| 00:00.37 | davidc__ | heh.. people don't just have flyback on ignore? |
| 00:03.55 | Mirell | Apparently not. |
| 00:36.27 | joe_bleau | Some of us don't know how to do that. |
| 00:40.48 | Mirell | joe_bleau: /ignore flyback- |
| 00:42.52 | davidc__ | or in my client, Preferences, ignores, "flyback" |
| 00:44.41 | cajaka | Its funny my Irc client has spell check but lacks ignore |
| 00:46.19 | davidc__ | what client is that? |
| 00:47.38 | cajaka | Dirc, I have had it for years. I dont need to ignore people much so its not a problem |
| 00:56.06 | *** join/#edev Crofton (n=balister@70.88.228.205) |
| 01:55.48 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-120.pdx1.dsl.speakeasy.net) |
| 01:58.15 | RobotGuy | Hi Mirell |
| 02:04.56 | ds2 | hmmm |
| 02:41.23 | *** join/#edev Dr{Wh0} (n=mathewss@71-222-24-22.ptld.qwest.net) |
| 04:23.30 | *** join/#edev redblue (i=star@ppp137.108-253-207.mtl.mt.videotron.ca) |
| 04:28.45 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 05:00.26 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 05:03.43 | Mirell | Ugh. |
| 05:03.48 | Mirell | Hate job. |
| 05:19.28 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-120.pdx1.dsl.speakeasy.net) |
| 05:19.41 | RobotGuy | ~seen prpplague |
| 05:19.43 | purl | prpplague <n=dave@mail.americanmicrosystems.com> was last seen on IRC in channel #edev, 5h 51m 19s ago, saying: 'later folks'. |
| 05:22.40 | *** join/#edev gandhijee (n=akp@124.125.79.212) |
| 05:25.51 | Mirell | 'lo RobotGuy |
| 05:25.57 | RobotGuy | Greets, Mirell! |
| 05:29.18 | *** join/#edev CP|home (n=nobody@www.crouse-house.com) |
| 05:30.23 | RobotGuy | I am starting to learn Squeak. |
| 05:36.51 | Mirell | Hmm... |
| 05:36.52 | Mirell | I'm tempted to get a crypt for my funeral plot. |
| 05:37.05 | Mirell | This is not meaning I plan to die until at least 60 years from now, |
| 05:37.11 | Mirell | But looking at life insurance currently. |
| 05:38.03 | RobotGuy | What is the lightest, smallest laptop type computer with the best display (at least 800x600 or 1024x768). Lowest cost is better. |
| 05:38.15 | Mirell | Hmm. |
| 05:38.21 | Mirell | The Asus Eee is pretty small, |
| 05:38.25 | Mirell | But unusable to me. |
| 05:38.29 | RobotGuy | Why? |
| 05:38.39 | MonMotha | The Lenovo X series is pretty light |
| 05:38.45 | MonMotha | XGA or SXGA+ on the tablet version |
| 05:39.01 | RobotGuy | Sounds more expensive too. |
| 05:39.01 | Mirell | RobotGuy: Tiny, tiny keyboard. |
| 05:39.18 | MonMotha | it's not too bad, but it's >$1k |
| 05:39.24 | Mirell | RobotGuy: I'm actually tempted by http://www.dell.com/content/products/productdetails.aspx/xpsnb_m1330?c=us&cs=19&l=en&s=dhs&~tab=bundlestab myself. |
| 05:39.30 | RobotGuy | Mirell, It has to have a decent keyboard and be able to run full Linux (not uCLinux). |
| 05:39.43 | Mirell | RobotGuy: Then definitely not the Asus Eee. |
| 05:40.44 | RobotGuy | Too expenive for me. |
| 05:41.06 | RobotGuy | I was looking at an XO, but I hate the stupid kiddie keyboard. |
| 05:41.06 | Mirell | You have an issue with all three requirements then. |
| 05:41.12 | Mirell | Small, Low Cost, Good Keyboard. |
| 05:41.17 | Mirell | Pick 2. |
| 05:41.28 | RobotGuy | And LINUX. :) My choice of distro. :) |
| 05:41.41 | Mirell | That is not an issue. |
| 05:42.03 | RobotGuy | Let's try Low cost, good keyboard. |
| 05:42.17 | Mirell | Oh, well, that's easy. |
| 05:42.41 | Mirell | Dell has some good offerings, heck, even Wal-Mart. |
| 05:42.43 | RobotGuy | I have to be able to pack it with me on my wheel chair, perhaps in a backpack. |
| 05:43.36 | Mirell | Ahh |
| 05:44.12 | RobotGuy | I would expect to get a good quality laptop bag for it. |
| 05:45.49 | RobotGuy | I missed an excellent opportunity to really show hammer off last night because I had nothing to connect it to. |
| 05:46.19 | Mirell | Yah...hmm... |
| 05:46.45 | Mirell | You can *use* an Asus Eee and the ilk, |
| 05:46.53 | Mirell | But you won't be able to touch type at all. |
| 05:47.29 | MonMotha | I'm looking at a Lenovo T61p |
| 05:47.32 | RobotGuy | Well, I could get an inexpensive external keyboard. |
| 05:47.38 | MonMotha | only thing I can find with WUXGA/UXGA |
| 05:47.55 | Mirell | Then if you can do that, there you go. |
| 05:49.13 | Mirell | I go to sleep show I can show up for work at 7:00AM out of spite. |
| 05:49.22 | RobotGuy | G'Night Mirell |
| 05:49.56 | Mirell | 'night. But yah, Asus Eee is $400, and it's a decent little Linux box. |
| 05:50.04 | Mirell | Just the keyboard is tiny, so no fast-typing at 'tall. |
| 05:50.19 | RobotGuy | I will look at that one. Thanks |
| 06:31.13 | gandhijee | hey, anyone here know of any hotswap IDE cards? |
| 06:58.37 | ds2 | there is the CloudBook coming soon |
| 06:58.40 | ds2 | intel free |
| 07:47.53 | *** join/#edev jbevren (n=jbevren@dumt-04-044.dsl.netins.net) |
| 08:18.21 | *** join/#edev landley (n=landley@cpe-72-177-114-214.austin.res.rr.com) |
| 08:47.39 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 08:57.20 | *** join/#edev huaynac (n=huaynac@61.6.65.198) |
| 09:56.49 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-120.pdx1.dsl.speakeasy.net) |
| 10:00.28 | *** join/#edev lyakh (n=lyakh@p57BD2A2D.dip0.t-ipconnect.de) |
| 10:12.47 | RobotGuy | Does anyone know if the rootfs is actually stored on Hammer in compressed or uncompressed form? |
| 10:13.28 | RobotGuy | Hi landley |
| 10:13.55 | RobotGuy | Greets, Mirell |
| 10:13.58 | landley | Hi robotguy. |
| 10:14.12 | landley | rootfs is usually stored compressed. |
| 10:14.25 | landley | it's gzipped as part of the build process. |
| 10:14.27 | landley | rootfs == initramfs. |
| 10:14.35 | RobotGuy | WHEW, I was hoping that is the case. I already have a 6.6 Meg compressed rootfs. |
| 10:14.37 | landley | See the "usr" subdirectory of the linux source code. |
| 10:15.08 | landley | Because I wanted to read it, and there wasn't any, so I had to go fish... |
| 10:15.09 | RobotGuy | Cool, landley |
| 10:16.00 | landley | I also wrote http://www.timesys.com/timesource/initramfs.htm and the one before it. |
| 10:16.11 | landley | Which is probably far more info than you wanted. :) |
| 10:17.10 | RobotGuy | Something is not right, because changing the .size value does not seem to gain me space for my rootfs. |
| 10:17.57 | *** join/#edev redblue (i=star@ppp021.108-253-207.mtl.mt.videotron.ca) |
| 10:19.13 | RobotGuy | It's in the linux-2.6.22/arch/arm/mach-s3c2410/mach-tct_hammer.c file. |
| 10:21.02 | RobotGuy | If rootfs gets unpacked into RAM, I am in big trouble. |
| 10:26.18 | RobotGuy | Well, my largest rootfs image is 6.6 Meg compressed and it still needs at least one more package added. |
| 10:42.23 | landley | rootfs is an instance of ramfs. |
| 10:42.34 | landley | It scales dynamically as you write or delete files. |
| 10:43.02 | landley | It's stored in the kernel image as a gzipped cpio archive. Basically a simpler tarball. |
| 10:43.14 | landley | It's extracted into a ramfs instance early in the boot process. |
| 10:43.16 | RobotGuy | That doesn't really seem to answer my question, but I may be missing something. |
| 10:43.31 | RobotGuy | Oh, so it IS extracted. Then I am in trouble. |
| 10:43.32 | landley | "if rootfs gets unpacked into ram..." |
| 10:43.34 | landley | yes, it does. |
| 10:43.45 | landley | I can point you at the code that does it, even. |
| 10:43.50 | landley | You might want to try squashfs instead. |
| 10:44.44 | landley | Have a simple ramfs init script mount a squashfs image out of your flash and switch_root into it. |
| 10:44.51 | landley | Downside of that is it's read only, but it stays compressed. |
| 10:45.52 | RobotGuy | Ack, my fourth rootfs image would be 36 Meg unpacked. :( :( Not going to fit.. |
| 10:47.18 | RobotGuy | I would not have a clue how to do that. |
| 10:47.41 | RobotGuy | If it's unpacked into RAM then it won't fit anyway. |
| 10:51.27 | RobotGuy | I know I probably don't understand this very well. |
| 10:54.22 | RobotGuy | Do I have to set .size according to the unpacked rootfs size or compressed size? |
| 10:55.02 | landley | hmmm? |
| 10:55.06 | RobotGuy | Do I have to set .size according to the unpacked rootfs size or compressed size? |
| 10:55.21 | landley | Where do you set .size? |
| 10:55.34 | RobotGuy | It's in the linux-2.6.22/arch/arm/mach-s3c2410/mach-tct_hammer.c file. |
| 10:55.35 | landley | The build calculates the size... |
| 10:55.52 | landley | Huh, hadn't seen that bit. |
| 10:56.00 | RobotGuy | prpplague said I need to change that and make it larger. |
| 10:56.16 | landley | I'll take your word for it. I have no idea what that bit is doing. |
| 10:56.33 | landley | It's probably more a question of loading the kernel from flash into memory in the first place... |
| 10:56.58 | landley | There's the bootloader code, then the decompression code, then you call start_kernel() or some such... |
| 10:57.05 | RobotGuy | Well, something is wrong because I still get messages I have over run the end of device when the kernel boots. |
| 10:57.07 | landley | I dunno where this platform's .size fits in. |
| 10:57.27 | landley | End of what device? |
| 10:57.31 | landley | I dunno what it's doing. |
| 10:59.29 | RobotGuy | I will be able to get the message in a few - am reflashing Hammer now. |
| 11:06.28 | RobotGuy | landley, http://www.pastebin.ca/849317 |
| 11:09.35 | landley | Sounds like something is specifying where the end of the flash chip is, and getting it wrong.. |
| 11:09.54 | landley | Possibly only part of the flash is mapped? |
| 11:09.58 | landley | Dunno, this is a hardware thing. |
| 11:10.02 | landley | Ask prpplague when he wakes up. |
| 11:12.47 | RobotGuy | I am using the stock openocd hammer.cfg file to flash the Hammer. |
| 11:12.58 | RobotGuy | It's actually a software thing too. |
| 11:25.21 | *** join/#edev Crofton (n=balister@70.88.228.205) |
| 12:18.20 | *** join/#edev gandhijee (n=akp@124.125.79.212) |
| 12:22.24 | *** join/#edev dijenerate (n=dijenera@72.51.111.182) |
| 12:49.45 | *** join/#edev dijenerate (n=dijenera@66.54.118.16) |
| 12:58.00 | *** join/#edev gandhijee (n=akp@124.125.79.212) |
| 13:08.01 | *** join/#edev lyakh (n=lyakh@p57BD2A2D.dip0.t-ipconnect.de) |
| 13:38.46 | *** join/#edev sjhill (n=sjhill@real.realitydiluted.com) |
| 13:53.24 | RoboAway | +/clear |
| 14:00.08 | *** part/#edev RobotGuy (n=RobotGuy@dsl093-038-120.pdx1.dsl.speakeasy.net) |
| 14:01.02 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-120.pdx1.dsl.speakeasy.net) |
| 14:07.57 | *** join/#edev Redhatter-LM2E (n=stuartl@gentoo/developer/redhatter) |
| 14:10.59 | *** join/#edev T0mW (n=Tom@24.238.86.158.res-cmts.sth.ptd.net) |
| 14:17.06 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 14:18.40 | *** mode/#edev [+o prpplague] by ChanServ |
| 14:26.57 | RobotGuy | G'Morning prpplague |
| 14:27.08 | SpeedEvil | Morning. |
| 14:27.13 | prpplague | remove the G and i'll respond, hehe |
| 14:27.58 | RobotGuy | I have a question so I can understand how rootfs works better. |
| 14:28.38 | prpplague | ok, i have a ton of work today, so i won't be able to conversate(is that a word?) much today |
| 14:28.43 | RobotGuy | Does rootfs get unpacked into the initial ramdisk? |
| 14:28.49 | SpeedEvil | I can answer general questions |
| 14:29.03 | SpeedEvil | but not related to hammer, as I haven't looked at the software side. |
| 14:29.05 | RobotGuy | And what does .size control exactly? |
| 14:29.49 | prpplague | RobotGuy: it controls the size of the system area in flash that the kernel is aware of |
| 14:30.08 | RobotGuy | prpplague, The size of the compressed rootfs, right? |
| 14:30.17 | prpplague | RobotGuy: no |
| 14:30.41 | RobotGuy | Hmmmm. |
| 14:30.55 | prpplague | RobotGuy: picture it as a hard drive |
| 14:31.01 | *** join/#edev hapchap (n=hapchap@mail.americanmicrosystems.com) |
| 14:31.14 | RobotGuy | OK |
| 14:31.20 | prpplague | RobotGuy: based on the definitions in the machince file you have two partitions |
| 14:31.43 | prpplague | RobotGuy: the kernel doesn't care whats in the parititons or what file systems are on the partition |
| 14:31.58 | prpplague | RobotGuy: it just needs to know where they begin and how big they are |
| 14:32.25 | RobotGuy | OK, I have a 6.6 MB compressed rootfs, and it needs at least one more thing. |
| 14:32.30 | prpplague | hapchap: morning crack baby |
| 14:33.12 | hapchap | prpplague: :-/ |
| 14:33.59 | RobotGuy | I have .size set to 8 MB. |
| 14:34.11 | prpplague | RobotGuy: ok, so you need 256k for the bootloader , plus 1024k for the kernel, and 7168k for the rootfs image |
| 14:34.27 | RobotGuy | OK |
| 14:34.56 | RobotGuy | I also have a somewhat modular kernel. |
| 14:35.37 | RobotGuy | I2C, SPI, USB, etc. |
| 14:37.42 | RobotGuy | No idea how to install the modules. |
| 14:38.32 | prpplague | RobotGuy: you can just copy them to the rootfs image |
| 14:38.46 | RobotGuy | I am still getting messages telling me I have run over the end of the device at bootup. |
| 14:39.24 | RobotGuy | To lib/modules/<kernel version> ? |
| 14:39.26 | prpplague | RobotGuy: yep, there is one other item you have to change |
| 14:39.37 | prpplague | RobotGuy: yea, just copy the modules there |
| 14:39.59 | prpplague | RobotGuy: reset the carrier board and when you see the bootloader start, do a ctrl-c |
| 14:40.47 | RobotGuy | Did that |
| 14:41.05 | prpplague | RobotGuy: do a printenv command |
| 14:41.25 | prpplague | RobotGuy: you should see the ramdisksrc variable |
| 14:41.27 | RobotGuy | OK did that |
| 14:41.37 | prpplague | RobotGuy: see the +1024k on the end? |
| 14:41.49 | RobotGuy | ramdiskaddr *= 0x30800000 |
| 14:41.55 | prpplague | RobotGuy: thats the size to copy from flash to ram |
| 14:42.07 | prpplague | RobotGuy: no, ramdisksrc |
| 14:42.30 | RobotGuy | ramdisksrc *= nor:1280k+1024k |
| 14:42.33 | prpplague | right |
| 14:42.36 | prpplague | RobotGuy: see the +1024k on the end? |
| 14:42.39 | RobotGuy | Yes |
| 14:42.41 | prpplague | RobotGuy: thats the size to copy from flash to ram |
| 14:42.53 | prpplague | RobotGuy: you need to change that to the size of your compressed image |
| 14:42.53 | RobotGuy | Kernel? |
| 14:43.12 | RobotGuy | Oh, so that has to be close to 6.6 MB. |
| 14:43.39 | prpplague | RobotGuy: you set that by doing: setenv ramdisksrc "nor:1280k+7168k" |
| 14:44.03 | RobotGuy | OK, I need to figure my rootfs compressed size and set it there. |
| 14:44.06 | prpplague | RobotGuy: then do a printenv to make sure it has recorded the change |
| 14:44.20 | prpplague | RobotGuy: you told me it was 6.6mb |
| 14:44.27 | RobotGuy | It is |
| 14:44.44 | prpplague | nothing to figure out then |
| 14:45.23 | prpplague | RobotGuy: are you saying the image is 6.6mb or the rootfs when mounted is 6.6mb ? |
| 14:45.43 | prpplague | RobotGuy: i.e. what is the actual size of rootfs.ext2.arm.gz ? |
| 14:45.49 | RobotGuy | The compressed rootfs is 6.6 MB |
| 14:46.00 | prpplague | RobotGuy: i.e. what is the actual size of rootfs.ext2.arm.gz ? |
| 14:46.26 | prpplague | RobotGuy: is that the actual size of rootfs.ext2.arm.gz? |
| 14:46.27 | RobotGuy | -rw-r--r-- 1 linuxguy staff 5.8M 2008-01-10 02:54 rootfs.arm.ext2.gz |
| 14:46.42 | RobotGuy | It shrunk a little |
| 14:46.59 | prpplague | RobotGuy: ahh ok, i'd just use 7168k, give you some room |
| 14:48.01 | RobotGuy | So setenv ramdisksrc "nor:1280k+7168k" ?? |
| 14:49.02 | RobotGuy | The image will grow a bit. I still need to eventually get Squeak in there. |
| 14:50.22 | RobotGuy | prpplague: Is that right? |
| 14:50.30 | prpplague | RobotGuy: yep |
| 14:50.42 | RobotGuy | OK! Do I just exit apex now? |
| 14:50.50 | prpplague | RobotGuy: reboot |
| 14:52.09 | RobotGuy | OK, here goes reflashing with the large rootfs now. |
| 14:52.23 | RobotGuy | I bet I have to set the value each time I reflash.. |
| 14:52.33 | prpplague | RobotGuy: yep |
| 14:52.46 | prpplague | RobotGuy: unless you recompile apex with the change |
| 14:52.49 | RobotGuy | I thought so. :D Now I know how to do it. :D Thanks much! |
| 14:53.16 | prpplague | RobotGuy: you don't have to use the jtag app every time |
| 14:53.26 | prpplague | RobotGuy: you can just use apex to load just the rootfs |
| 14:53.49 | RobotGuy | Hmmmm. We have docs on APEX, right? IF so I can read those. |
| 14:56.28 | prpplague | http://www.elinux.org/Hammer_How_to_Reflash_Apex_and_kernel |
| 14:56.44 | RobotGuy | Right |
| 14:57.04 | prpplague | RobotGuy: just do the rootfs section and substitue for the 1024k your 7168k |
| 14:57.57 | RobotGuy | Gotcha |
| 15:00.02 | RobotGuy | I just did cp -rpd /src/hammer/linux-2.6.22/{drivers,fs,ipc,kernel,mm,net,sound} . (where , is root/lib/modules/2.6.22) |
| 15:01.38 | RobotGuy | I'll figger it out. |
| 15:02.24 | prpplague | RobotGuy: thats another issue, i'll have to go over how to mount a loop image |
| 15:03.32 | RobotGuy | prpplague: OK, when ya have some time. I want to learn all this real good. |
| 15:06.43 | RobotGuy | OH CRAP |
| 15:07.47 | RobotGuy | 53 MB compressed rootfs. I think I copied something I do not need. |
| 15:14.43 | *** join/#edev CP|home_ (n=nobody@www.crouse-house.com) |
| 15:15.04 | prpplague | RobotGuy: uh yea |
| 15:18.01 | RobotGuy | I may have found another possible problem. I can't do make modules_install as non-root. I think that has to be done to actually get modules created. |
| 15:22.22 | *** join/#edev g1powermac (n=g1powerm@unaffiliated/g1powermac) |
| 15:27.56 | RobotGuy | I got it worked around. |
| 15:32.37 | prpplague | RobotGuy: make modules is all thats needed to build the modules |
| 15:41.23 | RobotGuy | Hmmmmmmmm |
| 15:43.31 | *** join/#edev emil|k (n=emil@f054012170.adsl.alicedsl.de) |
| 15:46.41 | RobotGuy | Damn, kernel panic. Oh well, will work on that with you later, prpplague. |
| 15:47.28 | RobotGuy | Well, whenever. :) |
| 15:48.30 | RobotGuy | I have to take off for a bit, so will be back later. |
| 15:48.47 | *** part/#edev RobotGuy (n=RobotGuy@dsl093-038-120.pdx1.dsl.speakeasy.net) |
| 15:58.33 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 15:58.33 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 16:14.03 | Mirell | Okay, my SQL evidently sucks. |
| 16:24.12 | *** join/#edev ico2 (n=ico2@modem-3740.cheetah.dialup.pol.co.uk) |
| 17:07.36 | chouimat | hi |
| 17:08.43 | *** join/#edev ico2 (n=ico2@modem-3740.cheetah.dialup.pol.co.uk) |
| 17:10.18 | *** join/#edev antixogh (n=antixogh@cpe-67-11-218-157.satx.res.rr.com) |
| 17:10.55 | antixogh | have any of you heard or worked on putting linux on a sansa e200? |
| 17:25.00 | *** part/#edev antixogh (n=antixogh@cpe-67-11-218-157.satx.res.rr.com) |
| 17:28.24 | Mirell | Hmm.. |
| 17:36.11 | prpplague | oh noooooo, white wire fix |
| 17:40.44 | CosmicPenguin | white? |
| 17:40.46 | CosmicPenguin | what happened to blue? |
| 17:59.04 | prpplague | CosmicPenguin: we use either white or black, hehe |
| 17:59.20 | CosmicPenguin | way to stick with tradition |
| 18:11.08 | *** join/#edev dijenerate (n=dijenera@72.51.111.182) |
| 18:24.34 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 18:27.38 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 18:55.01 | *** join/#edev landley (n=landley@adsl-76-197-33-46.dsl.austtx.sbcglobal.net) |
| 18:56.04 | prpplague | landley: greetings |
| 19:00.22 | landley | How goes stuffness and thingies? |
| 19:05.37 | prpplague | landley: it goes |
| 19:05.59 | landley | ah. |
| 19:06.20 | landley | I'm told it's good for headphones and keyboards and such. |
| 19:06.31 | landley | Cheaper and lower power than 802.11 for that, anyway. |
| 19:10.29 | *** join/#edev ico2 (n=ico2@modem-3740.cheetah.dialup.pol.co.uk) |
| 19:12.37 | CosmicPenguin | I am so addicted to my bluetooth headphones |
| 19:12.47 | CosmicPenguin | I run them through a USB dongle and a2dpd |
| 19:12.53 | CosmicPenguin | works like a million bucks |
| 19:14.38 | prpplague | landley: yea hapchap and i are discussing maybe hacking up a new embedded hcid |
| 19:15.00 | prpplague | landley: as the current one is totally unusable in embedded systems |
| 19:15.08 | prpplague | landley: it requires massive amount of dbus stuff |
| 19:17.10 | landley | hcid? |
| 19:17.18 | landley | human computer interface... device? |
| 19:17.26 | SpeedEvil | yes |
| 19:22.52 | hapchap | landley: no, human computer interface daemon |
| 19:23.33 | hapchap | landley: currently there is no decent way to handle pins and link keys for bluetooth connections with any of the newer bluez-utils |
| 19:24.16 | landley | Ah. |
| 19:24.39 | hapchap | hcid > 2.25 depends on dbus very heavily, dbus depends on libxml2 for ALL messages, and libxml2 compiled with uclibc is still > 3MiB |
| 19:27.39 | *** join/#edev samuelr (n=samuelr@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 19:29.34 | landley | hapchap: is there a spec for this thing? |
| 19:29.40 | landley | I still dunno what hcid actually _does_. |
| 19:36.43 | hapchap | landley: hcid handles all items for bluetooth that should not be handled in the kernel (i.e. caching and handling pin requests and replys) |
| 19:37.32 | landley | ok |
| 19:38.10 | hapchap | landley: the man page for it is kind of lax on specifics, but to do _any_ bluetooth connections that are not the _most_ basic, hcid is needed |
| 19:38.48 | hapchap | i.e. i can set up a simple outgoing non auth'ed not encrypted rfcomm connection without it, but thats about it |
| 19:39.09 | hapchap | there is a movement to rename it bluetoothd to clarify what it really is |
| 19:40.07 | hapchap | and i misspoke when i agreed with SpeedEvil, it is "Host Controller Interface" daemon |
| 19:40.46 | SpeedEvil | Sorry - i've been reading other stfuf, which confused me |
| 19:41.00 | hapchap | SpeedEvil: ;-) |
| 19:56.12 | Mirell | Yay, I completed my first real task. |
| 20:00.39 | sjhill | CosmicPenguin: got any contacts at RIM? |
| 20:02.02 | CosmicPenguin | at RIM? |
| 20:02.07 | CosmicPenguin | not a chance |
| 20:02.21 | CosmicPenguin | You angry at your blackberry? |
| 20:02.33 | sjhill | everybody's a comedian |
| 20:03.17 | CosmicPenguin | I don't get the joke? |
| 20:03.17 | sjhill | CosmicPenguin: RMI, sorry |
| 20:03.21 | CosmicPenguin | ohhhh... |
| 20:03.29 | sjhill | see, we're both confused...my bad |
| 20:03.29 | CosmicPenguin | no, I don't |
| 20:03.31 | CosmicPenguin | and it makes me angry |
| 20:03.35 | sjhill | ?! |
| 20:03.48 | sjhill | hows come? |
| 20:04.04 | CosmicPenguin | One day, everything just became a black hole, and I haven't heard anything since |
| 20:04.10 | CosmicPenguin | no e-mail messages, no commits, nothing |
| 20:04.29 | sjhill | wow |
| 20:04.36 | CosmicPenguin | and it makes me angry Linux |
| 20:05.01 | CosmicPenguin | erm - what? |
| 20:05.05 | sjhill | you didn't have any buddies that went there? |
| 20:05.41 | CosmicPenguin | There was only one person I know that went to Raza, and she was a marketing type, I think |
| 20:05.45 | CosmicPenguin | she's not there anymore |
| 20:05.52 | sjhill | sheesh |
| 20:06.28 | CosmicPenguin | I wish i had somebody I could at least push questions to |
| 20:06.32 | CosmicPenguin | I get at least one or two a quarter |
| 20:10.40 | SpeedEvil | I'm not answering two questions for only a quarter. |
| 20:11.11 | sjhill | hmmm |
| 20:11.22 | sjhill | so Open On-Chip Debugger is only for ARM |
| 20:11.24 | sjhill | that sucks |
| 20:11.43 | SpeedEvil | it's dependant on on-die JTAG features |
| 20:11.48 | *** join/#edev Redhatter-Out (n=stuartl@gentoo/developer/redhatter) |
| 20:11.57 | SpeedEvil | that are not on any(?) other processors. |
| 20:12.07 | SpeedEvil | As I understand it at least. |
| 20:20.28 | ds2 | prpplague: an advance thank you for even considering to fix the bluetooth stuff |
| 20:20.52 | ds2 | dbus == EVIL |
| 20:28.01 | landley | dbus is indeed evil. |
| 20:28.24 | landley | prpplague: if you can get me some specs I might try to add a blued to toybox. |
| 20:28.31 | landley | Right now I have no idea what it's supposed to _do_. |
| 20:28.33 | hapchap | ds2: i dont think dbus in and of itself is actiall evil |
| 20:28.44 | hapchap | s/actiall/actually |
| 20:29.05 | landley | It's not dbus itself that's evil, it's what it encourages. |
| 20:29.19 | landley | It's its own category of "dll hell", only more dynamic. |
| 20:29.22 | hapchap | ds2: just how much it has been intertwined with things is the issue |
| 20:29.39 | prpplague | landley: yea, thats kinda what i was thinking |
| 20:29.43 | landley | It's a fairly elegant solution to doing something that may turn out to be deeply inadvisable in the long run. |
| 20:29.49 | prpplague | landley: adding blued to toybox |
| 20:30.05 | landley | Hey, if you want to tackle it... :) |
| 20:30.19 | landley | I'm in the process of rearchitecting toybox a bit to simplify adding commands. |
| 20:30.37 | landley | I think I can make it so you just drop a command's .c file in the toys directory, with some specially formatted comments at the top. |
| 20:30.43 | landley | And the other two files get automatically generated from the *.c files. |
| 20:43.58 | *** join/#edev Heet (n=lucas@a82-92-5-118.adsl.xs4all.nl) |
| 21:01.33 | CosmicPenguin | Does anybody know if the fbdev team keeps a public git tree? |
| 21:33.53 | *** join/#edev ico2_ (n=ico2@modem-1956.cougar.dialup.pol.co.uk) |
| 21:34.41 | prpplague | landley: neat, that brings up the question i tried to ask the other day, what are you going to do when there are 1000 apps in the the toys directory? |
| 21:36.15 | landley | Do you expect there to be 1000 interesting apps? |
| 21:36.40 | landley | ls -l /bin | wc on Kubuntu gives 107. |
| 21:36.49 | CosmicPenguin | I was just going to say that |
| 21:37.13 | landley | I admit "ls -l /{,usr}/{bin,sbin} | wc" gives 2140... |
| 21:37.33 | ds2 | there isn't really a good reason for DBUS to exist and the price in size/memory/performance is just too much |
| 21:37.35 | landley | But that includes all the stubs Ubuntu installs for "this command isn't installed yet, run aptitude install toadwrangler" |
| 21:38.02 | landley | I _can_ split up the toys directory into subdirectories or some such, I just await a decent reason to do so. |
| 21:38.30 | landley | On ubuntu, /usr/bin has 1602 entries in it, and they didn't break it into subdirectories. |
| 21:38.39 | landley | In fact, /bin and /usr/bin were only split due to a historical accident. |
| 21:38.54 | hapchap | landley: oh due tell |
| 21:39.00 | hapchap | s/due/so |
| 21:39.00 | landley | In 1971, Ken and Dennis filled up their first RK05 disk pack and had to leak the OS into the second disk, which up until then had contained the user home directories. |
| 21:39.19 | landley | When they got a third disk, they moved the user directories to /home and kept /usr for the os. |
| 21:40.00 | landley | Unix old hands will spout some mumbo jumbo about "/bin is needed for early boot, /usr is commands needed after you've mounted all partitions". |
| 21:40.20 | ds2 | isn't that something intro'ed with BSD or SunOS? |
| 21:40.38 | landley | And the response is "ever stopped and thought about what initrd means? Or initramfs? You want to mount a /usr partition that doesn't match your shared libraries in /lib? What are you, nuts?" |
| 21:40.57 | landley | ds2: The retroactive justification for the split, yeah. |
| 21:41.25 | landley | The split dates back to before Unix ever made it out of Bell Labs, and has to do with them simply running out of space on 2.5 megabyte disk packs attached to a PDP-11. |
| 21:41.49 | landley | Easier all around... |
| 21:42.03 | ds2 | I don't mean the split.. I mean the convention that /bin is needed in early boot where /usr may not be available |
| 21:42.16 | landley | "early boot" is initramfs. |
| 21:42.25 | ds2 | that's if you have an initramfs |
| 21:42.26 | prpplague | landley: yea, i'm not expecting a super large number, i was just curious if there was a long term plan to break it down into sections similiar to the way busybox has |
| 21:42.33 | landley | If /usr isn't available, then you need to either statically link /bin or you need your shared libraries in /lib rather than /usr/lib. |
| 21:42.52 | landley | And then /usr has to match those shared libraries. Version skew on network mounts makes system unhappy. |
| 21:42.52 | ds2 | right... /bin and /sbin, IMO, should be statically linked |
| 21:43.16 | ds2 | <PROTECTED> |
| 21:43.19 | landley | <PROTECTED> |
| 21:43.28 | prpplague | interesting - http://www.linuxdevices.com/news/NS8243966800.html |
| 21:43.48 | landley | ld-linux.so.* is in /lib. |
| 21:44.05 | ds2 | prpplague: and it is Intel free too |
| 21:44.06 | landley | libc is in /lib. |
| 21:44.11 | landley | Unimportant little things like that. |
| 21:44.13 | ds2 | I like it better then the EeePC |
| 21:44.14 | prpplague | landley: uh wait, i think we got talking about the wrong directories |
| 21:44.27 | prpplague | landley: i'm think more for the organization of the souce code |
| 21:44.31 | ds2 | landley: that is irrelevant if /bin, /sbin are static |
| 21:44.34 | landley | prpplague: anyway, my approach to the split is the same approach I'm taking to most things with toybox. |
| 21:44.37 | landley | Wait for something to need it. |
| 21:44.43 | landley | Infrastructure in search of a user is evil. |
| 21:44.43 | prpplague | landley: okie dokie |
| 21:44.53 | landley | This is partly an exercise to see what I can get away with _not_ doing. |
| 21:44.59 | ds2 | but then in this picture, glibc is just plain in compatible |
| 21:45.50 | landley | via c7 is last of the <strike>mohicans</strike> 32-bit x86 processors. |
| 21:46.05 | landley | Even via is finally going 64 bit later this year. |
| 21:46.08 | CosmicPenguin | now I have to smack you |
| 21:46.32 | landley | My main gripe about ultraportables like that is that the keyboards tend to suck. I'll have to see with this one. |
| 21:46.45 | landley | ds2: who said glibc? I said libc. |
| 21:46.47 | landley | uClibc is a libc. |
| 21:46.49 | landley | klibc is a libc. |
| 21:46.56 | landley | newlib is a libc. |
| 21:47.35 | ds2 | landley: Just saying, glibc is incompat. I much prefer the older libc5 if I am stuck doing porting work |
| 21:47.49 | landley | My point is that the / vs /usr split/duplication is a historical artifact that people got cognitatve dissonanced into justifying, and they now cling to it despite it having clearly outlived its purpose. |
| 21:47.50 | ds2 | uClibc is a nice idea but I think it is loosing its way |
| 21:47.55 | landley | Early boot is done through initramfs. |
| 21:48.02 | landley | ds2: it's lost its way. |
| 21:48.09 | landley | I may do a fork if I get really bored. |
| 21:48.21 | ds2 | revive libc5! |
| 21:48.25 | landley | ds2: no. |
| 21:48.30 | landley | libc5 isn't interesting. |
| 21:48.45 | ds2 | I want my static binaries to be really static and that is just impossible with current glibc :/ |
| 21:48.55 | landley | I know, that's why I use uClibc. |
| 21:49.13 | landley | The version in firmware linux is 0.9.29 with a few "blow to the head" patches to make it behave. |
| 21:49.30 | ds2 | the thing I don't like with uClibc is I just can't change the symlink and have everything use it |
| 21:49.46 | landley | Peter mazinger has done a lot of work on the gentoo embedded uClibc version, they've got a bucket of patches they apply. |
| 21:50.03 | landley | ds2: you can't always do that between different versions of glibc anymore. |
| 21:50.33 | ds2 | *nod* hence why I like libc5 |
| 21:50.53 | landley | ds2: Ok, back up. I'm confused. |
| 21:51.01 | landley | You can't just change the symlink between different versions of glibc. |
| 21:51.08 | landley | How do you just change the symlink to go from glibc to libc5? |
| 21:51.22 | ds2 | with libc5, versions were compatible |
| 21:51.34 | ds2 | with libc6 (aka glibc), we have this big mess that isn't compatible |
| 21:51.37 | ds2 | I want order! |
| 21:51.38 | landley | Yeah, and the reason they're still compatible is they stopped doing new development years ago. |
| 21:51.47 | landley | glibc sucking doesn't mean libc5 isn't dead. |
| 21:52.03 | landley | Where's the blackfin version of libc5? |
| 21:52.04 | ds2 | prehaps |
| 21:52.10 | landley | Where's the arm eabi version? |
| 21:52.11 | ds2 | work stuff, bbl |
| 21:52.22 | landley | Have fun... |
| 21:53.25 | landley | prpplague: the everex thing would be more interesting if it was $200 instead of $400. |
| 21:53.35 | prpplague | landley: indeed |
| 21:53.40 | prpplague | landley: and arm based, hehe |
| 21:54.44 | landley | Apropos of nothing, I just find it funny. |
| 22:00.49 | prpplague | hapchap: is simple_send in cvs? |
| 22:11.41 | *** join/#edev redhomey (n=neilhoma@89.100.83.14) |
| 22:19.28 | *** part/#edev redhomey (n=neilhoma@89.100.83.14) |
| 22:35.03 | Mirell | Ugh. |
| 22:56.56 | *** join/#edev RobotGuy (n=RobotGuy@dsl093-038-120.pdx1.dsl.speakeasy.net) |
| 23:04.03 | prpplague | RobotGuy: you didn't rebooot |
| 23:04.28 | RobotGuy | prpplague: Oooooh, silly me! |
| 23:04.37 | prpplague | RobotGuy: boot and reboot aren't the same |
| 23:05.46 | RobotGuy | prpplague: Oh, OK. Does reboot work at the apex prompt? |
| 23:06.03 | prpplague | RobotGuy: no, just push the reset button on the carrier |
| 23:07.29 | RobotGuy | prpplague: OK |
| 23:11.30 | RobotGuy | I bought the terminal strips I need to convert the carrier to dual power. |
| 23:11.50 | RobotGuy | I have one regulator if I can find it. |
| 23:11.56 | ds2 | hmmm |
| 23:12.02 | RobotGuy | I ordered 5 each of the 5.0V and 3.3V regulators. |
| 23:12.51 | RobotGuy | I just found out my friends has finished milling the new decks for W.A.L.T.E.R. !! |
| 23:13.27 | prpplague | RobotGuy: nice |
| 23:14.07 | RobotGuy | I think I uploaded a 3D PDF of how W.A.L.T.E.R. will look with the new decks to PARTS. |
| 23:15.20 | *** join/#edev GPSFan (n=kenm@12.10.255.246) |
| 23:15.20 | *** mode/#edev [+o GPSFan] by ChanServ |
| 23:17.09 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 23:17.09 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 23:21.03 | RobotGuy | prpplague: I do apex> setenv ramdisksrc *= "nor:1280k+7168k" and then printenv and I get ramdisksrc = *= nor:1280k+7168k |
| 23:21.22 | prpplague | RobotGuy: yep, that looks correct |
| 23:21.32 | RobotGuy | Now just reset? |
| 23:21.39 | prpplague | yep |
| 23:21.54 | prpplague | RobotGuy: you should see it copy the ramdisk |
| 23:21.57 | RobotGuy | # copy $kernelsrc $bootaddr |
| 23:21.57 | RobotGuy | # copy nor:256k+1024k 0x30008000 |
| 23:21.57 | RobotGuy | 1048576 bytes transferred |
| 23:21.57 | RobotGuy | # copy $ramdisksrc $ramdiskaddr |
| 23:21.58 | RobotGuy | # copy *= nor:1280k+7168k 0x30800000 |
| 23:21.58 | RobotGuy | Error -9 (parameter error) |
| 23:22.00 | RobotGuy | apex> |
| 23:22.14 | ds2 | prpplague: if you go ahead with the bluez-lite type project, I would be interested in contributing |
| 23:22.29 | prpplague | RobotGuy: hehe, sorry i didn't notice you syntax |
| 23:22.52 | prpplague | RobotGuy: its: setenv ramdisksrc "nor:1280k+7168k" |
| 23:23.12 | prpplague | ds2: okie dokie, i'll keep that in mind |
| 23:23.53 | RobotGuy | Still got a kernel panic. |
| 23:24.09 | prpplague | RobotGuy: pastebin.ca the log |
| 23:24.14 | RobotGuy | prpplague: OK |
| 23:24.33 | CosmicPenguin | I remember when bluez _was_ light |
| 23:24.37 | CosmicPenguin | I still like it though |
| 23:24.50 | prpplague | CosmicPenguin: yea, its not specifically bluez |
| 23:24.57 | prpplague | CosmicPenguin: its the hcid thats the issue |
| 23:25.08 | CosmicPenguin | Well - you have to play to your strengths |
| 23:25.09 | CosmicPenguin | I guess |
| 23:25.24 | CosmicPenguin | Lots of people want to plug their phone and headset and headphones into Linux |
| 23:25.26 | RobotGuy | prpplague: http://www.pastebin.ca/850054 |
| 23:25.29 | ds2 | the worse offense, IMO, is the new HCID requires working dbus to even pair |
| 23:25.39 | *** join/#edev Crofton|work (n=balister@66.207.66.26) |
| 23:25.39 | CosmicPenguin | and nearly all of them don't care about you |
| 23:26.01 | ds2 | none of thatput a pin in hcid.conf and be done with (fine for embedded) |
| 23:26.13 | hapchap | ds2: agreed |
| 23:26.18 | prpplague | CosmicPenguin: thats not entirely correct, if you have a linux based phone, you need a lite hcid to operate |
| 23:26.25 | hapchap | all i need is a single outgoing rfcomm connection with a pin |
| 23:26.26 | ds2 | the current expectation to pop up a UI and have to type a PIN, is IMO ludicirous |
| 23:26.31 | prpplague | CosmicPenguin: the one thats used in the openmoko is giant |
| 23:26.54 | ds2 | I'd say use bluez 2.x but there are stuff that needs bluez 3 :( |
| 23:26.59 | prpplague | RobotGuy: you left off the apex log portion |
| 23:27.05 | ds2 | <--- fought the bluez battle late last year |
| 23:27.05 | hapchap | look at all of the diagrams on bluez.org about how it is structured... every part of it it expecting a windowing environment |
| 23:27.06 | CosmicPenguin | you have to admit that the majority of Bluetooth users under Linux use a full fledged desktop |
| 23:27.20 | CosmicPenguin | Thats just too bad for embedded, but its the facts |
| 23:27.37 | hapchap | ds2: yea, we have ended up with just using libbluetooth 3.x and hcid 2.25 |
| 23:27.57 | hapchap | ds2: do far we have no problems with our little rfcomm connection |
| 23:28.03 | ds2 | hapchap: does that work with the a2dpd/headsetd stuff? |
| 23:28.14 | hapchap | ds2: havent even tried |
| 23:28.18 | prpplague | CosmicPenguin: but you have to also agree that linux is making alot more progess on embedded devices like phones that it is for desktop |
| 23:28.19 | ds2 | ah |
| 23:28.30 | prpplague | s/that/than |
| 23:28.31 | hapchap | ds2: we dont even have a full audio codec on the device we are using it on |
| 23:28.55 | ds2 | hapchap: gotcha... what physical BT device are you using, if you don't mind me asking |
| 23:29.11 | hapchap | ds2: a cs bluetooth dongle |
| 23:29.16 | CosmicPenguin | prpplague: your point being? Is it in the project requirements for bluez to play in embedded? Probably not |
| 23:29.20 | prpplague | RobotGuy: can you post the apex log? |
| 23:29.20 | ds2 | USB? |
| 23:29.25 | RobotGuy | prpplague: http://www.pastebin.ca/850059 |
| 23:29.28 | hapchap | ds2: yea |
| 23:29.35 | CosmicPenguin | Do they have embedded developers, embedded customers? |
| 23:29.46 | ds2 | CosmicPenguin: somewhere in there, there is aclaim of gumstix support so... yes it is. |
| 23:29.47 | prpplague | CosmicPenguin: again, its not bluez thats the issues, its the associated hcid |
| 23:30.03 | CosmicPenguin | The last thing you get to be is upset because _ANY_ open source project plays to their strengths |
| 23:30.17 | ds2 | prpplague: it is in other places too... a2dpd is another dbus infected thing |
| 23:30.43 | *** join/#edev ico2_ (n=ico2@modem-2676.chimpanzee.dialup.pol.co.uk) |
| 23:31.01 | prpplague | CosmicPenguin: not upset, just explaining that a whole area is being ignored that would seem to make the most sense to be able to work with |
| 23:31.10 | CosmicPenguin | you're upset, because you bring it up daily |
| 23:31.28 | hapchap | i dont really have a problem with it for most devices, but ours is just _very_ basic... no X, no audio codec, just a good ole console and greyscale framebuffer |
| 23:31.42 | CosmicPenguin | here's the bottom line - 99% of people who use bluetooth take full advantage of a little icon at the top of their desktop that very nicely pops up a dialog box when a new device is discovered |
| 23:31.43 | hapchap | dbus has some good things |
| 23:31.45 | ds2 | CosmicPenguin: I think they should have made it a ./configure option instead of forcing it on people |
| 23:31.52 | CosmicPenguin | Thats a very useful feature |
| 23:31.54 | CosmicPenguin | I use it |
| 23:32.31 | hapchap | ds2: like in 2.25 |
| 23:32.31 | CosmicPenguin | Thats what their market is |
| 23:32.31 | ds2 | hapchap: yep |
| 23:32.31 | hapchap | CosmicPenguin: true |
| 23:32.31 | prpplague | RobotGuy: post a copy of the rootfs on imagebin.ca |
| 23:32.31 | CosmicPenguin | Maybe they could have done something different - maybe they should have |
| 23:32.46 | CosmicPenguin | But its unlikely that they had reason or motivation to do something differently |
| 23:32.56 | CosmicPenguin | Heck - they probably didn't even think about it |
| 23:33.13 | prpplague | CosmicPenguin: indeed, hence the reason we are talking about address it |
| 23:33.15 | CosmicPenguin | Thats too bad - but thats open source |
| 23:33.18 | ds2 | you are probally right |
| 23:33.31 | prpplague | CosmicPenguin: we just have a problem to solve |
| 23:33.34 | CosmicPenguin | indeed |
| 23:33.48 | hapchap | and honestly, we solved it by mixing and matching versions |
| 23:34.13 | CosmicPenguin | It is a problem - but I don't think it should reflect poorly on the bluez developers |
| 23:34.18 | ds2 | another thing is, from what I have seen, a lot of embedded guys are not even familiar with bluetooth to have an opinion :/ |
| 23:34.25 | prpplague | CosmicPenguin: i agree |
| 23:34.30 | prpplague | CosmicPenguin: 100% |
| 23:34.58 | hapchap | prpplague: im out... see you tomorrow |
| 23:35.02 | prpplague | hapchap: later |
| 23:35.21 | RobotGuy | prpplague: Posting now |
| 23:35.29 | prpplague | RobotGuy: ok |
| 23:37.36 | prpplague | RobotGuy: heading home, be back in 20 |
| 23:40.01 | RobotGuy | Dang, I can't upload this to imagebin.ca |
| 23:54.53 | *** join/#edev dijenerate (n=dijenera@72.51.111.182) |
| 23:58.16 | *** join/#edev FireEgl (i=FireEgl@Sebastian.Atlantica.US) |