**** BEGIN LOGGING AT Mon Jul 23 02:59:56 2007
Jul 23 03:23:42 <CIA-12> nbd * r8119 /trunk/ (include/netfilter.mk package/kernel/modules/netfilter.mk): do dynamic kernel config changes for netfilter as well
Jul 23 03:59:52 <sn9> crazy_imp: openwrt does not have the sitecom flash mapper
Jul 23 04:00:33 <crazy_imp> sn9: but we could add it, right?
Jul 23 04:01:39 <sn9> i'm trying to figure out how that could be done without creating a new port for every board of the same platform
Jul 23 04:05:26 <crazy_imp> sn9: something like "#ifdef foo" in the code and then set foo via the profile, if it's possible
Jul 23 04:05:57 <sn9> but where would "foo" be defined?
Jul 23 04:06:05 <sn9> nvm,
Jul 23 04:16:37 <crazy_imp> sn9: maybe you can set them with something like "MAKE_FLAGS += ..." in the profile? but dunno if you can set such things there
Jul 23 04:16:54 <sn9> i wasn't thinking
Jul 23 07:18:13 <CIA-12> rwhitby * r8120 /trunk/target/linux/ixp4xx-2.6/config/default: Enabled IXP4xx beeper
Jul 23 07:23:38 <sn9> rwhitby: do you by any chance have a 2MB rdc device?
Jul 23 08:25:07 <CIA-12> rwhitby * r8121 /trunk/package/kernel/modules/block.mk: Fixed required KCONFIG symbols for block IDE modules
Jul 23 09:02:55 <[florian]> sn9: ping ?
Jul 23 09:03:02 <sn9> ack
Jul 23 09:07:49 <sn9> [florian]: is the ar525w your only rdc device?
Jul 23 09:09:03 <[florian]> sn9: absolutely
Jul 23 09:09:20 <[florian]> sn9: you have probably noticed that the .21 upgrade break it (again)
Jul 23 09:09:27 <sn9> yes
Jul 23 09:09:49 <[florian]> sn9: I read that peter h. anvin is rewritting the x86 code, so that I will probably wait for his changes before digging again
Jul 23 09:10:17 <sn9> crazy_imp asked if testing on a sitecom wl-153 would help, and i looked into it
Jul 23 09:10:22 <[florian]> sn9: I suspect 1) the lzma decompressor to be broken 2) a changes in the time code that was not forward ported
Jul 23 09:11:42 <sn9> do you know of any ETA on the x86 rewrite?
Jul 23 09:13:23 <[florian]> sn9: apparently, it has already been merged to linus' tree : http://kerneltrap.org/node/11754
Jul 23 09:20:33 <[florian]> sn9: about the sitecom, it is not really a model-specific issue, rather the rdc processor which is crappy by nature
Jul 23 09:21:08 <[florian]> sn9: also, datasheet says it has a cmos, but I really doubt it
Jul 23 09:21:26 <sn9> i'm temporarily forking r7314 for continued development and will submit a patch shortly. it should cleanly apply to the trunk
Jul 23 09:22:51 <sn9> afterwards, i will try including git-newsetup.patch with the new kernel
Jul 23 09:23:41 <sn9> the kernel is not a model-specific issue, no. but the current build scripts might be
Jul 23 09:25:10 <[florian]> you think a stock kernel could run ?
Jul 23 09:26:01 <sn9> the flash mapping on the wl-153 is apparently strange, and i would need to see a dump of the entire flash chip retrieved via jtag to adjust the build scripts properly
Jul 23 09:26:38 <sn9> probably not, but it can be a starting point
Jul 23 09:27:30 <CIA-12> juhosg * r8122 / (15 files in 5 dirs):
Jul 23 09:27:30 <CIA-12> [adm5120] cleanup patches
Jul 23 09:27:30 <CIA-12>  * nuke patches for 2.6.21
Jul 23 09:27:30 <CIA-12>  * refresh patches for 2.6.22
Jul 23 09:29:35 <[florian]> sn9: I know there was quite a lot of problems with the kernel config first, then I am very unsure about the relationship between size, cmos ...
Jul 23 09:31:54 <sn9> i also hope to contribute to ports for other devices: ar625w (marvell arm), wrtp54g (ar7), etc.
Jul 23 09:32:16 <[florian]> feel free to do so :)
Jul 23 09:32:31 <sn9> one at a time, at least for now
Jul 23 10:07:36 <CIA-12> juhosg * r8123 /trunk/package/: ignore dirs created with 'make package/symlinks'
Jul 23 10:32:38 <CIA-12> juhosg * r8124 /tags/kamikaze_7.07/package/haserl/: haserl in /packages/utils since [8006]
Jul 23 12:03:42 <glp> noz: ping
Jul 23 12:12:11 <CIA-12> florian * r8125 / (4 files in 4 dirs): Allow squashfs images for au1000, use the automatic rootfs split
Jul 23 12:26:49 <sn9> blogic: i'm wondering why the TuxCase, which is plastic, shipped in an anti-static bag...
Jul 23 12:38:40 <blogic> sn9: :-) why u asking me ? :-)
Jul 23 12:39:17 <blogic> sn9: its a guy called roberto fusali or so, that does the shipping from what i know
Jul 23 12:41:02 * sn9 tries to take his foot out of his mouth...
Jul 23 12:41:40 <blogic> :P no worries
Jul 23 12:41:54 <blogic> but its a fair question
Jul 23 12:43:12 <sn9> i was just curious
Jul 23 12:43:31 <sn9> although it may have been imagecraft that put it in the bag
Jul 23 12:44:39 <CIA-12> florian * r8126 / (2 files in 2 dirs): Make mac80211 compile even if CONFIG_LEDS_TRIGGERS is not defined in the kernel
Jul 23 12:46:51 <blogic> sn9: usa ?
Jul 23 12:46:58 <sn9> right
Jul 23 12:48:03 <blogic> ok, dunno, i normally get my samples directly from italy and they tend to be dev samples send from the fabrication plant directly o i have no idea of the "normal" method of packaging
Jul 23 12:49:29 <sn9> i found it a bit frustrating that both imagecraft and microcontrollershop require freight shipment, even though both are within 100km of here
Jul 23 12:57:24 <CIA-12> florian * r8127 / (2 files in 2 dirs): Add supporg for nls-iso8859-2 (#2123)
Jul 23 13:33:53 <CIA-12> juhosg * r8128 / (5 files in 5 dirs):
Jul 23 13:33:53 <CIA-12> [adm5120] board specific fixes
Jul 23 13:33:53 <CIA-12>  * fix detection of RB-133C, thanks to Christophe Lucas
Jul 23 13:33:53 <CIA-12>  * fix detection of RB-153
Jul 23 13:33:53 <CIA-12>  * add support for WP54G-WRT to the lzma-loader
Jul 23 13:33:53 <CIA-12>  * fix adm5120_mtd.c for WP54G-WRT in 7.07
Jul 23 13:33:55 <CIA-12>  * now it's ready for the release of 7.07
Jul 23 13:34:28 <sn9> crazy_imp: do you have a jtag connection to your sitecom?
Jul 23 13:39:30 <crazy_imp> sn9: dunno, maybe the row of solder points named JP1 is jtag, i could check that later if they are connected to the r3211
Jul 23 13:39:53 <sn9> the flash mapping on the wl-153 is apparently strange, and i would need to see a dump of the entire flash chip retrieved via jtag to adjust the build scripts properly
Jul 23 13:40:21 <sn9> it might not need a separate driver; the existing one might be suitable
Jul 23 13:41:34 <crazy_imp> ok, i'll try it :)
Jul 23 13:46:02 <CIA-12> juhosg * r8129 /trunk/target/linux/adm5120-2.6/files/drivers/mtd/maps/adm5120_mtd.c: [adm5120] adm5120_mtd.c is no longer in use
Jul 23 14:04:36 <CIA-12> nbd * r8130 /packages/utils/mksh/ (Makefile patches/100-dot_mkshrc): update mksh to R29g
Jul 23 14:10:29 <crazy_imp> sn9: i have tdo, tdi, tck and tms, all in jp1 :)
Jul 23 14:12:09 <sn9> great
Jul 23 14:12:42 <sn9> get me a dump of the entire 2MB flash -- assuming it's the stock firmware
Jul 23 14:16:13 <crazy_imp> sn9: which jtag package should i use? jtag-0.6-cvs-20051228 or jtag-brecis-ok.zip ?
Jul 23 14:16:28 <sn9> which cable do you have?
Jul 23 14:16:57 <CIA-12> nbd * r8131 / (4 files in 4 dirs): disable reserved blocks for checkpointing in yaffs2 on rb532 and adm5120
Jul 23 14:17:24 <crazy_imp> sn9: wiggler without the srst_n line
Jul 23 14:20:07 <sn9> the brecis patch seems preferable, but i have yet to try it
Jul 23 14:21:48 <crazy_imp> [florian]: which jtag tool should i use for rdc?
Jul 23 14:25:21 <[florian]> crazy_imp: I do not know,  openwince is ok
Jul 23 14:26:22 <crazy_imp> [florian]: well, both i listed abouve are openwince (at least i think they are :D), which one do you use?
Jul 23 14:26:46 <[florian]> crazy_imp: I have not been able to use jtag on the rdc, there was no need for it
Jul 23 14:28:20 <crazy_imp> [florian]: hehe, ok :). think i'll try the brecis one
Jul 23 14:46:51 <[florian]> crazy_imp: how is it going ?
Jul 23 14:48:00 <crazy_imp> [florian]: installing the openwince includes right now
Jul 23 14:52:25 <nbd> i think i might start maintaining a fork of openwince soon
Jul 23 14:52:28 <nbd> since the project is dead
Jul 23 14:52:41 <nbd> i have already written atheros SoC spiflash support for it
Jul 23 14:53:02 <sn9> nbd: #openjtag
Jul 23 14:53:18 <nbd> sn9: is there some activity again?
Jul 23 14:53:33 <sn9> hard to tell so far
Jul 23 14:54:15 <crazy_imp> hmmm, bad, the jtag stuff fails to build :/
Jul 23 14:54:35 <crazy_imp> afk, will try to find out whats going wrong later
Jul 23 14:55:56 * loswillios forks nbd to let him maintain more projects
Jul 23 14:59:39 <thepeople_work> while [ $stufftobedone = true ]; do
Jul 23 14:59:39 <thepeople_work>        fork nbd
Jul 23 14:59:39 <thepeople_work> done
Jul 23 14:59:56 <nbd> fork nbd &
Jul 23 14:59:58 <nbd> ;)
Jul 23 15:00:24 <nbd> my $barf = NBD::ForkBomb;
Jul 23 15:02:18 <sn9> [florian]: http://homepage.mac.com/danielg4/rdc.patch
Jul 23 15:02:51 <sn9> see how cleanly it might apply to the trunk
Jul 23 15:03:09 <[florian]> sn9: thanks
Jul 23 15:03:15 <[florian]> sn9: will give it a shot tonight
Jul 23 15:03:27 <sn9> this is the .19 fork
Jul 23 15:04:31 <sn9> if it cleanly applies, might as well commit it first
Jul 23 15:04:37 <[florian]> sn9: any reason you choose to put theo r6040 enet driver as module ?
Jul 23 15:04:43 <sn9> YES
Jul 23 15:05:10 <sn9> it has to load after rt61.ko (see code)
Jul 23 15:05:34 <[florian]> what if it does not ?
Jul 23 15:05:52 <sn9> the r6040 has no eeprom
Jul 23 15:06:33 <sn9> it's supposed to piggyback off the MAC of the wifi
Jul 23 15:07:15 <[florian]> oh I get it, it was working fine in-kernel with me
Jul 23 15:07:42 <sn9> ra0 has to be brought up at least once before the r6040 insmod, too
Jul 23 15:07:43 <[florian]> can't we ask r6040 to read the mac address without rt61 being loaded ?
Jul 23 15:07:55 <sn9> how?
Jul 23 15:10:10 <[florian]> by reading in the ralink chip resource (if possible)
Jul 23 15:11:16 <sn9> then what if an rdc device uses non-ralink wifi?
Jul 23 15:12:03 <[florian]> is there any ?
Jul 23 15:12:20 <sn9> probably, and if not, there will be
Jul 23 15:14:47 <sn9> btw, any idea why with the stock firmware, eth0 is lan and eth1 is wan, but with rdc-2.6, it's vice-versa?
Jul 23 15:15:14 <[florian]> absolutely no idea, I guess the stock mapping is more intuitive
Jul 23 15:15:43 <sn9> where did you originally take the r6040 driver from?
Jul 23 15:16:26 <crazy_imp> re. what will happen if i remove the ralink card? it wont do anything?
Jul 23 15:16:39 <sn9> crazy_imp: no idea
Jul 23 15:16:51 <[florian]> from the airlink ar525w tarbal
Jul 23 15:17:11 <sn9> hmm, maybe it's a 2.4/2.6 difference, then
Jul 23 15:18:11 <[florian]> I guess so
Jul 23 15:18:15 <sn9> the commit log mentions something about "NAPI polling"
Jul 23 15:18:24 <[florian]> absolutely
Jul 23 15:18:33 <[florian]> but I think the order the ports get registered is different
Jul 23 15:18:48 <sn9> so there were two versions of the driver?
Jul 23 15:21:12 <[florian]> yes, one for 2.4, another for 2.6
Jul 23 15:22:12 <[florian]> I do not exactly remember where I did get the 2.6 version tough
Jul 23 15:23:00 <sn9> https://dev.openwrt.org/changeset?new=trunk%2Ftarget%2Flinux%2Frdc-2.6%406805&old=trunk%2Ftarget%2Flinux%2Frdc-2.6%406802
Jul 23 15:23:07 <sn9> take a look at the very bottom
Jul 23 15:23:36 <sn9> i think that explains it, and answers what to do about the MAC
Jul 23 15:30:17 <[florian]> some code was dropped
Jul 23 15:32:20 <sn9> yeah, code to handle the switch chip, presumably not used in whatever device you got the 2.6 code from
Jul 23 15:32:30 <[florian]> I guess so
Jul 23 15:32:41 <sn9> it should be ported back
Jul 23 15:33:11 <[florian]> yes, but as pasrt of the kmod-switch preferably
Jul 23 15:34:28 <[florian]> I will make some tests toight
Jul 23 15:34:32 <[florian]> tonight
Jul 23 15:34:37 <[florian]> sorry for making so many typos
Jul 23 15:37:42 <[florian]> I wonder if we cannot patch/add a new phy driver for the r6040
Jul 23 15:38:27 <juhosg> [florian]: whats up with the french cows? ;)
Jul 23 15:39:54 <[florian]> juhosg: as you can see, I have been caught by one
Jul 23 15:40:51 <juhosg> [florian]: :P
Jul 23 15:43:04 <crazy_imp> ?
Jul 23 15:44:55 <[florian]> I use to say that I sometimes type like a french cow
Jul 23 15:45:37 <crazy_imp> hehe
Jul 23 15:46:16 <sn9> mou?
Jul 23 15:46:25 <[florian]> meuh here
Jul 23 15:46:32 <sn9> ah
Jul 23 15:50:54 <[florian]> sn9: will you be online tonight ?
Jul 23 15:51:57 <sn9> i will leave in less than three hours, but i expect to be back later (night, your time)
Jul 23 15:52:45 <[florian]> thanks
Jul 23 15:53:11 <sn9> oops, make that less than two hours
Jul 23 16:14:20 <sn9> [florian]: update: i will be back evening, your time. i leave in 90 min
Jul 23 16:15:59 <crazy_imp> http://pastebin.ca/630935
Jul 23 16:17:33 <sn9> try the brecis
Jul 23 16:18:58 <crazy_imp> same there i think, but i'll try it again
Jul 23 16:21:00 <sn9> also apply this: http://www.dm.ufscar.br/profs/waldeck/jtag-brecis-to-waldeck.diff
Jul 23 16:24:42 <crazy_imp> ok
Jul 23 16:28:17 <crazy_imp> sn9: http://pastebin.ca/630955
Jul 23 16:31:08 <sn9> is that with the waldeck patch?
Jul 23 16:31:33 <crazy_imp> yes
Jul 23 16:32:09 <sn9> what distro? gcc version? gettext version?
Jul 23 16:35:48 <sn9> crazy_imp: wait, why do you have openwince both in ~ and in /usr/local?
Jul 23 16:40:40 <crazy_imp> sn9: gentoo, gcc 4.2.0, gettext 0.16.1. i installed the openwince includes according to the readme, the ./include-0.4.2 dir was within the zipfile
Jul 23 16:41:30 <sn9> before or after patching?
Jul 23 16:42:32 <crazy_imp> before
Jul 23 16:42:50 <sn9> ok, that makes no sense
Jul 23 16:43:19 <sn9> are the includes patched?
Jul 23 16:43:58 <crazy_imp> the includes i've installed on my gentoo or the ones in the brecis dir?
Jul 23 16:44:08 <sn9> installed
Jul 23 16:44:19 <crazy_imp> sn9: no
Jul 23 16:44:52 <crazy_imp> sn9: only http://pastebin.ca/630972
Jul 23 17:21:06 <crazy_imp> sn9: ideas?
Jul 23 17:21:48 <sn9> trying to build now
Jul 23 17:22:12 <crazy_imp> ack
Jul 23 17:36:04 <CIA-12> pavlov * r8132 /packages/libs/net-snmp/Makefile: allow special types
Jul 23 17:37:17 <crazy_imp> it would be nice to have a jtag package in the toolchain :)
Jul 23 17:39:21 <CIA-12> pavlov * r8133 /packages/libs/yapsnmp/ (. Makefile patches/ patches/000-compile_fix.patch): yapsnmp library for python
Jul 23 18:35:30 <CIA-12> florian * r8134 /trunk/target/linux/rdc-2.6/files/drivers/net/r6040.c: Add back the port priority for r6040
Jul 23 18:36:27 <[florian]> sn9: I am back online
Jul 23 19:15:39 <[mbm]> anyone ever hear of an airlink nas 350? (anas350)
Jul 23 19:16:21 <[mbm]> x86 compatible chipset, ide, sata and ethernet
Jul 23 19:16:31 <[mbm]> runs a 2.4.28 kernel and busybox
Jul 23 19:16:41 <crazy_imp> nice
Jul 23 19:17:21 <[mbm]> linux seems to think it's a 486 cyrix
Jul 23 19:18:03 <[mbm]> plugged in a linux drive with a symlink to /
Jul 23 19:18:11 <[mbm]> and I'm poking aorund the filesystem
Jul 23 19:19:12 <[mbm]> CPU: Cyrix Cx486SLC
Jul 23 19:19:26 <[mbm]> r6040: RDC R6040 RX NAPI net driver, version 0.15 (26Sep2006)
Jul 23 19:19:29 <crazy_imp> hehe, thats quite a big hole in the system
Jul 23 19:19:56 <[mbm]> yeah, even funnier is that the root filesystem is writable
Jul 23 19:20:09 <[mbm]> some sort of ext2 ramdisk
Jul 23 19:21:42 <crazy_imp> [mbm]: it's an rdc device i guess, right? my wl-153 prints "CPU: Cyrix Cx486SLC" while booting too :)
Jul 23 19:21:44 <[mbm]> 4M flash, 32M ram
Jul 23 19:22:19 <[mbm]> chip is labeled "AMRISC 20000"
Jul 23 19:23:06 <[mbm]> looks like I found a serial port
Jul 23 19:23:43 <crazy_imp> if you see the loginutil from the sitecom i'll laught :D
Jul 23 19:24:14 <[mbm]> buncha stuff labeled with amit_* prefixes
Jul 23 19:25:06 <[mbm]> amit seems to be some tiwanese electronics shop
Jul 23 19:25:15 <[mbm]> www.amit.com.tw
Jul 23 19:31:43 <crazy_imp> [mbm]: their mgb100 looks like a fine replacement for the wl-hdd from asus
Jul 23 19:37:18 <[mbm]> mumble... doesn't seem to be any bootloader, and my serial port is only working in one direction
Jul 23 19:38:30 <Bartman007> device --> computer I assume? :-)
Jul 23 19:38:37 <[mbm]> heh yeah
Jul 23 19:41:04 <[mbm]> ah, there we go
Jul 23 19:42:12 <[mbm]> BusyBox v1.00-rc2 (2007.01.04-09:51+0000) multi-call binary
Jul 23 19:48:32 <crazy_imp> [mbm]: ever heard something about the zetera nas stuff?
Jul 23 19:51:46 <CIA-12> juhosg * r8135 /trunk/target/linux/adm5120-2.6/files/drivers/mtd/maps/adm5120-flash.c: [adm5120] cosmetic changes in flash driver
Jul 23 19:53:39 <[mbm]> crazy_imp: nope
Jul 23 19:55:46 <[mbm]> hmm lzma'd ramdisk .. cute
Jul 23 19:56:41 <crazy_imp> [mbm]: just asking because i have a nas here (sc101 from netgear) running vxworks and there i no ide controller on the board (maybe included in the cpu, tnetv2005). zetera lists it as licensed product and there is some zetera stuff visible while booting
Jul 23 19:56:47 <[mbm]> hmm some crazy stuff at the end of flash
Jul 23 19:56:49 <[mbm]> KBD END
Jul 23 19:56:50 <[mbm]> MESSAGE BEGIN
Jul 23 19:56:50 <[mbm]> mnBIOS V0.042 (C)2004 RDC inc with IDE
Jul 23 19:56:50 <[mbm]> KB BASE MEMORY
Jul 23 19:57:10 <[mbm]> didn't see the rdc logo on the chips
Jul 23 19:57:22 <crazy_imp> but i don't know how it's accessing the drives
Jul 23 19:58:00 <crazy_imp> [mbm]: cheap unlicensed clone of the rdc chips? :D
Jul 23 20:07:33 <sn9> [mbm]: your device is rdc, definitely. and with that much ram and flash, the port needs to support it too
Jul 23 20:07:55 <sn9> where'd you get it?
Jul 23 20:11:00 <[florian]> sn9: it can also be amrisc2000, which is really similar to rdc
Jul 23 20:11:22 <sn9> any idea what the differences are?
Jul 23 20:12:18 <[florian]> nope, just I read other the internet that they are two chips for a similar product
Jul 23 20:13:18 <sn9> <[mbm]> chip is labeled "AMRISC 20000"
Jul 23 20:13:47 <[mbm]> bought it at frys (local electronics shop)
Jul 23 20:13:57 <[mbm]> cheap too
Jul 23 20:14:28 <[florian]> I guess they bought the core and marketed it under a different name
Jul 23 20:14:59 <[florian]> I admit the r6040 driver makes it pretty clear it is a rdc-like device
Jul 23 20:15:01 <sn9> which fry's? not the fremont one...
Jul 23 20:15:45 <Bartman007> a nas with ide and sata for $70? wow.
Jul 23 20:16:08 <sn9> it was $70??
Jul 23 20:16:15 <Bartman007> sn9: based on the website.
Jul 23 20:16:20 <[florian]> sounds interesting
Jul 23 20:16:20 <[mbm]> sn9: sounds about right
Jul 23 20:16:35 <[florian]> sn9: put back the port priority stuff
Jul 23 20:16:51 <sn9> Bartman007: the frys.com website bears little relation to the brick-and-mortar stores
Jul 23 20:17:21 <sn9> $70 including a hard disk?
Jul 23 20:17:23 <Bartman007> sn9: yes, but prices are typically within $10 unless it is a returb, refurb, or a special sale.
Jul 23 20:17:28 <[mbm]> can't figure out the mnBIOS at the end of flash, I don't see it on serial at boot
Jul 23 20:17:33 <Bartman007> s/returb/return/
Jul 23 20:18:05 <sn9> [mbm]: other rdc devices used redboot
Jul 23 20:22:50 <crazy_imp> [mbm], sn9, [florian]: hmm, according to this forum post ( http://wl500g.info/showpost.php?p=59130&postcount=133 it's german) those chips are nearly the same, but the poster was unable to write on the whole flash, only to 1mb. maybe because this tool was for another flash chip (could that cause such a problem?).
Jul 23 20:29:17 <sn9> crazy_imp: comment out line 46 of src/flash.c and it will build
Jul 23 20:31:30 <crazy_imp> sn9: this line "#include "flash.h"" ? doesn't work :/
Jul 23 20:35:53 <sn9> worked for me
Jul 23 20:37:11 <sn9> you used the include pkg from the zipfile, right?
Jul 23 20:37:44 <crazy_imp> sn9: dunno which one it used, cleaned the complete dir right now
Jul 23 20:37:58 <crazy_imp> sn9: so i should point configure this time to the internal one?
Jul 23 20:38:09 <sn9> you have to specify the include dir to ./configure
Jul 23 20:38:23 <sn9> the one with the version number
Jul 23 20:38:43 <crazy_imp> sn9: iirc it found the one on my system
Jul 23 20:39:40 <crazy_imp> sn9: but lets see what happens if configure it this way: ./configure --with-include=./include-0.4.2/
Jul 23 20:41:06 <crazy_imp> ok, still the same bug, time to test it with the one from my system
Jul 23 20:41:21 <sn9> must be a full pathname
Jul 23 20:42:10 <crazy_imp> sn9: ok. configuring again with the internal one
Jul 23 20:42:49 <crazy_imp> sn9: :) it's building much more now
Jul 23 20:42:52 * crazy_imp waits
Jul 23 20:43:06 <crazy_imp> finished without an error
Jul 23 20:45:18 <crazy_imp> sn9: so, all i need to do is adding some pins to the board and some cables :). hope my parallelport is in the right mode, its there a way to test see in which one he is? don't want to reboot
Jul 23 20:46:03 <sn9> dmesg
Jul 23 20:47:46 <crazy_imp> sn9: and what should i grep for?
Jul 23 20:49:19 <crazy_imp> sn9: is this ok? -> parport0: PC-style at 0x378 (0x778), irq 7, dma 3 [PCSPP,TRISTATE,COMPAT,EPP,ECP,DMA]
Jul 23 20:49:52 <sn9> perfect
Jul 23 20:50:00 <sn9> epp+ecp mode
Jul 23 20:51:12 <sn9> nbd: you know about this, right? http://openocd.berlios.de/web/
Jul 23 20:51:54 <nbd> yes
Jul 23 21:07:32 <sn9> <[florian]> sn9: put back the port priority stuff
Jul 23 21:07:32 <sn9> any difference?
Jul 23 21:16:48 <crazy_imp> are there any restrictions for the cable lenght if i use a wiggler? would like to use 1m
Jul 23 21:19:40 <sn9> that's rather long; try extending the parallel side first
Jul 23 21:22:27 <crazy_imp> sn9: hmmm, no, don't have the plugs to build a cable, think i'll try to use 0.5m and creep below my desk
Jul 23 21:24:01 <sn9> ordinary parallel cable -- no need to make one
Jul 23 21:24:45 <crazy_imp> sn9: i only have some with a centronics plug on one end
Jul 23 21:24:53 <sn9> ah
Jul 23 21:25:19 <crazy_imp> and for the other one i would need an genderchanger (and dunno if it's crossed inside)
Jul 23 21:25:39 <sn9> a 25-pin scsi cable should also work (such as one from a scsi zip drive)
Jul 23 21:27:12 <crazy_imp> sn9: ha, good idea, forgot the cable from my iomega :) - so i can use it?
Jul 23 21:29:04 <sn9> yes, both the parallel and the scsi version of the cable should work
Jul 23 21:29:38 <sn9> just need a gender adapter
Jul 23 21:32:22 <crazy_imp> sn9: don't need one, it's male-female and i used a male connector for my wiggler board so i could plug it straight on the parport of my pc :)
Jul 23 21:33:21 <sn9> all iomega cables i've seen have been male-male
Jul 23 21:35:12 <crazy_imp> but not this one, could make a picture of it ;)
Jul 23 21:36:11 <sn9> no, you're right. i just took a look at my zip250 and it's m-f
Jul 23 21:39:38 <crazy_imp> good old hardware :)
Jul 23 21:40:12 <crazy_imp> hm, now i could use something like 10cm
Jul 23 21:41:20 <sn9> should be fine
Jul 23 22:04:59 <CIA-12> nbd * r8136 / (2 files in 2 dirs): merge madwifi trunk changeset 2597
Jul 23 22:10:09 <CIA-12> blogic * r8137 /trunk/target/linux/amazon-2.6/ (71 files in 24 dirs): initial merge of infineon code for amazon, pci is still broken a bit. a big thank you goes to infineon for providing info and reference code
Jul 23 22:10:21 <CIA-12> nbd * r8138 / (2 files in 2 dirs): fix mac80211 build for platforms that have led support disabled
Jul 23 22:18:31 <crazy_imp> sn9: i'm ready to jtag it. should i turn the jtag stuff on before i power up the board or can i boot it and start jtag later?
Jul 23 22:21:59 <blogic> jtag should trigger a reset when it powers up
Jul 23 22:22:21 <blogic> and if you power the jtag first, it should wait on the target vcc and then trigger a reset
Jul 23 22:23:01 <crazy_imp> ok, so i'll just power it up and see what i'll do next
Jul 23 22:27:02 <blogic> trail 'n error
Jul 23 22:27:17 <crazy_imp> can i select wiggler as cable if i have only connect tms, tck, tdo and tdi?
Jul 23 22:27:33 <blogic> dunno
Jul 23 22:27:59 * crazy_imp pokes sn9
Jul 23 22:28:00 <blogic> as i said, there is only 1 way to find out, trail 'n error
Jul 23 22:31:57 <crazy_imp> hmmm, nothing returns after detect.... think i'll recheck all connections
Jul 23 22:34:20 <crazy_imp> they are all high (3.3V) - normal?
Jul 23 22:43:00 <crazy_imp> eveything should be fine, voltage is there and i'm quite sure i connected the wires to the right pins of my wiggler
Jul 23 22:56:39 <sn9> crazy_imp: well
Jul 23 22:56:40 <sn9> ?
Jul 23 22:58:51 <crazy_imp> sn9: still nothing :/
Jul 23 23:01:12 <sn9> are you using DMA or PrAcc?
Jul 23 23:01:28 <crazy_imp> sn9: parport0: PC-style at 0x278 (0x678), irq 5, dma 3 [PCSPP,TRISTATE,COMPAT,EPP,ECP,DMA]
Jul 23 23:01:44 <crazy_imp> (i changed the address, but nothing happend)
Jul 23 23:02:20 <sn9> no, it should be 378/7
Jul 23 23:02:26 <sn9> not 278/5
Jul 23 23:02:51 <crazy_imp> sn9: is it hardcoded?
Jul 23 23:03:01 <crazy_imp> i can change it back again
Jul 23 23:03:34 <crazy_imp> i also have the choice between 1.7 and 1.9 but dunno which option it was
Jul 23 23:03:52 <crazy_imp> and i can set dma to a smaller value
Jul 23 23:04:01 <sn9> dma should be 3
Jul 23 23:04:17 <crazy_imp> ok
Jul 23 23:04:17 <sn9> can leave version at 1.9
Jul 23 23:04:45 <crazy_imp> ok, so i'll just change the address and the irq back
Jul 23 23:13:37 <crazy_imp> still the same
Jul 23 23:13:59 <sn9> crazy_imp: epp mode has to be enabled, but epp+ecp is better, since you don't have to change things around to use it
Jul 23 23:14:17 <sn9> what are you typing at the "jtag" proimpt?
Jul 23 23:15:19 <sn9> it should be "cable ppdev /dev/parport0 WIGGLER"
Jul 23 23:15:38 <crazy_imp> i hooked up my led-loadmeter to the cable and 2 leds are bright (one changes if i select DLC5 instead WIGGLER). if i type detect it starts to flicker really fast
Jul 23 23:15:59 <crazy_imp> sn9: i type "cable parallel 0x378 WIGGLER"
Jul 23 23:16:08 <sn9> that's wrong
Jul 23 23:16:26 <sn9> you should use ppdev instead of parallel
Jul 23 23:16:58 <sn9> parallel needs root access
Jul 23 23:17:09 <crazy_imp> sn9: hmmm, ok, but there is no parport0 device in /dev here :/
Jul 23 23:17:21 <crazy_imp> sn9: i tried it under root
Jul 23 23:17:27 <sn9> check dmesg
Jul 23 23:18:07 <crazy_imp> sn9: the dmesg entrie is still there (now again with 0x378/7)
Jul 23 23:18:30 <crazy_imp> but i don't have the parport-device in /dev
Jul 23 23:20:28 <crazy_imp> sn9: and it's accessing the parport, the leds are changing
Jul 23 23:27:17 <crazy_imp> sn9: if i enter discovery it's printing "Detecting IR lenght ... -1" and in the next line "Error: Invalid IR length!"
Jul 23 23:33:59 * crazy_imp builds an old version of the openwince jtag software from portage right now
Jul 23 23:47:04 <crazy_imp> same...
Jul 23 23:48:23 <crazy_imp> all i can see if i run detect and hook my multimeter up to the 74hc244 is something about 100kHz on the TCK line (both side of the ic, so the ic should be fine)
Jul 23 23:49:54 <crazy_imp> or is the jtag tool not aware of the rdc chip?
Jul 24 00:47:42 <sn9> crazy_imp: are you still there?
Jul 24 00:47:57 <sn9> if so, type the following into irssi:
Jul 24 00:47:59 <crazy_imp> sn9: jupp
Jul 24 00:48:16 <sn9> /exec -out dmesg|grep parp
Jul 24 00:48:26 <crazy_imp> parport: PnPBIOS parport detected.
Jul 24 00:48:26 <crazy_imp> parport0: PC-style at 0x378 (0x778), irq 7, dma 3 [PCSPP,TRISTATE,COMPAT,EPP,ECP,DMA]
Jul 24 00:48:43 <sn9> modprobe ppdev
Jul 24 00:49:11 <crazy_imp> ok, now i have /dev/parport0 :)
Jul 24 00:49:25 <sn9> now, at the jtag prompt:
Jul 24 00:49:34 <sn9> cable ppdev /dev/parport0 WIGGLER
Jul 24 00:49:41 <sn9> detect
Jul 24 00:49:50 <sn9> pastebin the results
Jul 24 00:50:37 <crazy_imp> sn9: http://pastebin.ca/631499
Jul 24 00:52:20 <sn9> detect shows nothing?
Jul 24 00:53:51 <crazy_imp> sn9: indeed, nothing :(
Jul 24 00:54:08 <sn9> are you sure the device is getting power?
Jul 24 00:57:32 <crazy_imp> sn9: yes, because i "see" the same frequency with my multimeter on the tck pin on the parport and on the tck pin on the device, so the 74hc244 should be working
Jul 24 00:58:50 <sn9> and no other jtag instance is running?
Jul 24 01:00:48 <crazy_imp> sn9: no, only this one
Jul 24 01:01:53 <sn9> and detect failed before, too?
Jul 24 01:02:27 <crazy_imp> sn9: first time i use it
Jul 24 01:03:25 <crazy_imp> hmmm, might found my failure, missed one 100ohm resistor for the BUSY-pin, which represents tdo.......
Jul 24 01:12:23 <crazy_imp> *grml*
Jul 24 01:12:27 <crazy_imp> still the same
Jul 24 01:12:38 <sn9> nothing on detect?
Jul 24 01:13:22 <crazy_imp> no
Jul 24 01:13:56 <crazy_imp> also still nothing going on the other pins (at least nothing i could notice)
Jul 24 01:14:34 <sn9> both with parallel and ppdev?
Jul 24 01:14:59 <crazy_imp> same
Jul 24 01:21:24 <crazy_imp> lets try something else
Jul 24 01:21:34 <sn9> http://cyberelk.net/tim/parport/archive/current/4373.html
Jul 24 01:22:45 <crazy_imp> so, i should reboot and remove <?> from the options?
Jul 24 01:23:17 <crazy_imp> if that doesn't work i'll try a unbuffered one
Jul 24 01:23:26 <sn9> use epp
Jul 24 01:23:32 <crazy_imp> ok, brb
Jul 24 01:30:12 <crazy_imp> sn9: bad news, can only select "normal, bi-directional or ecp+epp" :(
Jul 24 01:31:53 <sn9> i've never seen that
Jul 24 01:32:31 <crazy_imp> i'm booting my good old p2 right now, maybe i can get it running there
Jul 24 01:33:27 <crazy_imp> parport0: PC-style at 0x378, irq 7 [PCSPP,TRISTATE,EPP]
Jul 24 01:33:29 <crazy_imp> :)
Jul 24 01:44:57 <crazy_imp> still the same there....
Jul 24 01:47:00 <sn9> you're sure you traced the header leads to the correct pins on the rdc chip per the datasheet?
Jul 24 01:47:45 <crazy_imp> sn9: quite sure, used this datasheet R8610_D06_20051003.pdf
Jul 24 01:48:20 <sn9> does the device have the 3211 or the 3211-G?
Jul 24 01:48:48 <crazy_imp> -g
Jul 24 01:49:01 <sn9> rdc has separate datasheets for the 8610 and the 8610-G
Jul 24 01:50:03 <sn9> not sure what the difference between datasheets is
Jul 24 01:50:33 <sn9> the difference in chips is supposed to be RoHS
Jul 24 01:52:28 <crazy_imp> i'll take a look
Jul 24 01:54:24 <crazy_imp> sn9: nothing important at the moment, the jtag pins are the same
Jul 24 01:55:05 <sn9> there are two different pinouts for the same chip in the same datasheet for some reason
Jul 24 01:56:10 <crazy_imp> sn9: because one is for "MDC pulled high" and one for low
Jul 24 01:57:12 <sn9> did you determine which way the sitecom is?
Jul 24 01:59:09 <crazy_imp> if i find a definition of MDC in the datasheet i'll do
Jul 24 01:59:17 <sn9> it's a pin
Jul 24 02:00:13 <crazy_imp> 137?
Jul 24 02:00:28 <sn9> i don't remember the number offhand
Jul 24 02:00:41 <crazy_imp> MDC/PCI_REVERSE
Jul 24 02:00:55 <sn9> i guess so
Jul 24 02:02:16 <crazy_imp> sn9: page 18: MDC: MII management data clock is sourced by the R8610-G to the
Jul 24 02:02:19 <crazy_imp> external PHY devices as a timing reference for the transfer of
Jul 24 02:02:21 <crazy_imp> information on the MDIO signal.
Jul 24 02:02:44 <crazy_imp> don't think it would cause problems with jtag
Jul 24 02:03:06 <sn9> except if it alters the pinout
Jul 24 02:04:21 <sn9> how many pins is the jtag header, again?
Jul 24 02:04:31 <sn9> 5? 6?
Jul 24 02:04:44 <crazy_imp> 6, 4 jtag + gnd and vcc
Jul 24 02:04:51 <sn9> SIP?
Jul 24 02:05:42 <sn9> is it a Single Inline Package header?
Jul 24 02:05:54 <crazy_imp> sn9: the flash? no, it looks like the roms from a wrt http://openwrt.vcp-springe.de/wl-153/100_2816.jpg
Jul 24 02:06:07 <sn9> the jtag header
Jul 24 02:06:47 <crazy_imp> sn9: take a look at the image^^. JP1 is the jtag header
Jul 24 02:07:18 <sn9> is it the same pinout that i traced for the ar525w?
Jul 24 02:07:51 <crazy_imp> if you start from the jp1 mark, it reads: vcc, gnd, tck, tdo, tdi, tms.
Jul 24 02:09:39 <sn9> very odd, because the ar525w used the same header, but with a different pinout
Jul 24 02:09:46 <sn9> http://wiki.openwrt.org/OpenWrtDocs/Hardware/Airlink101/AR525W
Jul 24 02:10:19 <crazy_imp> sn9: what is Vddh ?
Jul 24 02:10:19 <sn9> granted, that info is only from my eyes, and not from trying it
Jul 24 02:10:38 <sn9> jtag-ese for Vcc
Jul 24 02:10:53 <crazy_imp> and Vss?
Jul 24 02:10:57 <crazy_imp> = gnd?
Jul 24 02:11:02 <sn9> yes
Jul 24 02:11:57 <crazy_imp> can i kill it if i get the jtag cables wrong? i'll just try your pinout
Jul 24 02:12:37 <sn9> if it's unbuffered, yes, you can, and also your parport hardware
Jul 24 02:12:59 <crazy_imp> so i'm fine with my wiggler?
Jul 24 02:13:40 <sn9> with the wiggler, i'd chance it, but verify the device boots with the jtag diisconnected both before and after trying
Jul 24 02:17:24 <crazy_imp> again nothing :(
Jul 24 02:17:25 <sn9> btw, are you soldering wires directly to the board? i made that mistake at first with the serial header
Jul 24 02:21:09 <crazy_imp> no, i use some kind of pins with a "cup" at the other end (and soldered the wires in there)
Jul 24 02:21:46 <sn9> good
Jul 24 02:21:49 <crazy_imp> think they are called "ic-socket"
Jul 24 02:21:59 <crazy_imp> sn9: http://www.reichelt.de/bilder//web/bauelemente/C140/SPL.jpg
Jul 24 02:22:34 <sn9> what are you sticking in the socket?
Jul 24 02:22:49 <crazy_imp> i soldered the cables in there
Jul 24 02:23:40 <sn9> ic sockets are for soldering to boards; they are not meant to hold solder
Jul 24 02:23:59 <crazy_imp> i know :D
Jul 24 02:29:05 <sn9> [mbm]: regarding the device you got: http://forum.openwrt.org/viewtopic.php?id=6827
Jul 24 02:31:43 <sn9> crazy_imp: look in the datasheet for "STRAP 11"
Jul 24 02:32:03 <sn9> it might be pin 40
Jul 24 02:34:34 <crazy_imp> ouuuuuch, don't like to count those damn small pins :(
Jul 24 02:34:44 <crazy_imp> but it's a good idea :)
Jul 24 02:36:18 <sn9> on the ar525w, each tenth pin is marked with a longer line than the lines marking pins halfway in between
Jul 24 02:37:10 <crazy_imp> sn9: lucker :P
Jul 24 02:50:21 <crazy_imp> hehe, found it, easy because it's next to a gnd pin :)
**** ENDING LOGGING AT Tue Jul 24 02:59:57 2007