01:20.04 | *** part/#htc-linux davidchow (n=davidcho@cm203-168-216-8.hkcable.com.hk) |
06:56.44 | *** join/#htc-linux rob_w (n=bob@p213.54.204.107.tisdip.tiscali.de) |
07:15.12 | *** join/#htc-linux awelux (n=awelux__@dslb-084-058-161-064.pools.arcor-ip.net) |
08:20.54 | *** join/#htc-linux toi_ (n=peter@d54C27365.access.telenet.be) |
08:25.46 | *** join/#htc-linux goxboxlive (n=goxboxli@ti500710a080-7302.bb.online.no) |
09:37.51 | rob_w | hey cr2 |
09:40.01 | *** join/#htc-linux florian_kc (n=fuchs@dns.kernelconcepts.de) |
09:40.38 | florian_kc | good morning |
10:33.17 | goxboxlive | morning all |
10:34.45 | florian_kc | hi goxboxlive |
10:34.53 | goxboxlive | hello sir |
10:35.23 | florian_kc | that reminds me to look at your patch :-) |
10:35.36 | goxboxlive | no rush :-) |
10:36.24 | goxboxlive | Florian, there is also a failure in the package manager. It starts, find packages, dowmload them, but it failes installing them. |
10:37.26 | florian_kc | goxboxlive: right, oe is readonly this week anyway |
10:37.47 | goxboxlive | yes thats right because of the monotone update |
10:37.51 | florian_kc | goxboxlive: hrm nasty. sounds liek this thing needs some additional care |
10:38.57 | goxboxlive | I have a gpe-image here. I can upload it so you can test it yourself. Maybe there is something i do or dont do. |
10:41.23 | *** join/#htc-linux BoSJo (n=bosjo@130.80-203-23.nextgentel.com) |
10:43.25 | goxboxlive | Hi mickey|office did you get my mail? Did you understand it? I have also sent you one more, because of other packages wich have problems to compile with the new qte-common file. |
10:44.43 | mickey|office | no i didn't get any mail |
10:44.49 | mickey|office | which address are you using? |
10:45.34 | goxboxlive | handhelds |
10:45.46 | goxboxlive | the one you gave me here at IRC |
10:46.11 | mickey|office | mickeyl@handhelds.org ? |
10:46.14 | mickey|office | (notice the 'l') |
10:46.24 | goxboxlive | hold on i'll check |
10:46.29 | mickey|office | try mickey@tm.informatik.uni-frankfurt.de if that doesn't work |
10:48.25 | goxboxlive | i didnt know about the 'l' . I have sent them again, check now. |
10:48.47 | goxboxlive | there are 2 mails. |
10:52.56 | florian_kc | goxboxlive: that would be nice |
10:53.17 | goxboxlive | ok, use the provided zImage |
10:53.18 | florian_kc | goxboxlive: why did you comment out ALLOW_EMPTY_ipaq_pxa270 = 1 in the hh kernel config? |
10:54.01 | goxboxlive | because i added a line ALLOW_EMPTY_htcuniversal = 1 instead |
10:55.38 | florian_kc | goxboxlive: but that means that you change the behaviour for ipaq devices. |
10:57.01 | goxboxlive | I know, and that was what i am telling you in the mail. What shall we do, should we make a equal folder/subfolders but with fitting it to the Universal, or is it doable to let the bb file be mnore machine spesific |
10:57.18 | goxboxlive | what i mean is |
10:58.01 | goxboxlive | If it detect that it is a htcuniversal, it skipps those to patches and it use this: ALLOW_EMPTY_htcuniversal = 1 instead of this: ALLOW_EMPTY_ipaq_pxa270 = |
10:58.34 | florian_kc | goxboxlive: hmm in this case the settings won't conflict with each other, just have both in it. |
10:58.55 | goxboxlive | ok, thank you , i will. What about the patches? |
11:00.32 | florian_kc | goxboxlive: most of it looks pretty good... about the opie specific stuff we should ask mickey|office as i don't know much about it. |
11:01.20 | goxboxlive | florian_kc yes i agree. I have also sent mickey the monotone_diff. |
11:01.57 | florian_kc | ah perfect |
11:04.21 | goxboxlive | florian_kc The image is now uploaded at http://gnulinux.biz/files/Universal/tmp/ . Remember to add the pointrecall file in /etc |
11:04.32 | goxboxlive | And use the zImage in the same folder |
11:04.43 | florian_kc | goxboxlive: okay, many thanks |
11:05.00 | goxboxlive | welcome |
11:10.28 | rob_w | florian_kc, in Bug 540 there is a "blueangel and new himalaya diff .. can you have a look and maybe give it some love ? that brings big tslib support for both devices |
11:12.47 | florian_kc | rob_w: ok, let me see... |
11:15.28 | rob_w | have your time |
11:22.34 | florian_kc | rob_w: looks good, but needs to wait for oe |
11:23.08 | rob_w | even great |
11:23.51 | goxboxlive | mickey|office But actually, the apps i compile with the old qte-common.inc doesnt start, so they are acutally uselless. But They have to b ecompiles to managed to make a Opie-image |
11:39.07 | florian_kc | uh my universal is unlocked now |
11:40.43 | florian_kc | gah.. this software is *CRAP* |
11:44.52 | florian_kc | okay, is someone working on the gsm modem? |
11:45.45 | cr2 | hi |
11:46.08 | cr2 | florian_kc: what do you mean "working on the gsm" ? |
11:47.04 | florian_kc | cr2: making it usable |
11:47.23 | florian_kc | cr2: that original phone software is a major pain |
11:47.51 | cr2 | i can reset it, DCD goes up, but then "chat" does not get any response. |
11:48.45 | cr2 | it should be normal "at\r" stuff, no framing like on older phones. |
11:48.55 | cr2 | at least wince does it this way. |
11:49.39 | florian_kc | cr2: hmm... sounds like i should try to free up some time to take a look |
11:50.07 | cr2 | tracing gpios makes a more complex picture, but afaik it is not necessary for the modem interface to work. |
11:50.38 | cr2 | florian_kc: i can write a short summary what i know so far. |
11:51.24 | cr2 | it is easy to get "AT-Command interpreter ready" from the modem. |
11:51.27 | florian_kc | cr2: that would be cool, maybe somewhere in the wiki. |
11:51.38 | cr2 | ok. |
11:52.48 | cr2 | i should look for my handwritten notes. |
11:56.30 | florian_kc | rob_w: the permissions patch is in latest release already |
11:59.20 | mickey|office | goxboxlive: got them |
12:00.31 | cr2 | mickey|office: before i forget it. i have added the my locomo gpio/spi data into oz wiki. |
12:00.57 | mickey|office | cr2: thanks, cool |
12:06.21 | cr2 | here: http://wiki.openzaurus.org/FAQ#Zaurus_SL-5500_MMC.2FSDIO_technical_info |
12:07.48 | cr2 | i think it makes sense to write a 2.4 kernel module to show the register status in /proc, and to enable watching them. |
12:08.12 | cr2 | but i have a lot of other things to do now :) |
12:08.16 | mickey|office | i can imagine :)# |
12:09.11 | mickey|office | when i get my money back from the ebay scammer I will start to bid for a htcuni again |
12:11.10 | cr2 | yes. i think that with htcuniversal we can show that linux supports not only the devices that are 4 years old ;) |
12:12.53 | cr2 | i think htc can enhance it by soldering in the gps chip (which they left out) and +64MB ram, but it will remain the top of the line device for quite some time. |
12:13.04 | mickey|office | right |
12:13.20 | mickey|office | so you sound very confident to be able to get the gsm hardware working |
12:13.56 | cr2 | voice+sms yes. umts will be more difficult. see a780 ;) |
12:15.04 | cr2 | i don't know much about usb like Harald, but he is not moving very fast too. |
12:16.00 | mickey|office | that's because he doesn't have time since 2 months :) |
12:16.16 | mickey|office | and me neither :/ |
12:16.40 | cr2 | ok. |
12:17.30 | goxboxlive | hi cr2 |
12:18.25 | cr2 | htc did not change the base phone concept from original xda, and it seems that they use more simple solutions compared to motorola. makes our life easier :) |
12:18.30 | cr2 | hi goxboxlive |
12:19.00 | goxboxlive | Yesterday i managed to build images with OE without aplying any patches after baking. So now we are able to boot right out of OE. |
12:19.15 | cr2 | qte-mt ? |
12:19.20 | goxboxlive | :-) yes |
12:19.26 | cr2 | great. |
12:19.37 | goxboxlive | But i miss konq/E so much. |
12:20.00 | cr2 | mickey|office: why does konq-E use qt-E 3.3.5 ? |
12:20.55 | cr2 | i feel lost in this qte-10, mt-10, -12, 3.3.5 jungle ;-) |
12:22.00 | cr2 | goxboxlive: does konq-e build for the latest VGA zaurus ? |
12:22.52 | mickey|office | konq/E doesn't use 3.3.5, it needs qmake to build which we take from qte 3.3.5 |
12:23.03 | goxboxlive | yes mickey has built it , and he sent me the ipk file, but it refuse to start because it was linking to some zaurus icon i didnt have. |
12:23.27 | cr2 | ok. |
12:41.36 | goxboxlive | cr2: If you make the powermanagment driver will the battery last longer? When i am plying with it it draines out in q cuople of hours. Will the charging work? suspend/resume? |
12:43.46 | florian_kc | mickey|office: sounds nasty... |
12:44.28 | florian_kc | i feel lucky that it is unlocked now... |
12:45.02 | florian_kc | i should start to play with the bootloader. i don't want to boot into ppc all the time |
12:46.34 | cr2 | florian_kc: on universal ? |
12:47.39 | cr2 | goxboxlive: i need to dump more info about the battery. suspend/resume is too complex for now. |
12:47.54 | florian_kc | cr2: yes |
12:49.07 | cr2 | you can prepare the "boot card" which will put zimage into RAM and jump to it. expect more init code missing in the drivers, though. |
12:49.50 | florian_kc | yes right, but that's something i could try and find out what needs some care |
12:50.07 | cr2 | actually it is a shame we still don't have a m-sys doc g3 driver. it is used by 70% of all htc/ipaq. |
12:51.27 | cr2 | if your phone is not locked, then mybe you don't need to care about the crypto on load. |
12:51.44 | florian_kc | sounds scary |
12:53.20 | cr2 | when my uni was half-dead , i have loaded the binary blob into RAM and then the machine was hung. i didn't have a working zimage at that time. maybe now it will be more smooth. and we init 90% of the hardware ourselves now. |
12:53.58 | cr2 | but the blob was loaded and jumped to. it was the most important thing. |
12:55.01 | cr2 | with some effort you may port the hx4700 bootldr. if you are planning to do it, i can tell you more raw data. |
12:55.37 | cr2 | actually i have already tried, but gave up because i didn't have this raw data at that time ;) |
12:56.20 | cr2 | do you have an old or a new bootloader ? |
12:56.26 | cr2 | htc bootloader. |
13:15.27 | rob_w | so why do i get stuck on touchscreen even on .12 and not any picture even on .16 |
13:16.42 | cr2 | ? |
13:17.08 | cr2 | latest haret, hh.org CVS kernel ? |
13:17.45 | cr2 | hmm. CVS needs a (ro) patch. |
13:17.47 | rob_w | i used the blueangel_lcd.c which BabelQued gave me .. was from you he said .. it also looks good .. from what i can say |
13:18.06 | rob_w | but still i am stuck at the red line with my zImage i compile |
13:18.07 | cr2 | i have everything in CVS. |
13:18.27 | cr2 | try CVS +(ro) and make a diff. |
13:18.47 | rob_w | i noticed that the blueangel_* things are kinda outdated and current .12 cvs had improved versions .. |
13:18.58 | rob_w | thats for led`s , battery, kp |
13:19.27 | cr2 | they are the same ? |
13:19.43 | rob_w | no they wherent the same to my latest ..12 cvs |
13:19.54 | cr2 | but there is a .12 -> .16 adaptation necessary. |
13:20.19 | cr2 | hmm. strange. but you should boot first. the drivers come later. |
13:20.30 | rob_w | ok let me try this , so i check .16-hh4 , right and diff it to the one i use and should see some difference |
13:20.40 | cr2 | ok. |
13:20.50 | rob_w | cr2, when did you sync the .12 stuff to .16 ? |
13:21.35 | cr2 | long ago. |
13:22.19 | cr2 | battery, kp, and ts need a sane tsc2200 driver. |
13:22.38 | cr2 | leds can be corrected right now. |
13:22.51 | rob_w | i was able to have that in my .16 tree ,, compiles fine , and outside the hima dir |
13:23.09 | cr2 | compiles != works :) |
13:23.46 | rob_w | but it compiled like in .12 .. i didnt notice any input.h changes or so |
13:24.19 | cr2 | there are. if you will load the module, you will see. it was the same on universal. |
13:24.54 | cr2 | it was necessary to change the ts and keyboard driver. |
13:26.19 | rob_w | ok , lets to the cvs checkout |
13:26.53 | rob_w | i desperatly need a running gpe soon --- become more addicted again |
13:30.07 | rob_w | how can i monitor hh4 changes ? |
13:31.27 | cr2 | cvs up -dfP > /tmp/changes |
13:34.35 | cr2 | we need to find out how wince detects the RO cards on blueangel. |
13:36.45 | rob_w | dumpgpio when inserting on with ro and one without ?!? |
13:38.00 | cr2 | asicw3gpio |
13:38.28 | cr2 | i did it yesterday, but was not happy with the results. |
13:38.32 | cr2 | asic3gpio |
13:38.39 | cr2 | damn. |
13:38.46 | cr2 | wasic3gpio :) |
13:40.10 | rob_w | was it not clear ? |
13:41.47 | cr2 | not really, but you can try. it may be interesting to watch the SD status register too. (wreg/wreg16) |
13:44.03 | rob_w | i need to finish my cvs screw up .. i really need to be able to roll my own now .. otherwise is doomed to sit and listen only !! |
13:45.36 | rob_w | and ideally i want to use similar env`s like current oe uses .. for gcc at least , otherwise i will find myself again in a mess soon |
13:47.35 | cr2 | kernel is not really linked to oe. i never use oe for kernel compilation. |
13:54.16 | rob_w | do you want me to point me to your current gcc chain ? |
13:56.08 | cr2 | take the "standard" 3.4.1 at hh.org |
14:00.38 | rob_w | ok got it |
14:10.52 | rob_w | that takes ... |
14:16.35 | florian_kc | is there a way to change the cpu speed without plugging in ac power before boot? |
14:17.05 | cr2 | cpufreq module + some userspace tools ? |
14:17.38 | cr2 | afaik it is incomplete, because it does not change the voltage too. but i'm not sure. |
14:17.57 | cr2 | mickeyl should know better :) |
14:18.48 | *** join/#htc-linux skodde (n=skodde@unaffiliated/skodde) |
14:19.01 | florian_kc | ok |
14:38.43 | *** join/#htc-linux NickBee (n=Dheeraj@dialpool-210-214-81-147.maa.sify.net) |
14:54.35 | florian_kc | goxboxlive: not bad, seems to work basically... wireless si ok, bt fails. |
14:57.19 | cr2 | florian_kc: bt handling is a general oe problem. |
14:59.24 | florian_kc | cr2: yeah, that might be something we should consider to have in a hal service for mobile devices. |
15:00.55 | cr2 | the problem is that for usb 'modprobe something" creates a hciXXX, but for the serial devices you need a hciattach with some wild parameters. |
15:01.56 | rob_w | so there ,, i finished my zimage with cvs + ro changes + gcc 3.4.1 |
15:02.03 | rob_w | keep fingers crossed |
15:02.09 | cr2 | :) |
15:02.19 | florian_kc | cr2: right |
15:04.25 | *** join/#htc-linux Tomoe_ (i=touched@Sia.AnimeNfo.com) |
15:05.00 | cr2 | maybe htcuniversal_bt should upload bt firmware and create hciXXX instead of hciattach. does not look very nice though. |
15:06.01 | goxboxlive | florian_kc : Yes most of the stuff in GPE works. You se the only thing we have to do after building it is to add the pointrecal file. The BT seems wrong. Why is it hciattach -n ? If it is trying to use the BT script it should have used -S instead. But with -S there is no need to set the speed. |
15:06.46 | goxboxlive | So i think the best is to use hciattach /dev/ttyS1 texas 115200. And as i said yesterday, the BT scripts should be up to users if they want to use it or not. |
15:07.41 | florian_kc | goxboxlive: that would be a question for pb_ |
15:07.50 | florian_kc | ok |
15:09.42 | rob_w | heya |
15:09.48 | rob_w | i see something .. |
15:10.15 | rob_w | console font is a ugly 10 or 12 but yet i see something ,, maybe i should turn on boologo |
15:28.40 | goxboxlive | florian_kc I am trying to "bitbale |
15:29.02 | goxboxlive | bitbake blueprobe. But it fails. Do your faile? |
15:31.52 | rob_w | goxboxlive, give us errors |
15:32.22 | goxboxlive | hold on, i was doing something as root (stupi me) |
15:32.24 | goxboxlive | +d |
15:34.03 | goxboxlive | It seems like it misses the blueprobe.init file : http://www.openembedded.org/filebrowser/org.openembedded.dev/packages/blueprobe/blueprobe-0.17 |
15:35.02 | goxboxlive | take a look at the hx4700 patch. I made the htcuniversal.patch like the hx4700.patch but it claimes that hjx4700 patch cant fine blueprobe.init. In fact, it is missing in the directory as you can see |
15:41.24 | florian_kc | didn't i add this blueprobe stuff to upstream oe som time ago? |
15:42.20 | goxboxlive | yes maybe, havent looked in the folder before. |
15:42.23 | rob_w | hmm what option increases my lcd resolution on kernel level ? |
15:44.00 | cr2 | rob_w: ? |
15:44.39 | rob_w | i found on .12 we used mini4x6 font on fb console .. will try that now |
15:45.04 | rob_w | i found that my whole problem not seeing someting was probably a crash in tsc2200 which i compiled in .. |
15:47.05 | cr2 | hmm. yes, it ix 8x16 vga font now. |
15:47.47 | cr2 | <PROTECTED> |
15:48.50 | rob_w | ok now it looks better ... yet it flickers too much |
15:49.22 | rob_w | ah now is see " input_allocate_device() " messages |
15:55.18 | *** join/#htc-linux _bosjo (n=bosjo@130.80-203-23.nextgentel.com) |
15:56.40 | *** join/#htc-linux lpotter_ (n=ljp@c210-49-127-23.rochd1.qld.optusnet.com.au) |
15:58.37 | rob_w | ok ok , i getting where |
16:17.46 | goxboxlive | cr2: You asked me once about the /dev . Here it is: http://pastebin.ca/90342 |
16:19.47 | florian_kc | bbl |
16:21.15 | *** part/#htc-linux NickBee (n=Dheeraj@dialpool-210-214-81-147.maa.sify.net) |
16:32.07 | goxboxlive | florian_kc : I have fixed the bluetooth issue in gpe, there where something wrong in the sysconfig/bluetooth file |
16:35.32 | goxboxlive | hmm he is not here |
16:55.14 | rob_w | ok now ba_kp oppses |
16:55.19 | rob_w | oopses |
17:32.45 | rob_w | OK , i now have keyboard support for .16 on ba ! |
17:54.59 | rob_w | ok . now i am back in to buisness .. usb is also up .. |
17:56.09 | *** join/#htc-linux davidchow (n=davidcho@cm203-168-216-8.hkcable.com.hk) |
18:00.47 | rob_w | http://pastebin.ca/90448 is what i get when calling xtscal against /dev/input/event0 |
18:11.31 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
18:16.32 | BabelOued | hi |
18:16.42 | BabelOued | hi rob_w |
18:16.52 | BabelOued | i read log this afternoon |
18:16.59 | BabelOued | it works for you :) |
18:32.55 | *** join/#htc-linux BabelOued (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
18:45.03 | rob_w | yep |
18:45.16 | rob_w | brb |
18:46.40 | rob_w | i compiled a segfaulting tsc2200 in |
18:47.00 | BabelOued | hi rob_w |
18:48.38 | *** join/#htc-linux lamikr|laptop (n=chatzill@dyn04-247.dsl.spy.dnainternet.fi) |
18:49.20 | BabelOued | your problem was from gcc 3.4.4 ? |
18:56.06 | *** join/#htc-linux sylvain (n=sylvain@APuteaux-154-1-42-173.w83-199.abo.wanadoo.fr) |
19:03.44 | cr2 | rob_w: xtscal uses some silly code, since it gives unaligned traps. x86 is evil :) |
19:03.59 | cr2 | the crash comes from "Trying to free free IRQ39" |
19:04.15 | cr2 | this is obviously wrong. |
19:15.54 | *** join/#htc-linux florian (n=fuchs@p50838E4F.dip0.t-ipconnect.de) |
19:16.20 | florian | re |
19:16.47 | goxboxlive | hi florian |
19:17.11 | goxboxlive | florian : I have fixed the bluetooth issue in gpe, there where something wrong in the sysconfig/bluetooth file |
19:26.51 | cr2 | goxboxlive: can we unify the defconfig file with hx4700 to the biggest possible extent ? |
19:50.18 | rob_w | cr2, the blueangel_defconfig does miss quite some stuff , usb , ext3 , font etc |
19:50.39 | rob_w | cr2, do you have an idea why the screen flickers that much ? |
19:51.16 | florian | goxboxlive: ah cool, that's good news |
19:51.44 | goxboxlive | cr2: why do we want to do that? |
19:52.17 | goxboxlive | florian, yes it is in the /etc/sysconfig/bluetooth. Look at the BT speed line |
19:56.45 | goxboxlive | florian : This is in the bluetooth file: SPEED='[ -f "$BLUETOOTH_SCRIPT" ] && echo 921600 || echo 115200' It should be like this: SPEED=[ -f "$BLUETOOTH_SCRIPT" ] && echo 921600 || echo 115200 |
19:57.15 | goxboxlive | notice after = and after 115200 . Remove ' that one. |
19:57.51 | florian | ah too bed that we can't check that in now :-/ |
19:57.59 | florian | s/bed/bad |
19:58.09 | rob_w | but why IRQ39 ?? |
19:58.17 | goxboxlive | no porblem, no we know what's causing it, thats the best :-) |
20:00.12 | *** join/#htc-linux jgw (n=jake@c-67-171-225-35.hsd1.or.comcast.net) |
20:02.58 | rob_w | so which new input code is good to look at in .16 ? |
20:03.35 | cr2 | rob_w: universal :) |
20:04.52 | rob_w | good |
20:05.33 | rob_w | so there is no open and close anymore but : |
20:05.35 | rob_w | <PROTECTED> |
20:05.35 | rob_w | <PROTECTED> |
20:05.35 | rob_w | <PROTECTED> |
20:05.35 | rob_w | <PROTECTED> |
20:06.06 | cr2 | irq39 must be a gpio ? |
20:06.24 | cr2 | open() is not a standard driver function. |
20:07.07 | cr2 | the most important are probe (=init), suspend and resume. |
20:07.18 | rob_w | int ts_tsc2200_open( struct input_dev *dev ) |
20:07.44 | rob_w | needs to morph to someting probe like , right ? |
20:08.09 | cr2 | is it called once ? |
20:08.22 | cr2 | .probe is called only on driver insert. |
20:09.17 | rob_w | driver_register(&ts_driver); what is this actaully |
20:13.11 | rob_w | but unit ts and ba ts drift away on tsc2200 |
20:17.31 | cr2 | rob_w: if you want a clean driver then it is tsc2101. but it is quite complex. i think it is on top of the spi driver. which device uses it ? |
20:28.20 | rob_w | on ba we do only deal with the tsc2200 from ts |
20:28.24 | Funklord | someone here used LTO drives? |
20:28.40 | rob_w | Funklord, some |
20:29.06 | Funklord | rob_w: do you know if it's normal for the drive to stop at exactly 2:20min while reading? |
20:29.32 | Funklord | could it be some kind of tape mark? |
20:30.14 | Funklord | it stops and then continues |
20:30.42 | Funklord | for another 2:20min |
20:31.59 | rob_w | Funklord, feels like a setting or a package break on low mb && timeout running up |
20:32.27 | Funklord | because.. the other reason would be, that my system is getting buffer underruns while writing :( |
20:32.51 | Funklord | rob_w: package break? |
20:33.10 | Funklord | I'm using dd to copy raw data |
20:36.01 | Funklord | rob_w: I'm not very experienced with tape drives.. so if you can elaborate, I'd be greatful :) |
20:36.17 | Funklord | *grateful |
20:38.54 | jgw | huh just booted the latest image on my xv6700 |
20:38.54 | jgw | nice |
20:43.23 | Funklord | I'm so glad I didn't buy an lto3 drive, I can barely deliver 15Mb/s to this one (it needs 10Mb/s minimum) |
20:44.44 | Funklord | I mean, who has 600Gb of 15krpm scsi drives in a stripe setup ;P |
20:45.09 | jgw | me ;) |
20:45.15 | jgw | well >1TB |
20:45.26 | jgw | well >4TB ;) |
20:45.35 | Funklord | 1T of 15krpm drives? |
20:47.10 | Funklord | I'm doing all of this time consuming backup in order to save a 1.5T raid5 array =( |
20:47.31 | Funklord | controller went nuts |
20:49.46 | jgw | dell powervault 200s |
20:50.47 | Funklord | jgw: for a company? |
20:50.53 | Funklord | or private? |
20:51.26 | jgw | .edu |
20:53.31 | rob_w | go buy 2 stripe boxes then mirror over gigabit .. then if one box fails , start to run to get another |
20:54.52 | Funklord | heh.. think I'm gonna try 2 striped sata drives first |
21:05.25 | Funklord | rob_w: I didn't understand what you meant by package break earlier.. |
21:05.41 | Funklord | is that a kernel setting? |
21:05.59 | rob_w | well actually talking about network packages |
21:06.07 | rob_w | u go local , right , |
21:06.21 | rob_w | any complains from the scsi layer in some way ? |
21:06.25 | Funklord | yes, dd a file straight to the tape |
21:06.48 | Funklord | no messages whatsoever |
21:06.59 | Funklord | although I haven't switched on debugging |
21:07.15 | rob_w | so then maybe that is a cache limit on the tape .. as i gues dd will pump quick |
21:08.15 | rob_w | check what scsi speed the lto is attached |
21:08.42 | Funklord | it's using 80M/s (u160) |
21:08.48 | Funklord | Use very large values for blocks between tapemarks ? approximately 1 GB. The writing of a |
21:08.52 | Funklord | tapemark causes a start/stop on the tape, which significantly reduces the read/write speeds. |
21:08.53 | rob_w | if it is a adaptec card , go into the bios a set the nominal labeled speed of the lto |
21:08.57 | Funklord | that's a whitepaper I found |
21:09.40 | Funklord | a whitepaper about lto technology |
21:10.06 | rob_w | go set the scsi id on the scsi controller bios directly to 160 .. sometimes the drive say 320 and let the controller run into it , until it decreases |
21:10.22 | rob_w | the scsi id off the lto |
21:12.10 | Funklord | it is set to that, and that's what it detects automatically |
21:12.57 | Funklord | you mean domain verification, I think I switched that off |
21:13.55 | Funklord | now I found something on wikipedia: |
21:13.58 | Funklord | Data tracks are written in forward and reverse passes, also called wraps. It takes several wraps to completely fill a data band |
21:14.36 | Funklord | so.. if it needs to reverse, then it isn't so strange that it stops ;P |
21:22.59 | rob_w | no data coming from ts :-( |
21:31.06 | rob_w | Funklord, yet it should bring some mb .. wait until you use real life data , which is again splitted in files or such |
21:32.38 | Funklord | rob_w: what are you referring to? |
21:33.42 | rob_w | lto drives as such so make maybe 40 mb /s ?!? |
21:34.21 | Funklord | this one gives about 24M/s |
21:34.37 | Funklord | uncompressed |
21:34.47 | Funklord | with compression it'd be more ;P |
22:04.27 | rob_w | ok . i am done |
22:04.40 | rob_w | at least keyboard works normally now |
22:17.35 | *** join/#htc-linux g3gg0 (n=g3gg0@ppp-88-217-0-96.dynamic.mnet-online.de) |
22:30.56 | *** join/#htc-linux BabelOued_ (n=Fabrice@lun34-2-82-238-28-28.fbx.proxad.net) |
23:03.24 | *** part/#htc-linux BoSJo (n=bosjo@130.80-203-23.nextgentel.com) |
23:05.48 | *** join/#htc-linux fimbulvetr (n=danv@24.220.28.74) |
23:07.49 | *** join/#htc-linux sylvain (n=sylvain@APuteaux-154-1-42-173.w83-199.abo.wanadoo.fr) |
23:58.53 | lamikr|laptop | cr2: Are you still awake? |