irclog2html for #openjtag on 20050513

00:33.44*** join/#openjtag ka6sox-zzzz (~ka6sox@netblock-66-159-209-52.dslextreme.com)
01:30.23*** join/#openjtag Tiersten (~tman@Tiersten.nslu2-linux)
01:54.20*** join/#openjtag prpplague (~billybob@200.50.80.211)
03:48.29ka6sox-awayhi....
03:55.52ka6sox-awaymy guess is that everyone is really gone.
04:19.50[g2]Don't think so
04:20.21ka6sox-awayHi there.
04:21.07[g2]hello
04:21.15ka6sox-awayhow goes your project?
04:21.51[g2]I think the pad at R137 is totaly gone and I'm booting the unit to see how the other three connections are
04:22.14[g2]I've got to re-setup the serial because I've got APEX in there
04:23.32ka6sox-awayI'm looking forward to using APEX on my slugs.
04:23.44[g2]Althought this project looks like the JTAG part won't be a success, my soldering is improving tremendously
04:23.59[g2]APEX is nice
04:24.13ka6sox-awaycool
04:30.50[g2]Ok the slug is a live
04:31.15[g2]so I'm 3 out of 4 on the JTAG pin connections with 4 hoses
04:31.17[g2]hosed
04:31.26ka6sox-awayouch
04:33.14[g2]Ok... Redboot is back
04:34.04[g2]ka6sox-away, you've got JTAG right ?
04:42.39ka6sox-awayyes.
04:42.47ka6sox-awayon the Slug?\
04:43.36[g2]nod.
04:44.05ka6sox-awayya
04:44.16[g2]There's just a small issue or two with the irq handler not operating properly on the slug, so Ethernet and USB aren't working properly yet on the slug
04:44.26[g2]the avila board is 100% with APEX
04:45.14[g2]well I'm pretty sure.... Maybe the PCI is still an issue.  I'd have to plug in the minipci card, but the regular ethernet is 100%
04:47.43ka6sox-awayexcellent...
04:48.07[g2]Ok... that's setup here goes
04:52.01[g2]Ok.... 64MB is working, IXP is working...  PCI still has an issue which is probably the same as on the slug
04:57.58[g2]OK...  PCI is hosed on Avila too with APEX
05:00.05*** join/#openjtag dyoung-web (www-data@dyoung.nslu2-linux)
05:00.05beewoolie-afk[g2]: k.
05:00.16beewoolie-afkI
05:00.22beewoolie-afkm just about to do some exploration.
05:00.25[g2]big paste coming
05:00.39[g2]APEX 1.2.9_64 Meg Avila  E1000 PCI issue
05:00.39[g2]When modprobing E1000 with APEX we get
05:00.39[g2]root@avila:~# modprobe e1000
05:00.39[g2]Intel(R) PRO/1000 Network Driver - version 5.6.10.1-k2
05:00.39[g2]Copyright (c) 1999-2004 Intel Corporation.
05:00.40[g2]root@avila:~# ifconfig eth2 192.168.1.79
05:00.42[g2]SIOCSIFADDR: No such device
05:01.09[g2]Note. The PCI *doesn't* come up with APEX 1.2.9
05:01.15[g2]now here's the Redboot
05:01.31[g2]Redboot and E1000
05:01.31[g2]root@avila:~# modprobe e1000
05:01.31[g2]Intel(R) PRO/1000 Network Driver - version 5.6.10.1-k2
05:01.31[g2]Copyright (c) 1999-2004 Intel Corporation.
05:01.31[g2]PCI: enabling device 0000:00:01.0 (0140 -> 0143)
05:01.32[g2]e1000: eth2: e1000_probe: Intel(R) PRO/1000 Network Connection
05:02.03beewoolie-afk[g2]: Ah.  That's what I was thinking was going to happen.
05:02.07beewoolie-afkThe USB is also PCI.
05:02.13beewoolie-afkright?
05:02.16[g2]Yeah... That was the point
05:02.27[g2]Same issue
05:02.36beewoolie-afkThat's where I was planning to start.
05:02.43beewoolie-afkLet me take a pass at the PCI registers.
05:02.45[g2]Most Excellent
05:03.31[g2]Note. you can tell because the ehci/ohci interrupts are not visible in /proc/interrupts
05:03.59beewoolie-afkThere is the other thing which is, IIRC, the PCI may get timing from the GPIO stuff  I thought that I was setting that correctly, but maybe not.
05:04.24beewoolie-afkRight.  I looked at the USB driver and discovered that they're all PCI based.  That's why I thought it was the best place to start.
05:04.43[g2]Well the GPIO get setup by the kernel
05:04.46beewoolie-afkThe thing about PCI, IIRC, is that the devices must be 'detected' or somthing before the kernel gets going.
05:05.06beewoolie-afkAre you sure?  Usually those lines are configured in the bootloader.
05:05.21beewoolie-afkReally, I'd expect PCI to be configured in the kernel, tooo.
05:05.25[g2]yep. I'm pretty sure
05:05.32[g2]I changed them for the Avila kernel
05:05.51[g2]and I changed them from what they were set to in the Slug
05:06.08[g2]I don't know if it's ALL, but most
05:06.13[g2]there are only 16
05:06.28beewoolie-afkWhere is that in the kernel code?
05:06.29[g2]it's in the patch files
05:06.33beewoolie-afkOK.
05:07.22dyoung-webyou have a avila patchset?
05:08.10beewoolie-afkI don't.
05:08.15[g2]dyoung-web, I've had one for a couple months
05:08.24dyoung-webis it pushed someplace?
05:08.26[g2]ok 8 weeks
05:08.36[g2]nope, I'm the only one using it
05:08.54dyoung-webOh.
05:09.19[g2]it's actually only like 200 lines total
05:09.20beewoolie-afkI was only curious about that.  I'm going to focus on what redboot may be doing to get PCI going.
05:09.24beewoolie-afkk.
05:09.40[g2]I'll give it to any of you guys if you want it
05:09.56[g2]beewoolie-afk, I'm focusing on the PCI too.
05:10.03beewoolie-afkThe thing is, though, now that I think about it.  It is possible that the missing piece isn't part of the redboot patch, but part of redboot proper.
05:10.07beewoolie-afknp
05:10.14[g2]I've got the boots from both kernels and I'm gonna look at the PCI messages
05:11.21[g2]beewoolie-afk, Ok kernel wonder-boy... My limited understanding was that the kernel was supposed to setup everything except the memory
05:11.47beewoolie-afkOK.  Then why is PCI not working?
05:12.22beewoolie-afkI think that the kernel configures everything that makes sense for itto configure.  However, it is *not* responsible for ethernet MAC address, for instance.
05:12.24[g2]I was asking you a general kernel question because you are more familiar with this stuff in general
05:12.34beewoolie-afkThere are always some things that have to be done outside of the kernel.
05:12.44beewoolie-afkI'm not sure there is a hard-and-fast rule.
05:13.06beewoolie-afkThere isn't anything about PCI in the redboot patch.
05:13.24beewoolie-afkWhat I'll do is look at the registers themselves to see if there is something going on.
05:13.35[g2]I'll check the messages right now
05:13.44[g2]want an email copy ?
05:14.19beewoolie-afkOf the kernel message?  Actually, what I'd rather see are the /proc entries for the PCI stuff.
05:14.27beewoolie-afkIf there are any.
05:14.42[g2]I can check
05:15.51[g2]cat /proc/pci
05:15.51[g2]PCI devices found:
05:15.51[g2]<PROTECTED>
05:15.51[g2]<PROTECTED>
05:15.51[g2]<PROTECTED>
05:15.52[g2]<PROTECTED>
05:15.54[g2]<PROTECTED>
05:15.56[g2]<PROTECTED>
05:15.58[g2]That's Redboot
05:17.15beewoolie-afkInteresting.  The other devices don't show?  Such as the USB or the ethernet.
05:18.18[g2]that's the avila board
05:18.20beewoolie-afkThe FSCK'ing redboot protects the PCI registers, so I cannot read them from redboot.
05:18.40[g2]AND.... we are missing something on the PCI setup
05:19.02[g2]because we *don't* get the dmabounce message from APEX but we do from Redboot
05:19.08[g2]PCI: bus0: Fast back to back transfers disabled
05:19.08[g2]dmabounce: registered device 0000:00:01.0 on pci bus
05:19.28[g2]We get the three PCI messages before that, but not the dmabounce
05:19.48[g2]root@avila:~# cat /proc/pci
05:19.48[g2]root@avila:~#
05:19.56beewoolie-afkOK.  I'll boot the slug, too, to get the info.
05:23.30[g2]Ok.. This routine isn't getting called or the pci_bus_type isn't right
05:23.42[g2]static int ixp4xx_pci_platform_notify(struct device *dev)
05:23.42[g2]{
05:23.43[g2]<PROTECTED>
05:23.43[g2]<PROTECTED>
05:23.43[g2]<PROTECTED>
05:23.43[g2]<PROTECTED>
05:23.45[g2]<PROTECTED>
05:23.47[g2]<PROTECTED>
05:23.52beewoolie-afkIs this in the kernel?
05:23.58[g2]nod.
05:24.10beewoolie-afkHmm.  I may have found the code...
05:24.18[g2]line 321 in pci-common.c
05:24.32beewoolie-afkIt's really short.  I'm talking about redboot.
05:24.53beewoolie-afkcyg_hal_plf_pci_init(void)
05:25.00[g2]ah...
05:26.00beewoolie-afkI was assuming that the code we needed was in the patch.  That, it appears, is not the case.
05:26.12beewoolie-afkIt's going to take some time to pick this all appart.
05:26.24[g2]that may be "standard" Redboot
05:27.05beewoolie-afkYep.  It's standard as delivered by Intel to cygnus/redhat.
05:28.08[g2]actually, I can't image a section of code that would be a better candidate than PCI
05:28.19beewoolie-afkIt still doesn't explain the NPE problem, but, I fear, that may be deeper.  Except, that the NPE works on the Avilla, right?
05:28.34[g2]Yup!
05:28.56beewoolie-afkOK.  Well, I'm not going to fret that yet.
05:28.58[g2]me either...
05:29.03beewoolie-afkIf we can get USB working, then we're at least making progress.
05:29.23[g2]and that'll round out the Avila most likely
05:30.14dyoung-webDont Fret.
05:30.30[g2]I'm not Frett'n at ALL
05:30.42[g2]we are tremendously close
05:30.54dyoung-webI wish I could help.
05:31.07[g2]after USB on the slug, APEX well boot to usb-ethernet
05:31.20[g2]and we'll have disk access again
05:31.20dyoung-webI've been occupied hacking/whacking on this packet communication protocol.
05:32.21dyoung-webI should take a picture of this board actually...
05:32.30dyoung-webits a very good example how how you dont want to hack a chip in.
05:33.40beewoolie-afkHeh.  I need to look at my own wiki to remember how to put apex on this board.
05:33.55[g2]dd
05:34.13[g2]dd if=apex.bin of=/dev/mdtblock0
05:34.13beewoolie-afkHmm.  That's a though.
05:34.23beewoolie-afkI don't want to clobber all of redboot, though.
05:34.26[g2]THAT's EXACTLY how I do it
05:34.34[g2]before you do...
05:34.46[g2]dd if=/dev/mtdblock0 of=redboot.bin
05:34.52beewoolie-afkBut if I do that, I'll have to use JTAG to rewrite redboot, now?
05:34.57beewoolie-afks/now/no/
05:35.02[g2]actually you only need the last 144 btyes (less actually)
05:35.12beewoolie-afkHuh?
05:35.12[g2]nope...
05:35.20[g2]You've got serial right ?
05:35.24beewoolie-afkYeah.
05:35.40[g2]APEX boots the linux kernel to and rootfs
05:35.44beewoolie-afkI don't have an ethernet cable at the moment.
05:35.44[g2]but only serial works
05:35.49beewoolie-afkright.
05:36.00[g2]so login in and dd back to redboto
05:36.01beewoolie-afkIs there a serial receive function in the slug rootfs/
05:36.15beewoolie-afkrz?
05:36.21[g2]don't need it
05:36.27beewoolie-afkHmm?
05:36.41beewoolie-afkwhere is redboot?
05:36.42[g2]You're running jffs2 righit ?
05:36.48beewoolie-afkStock.
05:36.57[g2]Oh... Run OpenSlug
05:37.19beewoolie-afkOf course.
05:37.52[g2]the first 256K will be the same as everyone elses
05:38.04beewoolie-afkWait...I've got another idea.  I think this one is better...
05:39.25beewoolie-afkwhat I really want to do is look at the redboot registers.
05:39.25beewoolie-afkerr.  I want to look at the PCI registers as redboot has set them up.
05:39.54[g2]Can I distract you for a couple minutes ?
05:40.00beewoolie-afkYou can try.
05:40.17[g2]Ok. That routine only gets call from 1 line in pci-common.c
05:40.55[g2]That routine in turn is called by ixp4xx_setup
05:40.58beewoolie-afkYou are back to the kernel, right?
05:41.02[g2]nod.
05:41.24dyoung-webdont forget about the MAC address.  that will be the only differnt part.
05:41.25[g2]Like I said earlier, Only 1 of 2 things can be happening
05:41.42[g2]it's less than the last 144 bytes
05:41.45beewoolie-afkdyoung: Do you mean to me/
05:41.52beewoolie-afkI saved redboot, first thing.
05:42.04beewoolie-afkIt's also printed on the board.
05:43.15dyoung-webok cool.  I wouldnt want you wasting time like I did wondering where my mac address was.
05:45.32beewoolie-afkI might be wrong about it being printed on the board.
05:45.49[g2]no there's a sticker on the bottom
05:46.48beewoolie-afkBastards.
05:47.01dyoung-webgood luck gentlemen.  I'm hungry.
05:47.01beewoolie-afkUsually, those numbers are printed somewhere on the package.
05:47.13[g2]that too
05:48.19beewoolie-afkActually, it is on the packages.
05:48.27beewoolie-afkdyoung-web: happy eating.
05:48.36[g2]it's both places
05:48.38beewoolie-afkThe first part of the number is LKG.
05:49.16beewoolie-afkOK.  I'm in the money.  I've got APEX letting me read the PCI registers.
05:51.10dyoung-web... And now its Payback Time. :-)
05:51.27[g2]Got it !
05:51.33beewoolie-afkHells yeah!
05:51.56[g2]Redboot
05:52.00[g2]PCI: IXP4xx Using indirect access for memory space
05:52.00[g2]PCI: bus0: Fast back to back transfers disabled
05:52.00[g2]dmabounce: registered device 0000:00:01.0 on pci bus
05:52.04[g2]APEX
05:52.22[g2]PCI: IXP4xx is host
05:52.23[g2]PCI: IXP4xx Using indirect access for memory space
05:52.23[g2]PCI: bus0: Fast back to back transfers enabled
05:52.44[g2]besides the dmabounce notice any difference ?
05:53.00dyoung-webs/dis/en
05:53.09[g2]ding ding ding
05:53.47beewoolie-afkHonestly, I'm not following closely enought to see what you are seeing.
05:54.08[g2]honesty is important
05:54.32[g2]PCI: bus0: Fast back to back transfers *disabled*  in Redboot
05:54.45[g2]PCI: bus0: Fast back to back transfers *enabled* in APEX
05:54.48beewoolie-afkWanna give me a hint.  I'll be clear.  My plan is to init the PCI registers and see if that's enough,
05:55.01beewoolie-afkHmm.  Interesting default difference.
05:55.19[g2]well the bus doesn't get recognized
05:55.45[g2]well registered is probably a better way of saying it
06:00.37[g2]pci_read_config_word(dev, PCI_STATUS, &status);
06:00.43[g2]I think it's defined in there
06:00.55[g2]if (!(status & PCI_STATUS_FAST_BACK))
06:00.55[g2]<PROTECTED>
06:04.46beewoolie-afkHmm.  Guess what.  An uninitialize PCI bus is in big endian mode.
06:04.52beewoolie-afkI mean little endian mode.
06:05.49[g2]do we come up in the wrong endian mode get bad config mojo an the bus don't run :)
06:05.58[g2]s/do we/so w/
06:06.46[g2]The pullups on the expansion bus determine some stuff about the processor
06:07.06[g2]I'm wondering if they are not set properly for the NSLU2
06:07.52[g2]Although.... it's both Avila and NSLU2 so it's just a non-default config item we need to set
06:07.54[g2]correct ?
06:13.50beewoolie-afkDon't know.  I'm looking at the default config as described in the documentation.  I haven't reflashed yet.
06:13.55beewoolie-afkStill coding.
06:19.30*** join/#openjtag ka6sox-zzzz (~ka6sox@adsl-66-159-198-116.dslextreme.com)
06:27.15beewoolie-afkka6sox: morning.
06:27.26beewoolie-afkka6sox: read the logs?
06:27.52ka6soxum...not yet...wazzup?
06:28.11beewoolie-afkI think we're onto part of our slug boot problem with APEX.  PCI needs to be configured.
06:28.22beewoolie-afkI'm almost ready to test some new code.
06:28.30ka6soxsweet.
06:28.41ka6soxmemory is fine?
06:28.53[g2]memory is fine
06:28.58[g2]both 32MB and 64MB
06:29.03ka6soxexcellent
06:29.11[g2]64MB on the Avila
06:29.13beewoolie-afkAFAIK, memory has been fine for some time, no [g2] ?
06:29.18ka6soxgood.
06:29.23ka6soxso now onto the PCI issues.
06:29.32[g2]go dog go :)
06:29.37beewoolie-afkI'm not sure if this is what is in the way of NPE, tho.
06:29.50[g2]that's OK!
06:29.54[g2]One at a time
06:30.16[g2]if it turns out to be a bonus two-for-one we'll take it right ?
06:31.49ka6soxwhat do you guys want/need in a JTAG dongle/debugging tool?
06:32.44ka6soxnow that I'm done with Dyoung's board I want to get going on the Juice Board.
06:48.10beewoolie-afkka6sox: Um, can I get back to you on that?
06:48.24ka6soxsure!
06:55.02[g2]beewoolie-afk, I'm gonna have to bail
06:55.14[g2]THX for all the help tonight
06:55.15ka6soxnite [g2]
06:55.16beewoolie-afkand I'm just about to test.
06:55.20beewoolie-afkNite.
06:55.40[g2]How long do you think it'll be ? 5-10 minutes ?
06:56.02beewoolie-afkWhile I'm not teasing, I cannot really say.
06:56.16beewoolie-afkI'm a pretty carefule fellow with the flashing stuff.
06:56.43[g2]:)
06:57.00beewoolie-afkHmm.  Well, the code loads.
06:57.14[g2]what did the PCI messages say ?
06:57.21beewoolie-afkHmm.
06:57.27[g2]is Fast back to back enabled ?
06:57.34beewoolie-afkNot good.  It didn't boot properly from RAM.
06:57.41beewoolie-afkI need to go back and check some things.
06:58.42beewoolie-afkNo.  I set things up the same way as it was being done.  The fast back-to-back setting is cleared by default.
06:59.37[g2]so it needs to be disabled
06:59.57beewoolie-afkI mean, I'm doing the same setup as redboot which will clear that bit.
07:00.15beewoolie-afkBTW, do you remember how we have always gotten a dump at the top of the kernel boot with APEX?
07:00.19beewoolie-afkAre you still getting that?
07:00.23beewoolie-afkOr did we fix it?
07:00.57[g2]I don't recall dumps.
07:01.17[g2]But I haven't run from memory in quite some time
07:01.27beewoolie-afkBooting kernel at 0x00008000...
07:01.27beewoolie-afkUncompressing Linux........................................................... .
07:01.27beewoolie-afk690541F1:000000F5:00001AFA
07:01.28beewoolie-afk000F4E24-002B30A4>00008000
07:01.30beewoolie-afk002B30A4
07:01.32beewoolie-afk00008000: E1A0C000 E3A000D3 E121F000 EB000048  E33A0000 03A00070 0A000043 EB0008
07:01.34beewoolie-afk00008020: E3370000 03A00061 0A00003F EB00001F  E28FE01C E28AF00C C0008080 C01C60
07:01.36beewoolie-afk00008040: C01EBC34 C01C6FEC C01C6FE0 C01BAD14  C01BA000 E51FE024 EE010F10 EE1100
07:01.38beewoolie-afk00008060: E1A00000 E1A00000 E1A0F00E 00000000  00000000 00000000 00000000 000000
07:01.40beewoolie-afk00008080: E24F304C E89321F0 E3A0B000 E1540005  3484B004 3AFFFFFC E5869000 E58710
07:01.44beewoolie-afk000080A0: E3800002 E3C02002 E8880005 EA000132  E24F40B8 E2444901 E1A00004 E3A030
07:01.48beewoolie-afk000080C0: E2802901 E4803004 E4803004 E4803004  E4803004 E1300002 1AFFFFF9 E3C42F
07:01.50beewoolie-afk000080E0: E0883002 E7843922 E2840A03 E3C3260F  E5802000 E2800000 E4803004 E28331
07:01.52beewoolie-afkLinux version 2.4.22-xfs (root@sure_linux) (gcc version 3.2.1) #377 Fri Jul 2 04
07:01.54beewoolie-afkNot familiar?
07:02.21[g2]nope...
07:02.32beewoolie-afkOK.  It may have to do with the older kernel.
07:02.33[g2]I don't think I've ever seen that
07:03.06[g2]hey THX again..
07:03.14beewoolie-afkDon't wait up.  I'm going to turn pumpkin-side myself in a few minutes.
07:03.15beewoolie-afknp.
07:03.17[g2]I'll catch you tomorrow
07:03.19beewoolie-afkWe'll lick this one.
07:03.21beewoolie-afkK.
07:03.36[g2]If you make some big progress pls e-mail
07:03.43[g2]nite
07:27.42beewoolie-afkWell.  No dice.  I've done more looking into the redboot init code and I think I've found something else that needs to be done.
07:27.48beewoolie-afkPerhaps tomorrow.
07:31.55beewoolie-afkThe bad news is that it doesn't yet work.
07:49.00beewoolie-afkThe good news is that I think I know what is missing.  
07:49.02beewoolie-afknite
12:59.05*** join/#openjtag prpplague (prpplague@200.50.64.101)
15:53.14*** join/#openjtag [g2] (~g2@g2.nslu2-linux)
16:06.41ka6sox[g2]: is your JTAG broken did I read?
16:07.50[g2]I lifted a pad while attempting to solder JTAG to a slug
16:08.28[g2]It's on R137 iirc, which makes it difficult to fix
16:08.45[g2]but the slug works fine. Just no JTAG
16:09.07[g2]Also, I was trying to use JTAG on the Avila aboard and that was not working
16:09.43[g2]so I'm not sure which you are referring to a) the lifted pad on the NSLU2, or b) Avila JTAG using the digilent cable
16:49.53ka6soxa)
18:04.48*** join/#openjtag ep1220 (~ep1220@gate.epatec.at)
19:09.47*** join/#openjtag prpplague (~billybob@72.22.136.131)
19:19.38ka6soxmusta scared off beewoolie...
19:19.54ka6sox[00:03] <[g2]>nite
19:19.54ka6sox[00:03] <--[g2] has quit ("G'nite all")
19:19.54ka6sox[00:27] <beewoolie-afk>Well.  No dice.  I've done more looking into the redboot init code and I think I've found something else that needs to be done.
19:19.54ka6sox[00:27] <beewoolie-afk>Perhaps tomorrow.
19:19.54ka6sox[00:31] <beewoolie-afk>The bad news is that it doesn't yet work.
19:19.56ka6sox[00:49] <beewoolie-afk>The good news is that I think I know what is missing.  
19:19.58ka6sox[00:49] <beewoolie-afk>nite
19:31.10*** join/#openjtag sh (~one_chanc@202.163.76.41)
19:32.02*** part/#openjtag sh (~one_chanc@202.163.76.41)
23:04.05ka6soxseen beewoolie-afk?
23:52.32ka6sox~seen beewoolie-afk
23:52.39purlbeewoolie-afk <~beewoolie@florence.buici.com> was last seen on IRC in channel #openjtag, 16h 3m 37s ago, saying: 'nite'.
23:52.40qbotbeewoolie-afk <~beewoolie@florence.buici.com> was last seen on IRC in channel #openjtag, 16h 3m 29s ago, saying: 'nite'.
23:53.43prpplagueka6sox:evening
23:53.56ka6soxhi there...howse things!
23:54.20prpplaguenot too bad, didn't get much work done this week :(
23:54.30prpplaguegot a fb driver thats being a beoych'
23:54.34ka6soxah...why not?
23:55.09prpplagueka6sox: just could not seem to focus, had too much info
23:55.21prpplagueka6sox: gonna try and get some stuff done tomorrow
23:55.34ka6soxyou in on the Island now?
23:56.42prpplagueka6sox: yea, been here for a little over a month
23:57.10ka6soxmissed the Hurricane season....
23:59.50prpplagueka6sox: yea, starts again soon

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.