irclog2html for tuxscreen on 2001.10.12

irclog2html for tuxscreen on 2001.10.12 00:26:56 sammy hello , all 00:27:36 Tangent sammy : Hi 05:38:00 BZFlag nite 06:40:49 scanline is away: . o O (zzzzzzzzzzzz) 06:54:14 GoRK greetings 07:05:21 Lenolium I guess I'm here too late to ask to have two tuxscreens brought to Lineo tommorow, eh? 07:05:48 GoRK probably 10:40:42 mallum hi SpaceCoaster 10:43:17 SpaceCoaster mornin 10:43:27 mallum afternoon :) 10:43:44 SpaceCoaster hmmm 10:45:50 SpaceCoaster whatzup, do anything interesting? 10:46:03 mallum at work .... 10:46:08 mallum covertly working on xkbd 10:46:35 mallum trying to figure out how clip_masks work in xlib 10:49:01 SpaceCoaster xkbd looks neat, wow just 16k 10:49:09 mallum thx 10:49:17 mallum I've just rewritten it 10:49:25 mallum its only in cvs at the mo 10:49:42 mallum it does a lot more like AA fonts, resizing, more configurability etc 10:50:15 SpaceCoaster AA fonts? 10:50:17 mallum I was hoping t would be smaller as well, but I dont think it will be 10:50:20 mallum yep 10:50:35 mallum http://mallum.com/img/xkbd.png <-- in production 11:01:27 mallum gets clip_masks working and is happy .... 11:01:55 mallum is away: celebratory ciggarette 11:09:11 mallum is back (gone 00:07:15) 13:19:38 mallum hi prpplague 13:19:44 prpplague howdy all 13:19:48 prpplague hows it going mallum? 13:20:14 mallum good good 13:29:31 GPSFan Morning all 13:30:20 GPSFan I've been poking around on the Tux mainboard, Brutus schematics, and various data sheets and have found some interesting things. 13:31:27 GPSFan 1) I had a bad solder joint on IC801, a Ti 7733 reset generator. It was pin 8 VCC, but must have been making contact because my reset worked. 13:31:56 GPSFan It looks like my board was hand soldered so there maybe more bad soldering on other boards. 13:33:03 GPSFan 2) In paragraph 10.1 of the SA1100 spec it says that if multiple banks of memory are used ttha each must be identical in size & configuration. 13:34:08 GPSFan This means that since there are 2 banks of 4Mbytes already on the Tux, that the SODIMM can't be anything but an 8MByte device. 13:35:03 GPSFan This begs the question of how did Shaggy have a larger memory operating?? 13:35:54 prpplague GPSFan: someone spoke about removing the onboard ram... 13:35:58 GPSFan 3) The PCMCIA configuration is somewhat different from what is suggested in the Brutus doc's, there is only one set of buffers for the 2 sockets, and the address & data 13:36:30 GPSFan pins of both sockets are tied together. THe control signals are seperately buffered however. 13:36:58 AlHaz GPSFan: the 32meg sodimm works in inferno somehow 13:38:03 GPSFan The VCC1, VCC2 are tied together on each socket but brought seperately to the I/Oboard so a 5V mod could be done by seperating these lines and adding a VCC controller. 13:39:05 GPSFan AlHaz, yes I know people have reported larger memory working, but I'm jusat reporting what the SA1100 spec says. What that leaves us to do is figure out how it works. 13:40:21 GPSFan There is only one set of timing controll registers on the SA1100, so the memories must refresh the same no matter what the size. 13:45:26 Tangent AlHaz : I heard that the problem may not be the VCC switching matrix, but voltage level shifts required for 5V card to talk to 3.3V ARM 13:46:02 Tangent GPSFan : I worked out a suitable VCC controller using just two FETS (since we already have regulated 3.3 and 5V supplies) 13:46:48 GPSFan Tangent: there are 162245 buffers between the SA1100 and the PCMCIA slots, this is similar to what's shown in the Brutus docs' 13:47:12 GPSFan and Brutus supports 3.3 & 5V PCMCIA. 13:47:19 Tangent GPSFan : Oh.. cool.. so they really did just miss out the VCC switching matrix to save $1 or so 13:48:27 GPSFan Tangent: it looks like the Mainboard may have been set up to support multiple I/O boards because on the I/O board is where all the VCC's are tied together. By 0ohm resistors. 13:49:30 GPSFan BTW the Tux blockdiagram (cave_drawings) doesn't show those buffers, but I have rung the lines out with my trusty ohmmete. 13:49:45 Tangent GPSFan : Well.. the datasheet for the 162245 says it does exactly what we want... i.e. takes 5V logic signals too :) 13:50:46 Tangent GPSFan : Which I figure means that we're in business 13:51:03 GPSFan Tangent: I believe the I/O board could be eaisly modified to do a 3.3/5V PCMCIA, but the control signals have to be generated bu GPIO pins. 13:51:16 Tangent GPSFan : That is my belief too... 13:51:30 Tangent GPSFan : I looked at a schematic for a switiching matrix chip... 13:51:44 GPSFan Brutus uses a MAX 1600 13:52:01 Tangent GPSFan : The only bit of it we actually need is two low resistance FETS at the last stage that switch the +5 or +3.3V rails to the VCC pin 13:52:08 GPSFan I haven't looked at Assabet yet. 13:52:16 Tangent most of the switch matricies seemed a bit overkill for what we needed 13:52:28 Tangent checks out the MAX1600 13:53:43 Tangent GPSFan : Do we have 12V present on the Tux board? 13:54:34 Tangent And for that matter... do we care about controlling the VPP pin, or is that already handled? 13:55:26 GPSFan Tangent: We would only care about 12V if we wanted to support 12Vflash cards. 13:56:14 Tangent GPSFan : Check out the drive stage on the LTC314.... http://www.linear-tech.com/pdf/ic17.pdf 13:56:28 Tangent GPSFan : I've got a supply of fets here that I can solder right in... 13:56:40 Tangent GPSFan : All we'd need is an I/O line to control them with 13:56:48 GPSFan VPP1 &VPP2 are seperated on each PCMCIA card but tied to 3.3V on the I/O board 13:57:35 GPSFan The LT part looks less complicated than the MAX1600 13:57:54 Tangent Nod... but both parts seem to assume that 12V supply will be available 13:58:23 Tangent The LTC1314 just uses 3 fets to swing the VCC rails around tho... external to the chip itself... 13:58:49 Tangent Which I'm more than confident I can piece together here, as I've just spend the last month playing with FETs 13:59:41 GPSFan Look at the 1470 it doesn't use 12V. 14:00:01 Tangent GPSFan : Indeed... that looks like exactly what we'd need 14:00:24 Tangent Now all we need is a supplier 14:00:48 GPSFan Newark, possibly 14:00:58 GPSFan I'll go look 14:04:44 GPSFan Linear will sell direct for prototype quantities through their online store! 14:05:27 Tangent GPSFan : Cool... and here's the propper dataheet for it... http://www.linear-tech.com/pdf/lt1470.pdf 14:06:04 GPSFan Tangent: 8pin SO package, not bad to work with either. 14:06:22 Tangent GPSFan : Yeah... should be fine :) 14:06:31 Tangent GPSFan : Like fitting a modchip ;) 14:07:57 GPSFan Tangent: now we need 2 GPIO bits to do the switching. 14:08:21 Tangent GPSFan : Hopefully they shouldn't be too hard to fine 14:08:25 GPSFan and of course the right code to make it all work. 14:08:26 Tangent s/fine/find 14:09:26 Tangent GPSFan : If I understood how the iPAQ's auto-VCC selection worked, I'd rip off that idea :) 14:10:33 GPSFan Tangent: what about ripping off the drivers which should give a good indication of the underlying hardware. 14:11:17 Tangent GPSFan : I know the PCMCIA stuff on the iPAQ just requests 5.0V every time, and the sleeve hardware just knows what the card should be getting 14:12:22 GPSFan Tangent: Hmmmm how does it know, I wish I still had access to the PCMCIA spec, it might help. 14:13:13 Tangent GPSFan : Almost certainly would... I never understood how we're supposed to read the card VCC spec until after we apply power to it in the first place 14:14:56 GPSFan Tangent: I rember reading in the spec about that, I believe a compliant card was supposed to operate just enough to allow you to read all the tupples 14:15:19 Tangent GPSFan : Ah.. so we always start off at 5V then? 14:15:22 GPSFan then you would know what voltage to set the control bits to. 14:15:34 GPSFan or 3.3V? 14:16:05 GPSFan It's a pity that the PCMCIA org is so coin operated. 14:16:41 GPSFan The spec was a pay $$$ for and uncopyable (blact text on blue paper). 14:17:58 GPSFan s/blact/black 14:21:26 Tangent GPSFan : Yeah... I was trying to find the info online a week or so ago, and ran up against the $$$ barrier 14:22:49 GPSFan Tangent: I had 2 copies where I used to work, didn't bring one :>( 14:23:45 GPSFan Tangent: did you see my away message? I'm still trying to get the hang of IRC? 14:24:07 Tangent GPSFan : Nope.. didn't notice the Away message... 14:24:16 Tangent doesn't really understand IRC either 14:24:42 GPSFan Tangent: Hmmm, ??? 14:25:06 Tangent GPSFan : Still got friends where you used to work? 14:26:45 Tangent GPSFan_away : That's the easy solution :) 14:43:36 prpplague Tangent / GPSFan : ya the PCMCIA org is really big on getting their $$ 14:46:10 CosmicPenguin prpplague: Thats because they thing they're the Next Big Thing (TM)... and in a way, they are... 14:46:18 CosmicPenguin prpplague: Still, no excuse.... 14:53:59 Tangent I just amazed that given the amount of time PCMCIA has been around, the specs aren't freely available 14:54:58 AlHazWork they are freely available 14:55:02 AlHazWork for $150 in a nice hardback book 14:55:29 Tangent $150 != free 14:55:39 AlHazWork Tangent: hey, at least it's not an NDA 14:55:44 mallum Tangent: whens your birthday ? 14:55:54 Tangent mallum : June 21 14:56:04 AlHazWork and it's only $100 if you join their little club 14:56:19 Tangent PCMCIA socket owners club? 14:57:28 prpplague ebay got a nasty gram for the PCMCIA ppl about auctions on those specs about 18 months ago 14:59:14 Tangent So... freely available as long as you pay _them_ $150 then 14:59:54 mallum Tangent: can you wait that long ? 14:59:57 AlHazWork the CF spec is freely available, but big pieces are missing under the heading "see pcmcia spec" 15:00:10 Tangent mallum : Yeah... I can wait :) 15:00:12 prpplague AlHazWork: ya i notice that too 15:06:01 Tangent Aha.. found the stuff for AutoVCC 15:06:09 Tangent There's two new lines... VS1 and VS2 15:06:36 Tangent I the card is capable of 3.3V, VS1 is grounded, and VS2 is floating 15:07:11 Tangent for a 5V card, both are floating 15:11:26 AlHazWork that doesn't sound difficult to implement 15:12:11 Tangent Yeah.. all I need now is some way of knowing if a card is in the slot or not... 15:12:38 Tangent since otherwise we'll end up feeding 5V to a 3.3V card for a fraction of a second when it's first inserted 15:13:47 GPSFan Tangent: there are card detect lines thjat let you know a card is present. 15:14:24 Tangent Cool... so we can have VCC at 0V... detect a card inserted.. wait x milliseconds.. then switch the appropriate rail to VCC 15:14:43 AlHazWork the pcmcia spec says how many milliseconds, even 15:14:52 AlHazWork iirc it's <100 15:15:12 Tangent Cool.. this should be pretty easy :) 15:15:21 GPSFan Tangent: Usually ona PeeCee the card detect lines are tied to some type of interrupt, so the driver gets alerted when there is a card insertion, or extraction. 15:16:02 Tangent GPSFan : All I care about is some line that will go to 0V when the card gets plugged in, so I can trigger VCC circuit 15:17:08 Tangent Damn server splits 15:17:18 Tangent GPSFan : All I care about is some line that will go to 0V when the card gets plugged in, so I can trigger VCC circuit 15:19:11 GPSFan Tangent: Pin 36 is CD01# and Pin 67 is CD02# on the PCMCIA Socket, I'll see if I can find out where they go on the Tux. 15:19:57 Tangent The only 5V card I have to test with is an ancient Dlink PCMCIA BNC etherenet card... but it worked on the iPAQ :) 15:21:34 GPSFan Tangent: we could hack the I/O board to only supply +5v and if you only used a 5V card that should tell us if the technique will work. 15:21:45 pattieja Tangent: mmmmm. BNC. 15:21:50 GPSFan Which of course it should ;>) 15:22:08 Tangent GPSFan : Yep.. that's an easy hack that I'm happy to do... 15:22:40 Tangent As long as the PCMCIA driver won't dumbly fail because it thinks the hardware can only do 3.3V or something 15:23:25 AlHazWork most people with these phones will probably be content to use the same card in their phone for a long time 15:23:33 GPSFan Tangent: I've heard of people putting %V cards in the Tux and having them "almost" work. 15:25:03 GPSFan Tangent: if you look at the I/O board the 0ohm resistors are R338,340,337,339. I don't know which is which yet. It may take me a while to ring out 'cause I'v got to go out again soon:>( 15:25:31 GPSFan s/%/5 15:26:39 GPSFan I wouldn't mind if my Tux only did 5V cards. 15:26:58 Tangent GPSFan : I won't be tackling it tonight I'm afriad... I'm off for the weekend soon 15:27:55 AlHazWork GPSFan: especially considering that 3.3v-only cards are rare 15:28:02 GPSFan Tangent: I'll putz around this weekend and see if I can generate a partial schematic of the PCMCIA power card ident circuits. 15:28:20 Tangent GPSFan : Excellent.. that'd be most helpful :) 15:28:22 GPSFan 802.11b CF cards spring to mind. 15:29:25 CosmicPenguin AlHazWork: I would expect that trend to reverse sharply...... 15:29:40 AlHazWork CosmicPenguin: that could definately happen 15:29:48 Tangent GPSFan : You could always stick a couple of dip switches at the back of the Tux and have manual VCC selection 15:30:26 GPSFan Tangent: been there done that (not on Tux yet though) 15:30:30 AlHazWork CosmicPenguin: there are definately some out there 15:31:20 CosmicPenguin AlHazWork: Yeah.... I agree that full 3.3V PCMCIA cards a little stupid, but I think that CF will take over very soon... 15:31:25 GPSFan The problem with hard wireing the VCC to 5V is hot insertion/extraction is not quite as safe. 15:31:46 Tangent GPSFan : Yep 15:33:59 AlHazWork well, shaggy said auto-vcc wasn't hard or expensive to add 15:34:23 AlHazWork I think a lot of us are capable of aquiring an IC or two, soldering them to a daughterboard, cutting a few traces, etc 15:34:59 AlHazWork we need some more hardware guys working on this . . . . we need to get duck's friend curt more involved 15:35:05 Tangent AlHaz : It really does look like it'd need about 8 transistors to implement... 15:35:21 Tangent AlHaz : or a couple of LT1470's and a bit of code 15:36:20 AlHazWork if y'all come up with a pretty good diagram of how it should work, i could get that information to curt 15:37:21 Tangent AlHaz : Did you see the URL for the 1470 datasheet I posted a little while ago? 15:37:26 GoRK 802.11a mmmmmmmmmmm 15:37:39 Tangent GoRK : 54Mbps... nice :D 15:38:03 AlHazWork Tangent: mm, nope. what's a 1470? 15:38:37 Tangent AlHazWork : http://www.linear-tech.com/pdf/lt1470.pdf 15:38:50 Tangent PCMCIA VCC switching matrix chip :) 15:38:53 GoRK most of the people putting out the stuff have "proprietary" modes faster than 54mpbs though.. Intel has a 108mbps mode and proxim has a 100mpbs mode 15:38:54 AlHazWork ah 15:39:15 GoRK actually i think intel and proxim cards will talk together at 100mbps 15:39:24 Tangent GoRK : The stuff Intel stuff looks most tasty 15:40:36 AlHazWork i hope they standardize 15:40:53 AlHazWork i hope folks realize that 5ghz is going to be *worse rather than better wrt signal propogation 15:42:37 Tangent AlHaz : Fortuantely I'm old enough to remember Short Wave Radio... travel the world on a watt etc :) 15:42:46 AlHazWork Tangent: that chip looks like just the thing 15:43:05 AlHazWork higher ghz is better for point to point directional stuff 15:43:13 AlHazWork omnidirectional tho, it's not so good 15:43:16 GoRK except for weather 15:43:16 Tangent AlHaz : Apparently LT sell small quantities for sampling/prototyping too 15:43:23 AlHazWork Tangent: Hmmmm 15:43:43 AlHazWork Tangent: By any chance does Maxim make a similar chip? they *give away samples 15:44:03 GoRK anything above about 6ghz tends to have big problems with rain and fog. Temp. inversion is a problem down to around 1ghz 15:44:22 Tangent AlHaz : I'll check through some maxim datasheets... 15:44:37 GoRK looks at his pile of maxim serial voltage converters he got for free 15:44:40 AlHazWork yup, they do 15:44:43 AlHazWork http://dbserv.maxim-ic.com/quick_view2.cfm?qv_pk=1026 15:45:29 GoRK oOoOo looks like the maxim chip might let us handle 12V! Pc cards :) 15:45:51 GoRK on both sockets no less 15:46:51 AlHazWork the MAX160x series is all pcmcia power stuff 15:46:54 Tangent GoRK : We have no 12V supply on the Tux board 15:46:56 AlHazWork they have a few more 15:47:06 GoRK Tanget: that could be fixed with a DC/DC :) 15:47:09 Tangent GoRK : That's why most of these Maxim chips aren't really suitable 15:47:20 Tangent GoRK : Nod... that can be fixed 15:47:40 GoRK but you may not have to hook 12V up to it anyway 15:47:46 GoRK it may not require that you do ... 15:47:48 GoRK pokes 15:49:31 Tangent GoRK : 12V not required for 3.3V switch 15:49:40 GoRK yeah thats what i just read 15:49:43 Tangent GoRK : Which I guess means it is required for 5V 15:51:05 AlHazWork hard to figure what they mean 15:51:09 AlHazWork looking at the datasheet here 15:51:46 AlHazWork says something about bypassing the 12v line with a capacitor 15:52:43 Tangent AlHaz : I suspect that the 12V line is used to power the gates on the output fets when switching 5V onto VCC 15:52:58 GoRK is the pcmcia controller on the tux listed as compatible with this chip? 15:53:04 GoRK page 10 of the data sheet 15:53:13 Tangent The LT1470 has built in charge pumps to generate the FET gate voltage 15:54:30 AlHazWork GoRK: iirc the sa1100 *is the pcmcia controller 15:54:43 GoRK oh 15:55:08 Tangent Nod... I think that too 15:55:39 prpplague does atmel make an arm core? 15:55:48 Tangent prpplague : Intel 15:56:14 Tangent Well... technically, ARM, but intel licensed the StrongArm architecture from them 15:56:26 AlHazWork maxim claims built-in charge pumps too, but doesn't specify no 12v needed 15:56:55 Tangent If the maxim parts can be got for free, then there's little point in _not_ getting them :) 15:57:23 AlHazWork they'll give you qty2 of up to 8 parts 15:57:46 AlHazWork so you could order 2ea of all 4 dual slot parts 15:58:08 GoRK or you can probably call and ask them 15:58:16 AlHazWork yeah, you could definately *ask them 16:00:16 Tangent What lies do I need to tell them to make them give me free stuff... 16:00:32 AlHazWork you don't have to lie . . . 16:00:34 Tangent Presumably something better than... 'I want to hack my telephone' :D 16:00:55 AlHazWork http://dbserv.maxim-ic.com/sl_requests2.cfm?parts=MAX1600 16:01:11 AlHazWork or http://dbserv.maxim-ic.com/sl_requests2.cfm?parts=MAX1603 16:01:18 AlHazWork you just fill out the form and they arrive in the mail 16:01:26 AlHazWork got serial buffer chips for another project that way a few months ago 16:01:36 GoRK tell them you have a board design that was hacked together quickly and is missing 5V on the pcmcia and that you don't have 12V anywhere and need to know "if it is suitible for my application" 16:01:48 Tangent Dear Maxim, I want to hack my telephone.. Please send me samples of any PCMCIA voltage switching matrix chips you make because I'm too lazy to analyse the datasheets, and too cheap to buy one... Love Tangent :) 16:02:02 prpplague hmm this dlink unit is suppose to have an ARM core, but i don't see it 16:02:45 Tangent Oh cool.. the do Dallas stuff too :) 16:03:00 AlHazWork GoRK: that's precisely what you should tell 'em, prolly 16:03:27 AlHazWork if they ask how many you might be ordering tell 'em 4000 16:03:41 GoRK yeah i get all my little parts like that 16:03:54 GoRK something about planning on a productin run in 6mo 16:03:56 Tangent Cool.. I could get some more DS1820s :) 16:04:40 AlHazWork a friend of mine used to get all sorts of bits & pieces from sygnetics just by asking a lot of questions 16:04:51 AlHazWork until they said "Hey, why don't i send you samples & documentation" 16:06:44 AlHazWork I don't feel like this is abuse of Maxim IC - I know engineers who build their chips into production stuff specifically because they're so easy-going to work with 16:08:13 AlHazWork on the flipside, I know engineers who avoid motorola wherever possible because they're such complete pricks until you start talking about 10k units 16:30:35 Tangent I used to have a mate who worked for Motorola... Got me lots of free databooks etc :) 16:35:39 AlHazWork the wearables.org folks say they're using max1600 as their vcc/vpp switching network, but they don't have much in the way of design specs online 16:37:06 Tangent is it just me, or has www.freshmeat.net disappeared from the planet 16:37:45 AlHazWork it's fallen off the edge of the planet, yeah 16:44:04 Tangent I warned them not to build their hosting centre on the edge,but would they listen... 16:44:47 GoRK Russ: afternoon. Bzflag said you had some blob splash patches? are those in cvs (yet?) 16:44:55 sammy hello Russ . 16:45:07 Russ|werk not yet, its pretty simple 16:45:22 Russ|werk it uncompresses an image into memory, and then turns on the scd 16:47:06 GoRK do you have to put the image into blob or is there a 'flash splash' option? 16:47:20 Russ|werk its a tag in the paramater block 16:49:50 GoRK do you have a .diff perhaps? :) 16:50:16 GoRK i wanted to twiddle with it and the linux progress patch 16:50:34 Russ|werk I might have it up 16:50:55 GoRK in cvs? 16:51:09 Russ|werk http://russ.dhs.org/files/tuxpatches/splash.tar.gz 16:51:23 GoRK ah ha! awesome 16:51:25 Russ|werk I don't htink it applies cleanly right now 16:51:39 GoRK well i can probably work that out 16:52:08 GoRK thanks a lot 16:52:18 Russ|werk np 16:52:28 Russ|werk the mini_inflate file will go into blob 16:52:53 Russ|werk at some point, so its not a penalty 16:54:22 as_ Evnin all. 16:54:46 GoRK ok cool 16:56:46 scanline is back (gone 10:15:57) 17:04:41 SpaceCoaster hi folks, just got a SODIMM from crucial which hung my TuxScreen, they substituted a CT8M32E8T6 for the CT8M32E4T6 that I ordered, said something about "the part being replaced but they still have some of the other ones" 17:07:56 Russ|werk SpaceCoaster: can you turn debugging on in blob and see what blob does? 17:09:43 SpaceCoaster Russ|werk: never gets to blob, locks up solid 17:13:09 SpaceCoaster Russ|werk: they don't even have this part CTM832E8T6 on their web site, they said that the 8 was an 8k page, they also seem to think that they are equivalent sortof 17:14:01 SpaceCoaster but I made them pay the return postage for disagreeing with me :-) 17:17:03 Russ|werk it locks in blob 17:17:14 Russ|werk if you compile blob with debugging, you can get more info 17:17:49 Russ|werk blob does all the memory setup/timing/sizing 17:18:54 SpaceCoaster ok 17:20:39 scanline is away: off to the EE lab I go... 17:30:00 as_ Hiya BZ, did you get a chance to process my order? 17:32:35 Tangent goes away for the weekend 17:36:57 CosmicPenguin GoRK: Did you get the LPP working in 2.4.7+| 17:37:07 CosmicPenguin GoRK: Last I noticed, it was broken 17:39:30 boodle is away: lunch 17:49:05 boodle is back (gone 00:09:31) 18:00:02 CosmicPenguin [OT] Anyone know of a Windows Bluetooth stack that i can get ahold of in a timely matter? 18:18:30 other Length: -1 seconds 18:18:35 other ack 18:18:37 other wrong window. 18:29:38 GoRK CosmicPenguin: i havent fooled with it yet, but that is my aim. 18:31:54 CosmicPenguin GoRK: You know, for a technology that alot of people talk about, it is surprisingly vaporwarish.... 18:32:03 CosmicPenguin Oops, sorry.... 18:32:07 CosmicPenguin Ignore that.... 18:32:13 GoRK ? 18:32:34 CosmicPenguin GoRK: If you can fix LPP to work with anything more than 2.4.5, that would be awesome 18:33:20 GoRK well afaik in my extremely limited kernel hackery, there were some redundant functions removed from the core kernel during 2.4.5->2.4.6 that broke *a lot* of stuff 18:33:35 GoRK so it shouldnt be hard to move whatever lpp uses over to the other functions 18:34:41 GoRK their themes site tracks literally thousands of downloads. i imagine someone somewhere has it working anyway 18:35:08 CosmicPenguin GoRK: Surprisingly, I haven't found anything..... 18:35:10 AlHazWork GoRK: out of curiosity, were you able to jflash your tuxscreen back from the dead? 18:35:20 GoRK AlHaz: oh yeah. worked fine 18:35:44 AlHazWork good, good 18:35:45 GoRK AlHaz: and on the plus side i now have a unit that's safe to screw with blob et. al. on 18:35:53 AlHazWork yep 18:35:58 Russ|werk http://people.debian.org/~bossekr/ 18:36:24 Russ|werk oh... 18:36:29 Russ|werk thats only to 2.4.5 18:36:40 Russ|werk bummer 18:38:51 GoRK here's a dude that got it into 2.4.6+ .. that should apply up to 2.4.9.. I'm going to email him and see if he has the patch sticking around 18:39:42 CosmicPenguin GoRK: is that in the comments off of freshmeat.net? 18:40:19 GoRK yeah that guy Invisi 18:54:37 GoRK hmm only 3 failed hunks anyway 18:56:04 GoRK and they are easy to correct... 18:58:56 GoRK 1 down 2 to go 18:59:31 GoRK hmm the other two are both the logo itself. i think that just gets replaced 19:07:29 CosmicPenguin GoRK: Yep.... 19:07:57 CosmicPenguin GoRK: But is it the same type of file, somewhere I got the impression that the format of the logo has changed? 19:12:29 GoRK doesnt look *too* different.. there are some extra defines and apparently there can be more colors now 19:13:13 GoRK and there can be different logos for different arch's 19:27:36 andersee CosmicPenguin: was it you who was seeing signal problem? 19:30:56 BZAway to work 19:32:04 CosmicPenguin andersee: I was... did ya fix it? 19:39:23 scanline is back (gone 02:18:45) 19:41:29 andersee nope. 19:41:45 andersee CosmicPenguin: do you have a test care for me to reproduce it? 19:42:05 andersee CosmicPenguin: I was messing with signals, but they seem to be working for me... 19:43:48 CosmicPenguin andersee: I don't have anything outside of the GUI.... try this: 19:44:30 CosmicPenguin make a SIGCHLD handler, and fork a couple of children off and kill 'em. You should see the first one get handled, but the handler will disappear for the second one 19:44:37 GoRK ok think i got LPP in there.. now to see if it compiles up on arm... the Config.in had it up under ARCH_i386 but it didnt look like there was anything x86 specific in there i.e. no special vidio init or asm 19:45:04 CosmicPenguin I stuck in a work around that just re-registered the handler after the SIGCHILD was completed 19:46:35 andersee GoRK: the fonts used in the LPP are ugly as sin... 19:47:02 andersee CosmicPenguin: k, I'll try that 19:47:25 CosmicPenguin GoRK: They are.... 19:47:52 scanline is away: Last class of the day 19:50:47 CosmicPenguin :q 20:14:36 GoRK andersee: that's ok we can replace them :) 20:14:55 GoRK it compiled ... time to see how it went 20:14:57 andersee Hmm, is the font configurable? 20:15:12 andersee I've messed with LPP, but not too deeply 20:16:29 andersee I'd looked briefly at the font stuff, and it seemed to be hard coded in an array, if I recall rightly... 20:16:52 GoRK if it is, we can probably make a new font pretty easily 20:20:31 andersee True enough 20:26:27 GoRK it works 20:27:11 GoRK mostly 20:27:24 GoRK i have to make the initscripts use it but that's cool 20:27:59 GoRK and i forgot to apply russ's patch so i can't really see the screen anyway :) 20:31:16 CosmicPenguin GoRK: Patch that sucker, and post it, will ya? 20:31:25 CosmicPenguin Now we need some bad ass art... 20:31:44 GoRK CosmicPenguin: yeah hold yer horses lemme get it finished 20:31:46 GoRK :) 20:32:57 GoRK you have to put the boot logo in manually 20:33:19 GoRK since the #define's wrapping it in the linux_logo.h file have changed a lot 20:33:40 CosmicPenguin GoRK: Sorry, I'm about to throw this bluetooth card across the room, so I need some outlet for my energy... 20:34:05 GoRK i bet you can throw it farther than it can transmit 20:34:14 GoRK by an order of magnitude 20:34:28 CosmicPenguin GoRK: Probably by an order of a couple of magnitudes... 20:35:15 GoRK heh 20:35:33 GoRK let me diddle with these init scripts so i can see if it is doing its job 20:35:41 GoRK i cant tell if the bar filling up works yet or not 20:37:22 GoRK should i make this patch to apply before or after rmk and russ's patches? It should work on all architectures so im thinking before 20:38:15 andersee GoRK: I fixed the tool up somewhere to produce proper #defines.... 20:38:26 andersee Lemme go look 20:38:32 GoRK the lpp logo tool? 20:38:41 andersee has zillions of little hacks hidden on his system 20:38:44 andersee GoRK: yup 20:39:14 GoRK andersee: if you cant find it i can do it in about 1 min anyway.. dont spend a bunch of time digging for it :) 20:41:59 andersee found it. 20:42:07 GoRK fblogo.c ? 20:42:18 andersee I'd hacked that into a 2.4.6 tree back in early august 20:42:21 andersee yup 20:42:31 GoRK cool can you dcc it over here? 20:43:18 andersee GoRK: sitting on winder.codepoet.org in /tmp 20:43:22 GoRK ok thanks 20:59:39 scanline is back (gone 01:11:48) 21:01:05 andersee CosmicPenguin: uClibc/test/unistd/fork.c now tests SIGCHLD behavior... 21:01:15 andersee CosmicPenguin: everything looks fine on x86 21:01:28 andersee CosmicPenguin: but things look messed on arm.. 21:01:41 andersee Compiling vs uClibc: 21:01:41 andersee 21:01:41 andersee ../../extra/gcc-uClibc/arm-uclibc-gcc -fomit-frame-pointer --uclibc-use-build-dir -c fork.c -o fork.o 21:01:41 andersee ../../extra/gcc-uClibc/arm-uclibc-gcc -s -Wl,-warn-common --uclibc-use-build-dir fork.o -o fork 21:01:41 andersee strip -x -R .note -R .comment fork 21:01:42 andersee ./fork 21:01:43 andersee The child process sleeps 2 seconds... 21:01:45 andersee ThThe chie childld process sleeps 3 seconds... 21:01:47 andersee Parent: waiting for the child to die. 21:01:50 andersee process sleeps 4 seconds... 21:01:52 andersee Child exiting. 21:01:53 andersee I got a SIGCHLD 21:01:56 andersee Child exiting. 21:01:58 andersee Child exiting. 21:02:00 andersee Child process exited. 21:02:02 andersee Goodbye. 21:02:04 andersee 21:02:05 andersee ------- 21:02:07 andersee 21:02:09 andersee Compiling vs GNU libc: 21:02:12 andersee 21:02:14 andersee gcc -fomit-frame-pointer -c fork.c -o fork_glibc.o 21:02:16 andersee gcc -s -Wl,-warn-common fork_glibc.o -o fork_glibc 21:02:18 andersee strip -x -R .note -R .comment fork_glibc 21:02:20 andersee ./fork_glibc 21:02:22 andersee The child process sleeps 2 seconds... 21:02:24 andersee The child process sleeps 3 seconds... 21:02:26 andersee The child process sleeps 4 seconds... 21:02:28 andersee Parent: waiting for the child to die. 21:02:29 andersee Child exiting. 21:02:31 andersee I got a SIGCHLD 21:02:33 andersee Child exiting. 21:02:36 andersee I got a SIGCHLD 21:02:38 andersee Child exiting. 21:02:40 andersee I got a SIGCHLD 21:02:42 andersee Child process exited. 21:02:44 andersee Goodbye. 21:02:50 andersee CosmicPenguin: so anyways, I've got the wierdness reproduced now 21:04:14 andersee Odd that this is arch specific... 21:04:31 andersee Compiling vs uClibc: 21:04:31 andersee 21:04:31 andersee ../../extra/gcc-uClibc/i386-uclibc-gcc -fomit-frame-pointer --uclibc-use-build-dir -c fork.c -o fork.o 21:04:31 andersee ../../extra/gcc-uClibc/i386-uclibc-gcc -s -Wl,-warn-common --uclibc-use-build-dir fork.o -o fork 21:04:31 andersee strip -x -R .note -R .comment fork 21:04:32 andersee ./fork 21:04:33 andersee Parent: waiting for the child to die. 21:04:35 andersee The child process sleeps 2 seconds... 21:04:38 andersee The child process sleeps 3 seconds... 21:04:40 andersee The child process sleeps 4 seconds... 21:04:41 andersee Child exiting. 21:04:44 andersee I got a SIGCHLD 21:04:46 andersee Child exiting. 21:04:48 andersee I got a SIGCHLD 21:04:50 andersee Child exiting. 21:04:52 andersee I got a SIGCHLD 21:04:54 andersee Child process exited. 21:04:55 andersee Goodbye. 21:04:58 andersee So uClibc x86 is fine. 21:05:08 andersee Maybe its something in include/bits/signal* on arm 21:11:41 CosmicPenguin andersee: Thanks for looking into this.... 21:11:54 CosmicPenguin Is anyone here a PCMCIA / Card services guru? 21:26:24 GoRK BZFlag: will the blob code in the tuxscreen cvs work? The top of the Makefile.am has a warning from you that it won't.... ? 21:26:36 GoRK Btw, good evening 21:26:42 GoRK or late afternoon 21:28:59 BZFlag GoRK: no. 21:29:13 BZFlag use the blob cvs tree, not the tux blob cvs tree. 21:29:36 BZFlag the tux build root has a "make blob" that does the right thing. 21:30:37 GoRK ok. do i still have to check out a special tag? the buildroot was getting it from a tarball.. is that tarball up to date? 21:36:47 BZFlag up to the latest we recommend, yeah. 21:37:01 BZFlag if you want to try leading edge, then you better have jtag. ;-) 21:37:07 GoRK i do 21:37:33 GoRK i got russ's splash patches and was fooling around with linux progress patch, which i have mostly working now 21:38:14 GoRK all that's left i think is a tweak to the initscripts .. and the proc interface is still broken i think 21:38:37 BZFlag cool. 21:39:02 BZFlag you have SF write access? 21:39:23 GoRK to the tuxscreen project? 21:39:41 GoRK i have a developer account and im on a couple projects, yeah if that' what you mean 21:40:58 BZFlag want me to add you to tux so you can commit to cvs buildroot? 21:41:11 GoRK that would be very dandy 21:41:28 BZFlag don't break anythink or andersee will be on yer butt. ;-) 21:41:45 GoRK no worries 21:42:28 BZFlag "gork" added. 21:43:32 GoRK thanks. i still have a lot of cleaning up on this to do, so... 21:45:04 BZFlag no prob. at least you can start posting things that don't break other things. (or something like that) 21:45:46 BZFlag Things that are partly broken should be posted as SF patches till the work. k? 21:46:18 GoRK sounds fine 21:53:21 GoRK thanks a bunch 22:35:08 BZFlag as you have not gotten it yet? 22:35:16 BZFlag as2: you have not gotten it yet? 22:35:58 as2 Not gotten what? 22:36:47 as2 Shipping of the actual tuxscreens will take between 4 and 6 weeks I think surface. 22:37:10 BZFlag oh yeah, you're right it was surface. 22:37:46 as2 Air shipping was seriosly expensive. 22:38:19 as2 As long as its sent, thats cool, I wasnt sure youd sent it yet becuase I didnt get a reply to the email. 22:39:35 BZFlag sorry. busy. I did get it and sent it off. I recall it was short like $5 or somthing like that. hmm. might have been $9.60 actually. 22:40:15 BZFlag but my bank charged me $20 for the wire transfer =( so sending $9.60 would have cost me money. 22:40:48 BZFlag Anyone else that asks for wire transfer will have to cover the $20 charge. 22:41:23 as2 Sorry, I normally tell the bank to pay both sides of the transfer, I think I might have forgotten this time. 22:42:00 BZFlag well it would cost you $49.60 to get me caught up, don't worry about it. 22:42:20 BZFlag buy me a cd if you like. ;-) 22:42:24 as2 Cheers. 22:42:42 Russ|werk as2: go to astroglide.com and send him a free gift 22:42:49 as2 Haha, ok, I'll think of somthing to send. 22:43:00 as2 I'll have a look. 22:43:38 as2 Why was the money short btw, I thought the math was right? 22:44:23 as2 If your down $30 thats a serios mistake on my behalf. 22:44:52 BZFlag I don't recall. Perhaps the weight was a little off. they vary a bit. I'm listed the weights as higher on the web site to avoid issues in the future. 22:45:22 BZFlag as2: I'm selling 100s of them I chauk it up to experience. 22:45:47 BZFlag like today I spent over $1300 in postage. 22:46:04 BZFlag to discover that my paypal card still had a $1000 limit. 22:46:13 as2 Yikes. Postal company must like you. 22:46:24 BZFlag upped it to $3000 limit when I got in the office. 22:46:40 BZFlag yep. /me remembers the picture..... 22:46:42 as2 Paypal are doing there own credit card now? 22:47:33 BZFlag master card debit. though they now offer a visa credit card as well iirc 22:47:51 as2 Ok, I see the joke in Russ's comment now. ;P aimed at Russ. 22:49:47 BZFlag http://TuxScreen.net/img/postofficerun.jpg 22:50:23 BZFlag that was the shipment today almost fitting in my mustang. 22:52:12 as2 This will take some time to load on my connection. 22:54:14 BZFlag oh yeah, sorry. it's 2048x1536 ;-) 22:54:18 as2 Thats a very nice digital camera though, is that the ricoh? 22:54:29 BZFlag yep. Ricoh i700 22:54:59 BZFlag you can even see how badly my car needs washing. 22:55:22 as2 That sky is like an advert for that camera. 22:55:47 as2 Ive been staring at the sky for some time, when I get to the car, I'll let you know. 22:55:57 BZFlag heh, yeah. but you probably can't read the "Slashdot" on my hat. ;-) 22:57:25 as2 Not quite, now you say what it is I can see it. 22:57:39 as2 I cant get over how good the quality of that image is. 22:58:05 BZFlag print it out and put it on the fridge. ;-) 22:58:16 as2 The focus is pin sharp. 22:58:38 as2 Ok, I dont like the picture quite that much ;] 23:00:04 as2 Point taken about the post run though. 23:02:16 GoRK BZFlag: are you down under 800 phones yet? 23:02:37 GoRK hmmm i dunno if what im seeing here is a bug of the way things are supposed to function... 23:03:10 GoRK i know.. i can not display console messages to tty1 when lpp is on! 23:03:16 GoRK that'll fix all my problems 23:04:48 as2 Says here it has 9 quality settings, is that max quality BZ? 23:07:15 scanline is away: food 23:08:26 GoRK hmm thats no good. now i dont have any vt's 23:17:00 BZFlag GoRK: nope. getting close to 900 though. 23:18:50 BZFlag as2: it has 3 resolutions and 3 quality settings. that was the highest res and quality, yes. 23:19:26 BZFlag I have a 340M microdrive in mine now, so I'm not real worried about space. ;-) 23:20:33 as2 Haha. 23:20:38 as2 Cool. 23:21:09 as2 How much space does a max quality image take up? 23:23:46 GoRK hrm how come setting console=/dev/tty2 still prints stuff on the screen? 23:24:27 BZFlag well, that one is 1162668 bytes. it varies by content though. 23:24:50 GoRK on tty1 rather 23:25:55 BZFlag GoRK: no idea. should work. 23:26:12 BZFlag try console=none ? 23:27:29 GoRK i then get: serial console detected. Disabling virtual terminals. 23:27:43 GoRK it works right but.. blah 23:27:58 GoRK no vt's is not the intent 23:28:48 BZFlag agreed. you might try removing the console=* lines in the kernel config and flashing that kernel. 23:29:16 GoRK so you think maybe those lines are getting used even though i'm specifing the whole command line in blob? 23:29:21 GoRK err ... used also 23:30:52 GoRK nm i got it working better 23:30:52 RedBaron Howdy all 23:31:11 GoRK yeah i got it 23:31:38 GoRK now the only output comes from the initscripts.... which is correct since they will be rewritten to output to /proc/progress in this patch 23:32:12 BZFlag GoRK: not sure. at one point what you typed was getting appended to the built one one. I think now it replaces it completely, but I'd remove the built in options just to test. 23:32:36 BZFlag GoRK: cool. 23:33:50 GoRK i got tty2 as a console but if i specify a serial console it wont run the gettys 23:34:22 GoRK err if i specify two consoles 23:35:29 GoRK mmrf. that's kinda dumb.. oh well it is working. 23:36:13 scanline is back (gone 00:28:58) 23:37:26 RedBaron Question about blob before I try to flash my phone. Is there a difference between the v0.4 directory on tuxscreen.net/download and the 2.0.4 blob on sourceforge? 23:44:17 RedBaron .o 23:44:24 RedBaron whoops 23:45:21 BZFlag The tuxscreen-blob in the v0.4 directory is a built binary of the 2.0.4 blob release for the shannon (TuxScreen). 23:45:40 RedBaron BZFlag: Ah. Okay. 23:46:09 BZFlag I don't recall if tuxscreen-blob includes the latest 115200 baud console patch or not. 23:46:32 BZFlag so console is either 9600 or 115200. uploads are 115200 on both. 23:46:55 RedBaron okay. I recall something saying that the latest was both 115200 baud 23:47:55 BZFlag buildroot-tux in cvs is definitely 115200 baud. I just don't recall if the v0.4 snapshot was before or after that change. 23:48:09 RedBaron *nod* 23:55:48 RedBaron Woohoo. I am blob'ed. 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.