irclog2html for #elinux on 20051106

00:15.58*** join/#elinux Crofton (n=balister@66-207-66-26.black.dmt.ntelos.net)
00:40.02Genesisbonne nuit
02:01.42*** join/#elinux Soopaman (n=soopaman@HSE-Montreal-ppp143736.sympatico.ca)
02:07.18*** join/#elinux Soopaman_ (n=soopaman@CPE000d8891ac4c-CM000a739bc374.cpe.net.cable.rogers.com)
02:50.05*** join/#elinux hufnus (n=slonsiki@DSL135-071.labridge.com)
02:57.36*** join/#elinux we0 (n=weo@p5499E71A.dip.t-dialin.net)
03:18.41*** join/#elinux Sgt-Donan (n=Loutre@mto30-1-82-242-53-8.fbx.proxad.net)
04:22.46*** join/#elinux Soopaman (n=soopaman@HSE-Montreal-ppp143736.sympatico.ca)
07:47.32*** join/#elinux cdm (n=cdm@adsl-69-109-217-19.dsl.pltn13.pacbell.net)
09:41.08*** join/#elinux zinosat (n=Davide@adsl-ull-223-137.44-151.net24.it)
10:01.37*** join/#elinux CosmicPenguin (n=nobody@aus-ext-proxy02.amd.com)
11:00.42*** join/#elinux T0mW (n=Tom@24.238.70.125.res-cmts.sth.ptd.net)
11:53.00*** join/#elinux Soopaman_ (n=soopaman@CPE000d8891ac4c-CM000a739bc374.cpe.net.cable.rogers.com)
11:55.28*** join/#elinux darkschneider (n=gab@213-140-6-96.ip.fastwebnet.it)
12:03.24*** join/#elinux cantona (n=cantona@unaffiliated/cantona)
12:11.14*** join/#elinux Soopaman (n=soopaman@HSE-Montreal-ppp143736.sympatico.ca)
12:48.14*** join/#elinux vittorio (n=andreas@213.174.253.4)
12:53.34*** join/#elinux Genesis (n=genesis@dev.aimao.org)
12:55.15Genesisbonjour
12:58.41T0mWtoo early for the other guys to roll out of bed, I guess
13:53.22*** join/#elinux GPSFan (n=Ken@65.121.49.208)
14:30.50*** join/#elinux file[junkys] (n=jcolp@Toronto-HSE-ppp3863802.sympatico.ca)
15:00.18*** join/#elinux chouimat|Zzzz (n=dieu@r2351064.cidc.net)
15:01.32chouimatmorning
15:02.46T0mWchouimat: so so here, kind of murky out there today
15:02.56chouimatyup
15:03.03chouimatLOL www.dilbert.com
15:03.15T0mWhaven't been there to read the comix today
15:03.56chouimatT0mW: I start the reading comics and jokes website ...
15:04.06chouimatand then I go make a second pot of coffee
15:05.42T0mWchouimat: usually that is the first thing in the morning: I have a tab in firefox that brings up my morning reading.  
15:06.00T0mWchouimat: today I was eager to get back to work on the MMC DOS stuff
15:06.36T0mWfopen(), fread() work, now I need fwrite()
15:07.40T0mWas usual, puting something together (fwrite) is harder to do than taking it apart (fread)
15:09.03T0mWthis is driving me nuts, I hate warnings: warning: cast increases required alignment of target type
15:09.36T0mWI dunno why gcc won't re-cast something without bitching about it
15:09.41T0mWushort   * ptr = ((ushort *) &FileBuffers[Idx].Buffer);
15:09.44m4lu7Oh, i'm bloody confused here... want/need/desire -> build an initrd using the ext2 fs from buildroot as the rootfs...
15:10.16T0mWm4lu7: ask Erik, that is his stuff
15:10.57T0mWm4lu7: I don't know of any bootloaders that access the ext2 filesystem
15:10.59m4lu7my initrd builds, but is apparenlty looking at *this box*, the "host", as it's root, and consequently pulling it's kernel & modules into the initrd. dis' is not what i want
15:11.50m4lu7and i'm overcaffeinated, as usual, leaving my attention span about as long as a gnat's pecker
15:12.03chouimatT0mW: currently thinking of playing with freebsd6
15:12.38T0mWm4lu7: you've got something screwed in your dev environ, buildroot of Erik's is self contained and, if used out-of-the-box, never references anything outside of the build tree
15:13.29T0mWm4lu7: the only time I've run into problems like that was when I attempted to use a compiler other than the one inside the buildroot dev tree
15:14.00m4lu7T0mW, based on your comments, i'm most certain now that i'm not doing any of this correctly ;)
15:14.04T0mWchouimat: I'm not, I wasn't impressed with freebsd when I tried it a few years ago
15:15.02T0mWm4lu7: I don't do embedded dev on my main workstation anymore.  I have a P-III box with a 40G drive that I throw shit onto to do builds / dev work.
15:16.01T0mWm4lu7: that way I can save a few bash scripts that put the dev environ in place, wipe the box clean (new install of Mandrake), and re-run the scripts to build a new environ.
15:16.30T0mWm4lu7: it got too confusing with having various cross-builds on the same machine
15:16.37m4lu7yeh. i've contemplated setting up the target box w/dev tools, but it's the only one i've got onsite, and i don't want to screw it up.
15:16.58T0mWm4lu7: what's the target system?
15:17.03m4lu7the host machine (this) and the target (that) are close enough in arch anyway
15:17.15m4lu7T0mW, a wristwatch
15:17.17m4lu7;)
15:17.22T0mWyeah right
15:17.31T0mWX86?
15:17.32T0mWheh
15:17.38m4lu7yeah. x86
15:17.56T0mWheh, Dick Tracy's wrist radio
15:18.33m4lu7buildroot/busybox doesn't provide mkinitrd, as best i can tell, so chroot won't work (i'm (and this is where i believe my failure lies) using hosts mkinitrd
15:19.24chouimatT0mW: they freebsd 6 driver for my wireless cards (ralink rt2500 based) support the AP which the linux one doesn't support yet
15:20.16T0mWm4lu7: buildroot assumes that an external loader will take control of the system and put the kernel image into place in RAM.  There is no initrd for an embedded system as initrd relies on the presence of 1) CMOS blind loader of partition table boot record and 2) lilo or grub
15:20.52m4lu7T0mW, and that makes plenty o' sense, as my buildroot image works quite well sans-initrd
15:21.04m4lu7(dunno why i'm so fixed on initrd, anyway)
15:21.49T0mWm4lu7: having said that, there should not be any reason why you couldn't use the mkinitrd to build a bootable embedded image.  You may have to hack the source of mkinitrd to not use /lib/modules and such
15:22.32T0mWm4lu7: iirc, Linux initrd expects to be loaded into 0xc0000000 and executed from there
15:23.19T0mWchouimat: how about taking the BSD driver source and porting it to Linux, thus "poisoning" the kernel?
15:23.40T0mWchouimat: um
15:23.50chouimatT0mW: someone ported the usb driver to linux ... so I'm looking at porting the rest
15:23.57T0mWchouimat: prolly the driver is tied to the BSD stack?
15:24.04T0mWchouimat: oh
15:24.15chouimatT0mW: i'm reading the code
15:24.31T0mWwell, I'm going back to writing the delete_file for the DOS
15:53.12*** join/#elinux prpplague (n=billybob@200.50.80.24)
15:53.46*** mode/#elinux [+o prpplague] by ChanServ
16:03.30prpplagueT0mW`coding: ping
16:04.05T0mW`codingprpplague: morning Dave
16:04.20prpplagueT0mW`coding: afternoon here , hehe
16:04.28prpplagueT0mW`coding: whats the latest on your mmc/sd fat code?
16:04.55T0mW`codingprpplague: yeah, I dunno what time / day it is anyway.  I was wondering why they were playing church music on a Saturday...
16:05.02prpplaguehehe
16:05.45T0mW`codingprpplague: still working with it, right now I'm working on code to fwrite() and doing the TRUNC for the open()
16:05.59chouimathey prpplague
16:06.06prpplaguechouimat: howdy
16:06.18prpplagueT0mW`coding: was planning to look at it today some
16:06.41prpplagueT0mW`coding: i've been thinking long and hard on bootloader issues
16:06.58prpplagueT0mW`coding: every board we bring up, i have to repeat a ton of stuff for bootloaders
16:07.05T0mW`codingprpplague: look over the structures surrounding the fatlayer.c and mmc_hardware.c
16:07.12T0mW`codingprpplague: that will not change
16:07.30prpplagueT0mW`coding: ok thanks
16:07.53prpplagueT0mW`coding: i'm thinking of doing a very little multi-arch bootloader
16:08.00T0mW`codingprpplague: I'm still working in the internals of those two files.  But, the code I gave you can get you started until my confidence level is high enough that I can give you the good-code to read with
16:08.13prpplagueok
16:08.31T0mW`codingprpplague: that code has a bug where it sometimes won't read the partial of a last file sector.
16:08.39prpplagueT0mW`coding: i'm thinking of pulling some of the blob code and starting with that
16:08.58T0mW`codingprpplague: everything else works: fopen(), fclose() just a short read sometimes
16:09.19T0mW`codingprpplague: yeah, it is going to be under a BSD license
16:09.40T0mW`codingprpplague: "use it, hack it, but give me credit"
16:10.12T0mW`codingprpplague: I contacted that guy about the Circuit Cellar code...
16:10.30T0mW`codingprpplague: I wanted to know what license he had over it
16:14.31T0mW`codingprpplague: basically, he said that you could use it for hobby purposes but he wanted $2500 for an unlimited commercial license for the "real" fatLib code, not the stripped down thing that he sent to CC
16:15.10prpplagueT0mW`coding: yikes
16:15.23prpplagueT0mW`coding: i assume you didn't use any of it?
16:15.32T0mW`codingprpplague: no
16:15.37prpplagueT0mW`coding: ok good
16:15.48prpplagueT0mW`coding: hmm, bsd eh?
16:16.10T0mW`codingprpplague: I was going to, but looking at his write code I could see that he didn't TRUNCate files, he only created new ones.
16:16.10prpplagueT0mW`coding: i've got to do something about my bootloader issues
16:16.23prpplagueT0mW`coding: ahh
16:16.44T0mW`codingprpplague: my mmc_hardware.c is derived from Microchip sample mmc code for the PIC
16:16.55T0mW`codingprpplague: everything else is mine
16:16.55prpplagueT0mW`coding: ahh
16:18.18T0mW`codingprpplague: you know, DOS is not that hard to write.  I'm kind of pissed that I have to do it and that no one else is willing to contribute such a thing: greedy bastids
16:18.49prpplagueT0mW`coding: yea
16:19.07T0mW`codingback to debuggin
16:19.44prpplagueT0mW`coding: my problem right now is that uboot(imho) is too big, blob is rottting, and everything else takes too much time to get working
17:17.23*** join/#elinux joe_bleau (n=joe_blea@adsl-69-152-216-75.dsl.fyvlar.swbell.net)
17:17.33prpplaguejoe_bleau: hola
17:17.39joe_bleauHowdy
17:27.30T0mW`codingprpplague: crap.
17:28.09prpplagueT0mW`coding: i thought i smelled something
17:28.18T0mW`codingprpplague: this mmc stuff sux, darned thing won't read after a write.  unless I suck on the status register a few thousand times
17:28.39prpplagueT0mW`coding: that sounds odd
17:29.01T0mW`codingprpplague: it is the touristas that let their little doggies do their business at your place of business
17:29.57T0mW`codingprpplague: I do a write, read the status (it is 0x05, all is well), then when I try a CMD17, it gives me 0xff after issuing the command.  should have said 0x00 to say "okay, busy"
17:30.13prpplaguehmm
17:30.30T0mW`codingprpplague: the strange thing is that I can stop it, wait a few, then let it read and it still fails
17:30.52T0mW`codingprpplague: it appears that I am not finishing the write operation somehow
17:31.50T0mW`codingprpplague: this low-level interface to the MMC has been the real problem, coding the FAT driver is easy
17:37.59T0mW`foodprpplague: I wonder if I have to set the BLOCKSIZE or something during initialization?
17:38.12prpplagueT0mW`food: iirc no
17:38.24prpplagueT0mW`food: should be automatically assumed to be 512bytes
17:38.33T0mW`foodyeah
17:38.46T0mW`foodgot me puzzleD!
17:39.50prpplagueT0mW`food: i wish i had time to help, unfortunately, i have my own domesticated canine waste exciment to deal with
17:40.28T0mW`foodif I do something like: http://rafb.net/paste/results/SZo0Mw17.html
17:40.34T0mW`foodprpplague: it works
17:40.48T0mW`foodprpplague: slow, but it works if I suck on the status reg
17:43.32*** join/#elinux Soopaman_ (n=soopaman@CPE000d8891ac4c-CM000a739bc374.cpe.net.cable.rogers.com)
17:54.31*** join/#elinux George_ (i=george@host-83-146-61-45.bulldogdsl.com)
18:10.29*** join/#elinux file[laptop] (n=jcolp@Toronto-HSE-ppp3781532.sympatico.ca)
18:13.30*** join/#elinux George (i=george@kde/developer/gwright)
18:40.09*** join/#elinux Crofton (n=balister@hc6521c20.dhcp.vt.edu)
18:55.07T0mW`foodprpplague: ok
18:56.01T0mW`foodprpplague: it looks like when you get the first response after sending the block of data to be written does *NOT* the write operation "succeeded"
18:56.22T0mW`foodprpplague: the status code only means "yeah, I got all that, looks good"
18:57.06T0mW`foodprpplague: then it needs to go program the data, which takes a lot longer than I expected, lots of miliseconds later, it gives a status code for the entire operation.  
18:57.17T0mW`foodprpplague: THEN(!), I can read the drive
18:58.10T0mW`foodprpplague: chili almost done, kidney beans have 'cracked'
19:01.09jbevrenmmm
19:01.13jbevrenhow long has it been brewing? :)
19:01.58joe_bleauT0mw: which MMC command are you discussing?
19:03.31T0mW`foodjbevren: lets see, an hour maybe
19:03.59jbevrenhmm
19:04.01jbevrenneeds 3 more :)
19:04.14T0mW`foodjoe_bleau: CMD24
19:04.31T0mW`foodjoe_bleau: have you used CMD13 ? (SEND_STATUS)
19:04.31jbevrenso you ned to send cmd 0xff cmd 0xff to get mmc to work?
19:05.15joe_bleauI was reading back the 'data token' after the sector transfer, then I start polling for the non-busy to indicate that the write was complete
19:05.21T0mW`foodjbevren: no, just have to wait until it has finished programming is all.  I had thought that when you got the 1st status byte that it was finished
19:05.39jbevrenahh
19:05.44T0mW`foodall the 1st status tells you is that the command + data block you sent was ok
19:05.55jbevrenI want to toy with interfacing MMC into my own non-linux projects
19:06.06jbevrenahh
19:06.09jbevrenthen you wait for another 'ok'/
19:06.10jbevren?
19:06.21jbevrensend cmd, wait ok, wait busy, wait ok?
19:06.40T0mW`foodjoe_bleau: Yeah, I'm going to start the write operation, then leave.  But I'll need to check for IDLE before I start any new read / write ops.   that was why I was looking at CMD13
19:07.01joe_bleauHaven't played with CMD13.
19:07.08T0mW`foodjbevren: yeah, that works: cmd, get status, wait busy, return
19:07.32joe_bleauJust brought up some of my code...I was reading status after the transfer (no waiting, just reading), then waiting on the busy if status was ok
19:07.36jbevrenthen you retrieve your data.. ok :)
19:07.59T0mW`foodI don't want to take the time to wait around for the operation to complete as I can go on doing other tasks.  Just wait if the previous operation is still pending the outcome.
19:08.08joe_bleauSure, sounds good
19:08.12jbevrenah.
19:08.26jbevrendoes it matter if your driver is interruptable?
19:08.26joe_bleau(I was single threaded, so I didn't have anything better to do...)
19:08.59T0mW`foodjbevren: btw, if it cooks for 3 more hours it will be paste, not chili
19:09.20T0mW`foodjbevren: no, not interrupt driven, not needed really
19:09.22jbevrentom: true.  I dont add teh beans right away though
19:09.40T0mW`foodjbevren: I get in a hurry
19:09.57joe_bleauDid you try CMD25 (write multiple)?
19:10.02T0mW`foodNO!
19:10.08jbevrentom: yeah, the driver's pure PIO, but if its interruptable in the wait loop, the system wont get much impact hehe
19:10.16T0mW`foodI only have a single write operation most of the time
19:10.23jbevrenwrite_multi?
19:10.26jbevrenhmm
19:10.32T0mW`foodwrite single
19:10.36jbevrenI need to look at the commandset again heh
19:11.08joe_bleauI tried to optimize and clean up a few things last time I messed with my code, and managed to break the init!
19:11.30T0mW`foodjbevren: I may shift later into a write multiple scheme, but first, get the thing working and get me to understand the stumbling points.
19:11.45T0mW`foodjoe_bleau: btdt, several times
19:11.53jbevrenyeah
19:11.56jbevrensimple first.
19:12.16jbevrenmy target is an 8bit system, so I dont suspect a write_multiple would help a lot
19:12.28T0mW`foodjoe_bleau: at one point I had the mmc only working if I started the program from JTAG, if I started from reset / power on, it would fail.
19:12.41joe_bleauNow that sounds like a register config issue
19:13.14T0mW`foodyeah, the fun question was "is it the ARM or MMC registers that are trashed"?
19:13.24jbevrenhehe
19:13.24prpplagueT0mW`food: hey, btw, apex-1.3.9 has some mmc/sd code in it
19:13.33joe_bleauBTW, you are using the SPI command set, right?  (Not the MMC set.)
19:14.17T0mW`foodwell, working at the moment, time for some TV: breaktime!
19:14.30prpplagueT0mW`food: noooooooo
19:14.38prpplagueT0mW`food: get back here you old fart
19:14.45T0mW`foodprpplague: huh?
19:15.07T0mW`foodprpplague: I think I've got / had the apex source
19:15.08prpplagueT0mW`food: hehe, just pestering you
19:15.18joe_bleauGet one of those LCD panel that do picture in picture...watch TV in the corner while working.
19:15.26prpplagueT0mW`food: its only in the 1.3.8 or 1.3.9 release
19:15.43prpplaguei used to do that with my tv capture card
19:15.43T0mW`foodprpplague: I couldn't get apex to boot nothin'
19:15.56prpplagueT0mW`food: yea its a beoych to get working
19:16.07prpplagueT0mW`food: but i was mainly saying you could look at the mmc code
19:16.30prpplagueT0mW`food: its all working now with all of the lpd boards
19:17.03prpplagueT0mW`food: he's also added alot of debug commands too
19:17.36T0mW`foodperfect movie for "after coding" relaxation: Mad Max 2 - The Road Warrior
19:17.42prpplaguehehe
19:17.57prpplagueT0mW`food: i'm more for blade runner
19:18.04prpplagueor avalon
19:18.09T0mW`foodthat's the one with the gasoline plant in the desert?
19:18.23CroftonLeave the fuel!
19:18.29T0mW`foodprpplague: blade is ok, it is not a blade day though: mad max
19:18.50T0mW`foodprpplague: I wanna kill!
19:18.57T0mW`foodprpplague: I wanna hurt people!
19:19.03T0mW`foodprpplague: I want Mel!
19:21.29prpplaguehehe
19:22.49prpplaguedon't suppose anyone has experience with solar voltaic cells?
19:36.45*** join/#elinux ionitro (n=Miranda@198.2.203.62.cust.bluewin.ch)
19:46.46*** join/#elinux chouimat|laptop (n=root@r2351064.cidc.net)
19:53.29*** join/#elinux Soopaman (n=soopaman@HSE-Montreal-ppp143736.sympatico.ca)
20:28.13*** join/#elinux file[laptop] (n=jcolp@Toronto-HSE-ppp3863658.sympatico.ca)
20:46.55*** join/#elinux GPSFan (n=Ken@65.121.49.208)
21:04.06*** join/#elinux Soopaman_ (n=soopaman@CPE000d8891ac4c-CM000a739bc374.cpe.net.cable.rogers.com)
21:22.02*** join/#elinux Soopaman (n=soopaman@HSE-Montreal-ppp143736.sympatico.ca)
21:40.09*** join/#elinux Soopaman_ (n=soopaman@CPE000d8891ac4c-CM000a739bc374.cpe.net.cable.rogers.com)
22:08.49*** join/#elinux Soopaman_ (n=soopaman@HSE-Montreal-ppp143736.sympatico.ca)
22:34.51*** join/#elinux chewy (n=dieu@r2351064.cidc.net)
22:49.55*** join/#elinux Soopaman_ (n=soopaman@CPE000d8891ac4c-CM000a739bc374.cpe.net.cable.rogers.com)
23:28.30*** join/#elinux Soopaman_ (n=soopaman@HSE-Montreal-ppp143736.sympatico.ca)

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.