta. So what kind of printf() would you think could help inside closed source code, for people encountering problems while using absolutely same sw? Oct 29 12:42:33 PaulFertser: also I'm not sure about how close to NDA'd domains AGPS helper needs to operate and whether it's such a simple thing to disclose the source for it Oct 29 12:49:08 DocScrutinizer-8: the guys suspected that sometimes there's a computational overflow or bogus sign extension or something like that that prevented gps from working reliably. Also it's unclear about how and when ephemeris and almanac data gets saved and restored. Oct 29 12:50:10 heh, the second part seems to apply to fso as well ;-P Oct 29 12:53:32 DocScrutinizer-8: nope, in fso it's all crystal-clear. Moreover our gps chip doesn't need almanac since it has enough correlators to scan all the satellites at the same time if i understand that correctly. Oct 29 12:53:52 and it seemed to me as if nokia and fso both do a store to pickle even if there's no good fix Oct 29 12:54:12 It needs one - otherwise it will take 30s or more to get a position Oct 29 12:54:25 as it has to wait for the ephemeris Oct 29 12:54:40 duh, that's a aspect completely new to me Oct 29 12:54:54 SpeedEvil: downloading of ephemeris takes 30 sec in any case. If you have them, you'll have fast fix, if not, you need to downloade them. Oct 29 12:55:24 30s of consistent error-free signal Oct 29 12:56:11 SpeedEvil: if the chip can't receive signals from all the sats at the same time, it can use almanac to avoid wasting time on listening for signals it can't hear anyway. But since our gta02 gps chip is quite modern it doesn't seem it needs almanac for anything ever. Oct 29 12:56:59 hmm, that's well within the "accuracy" of "guessing" time by me (which alphaone objected). I got an estimated 25s each time for TTFF Oct 29 12:57:44 and I never seen a faster TTFF Oct 29 12:57:51 SpeedEvil: i was a bit surprised to hear that myself but then i did testing and 1 min without any almanac and ephemeris data is indeed enough to get first fix. While downloading full almanac requires 12.5 minutes. Oct 29 12:58:07 yeah, bastard, that one Oct 29 12:58:15 PaulFertser: you don't need almanac for first fix with good signal. Oct 29 12:58:15 alphaone: hey :D Oct 29 12:58:16 also it seems to make no diff if I rm *.pickle or not Oct 29 12:58:31 SpeedEvil: why might i need almanac at all? Oct 29 12:58:38 It can help later fixes with slightly worse signal if you haven't been using the GPS for a couple of days though Oct 29 12:59:00 ttff will get very bad in the face of signal big enough to give - say 5% bit errors. Oct 29 12:59:05 SpeedEvil: i'd need to download ephemeris. And for that i'd need a good signal anyway. Oct 29 12:59:06 small enough Oct 29 12:59:30 you just need a TOW packet, you don't actually need the ephemeris Oct 29 12:59:41 alphaone: can you please comment? ;) Oct 29 12:59:46 What on? Oct 29 13:00:07 alphaone: is almanac useful for anything on gta02 gps? Oct 29 13:00:33 Almanace + approx. time + coarse position is used so the chip knows which SVs are worth looking for. Oct 29 13:00:51 SpeedEvil: i don't think we support approximate location when only almanac and TOW is known. Oct 29 13:01:07 alphaone: it looks like our chip can look at all SV at the same time, no? Oct 29 13:01:11 Full Almanac is 12 something minutes, but *every* SV sends out data for *all* other SVs as well Oct 29 13:01:14 Not all Oct 29 13:01:20 alphaone: that actually is a point for pauls POV. if we can look for all SV same time Oct 29 13:01:31 alphaone: how many then? ;) Oct 29 13:01:40 But the Antaris is advanced enough to quickly find the SVs overhead even without almanac Oct 29 13:02:23 Ephemeris is only transmitted for the SV you are receiving data from. Oct 29 13:02:51 alphaone: so it looks like performance of our chip with and without almanac is about the same, no? Oct 29 13:03:22 PaulFertser: TTFF performance yes Oct 29 13:03:35 ephemeris is important Oct 29 13:03:58 As ephemeris is is required for navigation. Oct 29 13:04:06 well, I've seen a max of 9 sats being "used". So obviously antaris can handle 9 concurrently - at least Oct 29 13:04:41 (Actually the chip has an option to use almanac navigation as well and it's enabled, but you get a quite bad fix with it - in the km range) Oct 29 13:05:22 for clarification: ephem is the "calibration" data to improve accuracy and compensate effects like weather etc? Oct 29 13:05:30 nope Oct 29 13:05:45 ephemeris is the fine orbit data Oct 29 13:05:56 almanac contains coarse orbit data Oct 29 13:06:03 ephemeris is much more precise Oct 29 13:06:04 hmm, probaly that's what I meant Oct 29 13:06:30 Ionospheric parameters are saved in the HUI packet Oct 29 13:06:33 DocScrutinizer-8: every sat transmits its own ephemeris every 30 sec. And also every sat transmits almanac for all sats but it takes 12.5 min to gather all the data. Oct 29 13:06:38 Health/UTC/Ionospheric Oct 29 13:06:53 and ephem holds valid for a few hours, while alm is for weeks ahaed Oct 29 13:07:03 Well, it's faster if you receive almanac from more than one sat Oct 29 13:07:22 Almanac is valid for 1-2 weeks Oct 29 13:07:33 ephemeris about 2 hours Oct 29 13:07:38 2-4 Oct 29 13:07:58 alphaone: so can you confirm that unless using special "almanac" navigation mode we don't need almanac for our antaris? Oct 29 13:09:32 seems alm is enabling a TTFF faster than the 30sec needed to dl ephem Oct 29 13:09:47 Can't be Oct 29 13:09:58 once you got ephem, you can forget alm Oct 29 13:10:21 that's what I understand now Oct 29 13:11:14 DocScrutinizer-8: 16:03 < alphaone> PaulFertser: TTFF performance yes Oct 29 13:11:24 (the same) Oct 29 13:11:32 PaulFertser: We don't really "need" almanac, but it will be downloaded anyway over time Oct 29 13:11:49 umm Oct 29 13:12:10 alphaone: sure Oct 29 13:12:20 And it probably helps the antaris look for SVs that come into view over time without using all the power to search all available channels Oct 29 13:13:21 alphaone: ok, thanks for the clarifications :) Oct 29 13:13:54 why can't alm help to get down TTFF? Oct 29 13:15:16 aiui with alm and time you got coarse pos of every sat. so as soon as you correlate 3 of them, you could spit out a first fix of low accuracy, even without ephem Oct 29 13:15:47 There is a difference between can, and what's implemented Oct 29 13:15:55 sure Oct 29 13:18:13 then, after you got ephem of the "used" sats (and possibly also the other unused but visible ones), you get a more precise fix, as well as a sub-30sec TTFF even without alm Oct 29 13:20:22 uBlox is quite open with info. Is there any paper that elaborates on what's actually implemented in antaris? Oct 29 13:22:05 (btw N810 is roundabout 25% faster on TTFF than FR, on *one* test to compare done ~30min ago) Oct 29 13:23:15 DocScrutinizer-8: Yes, accuracy is in km range with almanac navigation Oct 29 13:23:23 antaris supports that Oct 29 13:24:12 yep, that's what we occasionally see on GPS powerup - gps pos being way off real location Oct 29 13:24:58 DocScrutinizer-8: http://u-blox.com/en/download-center.html?task=view.download&cid=85 Oct 29 13:25:06 I think somebody in TPE did some tests wich exhibited this behaviour Oct 29 13:25:23 DocScrutinizer-8: That is an indication of the ephemeris fix not applied actually Oct 29 13:25:25 alphaone: :-) thanks Oct 29 13:25:32 Position way off and then no fix for minutes Oct 29 13:27:29 hmm, the link seems not to work :-/ Oct 29 13:28:24 here it does, nevermind Oct 29 13:29:26 .chm WTF? Oct 29 13:30:59 DocScrutinizer: docz! Oct 29 14:27:44 PaulFertser: on N810 I just found starting maps app doesn't get a fix this time. Starting AGPS app (and forcing update, though it probably did this automatically on start) fixed it to TTFF <60sec Oct 29 14:28:32 DocScrutinizer-8: AGPS is cheating. Lucky you if you have wifi everywhere you go. But i doubt gps is needed then. Oct 29 14:30:04 freesmartphone.org: 03mickey 07cornucopia * rb5934e7100b7 10/fsogsmd/src/lib/atcommand.vala: fsogsmd: secure against null names in sim phonebook Oct 29 14:30:13 PaulFertser: as last successfull fix using maps app is about an hour ago, this seems to indicate the AGPS data somewhat is destroyed by stopping GPS (after a too short time, so no alm/ephem downloaded) Oct 29 14:31:28 DocScrutinizer-8: even without agps the fix time should be reasonable Oct 29 14:32:47 yes, agree. But you know as well the issue a faulty agps data can mess up the chip to a mere dysfunction Oct 29 14:33:06 DocScrutinizer-8: the problems were observed long before introduction of beta agps app. Oct 29 14:33:19 I know Oct 29 14:34:17 second start of maps app -> again no fix Oct 29 14:34:43 Well, sad but it proves nokia sw department still sucks. Oct 29 14:34:49 max_posedon: ^^^ Oct 29 14:34:57 so I might consider to add to the 150 comments of this bug Oct 29 14:35:58 DocScrutinizer-8: my friend bought an n810 recently with the aim of using it as sip-client with 4G (wimax) usb dongle. Oct 29 14:36:46 restarted map again. fix immediately Oct 29 14:39:08 he should have baught N810-wimax-edition then Oct 29 14:41:43 though I admit I don't know details beyond mere existence of tthat device name Oct 29 14:42:25 DocScrutinizer-8: he knows about its existence too :) Oct 29 14:42:45 maybe it simply includes a 4G usb dongle? ;-P Oct 29 14:45:11 The production of the Wimax Edition of the Nokia N810 was canceled in January 2009 Oct 29 14:45:56 all N810 are EOL Oct 29 14:46:04 afaik Oct 29 14:46:07 Also it's unclear if it can be used in other wimax networks at all. Oct 29 14:48:48 (eol) that's why I got a spare in time ;-) Oct 29 14:49:28 Wise decision indeed :) Oct 29 14:49:39 heh, wikireader dropped from #4 to #5 Oct 29 14:52:54 yep, primary device now has severe stains on kbd, after >a year of daily excessife use Oct 29 14:54:20 s/if/iv/ Oct 29 14:54:20 DocScrutinizer-8 meant: yep, primary device now has severe stains on kbd, after >a year of daily excessive use Oct 29 15:00:42 DocScrutinizer-8: it was already on #5 this morning (8h ago for me) Oct 29 15:02:26 hello, i try to run an application on my FR. i compiled the sample application with toolchain and that works great. but when i start the application on the FR "there where no output but exit code 2". can anybody help me? sorry for my english :) Oct 29 15:10:47 hi folks Oct 29 15:11:26 wow... anyone up yet? Oct 29 15:14:37 bxyrk: many (different timezone :) Oct 29 15:17:06 they shut up when u came in:) Oct 29 15:17:13 indeed... 10 in the am here.. i'm guessing some of you folks are closer to 4-5pm? Oct 29 15:17:44 really don't know where that numer came frm exactly... Oct 29 15:17:45 10 am here Oct 29 15:17:46 bxyrk: indeed Oct 29 15:18:44 yeah... JUST realiaed my g/f's old tracphone uses a battery i can use in emergency Oct 29 15:19:47 wow... just looking at my typing... i am tired.. sorry people... don't intend to look like i just learned to type Oct 29 15:28:32 yup, i was up for a while trying to find something on the internet to help me with my wsod Oct 29 15:31:54 not meaning to bother any of you guys... guess i'm sounding like that guy that pops in a whines for help huh? Oct 29 15:50:31 freesmartphone.org: 03mickey 07cornucopia * r10584363e872 10/libfsobasics/fsobasics/ (smartkeyfile.vala utilities.vala): Oct 29 15:50:31 freesmartphone.org: libfsobasics: smartkeyfile: add write Oct 29 15:50:31 freesmartphone.org: utilities: allow creating files during a call to write() Oct 29 16:09:21 hm... how to add the latest community updates link on to the main wiki page? Oct 29 16:09:37 there's {{news}} but I do not understand how to edit those Oct 29 16:12:20 pieterc: wiki.openmoko.org/wiki/Community_Updates Oct 29 16:12:34 can someone tell me if there is a version of angstrom version working for FR? Oct 29 16:15:36 2007.2 Oct 29 16:24:01 freesmartphone.org: 03mickey 07cornucopia * r10a19faa2598 10/fsogsmd/src/ (5 files in 2 dirs): fsogsmd: misc SMS work Oct 29 16:33:22 slaxxer: were you talking to me? Oct 29 16:40:55 pbaxter: http://www.angstrom-distribution.org/narcissus/ Oct 29 16:41:11 Martix: but it doesn't work for me now Oct 29 16:41:20 at the end it says that there are some problems Oct 29 16:43:15 it works for me: http://www.angstrom-distribution.org/narcissus/deploy/om-gta02/1de305/random-80834d3a-image-om-gta02.tar.bz2 Oct 29 16:45:48 Martix: can i ask you favour? Oct 29 16:46:40 pbaxter: ok Oct 29 16:46:54 Martix: can you go on pvt? Oct 29 16:53:09 Image not found, something went wrong :/ Oct 29 16:53:54 pbaxter: maybe its something wrong with some package Oct 29 16:54:05 maybe Oct 29 16:54:11 i'll re-try later Oct 29 16:54:14 thanks a lot Martix Oct 29 16:54:25 your welcome Oct 29 16:56:26 btw. if someone likes Jabber, feel free to join to MUC: openmoko@conference.jabber.org Oct 29 16:56:57 a pro cesky pisici mame taky: openmoko@muc.openmoko.cz :-) Oct 29 16:57:02 thanks Oct 29 17:20:43 *why* does it cost $79 USD to have an item shipped to australia Oct 29 17:20:47 *rip* off Oct 29 17:40:42 hi lindi- Oct 29 17:40:44 :) Oct 29 17:41:31 Hi DocScrutinizer Oct 29 17:41:42 I have worked on the cpufreq scaling Oct 29 17:42:01 on the gta02 and I saw that is possible to run at 440Mhz Oct 29 17:42:17 I don't exaclty know Oct 29 17:42:21 if it is necessaty Oct 29 17:42:25 to increase voltage Oct 29 17:42:29 to the cpu Oct 29 17:42:37 but I'm using ondemand Oct 29 17:42:38 panicking: is it possible to do the reverse? Oct 29 17:42:49 yes Oct 29 17:42:55 I can move from 50 to 440 Mhz Oct 29 17:43:03 so the frequency can drop Oct 29 17:43:03 interesting. Oct 29 17:43:13 I have two big issue Oct 29 17:43:19 resume/suspend path Oct 29 17:43:22 it crash Oct 29 17:43:29 I have some trouble with debug board Oct 29 17:43:40 and I think that the audio Oct 29 17:43:43 perhaps lock the speed to be the normal speed before suspend and on wake up Oct 29 17:43:55 Yes that is what I want to do Oct 29 17:43:59 but it must work Oct 29 17:44:03 because it retry Oct 29 17:44:07 to reset parameter Oct 29 17:44:10 panicking: well does serial port work? Oct 29 17:44:19 The modem seem work Oct 29 17:44:23 now I would like to test Oct 29 17:44:24 panicking: even after suspend/resume? Oct 29 17:44:36 the suspend resume crash Oct 29 17:44:41 so that is an other problem Oct 29 17:44:46 I use a complete patch Oct 29 17:44:49 that recalculate Oct 29 17:44:54 the refresh rate Oct 29 17:44:56 of the ram Oct 29 17:45:00 and the iotiming Oct 29 17:45:12 is a backport of ben patch with the vdram regulator Oct 29 17:45:15 change Oct 29 17:45:17 in that patch Oct 29 17:45:34 Is it clear Oct 29 17:45:41 I think that 440 110 55 Oct 29 17:45:46 it's a good speed up Oct 29 17:45:49 for the neo Oct 29 17:45:54 :) Oct 29 17:46:03 but I don't know exacltu if I need to increase Oct 29 17:46:06 panicking: not if it reduces battery life tho. Oct 29 17:46:09 the voltage for 440 Mhz Oct 29 17:46:19 panicking: well I'd like to see lower frequency :) Oct 29 17:46:24 the battery life is not a problem Oct 29 17:46:24 because Oct 29 17:46:25 panicking: to save energy Oct 29 17:46:26 in ondemand mode Oct 29 17:46:29 the frequency Oct 29 17:46:34 panicking: so fix suspend :) Oct 29 17:46:48 panicking: or the patch won't get much testing :( Oct 29 17:46:59 yes I'm going to gix Oct 29 17:47:02 panicking: where is the patch? Oct 29 17:47:04 fix Oct 29 17:47:06 but do you know if it possible Oct 29 17:47:09 to use 440 mhz Oct 29 17:47:15 with a low voltage Oct 29 17:47:15 ? Oct 29 17:47:23 It's in my machine Oct 29 17:47:30 I have another one Oct 29 17:47:34 that use fcse Oct 29 17:47:44 panicking: no idea, is it really important? glamo is slowing us down anyway Oct 29 17:47:54 Fast Context Switch Oct 29 17:47:54 of arm Oct 29 17:48:03 glamo run at 110mhz Oct 29 17:48:08 the memory Oct 29 17:48:15 so glamo performance increase Oct 29 17:48:18 of 10 mhz Oct 29 17:48:30 but suppose that you need for a short time Oct 29 17:48:31 more power Oct 29 17:48:47 and the system wil give you a 440 Mhz cpu clock Oct 29 17:48:54 and increase the overral performace of the machine Oct 29 17:48:59 Excellent - interesting work! Oct 29 17:49:15 if the sdcard isn't in use, is it possible to allocate more for glamo? Oct 29 17:49:21 or is that a hardware issue. Oct 29 17:50:00 I don't know exacltu how glamo work and how it's correlate Oct 29 17:50:03 with the sdcard Oct 29 17:50:08 they use the same memory Oct 29 17:50:15 but different region Oct 29 17:50:32 sure. question really is if the sdcard is disabled can you use glamo any faster... Oct 29 17:51:10 uhm... the glamo is the sdcard controller Oct 29 17:51:26 larsc: that's why i was asking :) Oct 29 17:51:28 ok. Oct 29 17:51:31 so no. Oct 29 17:52:15 larsc, you can reuse my work on the 2.6.3x Oct 29 17:52:18 kernel Oct 29 17:52:26 I need some test for android Oct 29 17:52:52 d1b: i don't understand your question then? Oct 29 17:52:56 because we need more cpu and better glamo performance Oct 29 17:53:12 panicking: for android? Oct 29 17:53:24 larsc: i think you did. Oct 29 17:53:27 for android too :) Oct 29 17:53:45 i was just asking if the sdcard isn't in use, can you get more performance out of glamo or is it by hardware design not possible. Oct 29 17:53:57 Ok, I don't understand why openmoko doesn't invest anymore on gta02/gta03 Oct 29 17:54:08 I think that the android distrib is good inaugh to invest on Oct 29 17:54:10 d1b: the sd card is not the bottleneck of the glamo, so no. Oct 29 17:54:20 btw Oct 29 17:54:24 larsc: i thought it was.... Oct 29 17:54:37 d1b: bottle neck is the memory bus Oct 29 17:54:38 So I will try to fix suspend resume Oct 29 17:54:47 and I must buy another debug board Oct 29 17:54:48 :) Oct 29 17:54:49 larsc: oh. Oct 29 17:54:58 :( because I have a broken one Oct 29 17:55:08 d1b: if you don't use the sd card of course there is less memory to be transfered and graphics will be faster Oct 29 17:55:26 ok. Oct 29 17:55:28 larsc, bat the glamo memory is memory mapped Oct 29 17:55:28 ? Oct 29 17:55:38 panicking: yes it is Oct 29 17:55:48 so the transfer is from to? Oct 29 17:55:58 from XXX to XXX Oct 29 17:56:14 yes Oct 29 17:57:06 There is somenthing that I don't understand but I think that for you Oct 29 17:57:08 is obvius Oct 29 17:57:17 I write to the framebuffer memeory Oct 29 17:57:22 and the chip send info to the display Oct 29 17:57:38 I can wrtie memory to the sdcard in a different region too Oct 29 17:57:43 hi Oct 29 17:57:58 The transfer is on the memory bus Oct 29 17:58:23 So if the system copy memory or send info to the sd Oct 29 17:58:25 is the same Oct 29 17:58:37 the perforce depends on the memory transfer Oct 29 17:58:53 larsc, another issue is the refresh time Oct 29 17:58:58 we are sure that it is correct Oct 29 17:59:20 Because the memory bandwith depends from it Oct 29 17:59:51 what refresh time? glamo memory? Oct 29 18:00:00 gta02 memory Oct 29 18:00:04 I have a refresh of Oct 29 18:00:36 17.5usec Oct 29 18:00:43 I take it from qi Oct 29 18:19:24 so I hope that this value is correct and not too much Oct 29 18:19:32 I will take a look to the datasheet Oct 29 18:19:35 of the memory Oct 29 18:56:45