**** BEGIN LOGGING AT Tue Feb 22 23:59:56 2005
Feb 23 08:54:52 <prpplague> qbot: seen beewoolie 
Feb 23 08:54:53 <qbot> beewoolie <~beewoolie@florence.buici.com> was last seen on IRC in channel #openjtag, 9h 20m 18s ago, saying: 'night.'.
Feb 23 11:20:43 <beewoolie-away> [g2]: Hey.  Where is the network driver in the 2.6.9 kernel?
Feb 23 11:22:30 <prpplague> beewoolie-away: boo
Feb 23 11:22:34 <beewoolie-away> Hey.
Feb 23 11:22:52 <prpplague> beewoolie-away: how do you want to handle patches for apex?
Feb 23 11:23:17 <prpplague> beewoolie-away: just email them to you?
Feb 23 11:23:28 <beewoolie-away> Send them to my email, yeah.
Feb 23 11:23:33 <prpplague> ok
Feb 23 11:24:01 <prpplague> beewoolie-away: the mmc driver has alot in common with your cpld-spi driver
Feb 23 11:24:13 <beewoolie-away> I'd expect it to.
Feb 23 11:25:04 <prpplague> beewoolie-away: the block device for the fat driver is still hardcoded, do you have plans to change that?
Feb 23 11:25:38 * prpplague is still working on understanding the drivers structures
Feb 23 11:25:54 <beewoolie-away> prpplague: Of course.  I've been tossing around a couple of options.  Nothing has made enough sense to change it.  The hard coding makes the driver useful while giving me time to mull over the options.
Feb 23 11:26:11 <beewoolie-away> prpplague: I already know that I need to pull the partition code out of the fat driver.
Feb 23 11:26:19 <beewoolie-away> prpplague: That will come first.
Feb 23 11:26:29 <prpplague> beewoolie-away: np, i didn't want hack if i knew something was in the pipe
Feb 23 11:26:55 <beewoolie-away> prpplague: The idea that I'm working with is letting the user describe it.
Feb 23 11:27:20 <beewoolie-away> prpplague: For the time being, we can add a platform override.  You can set a config_ option and test for it in the driver.  If this doesn't make sense, then just maintain a small diff for now.
Feb 23 11:27:32 <beewoolie-away> prpplague: Anyway, the idea would be that you could chain drivers.
Feb 23 11:27:32 <prpplague> ok
Feb 23 11:27:42 <prpplague> beewoolie-away: yea
Feb 23 11:27:58 <beewoolie-away> Something like this:  fat:/zImage<-cf:/1
Feb 23 11:28:03 <prpplague> beewoolie-away: i'm not seeing where you init the spi interface
Feb 23 11:28:19 <beewoolie-away> This would mean that you were asking to read a FAT filesystem on the first partition of the CF card.
Feb 23 11:28:27 <beewoolie-away> Unfortunately, I don't yet like it.
Feb 23 11:29:06 <prpplague> beewoolie-away: ahh so you changed your mind aobut the alias functions? i.e. spi.mmc.fat(1):/zImage ?
Feb 23 11:29:09 <beewoolie-away> prpplague: Init SPI interface?  what do you mean?
Feb 23 11:29:40 <beewoolie-away> prpplague: Not yet.  I'm looking for a declarative form, first.  Once it works, the aliasing makes it user-friendly.
Feb 23 11:29:41 <prpplague> beewoolie-away: setting up all the control registers for the spi
Feb 23 11:29:54 <beewoolie-away> prpplague: I don't need to.  I
Feb 23 11:30:03 <beewoolie-away> I'm using SPI in the CPLD.  There is no setup.
Feb 23 11:30:21 <beewoolie-away> what setup are you expecting to see?
Feb 23 11:30:44 <beewoolie-away> Actually, that isn't true.  I am using the SSP interface, but that is only test code.
Feb 23 11:30:45 <prpplague> beewoolie-away: ahh, the 79520 has a whole set of spi registers
Feb 23 11:31:04 <beewoolie-away> prpplague: I know.  I just don't use them at the moment.
Feb 23 11:32:01 <prpplague> beewoolie-away: i don't have the schematics for the lpd79524 stuff, so i must have mis-read the implementation of the cpld-spi driver
Feb 23 11:33:29 <beewoolie-away> np.
Feb 23 11:34:52 <prpplague> ahh ok the 79524 has the same phys address for spi as the 79520
Feb 23 11:36:40 <prpplague> beewoolie-away: that loopback feature is nice for testing
Feb 23 11:37:08 <beewoolie-away> I don't think I've used it.
Feb 23 11:43:13 <prpplague> beewoolie-away: guess i need to make a list of the calls needed by the fat fs
Feb 23 11:43:35 <beewoolie-away> Open, seek, read.
Feb 23 11:43:50 <beewoolie-away> It isn't a very complex interface.
Feb 23 11:43:59 <beewoolie-away> And seek can use the helper.
Feb 23 11:44:44 <beewoolie-away> the only catch is going to be that the fat driver, as is, requires that there be a partition table.  A n undesirable requirement.
Feb 23 11:47:01 <prpplague> no close?
Feb 23 11:48:14 <beewoolie-away> You can implement it if you like.  The CF driver uses the close helper.  There isn't a need to do so since I don't allocate memory and the drivers are, for the most part, stateless.
Feb 23 11:49:00 <prpplague> ahh
Feb 23 11:49:11 * prpplague looks at seek_helper and close_helper
Feb 23 11:51:56 <prpplague> beewoolie-away: i assume the drv-mmc.c should go in the mach dir right?
Feb 23 11:52:08 <prpplague> beewoolie-away: since it is machine specific right now
Feb 23 11:52:18 <beewoolie-away> If it writes registers, yep.
Feb 23 11:52:36 <beewoolie-away> I'm going to move the CF driver out when I factor out the host specific elements.
Feb 23 11:53:15 <prpplague> beewoolie-away: yea, thats what i was considering doing with the mmc
Feb 23 11:53:19 <beewoolie-away> If you can do the same for MMC, it might be worthwhile.  Generally, though, there isn't much reason to do that.  I don't want a lot of needless interfaces.
Feb 23 11:53:52 <prpplague> beewoolie-away: do you have a naming scheme for the drivers? the cf is "cf-lpd79520" 
Feb 23 11:53:55 <beewoolie-away> Code sharing isn't that important when the source is only a few hundreds of lines.  Stability is more important as is compactness.
Feb 23 11:54:21 <beewoolie-away> The driver names are setup so as to be unique *and* and reasonably unique prefixes.
Feb 23 11:54:46 <beewoolie-away> It doesn't make sense to put the specific part first.  Right?
Feb 23 11:54:55 <prpplague> beewoolie-away: so, "mmc-lh79520" would be acceptable?
Feb 23 11:54:55 <beewoolie-away> You driver would be mmc-XXX
Feb 23 11:55:04 <beewoolie-away> As long as that is really what it is.
Feb 23 11:55:22 <beewoolie-away> If it is really mmc-prplague's-platform, then it should reflect that.
Feb 23 11:55:28 <beewoolie-away> Like the init-* sources.
Feb 23 11:55:36 <prpplague> beewoolie-away: this is implementing standard lh79520 hardware
Feb 23 11:55:44 <beewoolie-away> Users won't usually have to type the whole thing.
Feb 23 11:55:52 <beewoolie-away> Then you got it right.
Feb 23 11:55:54 <beewoolie-away> :-)
Feb 23 11:56:45 <prpplague> beewoolie-away: the mmc is implemented identical to the app note on the sharp web site
Feb 23 11:57:23 <beewoolie-away> Thanks.
Feb 23 11:57:26 <prpplague> beewoolie-away: what direction are you headed for any ethernet  support??
Feb 23 11:58:05 <beewoolie-away> Well I got working ethernet hardware on...Monday. I haven't had any time to look at integrating the driver into APEX.  My kernel driver seems to work OK.
Feb 23 11:58:23 <beewoolie-away> I'd like to have ethernet and at least tftp.
Feb 23 11:58:33 <prpplague> beewoolie-away: yea, thats where we are headed
Feb 23 11:58:45 <prpplague> beewoolie-away: gonna use the cs8900 in 8-bit polled mode
Feb 23 11:58:53 <beewoolie-away> Really, though, USB is of more interest because it is a production-time feature.
Feb 23 11:59:02 <prpplague> beewoolie-away: yea
Feb 23 11:59:20 <beewoolie-away> Network in APEX is really a dev issue for the most part.  I know there are some applications, but it is a minority.
Feb 23 11:59:20 <prpplague> beewoolie-away: we are implementing a ftdi chipset
Feb 23 11:59:28 <prpplague> agreed
Feb 23 11:59:31 <beewoolie-away> Nice.
Feb 23 11:59:41 <prpplague> its just one of the items on our list
Feb 23 12:00:03 <beewoolie-away> Polled cs8900 would be a good addition.  I plan to get the interrupt code working, too.  It is in, but it doesn't work.  I haven't debugged it yet.
Feb 23 12:00:51 <prpplague> beewoolie-away: http://www.maxim-ic.com/appnotes.cfm/appnote_number/710
Feb 23 12:00:55 <prpplague> beewoolie-away: oops
Feb 23 12:00:58 <prpplague> beewoolie-away: wrong url
Feb 23 12:01:09 <prpplague> beewoolie-away: http://www.dlpdesign.com/usb/usb245.html
Feb 23 12:01:19 <prpplague> beewoolie-away: thats what we are using on the prototypes
Feb 23 12:02:04 <beewoolie-away> Not cheap. but I suppose it's complete and ready for prototyping.
Feb 23 12:02:30 <prpplague> beewoolie-away: yea for proto's its easy
Feb 23 12:02:57 <prpplague> beewoolie-away: $25 isn't that bad
Feb 23 12:03:01 <prpplague> beewoolie-away: for proto's
Feb 23 12:03:51 <prpplague> beewoolie-away: mostly doing research today, getting familiar with your driver structure, should start work on the spi-mmc tomorrow
Feb 23 12:04:35 <beewoolie-away> did you find the overview?
Feb 23 12:05:01 <prpplague> the README_overview?
Feb 23 12:05:07 <beewoolie-away> You found it.
Feb 23 12:05:55 <prpplague> yea
Feb 23 12:08:32 <prpplague> think i might test my mmc code from within my custom bootloader first
Feb 23 12:54:39 <prpplague> dyoung: yo! crack baby
Feb 23 12:55:03 <prpplague> dyoung: have you looked at the code for jelie? 
Feb 23 12:57:20 <[g2]> beewoolie-away, the network driver is a two modules (ixp400 and ixp425_eth)
Feb 23 12:57:47 <beewoolie-away> [g2]: It took a while to find it.  I was expecting it in the kernel tree.
Feb 23 12:58:03 <beewoolie-away> [g2]: I asked because it occurs to me that we shouldn't have to load microcode in the boot loader.
Feb 23 12:58:33 <beewoolie-away> At least, I don't think so.  The suspend/resume code ought to have the oomph to do so.
Feb 23 12:58:50 <beewoolie-away> But there doesn't seem to be support for suspend/resume.  Alas, I'm back where I started.
Feb 23 12:59:27 <dyoung> Jelie?  Nope dunno what one.
Feb 23 12:59:51 <prpplague> http://lapwww.epfl.ch/dev/arm/jelie/
Feb 23 13:01:11 <dyoung> nice.
Feb 23 13:01:16 <dyoung> I'll need to try that.
Feb 23 13:02:48 <beewoolie-away> As it has hard-coded SDRAM init, there will be some work to get it running on the SLUG.  His target uses SDRAM in a 4 chip config.
Feb 23 13:06:33 <beewoolie-away> Still, it is an interesting piece of work that should prove illuminating.
Feb 23 13:10:29 <[g2]> beewoolie-away, the only reason for the ixp drivers in the bootloader is a network recovery
Feb 23 13:10:48 <[g2]> I think a usb-serial-console works fine for recovery
Feb 23 13:10:55 <beewoolie-away> [g2]: are you sure?  I thought that we needed it to init the hardware?
Feb 23 13:11:22 <beewoolie-away> I'm not talking about a driver here.  I'm wanting to get the init complete so that openslug can boot.
Feb 23 13:11:32 <[g2]> beewoolie-away, you're talking about the ixp hw init
Feb 23 13:11:59 <[g2]> well openslug boots, it just doesn't run the ixp properly
Feb 23 13:12:15 <beewoolie-away> Right.  According to another discussion I was reading, the bootloader has to be responsible to the default MAC address, unless it is entered through an initrd.
Feb 23 13:12:35 <beewoolie-away> That's my point.  Can you see the ixp network device with ifconfig -a?
Feb 23 13:13:07 <beewoolie-away> I am wondering if it is only missing a MAC address.  If it is, I think I can load it without loading the firmware.
Feb 23 13:13:12 <beewoolie-away> s/firmware/microcode/
Feb 23 13:13:15 <dyoung> the eth light doesnt come on at powerup.
Feb 23 13:15:40 <beewoolie-away> dyoung: That would be expected without microcode.  That doesn't mean that the kernel driver needs the bootloader to load the microcode.
Feb 23 13:16:04 <beewoolie-away> But it may.  I'm trying to find out if I *have* to load the microcode.
Feb 23 13:16:44 <[g2]> dyoung I'm guessing the ixp drivers reload the microcode
Feb 23 13:16:47 <dyoung> well, it doesnt come on later either... but I can double verify that with otherwise factory firmware later.
Feb 23 13:17:41 <beewoolie-away> [g2]: The only reason I don't (yet) agree is I cannot find where in the driver source it does so. 
Feb 23 13:17:42 <[g2]> but the hardware does need to be initialized and it may need the microcode for that, OR we may be able to just hw enable the ixp in the kernel
Feb 23 13:17:59 <[g2]> beewoolie-away, load the microcode ?
Feb 23 13:18:26 <beewoolie-away> Right.  I didn't see it in the code.  I'm looking again.
Feb 23 13:19:50 <[g2]> beewoolie-away, do you know who gelloglue is ?
Feb 23 13:20:06 <beewoolie-away> Don't.
Feb 23 13:20:37 <[g2]> He's an intel guy working with the community on ipx425 issues
Feb 23 13:20:48 <beewoolie-away> Where do you find him?
Feb 23 13:23:39 <[g2]> http://ixp4xx-osdg.sourceforge.net/
Feb 23 13:25:44 <[g2]> beewoolie-away, did you find something in the code ?
Feb 23 13:27:36 <beewoolie-away> there doesn't seem to be anything in the driver to indicate that the CPU needs to be informed that the NPE is active.
Feb 23 13:27:40 <beewoolie-away> Or should be active.
Feb 23 13:28:08 <[g2]> They are probably assuming "it is alive!"
Feb 23 13:28:35 <[g2]> I'm they do think it's an IXP425 dev board right ? :)
Feb 23 13:28:55 <beewoolie-away> In English please.
Feb 23 13:29:23 <[g2]> I'm joking that they hacked the ixp425 mods in place
Feb 23 13:31:31 <beewoolie-away> [g2]: what I'm going to have to do is inspect the NPE registers after RedBoot has initialized it.
Feb 23 14:11:43 <prpplague> beewoolie-away: jtag dongle will go out tomorrow
Feb 23 14:12:50 <beewoolie-away> prpplague: sweet.  That will make it possible to determine the cause of my JTAG woes.  Thanks go to your boss.
Feb 23 14:13:04 <prpplague> beewoolie-away: hehe
Feb 23 14:13:11 <dyoung> hehe that was classic.
Feb 23 14:13:28 <dyoung> prpplague, did you boss take one look at the beewoolie donngle and have a fit?
Feb 23 14:13:38 <prpplague> dyoung: hehe yea
Feb 23 14:13:48 <dyoung> beewoolie-away, (I think its wonderful myself...)
Feb 23 14:14:04 <beewoolie-away> On the one hand, I'd like it to be my implementation board because it means that I haven't fubar'd my slug.  On the other hand, my vanity is strong.  :-)
Feb 23 14:14:15 <prpplague> hehe
Feb 23 14:14:32 <prpplague> beewoolie-away: i thought you had a bdi2k
Feb 23 14:14:45 <beewoolie-away> prpplague: I do.  No FW for xscale, though.
Feb 23 14:15:05 <prpplague> fw?
Feb 23 14:15:21 <beewoolie-away> Tiersten seems to think he can program flash with a BDI2000 that doesn't have the right FW.  My BDI won't recognize the device at all.
Feb 23 14:15:46 <prpplague> some sort of description file?
Feb 23 14:15:53 <beewoolie-away> Yeah, BDI sells FW for each arch.  I have the ARM7/9 FW.  It's about 1G for the xscale stuff.
Feb 23 14:16:05 <beewoolie-away> No, I think it's a program they download to the BDI hardware.
Feb 23 14:16:05 <prpplague> jeeze
Feb 23 14:16:47 <dyoung> I wonder if kas11 found jelie yet.
Feb 23 14:17:00 <beewoolie-away> It has been worth it because I'm being paid to do the work.  If someone would pay for the xscale work, I'd probably buy the FW.
Feb 23 14:17:06 <dyoung> she could probably pilfer some bits out of it if she decided to hack her FX.
Feb 23 14:17:12 <beewoolie-away> I'd rather get openjag working, though.
Feb 23 14:17:40 <prpplague> beewoolie-away: the trace32 seems to be pretty good
Feb 23 14:17:42 <beewoolie-away> ~seen kas11
Feb 23 14:17:42 <qbot> kas11 <~kas11@193-63.73-24.tampabay.rr.com> was last seen on IRC in channel #openjtag, 8d 23h 29m 5s ago, saying: 'as I recall, you have a problem when the id doesn't fit in 8 bits'.
Feb 23 14:17:55 <beewoolie-away> what is trace32?
Feb 23 14:18:00 <prpplague> beewoolie-away: the linux install sucks, but once you have it running its nice
Feb 23 14:18:34 <prpplague> beewoolie-away: http://www.lauterbach.com/frames.html
Feb 23 14:18:42 <Tiersten> beewoolie-away: I was wrong.  We do have the XScale FW
Feb 23 14:18:48 <dyoung> geez their webpage is horrid.
Feb 23 14:19:10 <prpplague> dyoung: yea
Feb 23 14:19:11 <beewoolie-away> Tiersten: ah.
Feb 23 14:19:28 <prpplague> dyoung: their installation instructions are equally bad
Feb 23 15:06:27 <prpplague> beewoolie-away: btw, the cmd-copy.c doesn't compile properly with 2.95.3
Feb 23 15:06:45 <beewoolie-away> Hmm.  Let me try it.
Feb 23 15:06:50 <prpplague> beewoolie-away: get some preprocessor erros with the -v section
Feb 23 15:07:13 <prpplague> beewoolie-away: i just commented those lines out
Feb 23 15:07:18 <prpplague> beewoolie-away: for now
Feb 23 15:10:17 <beewoolie-away> It looks like a compiler bug.
Feb 23 15:11:09 <beewoolie-away> Or more specifically, a preprocessor bug.
Feb 23 15:13:17 <beewoolie-away> it can be worked-around, though.  will do.
Feb 23 15:14:01 <prpplague> ok
Feb 23 15:14:10 <prpplague> beewoolie-away: just thought i'd let you know
Feb 23 15:24:54 <prpplague> beewoolie-away: is apex going up on the sharpsma.com site?
Feb 23 15:25:41 <beewoolie-away> prpplague: I am posting to a private forum on the Sharp site at the moment.  It is up to Sharp.
Feb 23 15:25:52 <prpplague> ahh
Feb 23 15:28:47 * prpplague heads home
Feb 23 15:28:52 <prpplague> later folks
Feb 23 18:34:01 <dyoung-web> I guess so.
Feb 23 18:34:01 <[g2]> beewoolie-away, what's up in 1.1.1 ?
Feb 23 18:34:14 <dyoung-web> Enhanced RAM Configuration.
Feb 23 18:34:29 <dyoung-web> for our fattened slugs.
Feb 23 18:34:37 <[g2]> cool.
Feb 23 18:34:52 <[g2]> beewoolie-away was asking about the chip configs
Feb 23 18:34:59 <dyoung-web> today?
Feb 23 18:35:08 <[g2]> last night
Feb 23 18:35:13 <[g2]> :)
Feb 23 18:35:17 <dyoung-web> missed it.
Feb 23 18:35:22 <[g2]> it's in the log
Feb 23 18:44:45 <dyoung-web> okay, gotcha.
Feb 23 18:44:48 <dyoung-web> I was there for that.
Feb 23 19:09:58 <ka6sox-web> dyoung-web: ping
Feb 23 19:13:12 <dyoung-web> [g2] around?
Feb 23 19:13:34 <dyoung-web> drat he left.
Feb 23 19:21:30 <beewoolie-away> Well, a little late.  Yeah, dyoung got it right. 1.1.1 has a change such that a fat slug may be configured by uncommenting one line in the config file.  At this point, we still need a build that is specific to the SDRAM configuration.
Feb 23 19:29:08 <ka6sox-web>                                                               he's baaaack!
Feb 23 19:29:57 <dyoung-web> [g2] around/
Feb 23 19:30:08 <[g2]> around
Feb 23 19:30:18 <dyoung-web> Okay dumb uestion foryou...
Feb 23 19:30:29 <dyoung-web> so I wrote this bb file....
Feb 23 19:30:46 <dyoung-web> one of the configure options is nslu2 specific.
Feb 23 19:30:55 <[g2]> hey beewoolie-away did you see http://genux.org/index.php ?
Feb 23 19:31:44 <dyoung-web> so I want to do sometihng like if MACHINE=nslu2 then EXTRA_OECONF="--enable-nslu2" otherwise dont do it.
Feb 23 19:32:00 <dyoung-web> in the context of the bb.  do you know an example offhand I could look at?
Feb 23 19:33:39 <[g2]> dyoung-web, I can find one real fast
Feb 23 19:35:22 <[g2]> base-files_3.0.14.bb matches
Feb 23 19:39:16 <dyoung-web> Thanks, so I can use that sort of syntax..
Feb 23 19:39:58 <dyoung-web> so if I want to add a package whats the protocol?
Feb 23 19:40:23 <dyoung-web> mail it to kergoth or something/
Feb 23 19:40:26 <[g2]> dyoung-web, a new package ?
Feb 23 19:40:31 <dyoung-web> yes,
Feb 23 19:40:43 <dyoung-web> give it to you?
Feb 23 19:40:48 <dyoung-web> change-sets it?
Feb 23 19:40:50 <[g2]> check it into the nslu2-linux bk repo and kergoth can pull upstream
Feb 23 19:40:57 <dyoung-web> is that kosher?
Feb 23 19:41:35 <dyoung-web> I didnt know if I was supposed to be pushing "any kine stuff" into the repo.
Feb 23 19:41:46 <[g2]> dyoung-web, sure. You'll have kergoth's wrath to deal with if it's hosed, but a new package is quite unlikely to hose stuff
Feb 23 19:42:01 <[g2]> which package ?
Feb 23 19:42:06 <dyoung-web> mt-daapd.
Feb 23 19:42:20 <dyoung-web> The only thing stopping me from upgrading my 2nd production slug to openslug today.
Feb 23 19:42:53 <[g2]> what does openslug need to support it ?
Feb 23 19:42:59 <dyoung-web> a package.
Feb 23 19:43:04 <dyoung-web> the package rather
Feb 23 19:43:18 <[g2]> kernel changes ?
Feb 23 19:43:23 <[g2]> other packages ?
Feb 23 19:43:27 <dyoung-web> Nope
Feb 23 19:43:31 <dyoung-web> well, other packages yes
Feb 23 19:43:40 <beewoolie-away> [g2]: Not much to see.
Feb 23 19:43:41 <dyoung-web> zlib gdbm and libid3tag
Feb 23 19:44:09 <ka6sox-web> samm71 is interested in building a Maxxed out Slug
Feb 23 19:44:33 <[g2]> he's the guy that did php/apache right ?
Feb 23 19:45:11 <[g2]> zlib builds
Feb 23 19:45:20 <ka6sox-web> dunno...
Feb 23 19:45:44 <[g2]> ka6sox-web, that was a statement
Feb 23 19:45:55 <beewoolie-away> [g2]: What makes genux interesting?
Feb 23 19:46:10 <[g2]> it's paid OE
Feb 23 19:46:14 <[g2]> with binary distro
Feb 23 19:46:19 <[g2]> gentoo based
Feb 23 19:46:20 <dyoung-web> all 3 of those packages build.
Feb 23 19:47:43 <ka6sox-web> he has the chips already.
Feb 23 19:48:28 <ka6sox-web> dyoung-web : will you be in the office still in 30 mins?
Feb 23 19:48:40 <dyoung-web> probably not
Feb 23 19:48:42 <dyoung-web> getting sleepy.
Feb 23 19:48:46 <dyoung-web> I have pcb at home though.
Feb 23 19:48:47 <ka6sox-web> oky
Feb 23 19:49:03 <ka6sox-web> okay will you be on then?
Feb 23 19:49:12 <dyoung-web> I could be.
Feb 23 19:49:21 <dyoung-web> ~kick dyoung
Feb 23 19:49:22 <qbot> dyoung-web: are you smoking crack?
Feb 23 19:49:30 <ka6sox-web> are you sleepy?
Feb 23 19:49:56 <ka6sox-web> I stayed up until about 3 am looking for you last night.
Feb 23 19:50:19 <ka6sox-web> nite dyoung
Feb 23 19:50:32 <beewoolie-away> nite
Feb 23 19:50:41 <[g2]> dyoung-web are you done with pkg devel for today ?
Feb 23 19:51:07 <dyoung-web> I'm temporarily hung up in testing it because sourceforge is telling me the source file aint there.
Feb 23 19:51:18 <dyoung-web> ie: sf is busted.
Feb 23 19:51:28 <ka6sox-web> crud
Feb 23 19:52:32 <[g2]> so which devices support mt-daapd
Feb 23 19:53:05 <[g2]> ~mt-daapd
Feb 23 19:53:05 <qbot> bugger all, i dunno, [g2]
Feb 23 19:53:46 <[g2]> digital audio a player daemon
Feb 23 19:53:51 <[g2]> a...
Feb 23 19:53:53 <dyoung-web> Ooh maybe not.
Feb 23 19:54:04 <dyoung-web> I found the source file on builds.
Feb 23 19:54:20 <dyoung-web> Well, iTunes does...
Feb 23 19:54:33 <dyoung-web> so there fore the roku can drive it too.
Feb 23 19:56:05 <dyoung-web> ka6sox, am I hallucinating or is a chip missing?
Feb 23 19:56:23 <ka6sox-web> uh...nope
Feb 23 19:56:36 <ka6sox-web> not as far as I know.
Feb 23 19:56:53 <dyoung-web> Hmm
Feb 23 19:57:31 <ka6sox-web> the big blob in the lower right corner is is the regulator.
Feb 23 19:57:51 <dyoung-web> ADS8344
Feb 23 19:57:52 <dyoung-web> ?
Feb 23 19:58:12 <ka6sox-web> hmmm....it was there in the 9am feed
Feb 23 19:58:26 <dyoung-web> oh, didnt get that one.
Feb 23 19:58:33 <ka6sox-web> I'll resend
Feb 23 19:58:36 <ka6sox-web> when I get in.
Feb 23 19:58:51 <dyoung-web> RoadRammer is irritating.
Feb 23 19:59:02 <dyoung-web> none of these mesages made it to that account.
Feb 23 19:59:02 <ka6sox-web> I sent it to both
Feb 23 19:59:21 <dyoung-web> I didnt get it at the other place either at 9 though.
Feb 23 19:59:28 <ka6sox-web> that why I"m sending it to both.
Feb 23 20:10:30 <dyoung-web> well it built.
Feb 23 20:10:43 <ka6sox-web> good
Feb 23 20:10:45 <dyoung-web> I guess I need to find a voulnteer or go ahead and upgrade it.
Feb 23 20:11:25 <ka6sox-web> the ADS and LT1641 are there.
Feb 23 20:11:30 <[g2]> dyoung-web, mt-daap built ?
Feb 23 20:11:35 <[g2]> mt-daapd
Feb 23 20:11:40 <dyoung-web> [g2] yeah, but not tested yet.
Feb 23 20:12:07 <dyoung-web> it was ridiculously easy to add it.
Feb 23 20:12:08 * [g2] thinks dyoung-web is warming up to OE
Feb 23 20:12:50 <[g2]> what's machine specific about the package build though ?
Feb 23 20:12:58 <dyoung-web> I didnt look. ;-)
Feb 23 20:13:07 <dyoung-web> I just know theres a --enable-nslu2 option
Feb 23 20:14:42 <dyoung-web> ah crap
Feb 23 20:14:46 <dyoung-web> I didnt even need it.
Feb 23 20:15:03 <[g2]> need what ?
Feb 23 20:15:12 <dyoung-web> --enable-nslu2
Feb 23 20:15:19 <[g2]> you probably don't need it
Feb 23 20:15:22 <dyoung-web> its does 2 things.
Feb 23 20:15:40 <dyoung-web> one is unslung specific.  another is probably a good thing though.  always_scan=0
Feb 23 20:16:25 <dyoung-web> But since oe is targeted toward embedded stuff, maybe I'll leave that scan stuff in as a default.
Feb 23 20:16:39 <[g2]> right we lose the unslung specific thing and add a patch that sets always_scan=0 for everybody
Feb 23 20:17:10 <[g2]> you beat me to it :)
Feb 23 20:18:12 <[g2]> dyoung-web, congrats
Feb 23 20:18:28 <dyoung-web> well, not yet.  someone needs to test it first.
Feb 23 20:19:08 <[g2]> beewoolie-away, this is quite interesting http://www.kurobox.com/forums/viewtopic.php?t=277
Feb 23 20:19:57 <[g2]> looks like it might be 2.4 based though
Feb 23 20:22:36 <rand> is ka6sox here, in the office, or away - I can't tell
Feb 23 20:22:45 <rand> probably just busy
Feb 23 20:23:44 <dyoung-web> And it turns out it may not be needed.  theres a parameter in the config file
Feb 23 20:23:59 <ka6sox> heya rand.
Feb 23 20:24:37 <rand> hi Mr. ka6sox
Feb 23 20:25:09 <ka6sox> what's up tonight?
Feb 23 20:38:13 <[g2]> dyoung-web, GL
Feb 23 20:39:19 <[g2]> dyoung-web, you can just bb openslug-image after you build the other modules and then the ipkg Packages info get updated
Feb 23 20:51:21 <rand> good night ka6sox
Feb 23 20:52:15 <ka6sox> night rand
Feb 23 22:11:34 <dyoung> back
Feb 23 22:18:23 <beewoolie-away> [g2], if you read a log, you'll see this.
Feb 23 22:19:29 <beewoolie-away> the kurobox thing isn't really so difficult.  I am not sure if it is that interesting, though.  Our plan to read the kernel from the filesystem is, IMHO, way more interesting.
Feb 23 22:23:43 <beewoolie-away> dyoung: lots of activity today.
Feb 23 22:23:54 <beewoolie-away> dyoung: did I miss anything important?
Feb 23 22:24:29 <dyoung> Hmm, not really.
Feb 23 22:24:47 <dyoung> a lot of that was to firm up the migration plan for unslung users to get to openslug.
Feb 23 22:25:14 <beewoolie-away> k.
Feb 23 22:26:41 <beewoolie-away> Did [g2] ever receive his fat slug?
Feb 23 22:27:08 <dyoung> not that I know of, but not sure.
Feb 23 22:27:31 <beewoolie-away> But you're going to get one.  Or make one?
Feb 23 22:40:47 <dyoung> I'm gonna get ka6sox to mod one for me because I dont want to fork out for a rework station rightn ow.  heh,.
Feb 23 22:41:46 <samms71> dyoung...ka6sox was telling me you are working on the fatslug project...
Feb 23 22:42:05 <dyoung> Kind of.
Feb 23 22:42:09 <dyoung> I dont actually have one yet.
Feb 23 22:42:13 <samms71> have you built a unslung image w/  greater than 32MB configured?
Feb 23 22:42:30 <dyoung> But I am helping bewoolie test out apex on the nslu2 target.
Feb 23 22:42:38 <dyoung> Yes.
Feb 23 22:42:42 <dyoung> I dont know if it works.
Feb 23 22:42:56 <dyoung> We attempted to boot it; but it freaks out.
Feb 23 22:43:23 <dyoung> but I dont know if thats because the memory map never got initialized.. or some other problem.
Feb 23 22:43:30 <samms71> ah...i have the memory...i don't want to install it until i get a unslung image that will use it... how much mem did you config?  how was the mem setup on the extra chipselect?
Feb 23 22:43:58 <dyoung> for that one, he replaced the 2x8M16 parts with 2x16M16 parts.
Feb 23 22:44:23 <dyoung> so I changed it to be 64M at config line.
Feb 23 22:44:49 <beewoolie-away> samms71: It isn't the unslung image that cares about the memory size, it's the bootloader.
Feb 23 22:44:56 <samms71> hmm. so the fatslug is not slugging yet...
Feb 23 22:44:58 <beewoolie-away> samms71: More specifically, there are two pieces.
Feb 23 22:45:33 <samms71> beewoolie...yeah.. i was hoping dyoung had built a redboot image that worked..
Feb 23 22:45:38 <beewoolie-away> samms71: The SDRAM controller has to be programmed to drive the memory.  That's only doable by the bootloader because it has to be done before anything runs in SDRAM
Feb 23 22:45:50 <beewoolie-away> No one has gotten redboot to work, IIRC
Feb 23 22:45:53 <dyoung> I did that too; but its not clear to me if it works or not.  I pretends to work.
Feb 23 22:46:01 <beewoolie-away> You can try APEX without reflashing.
Feb 23 22:46:16 <beewoolie-away> dyoung: Does the kernel think there is more RAM?
Feb 23 22:46:30 <dyoung> the kernel pretends that theres more ram.
Feb 23 22:46:30 <beewoolie-away> samms71: The nice thing about APEX is that it will tell you what it can find.
Feb 23 22:46:32 <dyoung> via cmdline
Feb 23 22:46:45 <beewoolie-away> dyoung: free says you've got more RAM
Feb 23 22:46:48 <beewoolie-away>  ?
Feb 23 22:46:58 <dyoung> I gave up because its impossible test anything without having one in hand
Feb 23 22:47:01 <beewoolie-away>  /usr/bin/free
Feb 23 22:47:03 <ka6sox-away> with redboot it sees the new ram.
Feb 23 22:47:11 <beewoolie-away> Then it works.
Feb 23 22:47:23 <ka6sox-away> but either its NOT passing the kernel args correctly or something else.
Feb 23 22:47:24 <dyoung> iirc, ka6sox loaded stuff into higher memory.
Feb 23 22:47:34 <dyoung> it doesnt pass kernel args anyways.
Feb 23 22:47:34 <beewoolie-away> Which kernel were you using?
Feb 23 22:47:42 <dyoung> so theres something weird going on.
Feb 23 22:47:49 <beewoolie-away> right.  The old kernels don't accept args.  
Feb 23 22:48:09 <ka6sox-away> I'd have to ask jacques...he sent me the last kernel I tried to boot.
Feb 23 22:48:46 <beewoolie-away> OK.  I think that redboot will also tell you what it finds.  Even though it doesn't pass the information, I'd guess it was somewhat reliable about telling you what it knows.
Feb 23 22:49:09 <ka6sox-away> loaded and verified himem
Feb 23 22:49:31 <ka6sox-away> so the issue is something else.
Feb 23 22:49:55 <ka6sox-away> when [g2] gets his slug he can start testing :)
Feb 23 22:49:55 <beewoolie-away> Most likely the kernel not accepting the memory size parameter.
Feb 23 22:50:03 <beewoolie-away> :-)
Feb 23 22:50:09 <ka6sox-away> beewoolie-away, I agree.
Feb 23 22:51:58 <beewoolie-away> I haven't been able to put any real time to this project.  When I do, I'll see what I can discover about running openslug.  I know that there is a conundrum.  If redboot isn't passing valid args, then we may be in a pickle.
Feb 23 22:52:09 <beewoolie-away> That is without an alternative boot-loader.
Feb 23 22:54:25 <samms71> hmm....no wonder i like electronics more than software. ;) it's easier..
Feb 23 22:55:04 <dyoung> I've kind of given up hope for redboot.
Feb 23 22:55:11 <dyoung> Apex is just so much nicer.
Feb 23 22:55:25 <dyoung> Just that one little thing to overcome.  heh.
Feb 23 22:56:05 <samms71> so is apex stable to use if you figure out the problem?
Feb 23 22:56:31 <beewoolie-away> samms71: I wouldn't flash it until we've overcome the issue with the NPE init.
Feb 23 22:56:42 <samms71> npe==?
Feb 23 22:56:45 <dyoung> samms71, are you serial enabled?
Feb 23 22:56:47 <beewoolie-away> samms71: You can test it from ram
Feb 23 22:57:11 <beewoolie-away> NPE Network Processing Engine.  It's what runs the ethernet MAC.
Feb 23 22:57:14 <samms71> yeah...i have the hardware..i was just waiting to mod the ram and serial port at the same time so I only have to crack the case once..
Feb 23 22:57:22 <samms71> beewoolie..thx...
Feb 23 22:57:32 <ka6sox-away> samms71 its easy to crack the case.
Feb 23 22:57:38 <ka6sox-away> and put back.
Feb 23 22:57:39 <ka6sox-away> :)
Feb 23 22:57:48 <beewoolie-away> Ah.  OK.  In your position, I'd probably wait.  We've got a week or two before this is settled out.
Feb 23 22:57:57 <samms71> it's more  a matter of doing it all at once when the baby is sleeping...;)
Feb 23 22:57:57 <dyoung> its also pretty easy to break your case if you gorilla handle it like I did.
Feb 23 22:58:27 <samms71> i figure if i break it, some nice plastic standoffs and it'll be retro-industrial pcb on my shelf w/ blinky lights..
Feb 23 22:58:40 <beewoolie-away> I hear you.  It's nice to get the ducks in a row.  Do you plan to let the serial cable be available when the box is closed?
Feb 23 22:59:28 <samms71> yeah, i'll snake it through the usb2 hole...i did a max3227 converter a while ago for another project so i have a bunch of them on my desk.
Feb 23 22:59:47 <beewoolie-away> If so, there isn't really a reason to wait to do the mods.  Once closed, you'll be able to do the upgrades later.
Feb 23 23:00:27 <beewoolie-away> Then again, if your concern is testing that the memory is recognized...I'd probably still wait.
Feb 23 23:00:36 <beewoolie-away> It won't be long.
Feb 23 23:00:46 <beewoolie-away> Night folks.
Feb 23 23:00:51 <samms71> beewoolie.true..i'll probably take the slug to work this week and get it done...i was telling ka6sox i got some 512Mb parts from micron (32x16) for my slug...want to see a 128MB slug
Feb 23 23:00:55 <samms71> night...thx
Feb 23 23:00:56 <dyoung> night bewoolie!
Feb 23 23:01:12 <beewoolie-away> 128MB  damn.
Feb 23 23:01:32 <dyoung> I wouldnt mind a MaxSlug.
Feb 23 23:01:54 <samms71> don't know if it'll even work...no reason it shouldn't but i'm not familiar w/ arm....MaxSlug..good name...FatSlug=64MB  MaxSlug=128MB
Feb 23 23:01:59 <dyoung> but I could live with 128M....
Feb 23 23:02:29 <dyoung> doesnt 32M16 stacked give you 256M total?  or were not planning to stack um?
Feb 23 23:02:35 <dyoung> can you even stack that part?
Feb 23 23:02:37 <dyoung> hmm
Feb 23 23:02:40 <samms71> Well, If i don't trash anything I'll have a couple of chips left over..i told ka6sox he can have two but i still have two left.
Feb 23 23:02:52 <dyoung> I see!
Feb 23 23:02:53 <samms71> i think you could stack it and run a seperate chipselect for the top part...
Feb 23 23:03:13 <samms71> not sure, can the arm handle that much memory? (i mean the linksys arm we're using)
Feb 23 23:03:16 <dyoung> I'm pretty sure it'll go on the linux side.
Feb 23 23:03:26 <dyoung> For openslug.
Feb 23 23:03:41 <dyoung> According to dsaxena, it should be doable with our 2.4 tree.
Feb 23 23:04:09 <samms71> whoa...SuperMaxSlug=256MB..damn that's more memory than my first three pc's combined.
Feb 23 23:04:18 <dyoung> would you run unslung on it?  or openslug?
Feb 23 23:05:14 <samms71> if openslug is stable, then openslug otherwise unslung. i'm running apache w/ php on my slug serving gallery (picture gallery) so I want it to work for that first..
Feb 23 23:05:48 <samms71> i'm hoping the extra memory let me than run radiusd, qmail and apache w/o having the little slug thrash.
Feb 23 23:06:15 <samms71> and some other things like twonky, nfs and samba
Feb 23 23:07:01 <dyoung> heh
Feb 23 23:07:05 <dyoung> "Some other things..."
Feb 23 23:07:15 <samms71> ;)
Feb 23 23:07:50 <samms71> i should just pony up the cash and buy a SFF pc w/ linux on it but getting the slug to do this stuff is so much more fun...
Feb 23 23:10:35 <dyoung> sff?
Feb 23 23:10:44 <dyoung> ah
Feb 23 23:10:47 <dyoung> small form factor
Feb 23 23:10:54 <dyoung> Yeah, it is fun!
Feb 23 23:12:05 <samms71> agreed..so is there a howto to get APEX loaded into a slug? i get into redboot then???
Feb 23 23:12:34 <dyoung> if you have serial, you can use redboot to xmodem it in.
Feb 23 23:12:38 <dyoung> the execute it from ram.
Feb 23 23:12:48 <dyoung> its on beewoolie's wiki.
Feb 23 23:13:03 <samms71> ok...i'll find it...it's on the apex wiki...?
Feb 23 23:14:24 <dyoung> http://wiki.buici.com/twiki/bin/view/Main/ApexBootloader
Feb 23 23:14:33 <dyoung> halfway down the page theres a nslu2 section.
Feb 23 23:14:53 <samms71> thx...ok...time for bed..night..
Feb 23 23:16:02 <dyoung> nitey
**** ENDING LOGGING AT Wed Feb 23 23:59:56 2005