| 00:03.32 | CosmicPenguin | bye bye amd stock |
| 01:10.20 | *** join/#edev prpplague (n=dave@72.22.153.82) |
| 01:10.55 | prpplague | i'm 36 years old and my favorite show is justice league hehe |
| 01:11.34 | agaffney | that show rocks :P |
| 01:12.30 | *** join/#edev redb1ue (i=star@ppp122.109-253-207.mtl.mt.videotron.ca) |
| 01:16.42 | prpplague | agaffney: i agree |
| 01:16.58 | prpplague | agaffney: its an adult show disguised as a kids cartoon |
| 01:17.21 | agaffney | I haven't watched it in a while, though, unfortunately |
| 01:17.27 | agaffney | I don't watch TV much anymore |
| 01:17.37 | agaffney | I just record a few things during the week |
| 01:17.41 | agaffney | and watch them at night |
| 01:29.16 | *** join/#edev catastr0p (n=catastr0@d14-69-178-65.try.wideopenwest.com) |
| 01:57.56 | *** join/#edev gandhijee (n=gandhije@c-24-147-213-182.hsd1.ma.comcast.net) |
| 01:59.26 | *** join/#edev catastrop (n=catastr0@d14-69-178-65.try.wideopenwest.com) |
| 03:29.16 | MonMotha | prpplague: you use Eagle for board layout, right? |
| 04:22.18 | *** join/#edev ccjoe (i=clrsrv@www.clearcore.com) |
| 09:19.53 | *** join/#edev AMDPenguin (n=nobody@svl-ext-proxy02.amd.com) |
| 11:07.13 | *** join/#edev redb1ue (i=star@ppp228.108-253-207.mtl.mt.videotron.ca) |
| 11:31.08 | prpplague | MonMotha: yea thats what i'm learning now |
| 12:03.52 | T0mW | does anyone here have this book "ARM system's Developer's Guide"? |
| 12:30.42 | *** join/#edev [1]gandhijee (n=gandhije@mail.win-ent.com) |
| 12:44.34 | *** join/#edev GPSFan (n=kenm@12.10.255.248) |
| 12:44.34 | *** mode/#edev [+o GPSFan] by ChanServ |
| 12:48.05 | *** join/#edev prpplague^2 (n=billybob@72.22.131.223) |
| 12:48.43 | prpplague^2 | ho ho ho |
| 12:49.32 | *** join/#edev prpplague^2 (n=dave@72.22.153.82) |
| 12:49.58 | T0mW | prpplague: morning Dave |
| 12:52.37 | prpplague | T0mW: morning |
| 12:52.44 | prpplague | hehe |
| 12:53.17 | T0mW | prpplague: that was me yesterday, I've recovered now |
| 12:53.32 | T0mW | prpplague: getting too old for these "all nighters" |
| 12:53.46 | prpplague | T0mW: hehe, my blood pressure is kinda high, i'm kinda weizing due to sugar cane pollen in the air, and i had some home made bean burrito's last night :) |
| 12:53.58 | T0mW | heh |
| 12:56.34 | prpplague | hmm |
| 12:57.02 | prpplague | jbevren: so would a plan to upgrade the ram for the pixter-mm violate the hack constrants? |
| 13:02.51 | jbevren | pixter: Perhaps the hack constraints could have two levels |
| 13:03.02 | jbevren | easy (no large modifications) and difficult (adding ram IC's) |
| 13:03.13 | jbevren | I'd like to add ram to mine, to be honest |
| 13:03.22 | jbevren | but I'd like to keep a level of pixter hacking that most everyone could do |
| 13:03.27 | jbevren | hehe |
| 13:03.49 | jbevren | what are your thoughts? |
| 13:05.19 | prpplague | jbevren: if it were a 1:1 upgrade, i'd go with it, but its not |
| 13:05.43 | jbevren | how many lines are the same? |
| 13:06.37 | prpplague | jbevren: well none of the pins match, you'd need to solder all of them |
| 13:08.04 | jbevren | ok |
| 13:08.34 | jbevren | I'm kinda miffed too, I got a pair of sdram IC's for the hack and lost them |
| 13:08.49 | jbevren | I found the TSOP breakout boards |
| 13:08.50 | jbevren | heh |
| 13:08.53 | jbevren | fat lot of good they do |
| 13:08.54 | jbevren | :D |
| 13:10.02 | prpplague | hehe |
| 13:11.47 | jbevren | grumble |
| 13:11.56 | jbevren | I believe I got a pair of 8megabyte chips |
| 13:12.04 | jbevren | plenty of room to play heh |
| 13:12.19 | prpplague | yea |
| 13:15.38 | *** mode/#edev [+o prpplague] by ChanServ |
| 13:18.40 | T0mW | heh, there are two patches that aeronix gave which shouldn't be there, they added two files: netsyms.c.orig & ep7312.h.old |
| 13:22.23 | T0mW | you know, patch files are nice to have, but not one huge patch |
| 13:22.24 | T0mW | I would rather see: |
| 13:22.37 | T0mW | 100-fixup-missing-init-call |
| 13:22.49 | T0mW | 101-add-wireless-device |
| 13:22.51 | T0mW | etc |
| 13:23.53 | T0mW | then it makes it far easier to move code onto a later version kernel |
| 13:26.32 | prpplague | T0mW: yea, you can break a patches out to single patches |
| 13:26.46 | T0mW | why does a buildroot 'make clean' destroy the existing cross-compiler? |
| 13:27.05 | T0mW | I would rather a 'make allclean' do that |
| 13:27.27 | T0mW | with 'make clean' only clear out the packages, kernel and such |
| 13:27.39 | prpplague | T0mW: yea thats been an issue for a long time, we've discussed fixes on many occasions just haven't agreed on a fix |
| 13:28.02 | T0mW | oh, so, that is coupled in with Erik's source then? |
| 13:28.22 | prpplague | T0mW: yea thats a function of the buildroot make |
| 13:28.29 | T0mW | prpplague: who else is involved in decisions on the buildroot distro? |
| 13:29.07 | T0mW | prpplague: you're the maintainer, right? |
| 13:29.16 | prpplague | T0mW: well basically we just post an issue on the bug track, then it assigns a case number, then you can post the number on #uclibc for discussion |
| 13:29.43 | prpplague | T0mW: there are about 5 people, vapier, landley, andersee, sjhill, and me when i have the time |
| 13:29.51 | T0mW | ok |
| 13:30.08 | T0mW | sjhill still lives? |
| 13:31.28 | T0mW | prpplague: well, the plan for the day is, take what I got so far and get it working, correctly, under a fresh copy of buildroot. |
| 13:32.09 | T0mW | prpplague: then, if I have the time, vet the aeronix patches and GPSFan patches, try to break them out into individual patch files |
| 13:33.44 | prpplague | T0mW: rule of thumb, after a patch is applied you should be able to do a fresh build without errors and create a usable binary |
| 13:36.11 | *** join/#edev dsmith (n=user@66.178.229.162) |
| 13:37.33 | dsmith | Howdy Folks. |
| 13:38.14 | jbevren | hm |
| 13:38.21 | dsmith | Does anyone know how to convice objcopy to split a file into 16bit wide sections? |
| 13:38.57 | *** join/#edev ade|desk (n=ADAVEY@194.200.143.249) |
| 13:39.02 | ade|desk | ello |
| 13:39.30 | ade|desk | why the invite ? |
| 13:40.11 | ade|desk | hehe, indeed |
| 13:43.56 | ade|desk | whats the difference between #edev and #elinux ? |
| 13:46.39 | jbevren | #edev isnt linux-centric |
| 13:46.44 | jbevren | although most chat here ends up being so |
| 13:47.13 | jbevren | hm perhaps its better to say #elinux focuses more on linux, and #edev focuses more on the device |
| 13:47.16 | jbevren | :D |
| 13:47.24 | prpplague | ade|desk: hehe i invited you |
| 13:48.32 | prpplague | ade|desk: its also more centered around hacking devices, kinda like a cross between elinux and handhelds.org , just a little saner(sp?) |
| 13:49.14 | prpplague | ade|desk: arjans was asking about tinyx/kdrive memory footprint |
| 13:49.15 | ade|desk | anyone here got and amstrad e3 emailer phone ? |
| 13:49.31 | prpplague | ade|desk: tried for awhile, but couldn't get one in the states |
| 13:49.40 | prpplague | ade|desk: nice dev board i hear though |
| 13:49.43 | jbevren | nod, amstrad stuff is hard to get |
| 13:50.13 | ade|desk | i haven't had time to pull it apart yet and my wife told me i couldn't get more than one :( |
| 13:50.29 | prpplague | hehe |
| 13:51.33 | ade|desk | the only other device i would llike is a motorola a780, ordered one but still haven't got it cos they are out of stock |
| 13:51.35 | jbevren | I love being single |
| 13:52.17 | ade|desk | jbevren: tried that, for a few of my earlier years ;) |
| 13:52.46 | ade|desk | now its wife and time to start a family |
| 13:53.22 | prpplague | ade|desk: try that for a few years and then you can go back to the good life, hehe |
| 13:57.54 | ade|desk | heh |
| 14:23.50 | prpplague | CosmicPenguin: ho ho ho |
| 14:32.24 | prpplague | jbevren: i think we can do this pretty well with xip |
| 14:32.34 | prpplague | jbevren: probably two file systems |
| 14:32.47 | prpplague | jbevren: one cramfs on some nor flash, and jffs2 on some nand |
| 14:35.54 | jbevren | prpplague: Sure |
| 14:36.11 | jbevren | prpplague: include the cramfs initial rootfs in the kernel binary |
| 14:36.18 | jbevren | that way it gets put on the nor flash with the kernel ;-) |
| 14:36.27 | jbevren | After that it locates and mounts the nand jffs2 |
| 14:36.54 | jbevren | maybe failure ot mount would give a shell for debug? heh |
| 14:38.07 | jbevren | hm ebay has a pair of amstrad e3's :D |
| 14:38.13 | jbevren | 140003683945 |
| 14:38.19 | jbevren | curious looking devices, they are |
| 14:39.01 | prpplague | jbevren: yea, thats what i'm thinking |
| 14:39.30 | jbevren | prpplague: I'm thinking it'd make a cool kitchen terminal heh |
| 14:39.51 | prpplague | hehe |
| 14:42.13 | jbevren | how come the brits get the cool stuff :( |
| 14:42.31 | jbevren | the acorn riscpc's, that phone, the unfair music laws.. |
| 14:42.32 | jbevren | wait |
| 14:42.35 | jbevren | exclude the third |
| 14:42.38 | jbevren | anyway brb |
| 14:43.29 | prpplague | hehe |
| 14:43.49 | prpplague | jbevren: strange , they seems to get the cool kids laptops too |
| 14:45.50 | *** join/#edev gandhijee (n=gandhije@mail.win-ent.com) |
| 14:51.34 | ade|desk | prpplague what kids laptops ? |
| 15:04.06 | prpplague | ade|desk: ebay Item number: 140002422746 |
| 15:10.24 | jbevren | prpplague: indeed |
| 15:10.32 | jbevren | if I trusted customs to let the phones go through, I'd bid |
| 15:10.58 | jbevren | hehe |
| 15:10.58 | prpplague | hmm anyone know if there is an easy way to turn off the initramfs probe in 2.6? |
| 15:11.13 | T0mW | prpplague: isn't that an ATAG? |
| 15:11.57 | T0mW | err, initrd |
| 15:14.12 | prpplague | T0mW: sorta, the problem is that the initramfs checks the ramdisk to see if it is a initramfs or initrd image |
| 15:14.18 | prpplague | T0mW: pretty long delay |
| 15:14.26 | prpplague | T0mW: i need that hardcoded for boot time |
| 15:14.45 | T0mW | heh, create a new ATAG definition? |
| 15:21.41 | prpplague | T0mW: naw |
| 15:21.53 | T0mW | <shrug> |
| 15:21.54 | prpplague | T0mW: looks like there isn't any code to handle it being preset |
| 15:22.19 | T0mW | if (ATAG.foo) return; |
| 15:23.08 | prpplague | yeam doesn't work that way |
| 15:23.18 | prpplague | this is part of the initial kernel bring up |
| 15:23.28 | prpplague | the atag already knows there is a ramdisk |
| 15:23.52 | prpplague | just doesn't know if its a initramfs or initrd |
| 15:24.05 | prpplague | normal method is to check the magic number |
| 15:24.34 | jbevren | prpplague: and you want it as an initrd? |
| 15:25.00 | prpplague | jbevren: it is an initrd, and it works, but i just don't want it to waste time doing the check |
| 15:25.10 | prpplague | jbevren: the check takes almost a full second |
| 15:25.19 | prpplague | jbevren: .9987 seconds |
| 15:25.39 | jbevren | prpplague: or the codespace ;-) |
| 15:25.54 | prpplague | jbevren: ?? |
| 15:26.25 | T0mW | prpplague: distclean is supposed clean up buildroot prior to packaging & distribution? |
| 15:26.26 | jbevren | prpplague: not doing the initramfs check (if you remove the code) can save a few bytes |
| 15:26.36 | jbevren | T0mW: yeah |
| 15:26.40 | T0mW | k |
| 15:26.41 | prpplague | T0mW: correct |
| 15:26.46 | jbevren | its supposed to make it as clean as it was when you unpacked it to begin with |
| 15:27.00 | prpplague | jbevren: not worried about the bytes, worried about the time |
| 15:27.09 | jbevren | prpplague: nod |
| 15:27.45 | prpplague | jbevren: alot of the probe's in the kernel can be disabled via the kernel boot arguments by passing known values |
| 15:27.56 | prpplague | jbevren: like the loops per jiffy |
| 15:28.21 | jbevren | bogomips calibration? |
| 15:28.21 | prpplague | jbevren: that calibration loop takes about 3 seconds on my dev board |
| 15:28.25 | prpplague | jbevren: yea |
| 15:28.32 | jbevren | wtf is bogomips used for anyway |
| 15:28.55 | prpplague | jbevren: its used to do very specific timed delays |
| 15:29.48 | jbevren | shrug. Isnt that waht timers are for? |
| 15:31.47 | prpplague | jbevren: yea, there are several, but there are some very specific delays that need this |
| 15:32.10 | jbevren | shrug |
| 15:32.15 | *** join/#edev gandhijee (n=gandhije@mail.win-ent.com) |
| 15:32.17 | jbevren | so measure your bogomips and hardcode it :) |
| 15:32.43 | gandhijee | should mdev/udev be called before modules are loaded or after |
| 15:32.44 | prpplague | jbevren: yea, just run through once, read the lpj and then set it as a kernel arg |
| 15:32.51 | gandhijee | or does it not matter? |
| 15:33.07 | prpplague | jbevren: my point being you can do that with lpj, you can't with the initramfs or initrd detection |
| 15:33.22 | jbevren | ohh |
| 15:33.24 | jbevren | right |
| 15:35.12 | prpplague | jbevren: yea that dropped a whole second |
| 15:36.35 | T0mW | make distclean does not cleanup buildroot to pristine readiness, it still leaves .config & .config.cmd there |
| 15:38.19 | prpplague | yea |
| 15:38.44 | T0mW | heh, 'make really-really-clean-really' ? |
| 15:39.39 | T0mW | prpplague: I know, "if it is broke, and you found it, then you fix it" |
| 15:45.40 | jbevren | T0mW: make distclean is broken in that package |
| 15:45.41 | jbevren | :) |
| 15:46.08 | T0mW | yes, it is, it leaves object files all over the place |
| 15:52.22 | jbevren | hehe |
| 15:52.33 | jbevren | find . -name *.o -exec rm -fv {} \; |
| 15:54.37 | T0mW | bah!@ |
| 15:54.51 | T0mW | this is all fscked up |
| 15:55.00 | chouimat | morning |
| 16:04.03 | jbevren | chou |
| 16:04.54 | EvilDevil | anybody knows about "ST20" software for linux? i've found a spare DVB-S STB based on the http://us.st.com/stonline/products/promlit/pdf/flsti5518-1101.pdf chip and i want to experiment with that thing ;) |
| 16:06.51 | prpplague | hmm, is initramfs images actually built into the kernel? |
| 16:06.59 | jbevren | yeah |
| 16:07.07 | jbevren | it becomes part of vmlinux |
| 16:07.31 | agaffney | isn't it just combined with the kernel image afterwards? |
| 16:08.25 | jbevren | agaffney: I believe so, but it does become part of vmlinux (?) |
| 16:09.12 | prpplague | hmm |
| 16:09.14 | prpplague | interesting |
| 16:09.32 | prpplague | thats kinda of a problem for us to use as we need to be able to upgrade kernel and ramdisk seperately |
| 16:09.52 | prpplague | agaffney: the wording i see is "linked in" |
| 16:10.18 | agaffney | it doesn't *have* to be combined |
| 16:10.22 | agaffney | it can still be separate |
| 16:10.55 | prpplague | agaffney: ahh |
| 16:11.26 | prpplague | agaffney: so i can use a initramfs image in the same manner as a initrd? |
| 16:11.33 | agaffney | yes |
| 16:11.45 | agaffney | they're basically the same thing with different packaging |
| 16:11.59 | prpplague | agaffney: well there is a little bit of difference in the internals |
| 16:12.25 | agaffney | initrd was a loopback ext2 (or whatever) and initramfs is a cpio archive (or possibly multiple cpio archives cat'd together) |
| 16:12.27 | prpplague | the last thing i have slowing down the system is the uncompressing of the ramdisk |
| 16:12.35 | prpplague | agaffney: yea |
| 16:14.23 | agaffney | are you asking if it'll work in your environment or if the kernel supports it? |
| 16:14.58 | prpplague | both, hehe |
| 16:15.22 | agaffney | I'm sure the kernel supports it |
| 16:15.26 | prpplague | looks like initramfs is probably a better choice for both size and speed |
| 16:23.06 | jbevren | hm |
| 16:23.12 | jbevren | XIP uncompressed initramfs? |
| 16:24.12 | prpplague | jbevren: hehe sorry two different projects |
| 16:24.37 | jbevren | prpplague: it could still apply ;-) |
| 16:24.51 | prpplague | jbevren: yea, we don't want the ramfs for the pixter |
| 16:24.58 | prpplague | we are already short on ram |
| 16:25.23 | jbevren | initROMfs? hehe |
| 16:25.34 | jbevren | nayway |
| 16:25.37 | jbevren | I'll eave you to do your work |
| 16:25.39 | prpplague | hmm, my zImage appears to have my initramfs in it |
| 16:47.10 | prpplague | uncompressing the initrd seems to be the last place left to tweak |
| 16:49.08 | prpplague | doesn't appear i'm gonna get below 4 seconds |
| 17:05.33 | prpplague | jbevren: interesting |
| 17:05.45 | prpplague | jbevren: pixter retains picture after battery removed |
| 17:09.04 | prpplague | hmm |
| 17:09.12 | prpplague | only seems to save one picture though |
| 17:11.27 | jbevren | might be a software limitation |
| 17:11.50 | jbevren | 160x160 at 8bits is 25k |
| 17:11.51 | prpplague | jbevren: looks like it has 90 preprogrammed audio sounds |
| 17:11.53 | jbevren | there's at least 32k |
| 17:12.00 | jbevren | cause it also stores your name |
| 17:12.06 | jbevren | the screensaver setting |
| 17:12.22 | jbevren | wonder if there's a little serial rom somewhere |
| 17:13.22 | *** join/#edev redblue (i=star@ppp059.109-253-207.mtl.mt.videotron.ca) |
| 17:13.34 | prpplague | jbevren: i'm thinking that one of those COB's might be one |
| 17:14.39 | prpplague | jbevren: the battery info is connected to one of the adc's |
| 17:15.36 | prpplague | jbevren: hehe, you can turn the melody on and off inside the test app |
| 17:17.23 | prpplague | jbevren: yea its got something |
| 17:17.51 | prpplague | jbevren: there are two buttons in the test app, one marked BBE and the other CBE |
| 17:18.04 | prpplague | jbevren: if you click on BBE you see for a second, "erasing" |
| 17:19.45 | jbevren | test app? |
| 17:19.57 | prpplague | jbevren: hold down the calibrate button and turn on the power |
| 17:20.08 | jbevren | mine just goes to the calibrate screen when I do that |
| 17:20.08 | prpplague | thought you knew that on |
| 17:20.20 | prpplague | jbevren: yea do the calibrate and then you get a test app |
| 17:20.24 | jbevren | thats why I had it jumpered |
| 17:20.25 | jbevren | oh |
| 17:20.26 | jbevren | well, |
| 17:20.28 | prpplague | doh |
| 17:20.29 | jbevren | I cant calibrate :) |
| 17:20.30 | prpplague | no ts |
| 17:20.35 | prpplague | hehe |
| 17:20.38 | jbevren | maybe I should take you up on the offer |
| 17:20.39 | jbevren | lol :) |
| 17:20.41 | prpplague | yea |
| 17:20.45 | prpplague | i was just thinking that |
| 17:20.55 | prpplague | wonder what BBE and CBE are |
| 17:20.56 | jbevren | feel free, I just dont feel right asking for stuff from a relative stranger |
| 17:21.09 | prpplague | hehe |
| 17:21.22 | prpplague | CBE doesn't appear to do anything |
| 17:28.31 | prpplague | ahh check this out |
| 17:28.57 | prpplague | it must be Based Block Erase and Cartridge Block Erase |
| 17:30.11 | prpplague | jbevren: hey, btw, are the cartridges, component side up or down? |
| 17:30.54 | jbevren | down |
| 17:31.00 | jbevren | if you put it in backwards, it doesnt bootstrap |
| 17:33.51 | prpplague | interesting |
| 17:34.02 | jbevren | probably becuase the address lines are shorted or something |
| 17:34.11 | prpplague | it appears that its running the test app from the cartridge |
| 17:34.11 | jbevren | but yeah, finding the onboard flash would be cool |
| 17:34.33 | jbevren | oh, and I'll eventually order a palm3c lcd |
| 17:34.35 | jbevren | eventually.. heh |
| 17:35.17 | prpplague | its got to be some sort of storage |
| 17:35.50 | jbevren | probably the same flash chip that holds the picture |
| 17:35.58 | jbevren | it has 3 CoB's right? |
| 17:36.20 | prpplague | yea three |
| 17:36.26 | prpplague | i suspect one is the audio |
| 17:36.28 | jbevren | one's firmware obviously |
| 17:36.36 | jbevren | I'd like to speculate that another is the power control |
| 17:36.40 | prpplague | yea, the big one is probably where the firmware its |
| 17:36.42 | jbevren | there's no regulators visible on the board |
| 17:36.47 | prpplague | jbevren: i doubt that |
| 17:36.51 | jbevren | hmm |
| 17:37.01 | jbevren | where's it get the vcc and vcore voltages? |
| 17:37.09 | jbevren | does the sharp have an onboard psu? |
| 17:37.18 | prpplague | jbevren: there are two ldo's on board |
| 17:37.26 | jbevren | oh |
| 17:37.27 | jbevren | I'm blind |
| 17:37.36 | jbevren | that's the two little chips by the caps? |
| 17:38.11 | prpplague | yea |
| 17:38.46 | jbevren | ok |
| 17:39.01 | jbevren | so what else is there in a simple computer |
| 17:39.09 | jbevren | address decoding is done by the arm |
| 17:39.27 | jbevren | the TI DAC can drive headphones according to the sheet iirc |
| 17:41.31 | prpplague | jbevren: it looks like the middle COB is the audio |
| 17:41.41 | prpplague | it appears to be connected to the dac |
| 17:41.55 | prpplague | jbevren: the top smaller COB appears to be the flash |
| 17:42.52 | jbevren | the one closest to the lcd port or the one in the middle |
| 17:42.53 | jbevren | hehe |
| 17:42.58 | jbevren | http://www.jbevren.net/pixter/pixterboard-overhead.jpg |
| 17:43.00 | jbevren | good reference picture |
| 17:43.25 | prpplague | jbevren: the one closest to the lcd port appears to be the storage |
| 17:45.41 | jbevren | hm ok |
| 17:45.48 | jbevren | yeah, it does have lots of lines going from the mcu to the blob |
| 17:45.53 | jbevren | so now how about the middle one :) |
| 17:46.24 | prpplague | that has some connections to the dac |
| 17:46.32 | prpplague | so i would guess that has the audio on it |
| 17:46.34 | jbevren | oh that reminds me |
| 17:46.46 | jbevren | under the yellow wire, between C50 and the middle blob is a SMT cap |
| 17:46.53 | jbevren | I was probing different locations the other night with my scope |
| 17:47.06 | jbevren | if I probe the blob-side of that cap, the pitch on the music changes slightly |
| 17:47.45 | jbevren | (I imagine either the capacitance or resistive load on my probe is changing the freq of some oscillator) |
| 17:48.25 | jbevren | its a 10Mohm probe though |
| 17:48.25 | jbevren | heh |
| 17:48.26 | jbevren | a few pf in capacitance iirc |
| 17:49.11 | prpplague | i really need to remove the parts from one of these and do some mapping |
| 17:49.16 | prpplague | its gonna take forever if we done |
| 17:49.19 | prpplague | dont |
| 17:49.27 | jbevren | grue |
| 17:49.29 | jbevren | er.. |
| 17:49.30 | jbevren | true |
| 17:50.01 | prpplague | mainly removing the processor so we can see the ball pads would be a big help |
| 17:50.36 | jbevren | nod |
| 17:50.50 | jbevren | or dissolve the epoxy away |
| 17:50.53 | jbevren | hehe |
| 17:51.12 | jbevren | it's possible to guess at a chip's function by examining the visual pattern on it |
| 17:51.23 | prpplague | yea |
| 17:51.28 | prpplague | but i don't like guessing |
| 17:51.31 | jbevren | I used to do that with pentiums hehe |
| 17:51.38 | jbevren | here's the cache |
| 17:51.40 | jbevren | this looks like the ALU |
| 17:51.41 | jbevren | hehe |
| 17:51.54 | jbevren | that small storage area is the microcode... ;-) |
| 17:54.08 | prpplague | [21474536.870000] RAMDISK: Compressed image found at block 0 |
| 17:54.09 | prpplague | [21474538.555000] VFS: Mounted root (ext2 filesystem). |
| 17:54.16 | jbevren | cool |
| 17:54.19 | prpplague | almost two whole seconds |
| 17:54.26 | jbevren | it has to uncompress |
| 17:54.30 | prpplague | jbevren: not so cool |
| 17:54.41 | jbevren | I was noting the coolness of getting root mounted |
| 17:54.43 | jbevren | irc lag there |
| 17:54.47 | prpplague | hehe |
| 17:54.54 | jbevren | so, uncompressed ramdisk? :) |
| 17:55.09 | jbevren | btw, have you tought bout using minixfs? It has lower overhead iirc |
| 17:55.17 | prpplague | uncompressed ramdisk means copying more during the bootloader init |
| 17:55.25 | prpplague | jbevren: yea |
| 17:55.27 | jbevren | prpplague: which is slower? hehe |
| 17:55.39 | prpplague | jbevren: the uncompressed copy |
| 17:55.43 | jbevren | hm |
| 17:55.59 | jbevren | cramfs? |
| 17:56.01 | prpplague | jbevren: thats why i'm thinking i might need to use a flash based rootfs, i.e. a read only jffs2 |
| 17:56.16 | jbevren | or a romfs (I know there is one..) :) |
| 17:56.22 | jbevren | you could point the kernel to its memory address and go |
| 17:56.29 | jbevren | but then you cant write to it |
| 17:56.45 | prpplague | yea |
| 17:57.52 | prpplague | not much i can at this stage of the game, just gonna have to live with with a 4.5 second startup time |
| 17:58.09 | prpplague | power-on to login |
| 17:58.52 | prpplague | if i could get rid of that initrd decompression, that would bring me down to around 2.25 seconds |
| 17:59.30 | prpplague | i think i have 1 second delay in the bootloader as well |
| 18:01.22 | jbevren | trying to get an instant-on device? |
| 18:01.25 | jbevren | (or as close as possible) |
| 18:01.35 | prpplague | yea |
| 18:01.49 | prpplague | hmm, no it looks to be 1/10 of second delay |
| 18:01.57 | ade|desk | prpplague what are you making in btw? |
| 18:02.12 | prpplague | ade|desk: for work, i'm doing a data collection handheld |
| 18:02.25 | prpplague | ade|desk: for hobby, i'm hacking the pixter multimedia at the momment |
| 18:02.57 | ade|desk | cool |
| 18:03.17 | ade|desk | home calls .. cya |
| 18:03.22 | prpplague | later |
| 18:05.10 | prpplague | jbevren: basically i'm looking at 2.08 seconds for just the kernel |
| 18:06.07 | jbevren | trying to keep it under 5? |
| 18:06.08 | prpplague | and about 1.68 of that is the ramdisk uncompression |
| 18:06.25 | prpplague | jbevren: i'd like to get it around 4 |
| 18:06.44 | prpplague | jbevren: if i can shave another .6 seconds i'd be happy |
| 18:07.24 | jbevren | would it make the user more comfortable if the fb init were earlier so that a splash could b epresented? |
| 18:07.53 | jbevren | I know the feeling though, the gp2x 1.x firmwrae took a good 30sec to boot |
| 18:08.03 | prpplague | jbevren: already doing that |
| 18:08.22 | jbevren | hrm |
| 18:08.35 | jbevren | dunno really |
| 18:08.40 | prpplague | jbevren: the bootload inits the fb, and dumps a splash screen, then when the kernel fb inits it dumps the same splash screen with the addition of "loading" |
| 18:08.45 | prpplague | jbevren: hehe |
| 18:08.54 | jbevren | sounds like youve trimmed what you can |
| 18:08.55 | prpplague | jbevren: just .05 seconds at a time, hehe |
| 18:09.01 | jbevren | hehe |
| 18:09.07 | jbevren | I remember when I used to do that on our p166 fileserver |
| 18:09.18 | jbevren | we had lilo to login in about 5 seconds |
| 18:09.34 | jbevren | thanks to some fast scsi disks |
| 18:09.34 | jbevren | hehe |
| 18:09.45 | jbevren | broke a few scsi rules by removing reset delays |
| 18:09.53 | prpplague | hehe |
| 18:10.58 | prpplague | if i made the flash the rootfs, i think it would be alot quicker |
| 18:11.19 | prpplague | but who knows, the jffs2 partition might take longer to mount |
| 18:12.08 | prpplague | i wish you could uncompress directly from flash |
| 18:13.11 | jbevren | here's an idea |
| 18:13.24 | jbevren | store your compressed ramdisk in flash separate from the kernel |
| 18:13.34 | jbevren | tell the kernel that the compressed ramdisk is at ${addr} |
| 18:13.42 | jbevren | that way the bootloder doesnt have to copy it |
| 18:13.49 | jbevren | the kernel should then find and uncompress it |
| 18:14.05 | prpplague | hmm can you do that? |
| 18:14.17 | jbevren | I dont know for sure to be honest. You coul duse the mtd driver to cheat possibly though |
| 18:14.36 | jbevren | possibly not. heh. |
| 18:14.55 | jbevren | I just remember on the dreamcast with redboot, you could tell redboot the discrete addres sof your kernel image as well as its ramdisk |
| 18:18.41 | dsmith | I think we are doing something like that. |
| 18:23.32 | prpplague | hmm, it doesn't like it not being in the sdram memory space |
| 18:23.53 | jbevren | grumble |
| 18:32.09 | prpplague | looking at the code now |
| 18:32.22 | prpplague | jbevren: overall i think we can do the pixter-mm with the 4mb |
| 18:35.17 | prpplague | hmm i can't seem to find where that error is generated |
| 18:39.22 | prpplague | hmm, there is a check to see if its in the ramdisk area |
| 18:44.16 | prpplague | jbevren: yea looks like thats not easily doable |
| 18:44.52 | prpplague | i think what i need to use is the compressed rom |
| 18:55.54 | prpplague | jbevren: ok so i give up for today |
| 18:56.12 | prpplague | jbevren: its gonna have to stay at 4.5 seconds |
| 18:56.12 | prpplague | oh well |
| 18:57.06 | jbevren | heh |
| 18:57.08 | prpplague | jbevren: anyway, i think we can do XIP kernel with a jffs2 to get started |
| 18:57.36 | prpplague | jbevren: then when get to the point we memory is an issue, we can move to an XIP rootfs |
| 18:57.40 | jbevren | xip kernel frees the ram the kernel image would normally consume so that saves a lot |
| 18:57.59 | prpplague | hmm |
| 18:58.02 | jbevren | sounds fun :) |
| 19:00.14 | jbevren | I have no interest in bots :) |
| 19:00.15 | jbevren | oop |
| 19:01.12 | CosmicPenguin | and they don't c are for you that much either |
| 19:07.19 | jbevren | CosmicPenguin: youre a bot, arent you |
| 19:07.26 | jbevren | hehe |
| 19:09.20 | CosmicPenguin | you would say that, wouldn't you? |
| 19:09.59 | jbevren | hehe |
| 19:09.59 | jbevren | you've been awful quite in the past week ro so |
| 19:09.59 | CosmicPenguin | I can' |
| 19:09.59 | CosmicPenguin | 't get a damn word in edgewise |
| 19:09.59 | jbevren | almost bot-like |
| 19:14.20 | prpplague | jbevren: my xip kernel image is 1.4mb |
| 19:14.35 | jbevren | gonna need some serious flash chips heh |
| 19:15.07 | prpplague | jbevren: naw, 16mb should be fine |
| 19:15.16 | prpplague | jbevren: plus whatever we use for extra storage |
| 19:15.21 | prpplague | jbevren: probably sd/mmc |
| 19:15.44 | jbevren | prpplague: thats what I meant by serious :) |
| 19:16.16 | prpplague | hehe, well actually we should be able to get away with 8mb |
| 19:16.29 | prpplague | just the chips i have are all 16 |
| 19:16.56 | prpplague | jbevren: technically if i built the control circuit i could boot from nand right now |
| 19:17.27 | jbevren | did you figure out how to run a nand bootloader? |
| 19:19.07 | prpplague | jbevren: if i understand you correctly, apex already supports booting from nand on the lh79524 |
| 19:19.23 | jbevren | oh cool |
| 19:22.48 | *** join/#edev EvilDevil_ (n=miau@p54A6F46B.dip.t-dialin.net) |
| 19:29.02 | prpplague | jbevren: well the xip kernel boots but hangs on some mtd stuff |
| 19:29.16 | prpplague | jbevren: gonna disable the mtd stuff for now, and just do a ramdisk for comparison |
| 19:29.52 | prpplague | jbevren: man it boots fast too |
| 19:30.51 | CosmicPenguin | http://www.linuxdevices.com/news/NS4030713830.html |
| 19:30.52 | prpplague | jbevren: i count about 3.5 seconds for the pixter-mm to boot |
| 19:31.37 | prpplague | cute |
| 19:31.56 | CosmicPenguin | I wouldn't mind getting one of those to replace my aging firewall |
| 19:32.17 | prpplague | I wouldn't mind getting one of those to replace my aging wife |
| 19:32.31 | prpplague | hehe |
| 19:34.09 | jbevren | hm |
| 19:34.16 | jbevren | shame they dont use 82559's |
| 19:34.19 | jbevren | it'll work tho |
| 19:36.52 | prpplague | jbevren: saved 1588k of ram with XIP |
| 19:37.09 | jbevren | thats quite a bit of ram |
| 19:37.44 | jbevren | you wouldnt be willing to email me your config files and info on the simulation setp you use would you? :) |
| 19:38.46 | prpplague | jbevren: i could but it really won;t do you any good, i'm testing on our s3c2410 device |
| 19:38.57 | jbevren | ohh |
| 19:42.01 | jbevren | brit: do you have an @reboot cron :) |
| 19:42.02 | jbevren | oop |
| 19:42.10 | jbevren | I need focus follows intent |
| 19:42.45 | prpplague | hehe |
| 20:11.09 | gandhijee | is there away to load firmware for a device without hotplug? |
| 20:11.54 | prpplague | gandhijee: thats a rather vague question |
| 20:12.16 | prpplague | gandhijee: what do you consider firmware? |
| 20:12.24 | gandhijee | ummm is there away to load firmware for a wifi card without hotplug |
| 20:12.37 | prpplague | gandhijee: yea |
| 20:12.50 | prpplague | gandhijee: most cards have a utility available to upload the firmware |
| 20:13.00 | gandhijee | can you enlighten me? |
| 20:13.01 | gandhijee | O |
| 20:13.07 | prpplague | i just did |
| 20:13.10 | prpplague | hehe |
| 20:13.21 | prpplague | gandhijee: just get the util to do the firmware upload |
| 20:13.28 | gandhijee | hmm =/ intels lame ass say i have to have hotplug |
| 20:14.54 | prpplague | gandhijee: well if hotplug can load it, then you can load it |
| 20:15.07 | prpplague | gandhijee: after all most of hotplug just calls helper applications |
| 20:15.29 | gandhijee | O |
| 20:16.35 | gandhijee | thanks |
| 20:17.43 | MonMotha | doing a firmware load by hand is annoying |
| 20:18.05 | MonMotha | it's timing sensitive, and the place you dump the firmware is some random place in sysfs that's only there for 5 seconds or so until the load times out |
| 20:18.11 | MonMotha | it's easier to just load up hotplug |
| 20:22.35 | prpplague | hehe |
| 20:26.45 | gandhijee | =/ hotplug keeps saying firmware event not supported |
| 20:29.51 | prpplague | jbevren: hmm, looks like the kernel is using about 2.5mb of ram even when doing XIP |
| 20:31.01 | prpplague | jbevren: oh wait that isn't correct |
| 20:31.20 | prpplague | jbevren: that 2.5 also includes includes init and getty |
| 20:36.30 | prpplague | doh, i erased my bootloader, and forgot to reflash it |
| 20:37.04 | gandhijee | that sucks |
| 20:37.19 | prpplague | hehe, more of just an inconvience |
| 20:38.36 | gandhijee | what bootload are you using? |
| 20:42.42 | prpplague | apex |
| 20:48.07 | jbevren | prpplague: that'll teach ya |
| 20:48.28 | gandhijee | i'll have to try that out for my 465 |
| 20:52.46 | prpplague | 465? |
| 21:07.36 | gandhijee | IXP |
| 21:07.44 | prpplague | ahh |
| 21:07.51 | gandhijee | the 425's big brother |
| 21:08.20 | prpplague | good old Inhell |
| 21:08.39 | gandhijee | yeah pretty much |
| 21:08.50 | gandhijee | i missed the patch for the 2.6 kernel |
| 21:09.07 | gandhijee | to get the NPE's working and ended up modifiying the damn files by hand |
| 21:10.02 | prpplague | jbevren: hehe i just booted with xip kernel and a 1.5mb ramdisk on 4mb ram |
| 21:10.13 | prpplague | jbevren: hehe wiht 428kb left over |
| 21:10.14 | jbevren | D: |
| 21:10.16 | jbevren | :D |
| 21:10.43 | jbevren | can a static busybox be made smaller? |
| 21:10.49 | jbevren | I know I can get one into 400k on x86 |
| 21:10.59 | prpplague | jbevren: thats not really an issue |
| 21:11.07 | prpplague | jbevren: as we don't plan on using a ramdisk |
| 21:11.13 | jbevren | true |
| 21:11.29 | prpplague | jbevren: so total with a booted system, so far we can expect about 1912k free |
| 21:12.09 | prpplague | MemTotal: 3884 kB |
| 21:12.09 | prpplague | MemFree: 428 kB |
| 21:12.21 | prpplague | Filesystem 1k-blocks Used Available Use% Mounted on |
| 21:12.21 | prpplague | /dev/ram0 1484 1202 206 85% / |
| 21:12.43 | prpplague | so 1484+428 would be what we should expect |
| 21:14.06 | jbevren | nod |
| 21:14.09 | jbevren | sounds good |
| 21:14.15 | jbevren | does that kernel have networking? |
| 21:14.45 | prpplague | no, just unix sockets |
| 21:15.39 | prpplague | i think over all we'll have about 2mb to work with easily |
| 21:16.08 | prpplague | CosmicPenguin: think you can get nano-x working with some apps under 2mb? |
| 21:16.51 | jbevren | is there a 'tiny' Xlib? |
| 21:17.05 | prpplague | jbevren: there is kdrive and a couple others |
| 21:17.18 | prpplague | jbevren: but their mem footprint is still pretty large |
| 21:17.25 | CosmicPenguin | yeah 2mb should be fine |
| 21:17.26 | prpplague | jbevren: we have some serious mem constraints |
| 21:17.46 | prpplague | CosmicPenguin: what do you figure the basic nano-x is gonna cost us in mem? |
| 21:18.07 | CosmicPenguin | half a meg |
| 21:18.13 | prpplague | lovely |
| 21:18.22 | prpplague | every tried running nano-x XIP ? |
| 21:18.28 | CosmicPenguin | no |
| 21:18.43 | prpplague | hehe |
| 21:19.18 | prpplague | CosmicPenguin: so you think about 2mb should give us some good functionality? |
| 21:19.47 | CosmicPenguin | it will give you possiblities |
| 21:19.54 | CosmicPenguin | I can't promise that great things will happen right out of the box |
| 21:20.05 | prpplague | hehe |
| 21:20.09 | jbevren | of course not. we only have 4M and no swap :) |
| 21:20.25 | jbevren | it'll be interesting seeing it work ;-) |
| 21:20.28 | CosmicPenguin | But you'll be able to play snaks |
| 21:20.29 | CosmicPenguin | es |
| 21:20.31 | CosmicPenguin | snakes even |
| 21:20.36 | CosmicPenguin | and really, isn't that why we program? |
| 21:21.08 | CosmicPenguin | oh, and mindsweeper rtoo |
| 21:21.16 | CosmicPenguin | with my kickass graphics |
| 21:21.22 | CosmicPenguin | because I'm such an artist |
| 21:21.23 | CosmicPenguin | not |
| 21:21.25 | jbevren | snakes rocks |
| 21:21.32 | jbevren | the version wher eyou have to eat the apple |
| 21:21.42 | CosmicPenguin | in this case, its just a dot |
| 21:21.49 | CosmicPenguin | but you can imagine its an apple if you want |
| 21:22.00 | jbevren | I made a snake application for the apple2 |
| 21:22.01 | gandhijee | lol |
| 21:22.06 | jbevren | whoppin 40x40 graphics on taht one |
| 21:22.24 | CosmicPenguin | could you wrap the snakes around the screen? |
| 21:22.26 | CosmicPenguin | because thats fun |
| 21:23.01 | jbevren | naw |
| 21:23.05 | jbevren | I was in middle school |
| 21:23.19 | CosmicPenguin | mmm... logo |
| 21:23.32 | CosmicPenguin | now there's an programming language for ya |
| 21:23.48 | prpplague | hehe |
| 21:24.57 | jbevren | logo can be pretty impressive |
| 21:24.59 | jbevren | its very forthlike |
| 21:27.46 | prpplague | i'd really like to get this up to 3mb of free space |
| 21:27.53 | prpplague | but i don't think thats gonna happen |
| 21:31.34 | prpplague | jbevren: i removed most of the stuff that won't be on the pixter from me test image |
| 21:31.49 | prpplague | do one more test for this evening |
| 21:37.36 | jbevren | cool |
| 21:37.56 | prpplague | MemTotal: 3888 kB |
| 21:37.56 | prpplague | MemFree: 516 kB |
| 21:38.00 | prpplague | a little better |
| 21:38.12 | jbevren | cool |
| 21:38.23 | prpplague | ok i'm headiing home |
| 21:38.24 | prpplague | later |
| 21:38.25 | jbevren | are you using an ext2 ramdisk or a cramfs? |
| 21:38.35 | prpplague | ext2 |
| 21:38.39 | jbevren | try a cramfs this evening |
| 21:38.40 | jbevren | :) |
| 21:38.40 | jbevren | see ya |
| 21:38.47 | prpplague | bbiaw |
| 21:39.50 | jbevren | ugh I -hate- single screen systems :) |
| 21:40.12 | davidc__ | heh, its hard for me to live without 2 |
| 21:40.15 | davidc__ | 3 is best ;) |
| 21:40.22 | jbevren | nod |
| 21:40.32 | jbevren | my new workstation will have two dual-dualchannel dvi cards |
| 21:40.33 | jbevren | mwahaha |
| 21:40.39 | jbevren | should arrive monday |
| 21:40.46 | jbevren | my old workstation lost its /usr partition |
| 21:40.56 | jbevren | any attempt to access it gives the system a flood of 0x51 0x04 errors |
| 21:41.19 | jbevren | I just need to get some OpenLDI to DVI adapters for my LCDs |
| 21:41.54 | davidc__ | 4 monitors? nice |
| 21:42.02 | jbevren | only 3 for now |
| 21:42.08 | jbevren | the OpenLDI adapters are $200ish |
| 21:42.16 | jbevren | but I (heart) my 1600SWs |
| 21:42.38 | jbevren | I wish I understood openLDI and DVI |
| 21:42.42 | jbevren | I'd attempt my own adapter heh |
| 21:43.01 | davidc__ | eh, should be doable with an FPGA + a DVI bridge |
| 21:43.04 | davidc__ | er, DVI phy |
| 21:43.07 | jbevren | nod |
| 21:43.16 | jbevren | philips sitll makes a LVDS phy too |
| 21:43.22 | jbevren | openLDI uses LVDS signaling |
| 21:43.42 | jbevren | dvi->phy->fpga->phy->openldi :) |
| 21:43.54 | jbevren | possibly with enough room for a 1600 pixel buffer |
| 21:45.10 | davidc__ | most fpgas will talk LVDS natively ;) |
| 21:45.20 | davidc__ | yeah, the buffer would be the only place you'd have issues |
| 21:45.24 | jbevren | heh very good point |
| 21:45.31 | jbevren | that said, buffering :) |
| 21:45.43 | davidc__ | you just need the TMDS phys for the DVI bits |
| 21:45.51 | jbevren | if I dont bother scaling (who cares, I'm going to use the native resolution), it should be ok |
| 21:45.51 | davidc__ | eh, if the protocols are similar enough, might be pretty easy |
| 21:46.05 | jbevren | openldi was made before dvi was |
| 21:46.11 | jbevren | shame it didnt become the standard heh |
| 21:46.23 | jbevren | the connector is easier to use, is probably more robust.. :) |
| 21:46.33 | davidc__ | and LVDS is nice and simp[le |
| 21:46.36 | jbevren | its a mini centronix plug like DFP connectors |
| 21:46.44 | davidc__ | though definitely higher EMI |
| 21:46.45 | jbevren | so whats with tmds signalling? |
| 21:47.04 | davidc__ | eh, its yet another signalling standard |
| 21:47.12 | jbevren | sigh :) |
| 21:48.15 | jbevren | I'll just look into PHYs :) |
| 21:48.30 | jbevren | center up anything taht's not a native resolution |
| 21:48.41 | jbevren | most video cards can scale their DVI output anyway |
| 22:02.39 | jbevren | out for the weekend |
| 22:02.40 | jbevren | see ya monday |
| 22:57.10 | *** join/#edev elvueno (n=elvueno@84-120-211-7.onocable.ono.com) |
| 23:12.24 | *** join/#edev redblue (i=star@ppp059.109-253-207.mtl.mt.videotron.ca) |
| 23:44.39 | *** join/#edev CosmicPenguin (n=nobody@svl-ext-proxy02.amd.com) |
