irclog2html for blob on 2001.10.09

00:45:12sammymorning all...
00:45:30prpplaguemorning sammy
00:47:18sammyprpplague : make rootfs ?
00:48:34prpplaguesammy: ya got that part down, not the problem is with busybox init
00:49:24prpplaguesammy: all i want is for init to drop me down to a ash prompt but can't seem to get it to do that
00:51:08sammyyou have tuxscreen ...
00:51:56prpplaguesammy: yes, but i'm working on another sa1100 platform right now
00:53:59sammyyeap , it should be ...
00:54:01sammyBZFlag : help to cal if I want to ordered 5 , how much should I pay ?
01:03:49BZFlagsammy: 2 2packs and 1 single.
07:37:00erikmmorning
07:37:04erikmhey WeirdArms
07:42:38WeirdArmshey erik
07:42:50WeirdArmsgot his Russell harshing for the week
07:45:11erikmWeirdArms: rmk is disappointed that you followed up on l-a-k
07:45:27WeirdArmsThat was a mistake
07:45:37WeirdArmsin judgement that is
07:45:55WeirdArmsI send an email saying I was sorry
07:46:16WeirdArmsit's proving tricking to impliment the idea that I was talking about though :(
07:46:33WeirdArmsI really wish he hadn't used the domains like he did
07:46:54erikmok, but he did
07:47:23WeirdArmsI know
07:47:28WeirdArmsI'm trying to work around it
07:47:34WeirdArmsit's just not working
07:48:53WeirdArmserikm: do you really want to come to Aust?
07:49:10erikmWeirdArms: that was something koen came up with yesterday
07:49:20erikmneeds coffee
07:49:23WeirdArmsoh?
07:51:27WeirdArmsWas my mail really bad?
07:51:32WeirdArmsI didn't think it would be a problem
07:54:15erikmI don't know about your mail, I don't have access to koen's mailbox :)
07:55:31WeirdArmsnot that one
07:55:36WeirdArmsthe l-a-k one
07:55:51WeirdArmsso koen wants to send you here?
07:55:55WeirdArmsis confused
07:58:09erikmI think rmk isn't too upset about the mail anymore, certainly if you appologised
07:58:26WeirdArmsok
07:58:28erikmno, koen doesn't want to send me down there
07:58:32WeirdArmsI don't want to annoy him
07:58:42WeirdArmsthen what was Koen talking about ?
07:58:52sammyhello erikm
07:58:58erikmkoen was just kidding. I thought you kewn him...
07:59:03erikmknew I mean
07:59:05erikmhello sammy
07:59:08WeirdArmsoh
07:59:11WeirdArmshehehe
07:59:25WeirdArmsI didn't pick up on it
07:59:35WeirdArmsis dense
07:59:44erikmbut /me has to update his resume anyway...
08:01:55erikmthinks about putting the SA1100 manuals on the LART site
08:03:06sammysorry I forget sent you a sa1100 manuals , and I think erikm do this is great ...
08:03:37WeirdArmserikm: could you, I can't believe they took them down
08:03:48sammyWeirdArm : still need it
08:04:20erikmreads the SA1100 manual legalese
08:07:56WeirdArmsI've got copies
08:07:57sammyin here , some sales tell me that Intel want let the sa-1100 layout ...
08:08:01WeirdArmsbut it's just stupid to take it down
08:08:07WeirdArmseven if they discontinue it
08:11:34RussI hate how documentation gets pulled with product
08:15:11sammyRuss : 24 ...
08:17:21sammyRuss :say something ...
08:17:32Russsomething
08:18:25sammya speech about birthday ...
08:18:35erikmRuss: you found something?
08:18:40Russyou're ahead of me
08:18:46Russits barely the ninth
08:22:32Russalmost there, and working on the girlfriend thing like you told me to
08:24:14sammystrange , why is good hanker always don't have girlfriend ?
08:24:39Russhmm....hanker?
08:24:58RussHanker Han"ker, v. i. imp. & p. p. Hankered; p. pr. & vb.
08:25:16Russ"1. To long (for) with a keen appetite and uneasiness; to have a vehement desire;" etc
08:26:15sammyI mean an expert guy
08:26:26erikmhacker
08:26:31Russah
08:26:33Russi see
08:26:38sammysorry ,,,,,,,,,,,,,poor
08:26:47sammyyou all know ....
08:26:58Russits ok, you should see my spanish
08:27:14erikmor "h4x0r" as "script writing juveniles" write it :)
08:50:39seletzhi all!
08:50:46erikmhi seletz
08:51:15seletzerikm: hi. shall we continue our discussion from yesterday?
08:51:35erikmok
08:51:51erikmthough I expect a SIGLUNCH soon...
08:52:19seletzerikm: SIGLUNCH has highest priority of course :)
08:53:30erikmyeah, we have to be quick or otherwise the students will take over the lunch room ;)
08:54:07seletzerikm: everywhere the same scenario :)
08:57:15erikmSIGLUNCH!
08:57:16seletzerikm: well, nevertheless, i found the source. its drivers/mtd/maps/sa1100-flash.c
08:57:24erikmseletz: oh, that one
08:57:25ed__SIG LUNCH CANNOT BE IGNORED
08:57:36seletzerikm: the structure is mtd_partition
08:57:38erikmseletz: LART doesn't use it
08:57:56seletzerikm_lunch: lart does not use mtd
08:58:11seletzugh. got preemted
08:58:15seletz:)
08:58:38seletz: sorry...
08:58:38seletzibot: whereis mtd_partition
08:58:53seletzibot: help me!
08:59:21seletzibot: help slashdot
08:59:44Slashdot - Updated 2001-10-09 07:57:41 | 100 Mbps Community Fiber Network: Howto | GPS Drawings | Slashback: Equivalence, Toilets, Hundredth | Matt Dillon On FreeBSD 5.0 VM System And More
08:59:44seletzibot: slashdot
09:00:17sammyseletz : you can see http://www.linux.mtd.infradead.org /
09:00:33seletzsammy: ?
09:00:51seletzsammy: ah, the mtd source.
09:01:26seletzsammy: well, now i am in my office & The Source Is With Me. So to say.
09:01:42sammyor http://www.linux-mtd.infradead.org/list-archive/
09:04:13sammyI am not sure, maybe it will help
09:07:38sammy_dinner ....
09:28:11seletzdoing cvs update
09:43:50seletzibot: help insult
09:44:21seletzibot: help main
09:45:00try http://linux-mtd.infradead.org/
09:45:00seletzibot: where is mtd?
09:45:38...but ibot is an infobot owned by Tangent with some assistance from BZFlag or stupid....
09:45:38seletzibot: you are a intelligent little brave bot!
09:46:40i haven't a clue, seletz
09:46:40seletzibot: where is erikm?
09:46:54seletzerikm is at lunch.
09:46:56bugger all, i dunno, seletz
09:46:56seletzibot: where is erikm?
09:47:17seletzstop playing
09:49:43erikm was last seen on #blob 52 minutes and 5 seconds ago, saying: seletz: LART doesn't use it [Tue Oct  9 09:57:38 2001]
09:49:43Russibot: seen erikm?
09:50:28seletzRuss: ahh. Handy little bot, he is.
09:50:35ed__plays the 'seen my-daddy' joke on ibot
09:50:44lxrbot was last seen on #blob 14 hours, 57 minutes and 13 seconds ago, saying: partition_map is Not used [Mon Oct  8 19:53:31 2001]
09:50:44Russibot: seen lxrbot?
09:51:31I haven't seen 'my-daddy', seletz
09:51:31seletzibot: seen my-daddy?
09:51:45I haven't seen 'daddy', seletz
09:51:45seletzibot: seen daddy?
09:52:00seletzhrmpf. got to work :))
09:52:05seletzcoding
10:00:48sammyhello lxrbot
10:01:51erikmhave a botsnack
10:01:57erikmlxrbot: have a botsnack
10:01:57lxrbotthanks erikm :)
10:02:02thanks erikm :)
10:02:02erikmibot: have a botsnack
10:02:02lxrbotthanks erikm :)
10:03:59seletzerikm: agreed to put POST routines for a platform in file PLATFORM.c, i.e. system3.c and call it in diag via initcall(), function system3_init()?
10:07:07erikmseletz: calls like what?
10:07:22erikmseletz: note that I will commit some GPIO code tonight
10:07:30erikmseletz: (it's needed for the PLEB)
10:07:41seletzerikm: well, use __initcall( system3_init, INIT_LEVEL_HTHERHARDWARE )
10:08:08erikmok, but what does it initialise?
10:08:42seletzerikm: nothing as of now. :)
10:09:39erikmwell, don't do it in that case
10:10:10seletzerikm: i thought of calling in system3_init() a simple version of a post routine.
10:10:10seletzerikm: always to run when diag loads and inits.
10:10:22erikmseletz: I don't think that's a good idea
10:10:47seletzerikm: you rather want to have commands, do you?
10:11:02seletzerikm: and do init => only init. i see.
10:11:17erikmseletz: something might be broken in the system. with a minimal system you can manually start individual tests to see exactly what fails and why
10:11:40seletzerikm: big point. Ok, then. i add commands.
10:11:49erikmseletz: sometimes the order of tests gives you a clue *why* it is broken
10:16:31erikmhehe: http://bbspot.com/News/2000/5/amd_moron.html
10:18:48seletzlooking
10:19:19sammysorry , is that step the same with make jffs and jffs2 ?
10:19:48sammyI mean image file
10:21:45erikmsammy: jffs is obsolete, use jffs2
10:22:57sammybut I can't make mkfs.jffs2 ?
10:23:36erikmoh? I had no problems with the latest code from CVS
10:24:05sammyin mtd/util/ is that type make mkfs.jffs2 ?
10:24:27sammyor what command should I type ?
10:24:51erikmmake -k did the trick for me
10:26:36seletzcoding
10:30:32sammyerikm : more detail , plz :-|
10:34:17erikmdccs mkfs.jffs2 to sammy
10:36:14seletzlxrbot: whereis SA1100_GPIOBASE?
10:36:15lxrbotSA1100_GPIOBASE is Not used
10:38:07erikmseletz: just use GPSR etc
10:38:14erikmseletz: it works from C
10:38:26seletzerikm: just found it.
10:38:33erikmseletz: and don't forget to #include <blob/sa1100.h>
10:38:36sammyis waiting receive but the transfer Rate = 0
10:39:06seletzerikm: ugh. new? looking ...
10:39:32erikmseletz: no, not new, renamed from <sa1100.h>
10:39:51erikmseletz: it's a wrapper to include asm/arch/SA-1100.h
10:40:22seletzahhhh. Why did i add those defines to memtest.h (was it memory.h)?
10:40:49seletzlike mdcnfg etc.
10:41:01erikmbecause in asseembly you can't use it
10:41:10erikmfor reasons I still don't know
10:41:21seletzok, i shut up.
10:41:23erikmand I don't even care to figure out :)
10:41:24seletz:)
10:41:26sammyerikm :it's faild ?
10:41:38sammy?
10:41:38erikm--- DCC SEND mkfs.jffs2 to sammy timed out - aborting.
10:41:56sammyis my problem ?
10:42:01erikmretries
10:42:15erikmsammy: if you are behind a firewall or a masquerading host: yes
10:42:40sammynope
10:48:37seletzerikm: may i add a sa1110.h like the sa1100.h in blob? Separate file?
10:48:49seletzs/1110/1111/g
10:49:02seletzi.e. companion chip
10:49:43sammycontinue...
10:49:53erikmseletz: ok with me
11:06:46seletzerikm: how to best include platform dep sources in configure.in for diag?
11:07:14erikmplatform dependent like what?
11:07:19seletzerikm: use AC_SUBST(DIAG_PLATFORM_SOURCES)
11:07:39seletzerikm: like test functions for my lcd display
11:07:57seletzerikm: diag/system3.c, as discussed above
11:08:08erikmseletz: the LCD is not platform dependent at all, all SA11x0 systems have that
11:08:56seletzerikm: well, yes, all sa11x0 _can_ have lcd displays.
11:09:14seletzerikm: all platforms are different in _what_ lcd they use.
11:09:22erikmseletz: I think you should make something like lcd.c with generic LCD code
11:09:27seletzerikm: tft cstn, stn ...
11:09:32erikmseletz: and put the platform specific stuff in lcd.h
11:10:01seletzerikm: ugh. that big. hmmm.
11:10:13erikmseletz: no, it's not big at all
11:10:31seletzerikm: well, i think about it just now ...
11:11:05erikmseletz: enabling the LCD controller is about 20 lines of code, I guess
11:11:36seletzerikm: 14 lines, actually :)
11:12:21seletzerikm: well, then. i add a lcd.c and throw in some generic functions.
11:12:38erikmseletz: so put the init values in a header file, and your code won't need *any* ifdefs
11:12:53erikmseletz: which makes it *very* easy to read
11:13:18seletzerikm: hmmm. linux kernel philosophy. hmm. yes.
11:14:03seletzerikm: i add LCCRxxx defines for a platform in a header file, base adresses for palette and video ram.
11:14:20seletzerikm: test pattern should be quite generic.
11:14:39seletzerikm: rows and columns quite so.
11:15:36seletzerikm: what about the different platform-dependent ways to enable the lcd controller? Each platform uses different CPLDs? Macros?
11:16:14erikmseletz: there are no platforms that do that
11:16:43seletzerikm: well, actually ouer platform does.
11:16:51erikmseletz: that's sick
11:17:02seletzerikm: and assabet does, too. (BCR)
11:17:16seletzerikm: i havent designed it :(
11:17:21erikmseletz: the assabet is sick as well, even the designer admits it
11:17:37erikmseletz: (but he had to do it because marketing insisted)
11:17:52erikmseletz: see for example the hack they used to clock the UDA1341
11:18:29seletzerikm: i know.
11:18:51seletzerikm: what does lart do?
11:19:26erikmseletz: LCD is directly connected to the correct GPIOs, sound chip uses an external oscillator
11:20:28seletzerikm: well, same as for our board (LCD thing.) but LCD has to be _enabled_ through setting a bit in a control register (CPLD).
11:21:32erikmyou mean, the backlight?
11:21:56seletzerikm: no. backlight is another bit.
11:22:18seletzlooking at schematics
11:24:51seletzerikm: well, we have one cpld xcr3128, which manages ethernet, irda, usb-host/slave, serial line switch, and, alas, an LCD_ON signal.
11:26:00seletzerikm: this CPLD has 4 registers, where operation is controlled. One register has a bit for this LCD_ON signal.
11:28:34sammysee you later , all
11:28:46seletzbye!
11:29:00sammychange place
11:38:48seletzdownload kernel
11:38:58erikmother window, please
11:39:12seletzerikm: :/
11:39:49erikmseletz: hmm, ok, what about this: make the LCD code generic, and add a system3.c file that will be conditionally compiled on system3
11:40:35seletzerikm: and system3.c defines some sort of lcd_enable()
11:40:48erikmyes
11:41:04seletzerikm: which is in turn called by lcd.c::lcd_test()
11:41:06erikmyes
11:41:47seletzerikm: we could also use some defines in lcd.h, say LCD_ENABLE=system3_lcd_en()
11:42:03seletzs/=/ /g
11:42:52erikmeven better:
11:43:09erikmmake an empty lcd_enable() and make it a weak symbol
11:43:43erikmnow if you want to override it for a platform, you just redefine it in system3.c
11:44:02seletzerikm: well, this would hide nicely all platform deps. similar platforms could use same functins. yes.
11:45:33seletzerikm: i'll take a look at rmk's sa1100-fb.c, there are all low-level enable/disable functions defined. we should use the same function set IMHO.
11:45:58seletzerikm: i.e. backlicht, contrast, brightness and so forth.
11:46:22erikmyes
11:47:19seletzok, i'll create a empty framework for evaluation :)
11:56:00erikmafk, bbl
11:56:32seletzpardon me?
12:00:23erikmAway From Keyboard, Be Back Later
12:01:41seletz:) that's nerd-speak, isnt it? :)
12:31:25sammyhello I am back ...
12:33:44seletzsammy: do you sleep at last _sometimes_ ?
13:06:10prpplaguehowdy howdy
13:06:21seletzhi!
13:07:28prpplagueanother day, another $.62 after taxes
13:10:28sammyseletz :no , I am not sleep , I just go home  :-)
13:17:43seletzerikm: just discovered that our sysop has pulled the plug from our M$$ Mail server due to nimda attacks :))
13:18:39seletzgetting coffee
13:18:39prpplaguewahoo, another m$ server down the drain
13:18:52seletz:^D
13:19:05seletzi'm immune, i run a decent os....
13:19:45prpplagueme too (except my wife)
13:20:54seletz3 weeks ago i asked them wether or not to shut down this M$ crap and set up a linux box, but they said no. Well, i guess they rather agreed....
13:23:27prpplagueseletz: we've been using a linux box for email for about two years now, managment keeps asking me why these stories about large corps email going down
13:23:52prpplagueseletz: they think its all hype since we've never been down
13:32:25erikmreturns
13:32:51erikmin our group we've been running sendmail for as long as I remember. never a single problem
13:37:49prpplagueerikm: just got a price back on that Linuette Box from HNT, they want $2300 for the dev kit
13:38:29prpplagueerikm: and anothe $2k for schematics and JTAG info
13:39:58erikmyeah, I read about the price somewhere on the web site
13:40:11erikmbut $2k for the schematics, that's rediculous
13:43:00prpplagueerikm: it looks nice, but i think i'll order a lart for that price
13:44:03seletzback
13:44:16seletzwoha! That's expensive!
13:44:33seletzwhat's in the package?
13:47:08prpplagueseletz: hmm, gold bullion, diamonds?
13:47:34seletz:)
13:47:44seletzerikm: how much is a creditLart
13:48:21erikmno idea
13:48:37erikmprotos are quite expensive, as allways
13:49:01seletzah, prototype yet.
13:49:31erikmwe do have prototypes, but jdb went ill, so they're not yet tested
13:49:35prpplaguei've got 8k to to spend before jan1
13:49:54seletzWell, i'll take it ... :)
13:51:31prpplagueseletz: only a tall good looking blonde or someone with a sa based product get this dough
13:52:25seletzprpplague: what proguct has it to be?
13:52:44seletzprpplague: sa1110+sa1111 enough?
13:53:26prpplagueseletz: what ya got?
13:53:30erikmdetects heavy board trading :)
13:55:06seletzprpplague: custom board, sa1110+sa1111, irda, smc91x96, usb host&client conn, lcd conn, big cpld, cf slot & slot 1 of pcmcia via test points. 32 mb ram, 16 mb flash.
13:55:29seletzprpplague: and designed to be low-power.
13:56:17prpplagueseletz: your design?
13:56:39seletzprpplague: well, design of my customer.
13:57:54prpplagueseletz: will we be able to by the board in qty later?
13:58:02seletzprpplague: they intend to use this kind of design for data logging purposes. an analog measurement hw will be connected via pcmcia slot 1.
13:59:02seletzprpplague: yes, they plan to produce it in numbers. And they wat to sell it as OEM dev boards.
13:59:32prpplagueseletz: what kinda $$ are talking for a dev board?
14:00:17seletzprpplague: no idea yet. I'm no marketeer tough.
14:01:17seletzprpplague: they pay me to bring it up running & keep it running. I will write drivers for the ad/IO hardware, for intelligent sensors and such.
14:01:55prpplagueseletz: tell your sales gurus to drop me an email dave123@abcsinc.com
14:02:14seletzprpplague: yup. will do.
14:03:59prpplagueseletz: sounds like you have a nice job
14:07:16seletzprpplague: well, getting paid for linux kernel hacking :^)))
14:07:43seletzprpplague: many things to learn and understand, tough.
14:08:51prpplagueseletz: same here
14:12:38seletzprpplague: just wrote an EMAIL to our CTO, hope that's enough.
14:13:04prpplagueseletz: thanks
14:32:15seletzerikm: when using __attribute__ ((weak)), is link-order important, yes? I.e. one sould first link the obj with weak symbols, then thoe one with "strong" ones, right?
14:32:44prpplagueerikm: is there a link on the lart pages for info on the creditlart?
14:35:59seletzerikm: ah, got the error: Don't define symbol weak in header file, but in source.
14:47:45erikmprpplague: no, not yet
14:47:52prpplagueerikm: np
14:50:16erikmwhat the heck, I just make the SA-1100 manuals available from the LART site
14:51:23ed__you're one crazy mutha !
14:51:43erikmed__: hey, the legal mumble doesn't say it's forbidden
14:52:40erikmand if intel is so crazy to pull them from their website, I'll just make them available. that will teach them
14:54:17ed__they pulled them?
14:54:20ed__blah
14:54:21erikmyes
14:54:40ed__hrmphf sillies
14:55:51prpplaguethats silly, you'd think they'd want people to find information on their products
14:57:26prpplaguelets see how many time i can flash my board in one day
14:58:39seletzerikm: Ok, I'll try to mail you the lcd stuff. Perhaps this won't work, so we have to dcc it, ok?
15:02:19erikmmail should work
15:05:50seletzerikm: mail sent. Well _your_ mail server is ok, i suppose, but is ours?
15:08:53erikmgets new mail
15:09:44seletzhappy
15:12:52seletzHmmmm. Is there a way to check wether or not i was thrown out of arm-kernel (besides bugging rmk)?
15:13:50erikmcheck with the mailing list archive
15:14:02erikmor better: with the GNU mailman pages
15:16:08erikmseletz: I suppose the lines with SerialOutputString("_system3_" ); in system3.c are debug messages?
15:16:35seletzerikm: sure. Its all a bit sketchy at the moment.
15:17:03erikmquestion why don't you make them void functions if they always return 0?
15:17:30seletzerikm: because they will return errors when i make them solid.
15:17:48erikmoh OK
15:18:17seletzon the phone
15:22:30seletzback
15:26:43seletzerikm: no complaints yet? :)
15:27:03erikmseletz: put lcd.h in blob
15:27:05prpplagueerikm: what are the top 3 reasons why you'd get sig4 errors on some arm apps?
15:27:20erikmprpplague: sig4 errors? never got those
15:27:27seletzmemory alignment?
15:27:34prpplagueerikm: Illegal instruction
15:28:29erikmprpplague: really, never got those.
15:28:39prpplagueerikm: hmm
15:28:45erikmprpplague: might be memory timing problems
15:29:24seletzerikm: isnt a user proces being killed by the kernel if it does unaligned memory acesses?
15:29:49erikmdepends
15:30:03erikmyou can ask the kernel to fixup the unaligned accesses
15:30:12erikmbut you'd better not and fix the code
15:30:40seletzCONFIG_ALIGNMENT_TRAP=y always on newer kernels. User Space is not affected by this, IIRC
15:31:23erikmthere was a trick to fixup userland as well, not sure if it is still in newer kernels or if rmk ripped it out
15:31:55erikmbut anyway. instructions are always aligned
15:31:59seletzIIRC the kernel now just kills those processes. This fix was considered _bad_ i think.
15:33:10seletzerikm: nevertheless, i do add the sources to cvs. agreed?
15:33:28erikmseletz: if you put lcd.h in include/blob
15:33:30seletzerikm: lcd.h -> include/blob of course
15:33:41seletz:)
15:33:42erikmseletz: and could you also make system3.c conditionally compiled?
15:34:19seletzerikm: yes. I do that via AC_SUBST() like the clock.c example, too?
15:34:37erikmyes
15:34:49seletzall right, then.
15:52:22erikmseletz: btw, you don't have to check in empty files anymore, I upgraded syncmail
15:52:34seletzerikm: ok.
16:46:35sammyerikm : ?
16:46:55erikmhmm?
16:47:12sammyafter I make image with jffs2 , how to put into LART ?
16:47:42erikmdownload and flash it as a ramdisk
16:47:44sammyjust like ramdisk or not ?
16:48:28sammyconfuse with the HOW TO ...
16:50:03sammyit say copy jffs.image /dev/mtd0 but I must turn on the lart with minicom , right ?
16:50:30erikm/dev/mtd0 == blob
16:50:35erikmyou want /dev/mtd2
16:50:53erikmbut you can also download the image in blob and just flash it
16:51:15sammylike ramdisk ?
16:51:21erikmyes
16:54:41sammyis try ...
16:55:25seletzerikm: cvs commit done.
16:55:38erikmseletz:
16:55:40erikmseletz: ok
16:56:28sammyamm... should I need to compiler kernel again with my root device going to be ?
16:58:30seletzsammy: i think you should say "noinitrd" as kernel param IIRC, because the kernel treats the jffs2 image as ramdisk.
16:58:51seletzsammy: "noinitrd root=/dev/mtdblockXXX" should do it.
16:59:15seletzsammy: where XXX is the partition number the kernel detects.
16:59:30seletzsammy: 2 if you use /dev/mtd2
17:00:07erikmboot console=ttySA0,9600 root=/dev/mtdblock2
17:00:21seletzsammy: btw, if you have a ramdis up and running with a correct set-up network,
17:01:03seletzsammy: then its way faster to flash under linux & download the image over ethernet.
17:02:42seletzerikm: did i break something? It works for me with build_all & make .f build_Makefile
17:03:26sammythat's my really hope to , but no one help me to layout the hardwave about ethernet on KSB ... :-(
17:04:17sammyso only I can do is step by step to go... +_+
17:05:02erikmseletz: /me cvs updates
17:05:17seletzsammy: same over here. I currently try to get basic HW things to run ...
17:05:25seletzerikm: i'll wait
17:05:55erikmsammy: next KSB design will have ethernet onboard instead of high end audio
17:06:07seletzsammy: lcd, then PS/2 keyboard, then smc91x96 ethernet chip
17:06:40seletzsammy: "high-end audio". hmmm. good for an mp3 player, eh?
17:06:46erikmsssh!
17:06:54seletz:)
17:06:58prpplagueerikm: i like the idea of KSB with ethernet
17:07:10erikmseletz: you forgot to commit blob/sa1111.h
17:08:03seletzugh.
17:08:06seletzone second ...
17:08:11sammyso I hope LART + KSB can work like some other development board ...and still work for it ..
17:08:51seletzerikm: done. sorry.
17:09:28erikmhmm, assabet breaks immediately
17:09:46erikmlooks
17:10:17seletzhmmm. I testet it, tough. :^/ do something, do something wrong and break something else :(
17:11:27erikmah, lcd.h is still empty
17:11:39seletz????
17:12:15seletzgrng....
17:12:35sammyadd it , be better then bootldr ...
17:12:55erikmsammy: we're already better than bootldr: blob is small
17:13:08seletzcommitted.
17:13:17erikmsammy: FYI: a bootldr binary is about 130k
17:13:55seletzgetting tired & angry
17:14:31sammyI know , do you remember someday before you tell me about blob , and at that time I am reading about bootldr ?
17:14:44sammythis is why I change to blob ...
17:15:09seletzsammy: for me it was worse: i changed from redboot.
17:15:09erikmlong long time ago, in a galaxy far far way, yes :)
17:16:07seletzerikm: is assabet still broken?
17:16:41erikmyes, but I fixed it
17:16:46seletz?
17:17:00erikmthe error in lcd.h is not nice
17:17:06erikmso I made it a warning
17:17:21sammybut I really hope that blob is more smaller , but still can do as much thing's as bootldr , even more then
17:17:47erikmsammy: blob will get more features in the near future
17:18:18erikmsammy: flash partition, parameters, ethernet, bootp, tftp
17:18:21seletzerikm: hmm, why did it threw an error? did you specify enable_all_features?
17:19:13seletzerikm: IMHO this should be an error, one could ruin LCDs with wrong settings.
17:19:36erikmhmm, true
17:19:50erikmlet me do it different
17:19:55seletzerikm: thats why i added a CONFIG_LCD_SUPPORT
17:20:01erikmhmm
17:21:22erikmrestores the original situation
17:21:42erikm.... but forgot the correct error message
17:21:51erikmmakes a new one:
17:21:54erikmCowardly refusing to break your LCD, please supply correct LCD parameters.
17:22:12seletzerikm: Thats polite ...
17:23:01sammyis that assabet always use Angelboot ?
17:23:03erikmI think lcd_test can be a static function
17:23:15seletzok
17:23:25seletzsammy: no.
17:23:37erikmsammy: assabet can be used with bootldr as well, but if you use assabet+neponset your only choice is Angel
17:23:53seletznot true.
17:24:16seletzassabet+neponset works fine with redboot&blob.
17:24:32erikmseletz: hmm, rmk told me that only angel works
17:24:43seletzyou only have to make sure that you use the right serial.
17:25:25seletzperhaps he thinks about static memory setup, which is done wrong with all bootloaders AFAIK.
17:26:08seletzi patched redboot to support assabet & it worked fine. All i did change was the MCSXXX settings and the SMROM settings.
17:26:24seletzso blob should be no problem at all.
17:26:48sammycan you sent the patch to me seletz ?
17:27:09erikmhmm, I tried blob on assabet last weekend and it hangs just after the memory configuration
17:27:22erikmso apparently there is something wrong over there
17:28:27seletzerikm: hmmm. I have to dig my settings out, if you're interested.
17:28:33erikmsure
17:28:47sammysome of my friend have a board just alike assabat , I think I can try  porting on it ...
17:28:58seletzerikm: i switched to our board 4 weeks ago.
17:29:15seletzerikm: before i used assabet for development.
17:29:21erikmsammy: problem is: almost all sa1110 boards look like assabet :)
17:30:19sammyyeah , this is also right ...
17:30:33seletzerikm: the problem with assabet is that there are number of versions of it.
17:30:49erikmyeah
17:30:49seletzevery version has its own quirks.
17:31:13erikmI have two of them, an early version and a recent version (january 2001)
17:31:23seletzi had to re-solder a wire on neponset to make it work.
17:31:42erikmoh, yeah, that bug
17:32:01seletzlooking for redboot patch
17:33:01erikmTimRiker: TimRiker has quit (Killed (services1.openprojects.net (Too many invalid passwords)) :)))
17:33:36erikmseletz: I got it
17:33:50erikmseletz: I'll define a couple of dummy LCD parameters
17:34:01erikmseletz: and I'll also define DONT_USE_LCD
17:34:17TimRikererikm: someone changed my password apparently.
17:34:34erikmseletz: the LCD test function will refuse to start when that's set
17:35:01seletzerikm: well, ok. kindof inversed my logic, eh? Well, thats ok too.
17:35:21erikmseletz: the idea of --enable-all-features is that it enables all features and still compiles
17:35:37seletzsammy: could send you a correctly configured tarball of ecos+redboot
17:35:43erikmseletz: hmm, wait
17:35:49erikmseletz: I'll do it in configure.in
17:36:10seletzerikm: well, i did it there ....
17:36:33seletzsammy: but it's 10MB files
17:36:44seletzsammy: tar gz.
17:37:02erikmseletz: yes, but I test for system3 over there, just like the the clock scaling will not be compiled for sa1110
17:38:02seletzerikm: wait a minute, i dont understand: lcd.c should be generic, right? so all platforms, right?
17:38:32erikmseletz: it's generic for platforms that have an LCD
17:39:03erikmseletz: but unfortunately lart doesn't it fails to compile with --enable-all-features
17:39:08seletzerikm: all platforms that --enable-lcd should add their defs in lcd.h all others not. and for those who dont, there is no error. Am i wrong?
17:39:12sammyseletz : ftp://211.75.21.252/  username:test  password:test
17:39:21erikmwarez!
17:39:25seletzerikm: ok. remove it from --enable-all-features
17:39:33erikmnah, I already fixed it
17:39:43seletzerikm: ok.
17:40:12seletztries ftp
17:40:31seletzbye
17:40:38seletzwrong window
17:41:47TimRikersammy: in other words: ftp://test:test@211.75.21.252/
17:42:13seletzsammy: ok 20 kb/s.
17:42:46seletzwell guys. i'm late already & my girlfriend is waiting :(
17:42:54sammythanx seletz
17:43:00seletzpreparing to go home
17:43:08erikmseletz: it builds
17:43:24seletzerikm: ok, at last. :(
17:44:06seletzsammy: kick me out from ftp when trasfer is done, ok?
17:44:07sammyTim : how can you know which way Iwant to got my tux ?
17:44:42seletzbye all!
17:44:45erikmbye
17:44:49sammybye.
17:44:59TimRikerglobal express mail is preferred, but airmail parcel post is cheaper.
18:08:33sammyerikm : about LCD device , what I should change is all in sa1100fb.c right ?
18:12:58sammyif that so , after I change , that I do is compiler the kernel again , and load it again ?
18:13:20erikmyes
18:13:53sammybut it's waste a lot of time in download and remake kernel ...
18:14:23sammydon't have another way ?
18:14:36erikmthere is no other way
18:14:53erikmthe memory for the framebuffer can only be allocated on boot time
18:15:19sammyyes , it have add the driver into blob :=)
18:17:29erikmslaps sammy
18:34:35sammywill go to bed after say night.
18:34:45sammynight all...
18:35:07prpplaguenight
19:15:57erikmgoes home
19:16:07prpplaguelater
19:16:10erikmlater

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