Wednesday, 2013-03-06

prpplagueho ho hum00:42
*** jayabharath <jayabharath!> has quit IRC01:02
*** davest <davest!~Adium@> has quit IRC01:20
*** calculu5 <calculu5!~calculus@gentoo/user/calculus> has joined #minnowboard04:36
*** calculus <calculus!~calculus@gentoo/user/calculus> has quit IRC04:39
*** philips_ <philips_!~philips@2001:4801:7817:72:d4d9:211d:ff10:2b63> has quit IRC05:47
*** philips_ <philips_!~philips@2001:4801:7817:72:d4d9:211d:ff10:2b63> has joined #minnowboard05:49
*** davest <davest!~Adium@> has joined #minnowboard06:15
*** davest <davest!~Adium@> has quit IRC06:19
*** davest <davest!Adium@nat/intel/x-curcxclogusapzkl> has joined #minnowboard06:20
*** davest <davest!Adium@nat/intel/x-curcxclogusapzkl> has quit IRC06:25
*** panto <panto!~panto@> has joined #minnowboard06:40
mranostaywelcome panto to the world of x86!06:41
pantoit's their world, we just live in it06:42
mranostaybe nice :)06:42
mranostaypanto: so how hard would it be have capemanager for SSDT?07:31
pantoit would be easier just ditching that sucker and going DT07:32
pantobut, x86 without BIOS level crap? that's unpossible...07:32
mranostayUEFI is mostly BIOS free07:33
mranostaypanto: i agree would be nice to cut out the old legacy crap :)07:34
mranostayesp from the die07:35
mranostaywho is honestly running real mode code in 2013? :)07:35
pantoyou'd be surprised...07:36
mranostayyeah i know07:37
mranostaypanto: some Fortran code running somewhere, right? :)07:37
pantoand not the latest released version either07:39
mranostayFreeDOS and the all the other stuff07:39
mranostaypanto: last time i ran "DOS" was to flash a BIOS version07:39
pantoI bet you'd run into trouble if your CPU didn't have all that legacy support in there07:40
pantoso, it's not so easy removing legacy stuff - better to leave them in07:40
mranostaybut stuff before i was born? :)07:40
pantothe world had managed to get along just fine before you came along punk :)07:41
pantothe truth hurts07:42
av500mranostay: if we drop all the stuff from before you are born, we'd not have much :)08:21
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has quit IRC09:17
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has joined #minnowboard09:21
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has quit IRC13:11
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has joined #minnowboard13:11
*** prpplague <prpplague!> has quit IRC14:43
*** jayabharath <jayabharath!> has joined #minnowboard14:46
*** calculu5 is now known as calculus15:37
*** davest <davest!~dcstewar@> has joined #minnowboard15:38
*** davest <davest!~dcstewar@> has left #minnowboard15:38
*** dvhart <dvhart!~dvhart@> has joined #minnowboard15:55
dvhartmorning all15:57
*** jackmitchell <jackmitchell!~Thunderbi@> has joined #minnowboard16:05
mdplet's go16:15
av500can I boot minnow with a stock linux distro?16:20
av500or is there no BIOS as we know it?16:20
dm8tbryou need something that boots from EFI16:27
dm8tbrwhich most distros should support by now16:28
av500so its EUFI?16:28
av500or UEFI16:28
dm8tbri think that latter thing16:29
dm8tbrat least thats how my FRI2 boots16:30
koenanything that supports (U)EFI should boot16:34
koenand 32 bit of course16:34
*** zenlinux_ <zenlinux_!~sgarman@> has joined #minnowboard16:40
*** prpplague <prpplague!> has joined #minnowboard16:43
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has quit IRC16:44
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has joined #minnowboard16:44
prpplaguedvhart / koen  updated wiki for the lures with pictures -
prpplaguepanto: you too17:18
dvhartlooking, and otp17:19
mdpnice, type-C look like "wings" from other boards17:20
dvhartwhat's with type-C?17:20
dvhartwhy go off the board?17:20
prpplaguedvhart: type-c is for smaller items such as connectors, camera modules, small lcds and adapters17:21
prpplaguedvhart: we've found that similar designs for the beagle as well as devices such as the papillio fpga board are very useful17:22
mdpdvhart: can you share anything about your ideas for how platform devices get instantiated on minnow? I was reading the drivers/acpi/ code and didn't see anything there yet to do that generically (like a device that uses spi), but maybe I missed it.17:22
mdpprpplague: papillio is what I was thinking of17:22
dvhartOK.... it just concerns me that the one form-factor that would be prone to bumping, flexing, bending, etc (as it hangs over open space) is the one that has no mounting support.17:22
prpplaguedvhart: yea, the connector is pretty solid, dave albert  will be checking during layout of X3 to determine if we have the pcb area to add one additional screw hole near the connector17:23
prpplaguedvhart: but in general it is pretty secure17:23
dvhartmdp, this is something I've only discussed at a high level with the ACPI and x86 architects at this point. So I don't have the details that I think you're looking for.17:23
dvhartmdp, I'm currently focused on the enablement of the current hardware, and will be moving on to that hopefully soon17:24
dvhartI'm eager to get that defined as well17:24
dvhartpeople are right to be eager to see how that will work17:24
mdpcool, I have this vague notion of how it would work with SSDTs and, yes, was thinking about a particular device I've interfaced to other boards..and how that will work here.17:25
mdpok, thanks17:25
dvhartmdp, from a high level, think OF descriptions17:25
dvhartit might even be able to use the same code17:25
pantodvhart, we've been through the same path before with beagle*17:25
pantojust OF is not enough17:26
dvhartpanto, ack17:26
dvhartpanto, anything in particular?17:26
pantodvhart, OF unmodified is only capable of static hardware definition17:26
mdpstatic is a start ;)17:27
prpplaguedvhart: let me know of any concerns and/or questions prior to tomorrows meeting17:27
pantowhen you add in expansion capabilities which use limited (and conflicting) resources of the SoC you'll run into problems17:27
dvhartpanto, maybe I'm misunderstanding you, but the SSDT is a static definition....17:27
pantodvhart, yes, that's my understanding too, and that's the problem17:27
pantoif you have spare time google for DT overlays17:28
dvhartpanto, will do17:28
dvhart(multiple SSDTs can be used FYI)17:28
dvhartprpplague, those are my concerns with the lures so far.17:28
prpplaguedvhart: ok thanks17:29
dvhartprpplague, oh, one more. Does type-C define the height of the connector?17:30
dvhartonly the first one *needs* to be tall to clear the Eth/USB17:30
dvhartBUT, that means all of them need to be17:30
prpplaguedvhart: type-B needs to clear the caps and headers17:30
dvhartthat or we need a bridge connector17:30
prpplaguedvhart: so there is a definition for the height on it17:31
prpplaguedvhart: 10mm17:31
prpplaguedvhart: the type-C board can be either 5mm or 10mm17:31
prpplaguedvhart: with a recommendation of 5mm17:31
dvharthow can type-C below lower than type-B?17:31
prpplaguedvhart: because type C does no go over the top of the board17:32
prpplaguedvhart: so it does not have to clear the caps or headers17:32
dvharts/C/A/ in my comments17:32
prpplaguedvhart: type B boards fit in between the taller headers such as the ethernet and usb stack17:33
prpplaguedvhart: so the board does not have to be taller17:33
dvhartright, type A is my concern17:33
dvhartare all type A lures 17mm between PCBs?17:33
prpplaguedvhart: yep17:34
prpplaguedvhart: or at least that is the spec i am writting17:34
dvhartnot sure of any way around it - other than maybe defining it as smaller and providing a riser17:34
dvhartbut.... yeah, ick17:34
prpplagueyea, this is the best / cheapest /easiest combo of devices17:35
dvhartWe could cut a lot of empty space in the top 3 type-A boards17:35
prpplaguedvhart: i am sorry i did not understand your statement17:35
dvhartnote that if the type A spec is 17mm, then people might make type-A with 17mm components making type-B not stack on type-A17:35
dvhartso a max component height needs to be specified which ensures you can stack any combination of ABC lures, 4 high.17:36
dvhartwhich comment was unclear? the one ending in "riser" ?17:37
prpplague<dvhart> We could cut a lot of empty space in the top 3 type-A boards17:37
prpplaguedvhart: yes the max height would need to be specified if they plan to make the boards stable17:37
av500why A at all? B is already pretty large17:38
dvhartprpplague, ah, only the first type-A board really needs 17mm clearance, those above it could have the same spacing as B.17:38
prpplagueav500: more space for bigger interfaces such as pcie, a laptop sata hd, those types of things17:39
prpplaguedvhart: correct17:39
dvhartso with 4 type A lures, if they could use B spacing above the first, it would reduce the size quite a bit.17:39
prpplaguedvhart: yea17:39
av500but the spacing is defined by the lure17:39
dvhartso if A used B spacing and we provided a "riser" which lifted the first A above the Eth/USB, we could alleviate that.17:39
dvhartbut... ewe.17:39
av500ah riser17:39
prpplaguedvhart: yea like i said, this is the best option for long term support17:40
av500spacer rather17:40
dvhartav500, yeah17:40
prpplaguedvhart: in most cases folks are going to use one maybe two lures together17:40
dvhartprpplague, simple is sometimes better, yes17:40
dvhartprpplague, agreed17:40
dvhartoptimizing for the corner case isn't a good plan17:40
prpplaguedvhart: and as long as we make some 0.1" header style type-A boards, other expansion is available17:40
dvhartand people who really care *will* do what I suggested anyway17:40
prpplaguedvhart: exactly17:41
av500dvhart: also, 17mm looks more impressive:
dvhartav500, hehe, nice17:41
dvhartoh, prpplague can the minnow itself stack at 17mm?17:42
* dvhart thinks so...17:42
prpplaguedvhart: yep17:42
dvhartminnow cluster = school17:42
prpplagueav500: lots of options to evaluate, but the key here is i need have a very solid and robust solution by EoD friday17:43
prpplagueav500: so there are probably a lot of solutions17:43
prpplagueav500: but we need something that is both economical to produce as well as provide enough versitility to people who want to use the board17:43
prpplagueav500: indeed17:44
av500prpplague: you know, as long as your send me one each, I'm happy with whatever you do :)17:45
prpplagueav500: hehe17:45
prpplagueav500: we know17:45
mranostayour loyality can be bought with free stuff18:09
dm8tbrdid someone say freebies? ;-รพ18:11
mranostayprpplague did18:11
zenlinux_prpplague, do you want to post the lure wiki page on G+ and I can re-share it?18:13
* mdp perks up at 'free'18:19
*** zenlinux_ <zenlinux_!~sgarman@> has quit IRC18:19
dvhartzenlinux, might be a bit early18:21
dvhartzenlinux, guessing Friday18:21
koenprpplague: 68MHz is a typo, right?18:41
koenprpplague: 720p is 72MHz already18:41
prpplaguekoen: most fpdlink-i receivers are limited to a max of 68MHz18:42
zenlinuxdvhart, sure, if you think so that's fine18:43
dvhartit's up to prpplague of course18:43
prpplaguekoen: let me double check the max18:43
prpplaguedvhart: i have no problems posting it now, i want as much feedback as possible before friday18:44
dvhartzenlinux, ^ there you go18:44
prpplaguekoen: 720p = 1280 x720 x 60 frames/sec = 5529600018:45
zenlinuxprpplague, ok, well if it's easier for you I can post something from the MinnowBoard account later today, or you can write up a post if you prefer18:45
prpplaguekoen: of course there is some blanking in there18:45
prpplaguekoen: but in general you should not have more than 68MHz18:46
prpplaguezenlinux: no worries, i am happy to post on G+ shortly18:46
zenlinuxok, no rush18:46
prpplaguekoen: max pll clock for the lvds82 is 68MHz18:49
* prpplague looks18:49
koen13468 ps = 74.25MHz18:49
koenI guess that leaves a huge blanking gap for audio, ethernet and other hdmi features18:50
prpplaguekoen: yea that is typical of 720p with audio18:51
prpplaguekoen: but for standard FPDLink-1 there is no audio or other oob transmissions18:51
koenbut I'm paranoid when it comes to timings and hdtvs :)18:52
prpplaguekoen: yes indeed, but again, this is for the fpdlink18:52
prpplaguekoen: you can't drive the fpdlink over 68MHz, there is nothing that will latch18:53
prpplaguekoen: so you really can't generically say 720p is x pixel clock18:56
koenbut if it follows the CEA stuff pretty much every TV works18:56
prpplagueyes again, for DVI and HDMI i agree18:57
koenbut as you say, it's just for lvds18:57
koenwe can pick the right LCD and not care18:57
prpplaguekoen: plus it _should_ be reading the EDID for proper configuration18:59
prpplaguekoen: yea i just checked the 45Khz hsync value is what is required for the addition audio features19:03
* prpplague is bleeding19:04
prpplaguegot a a sharp tortilla chip pointed into my cheek while eating lunch19:04
koenapply salsa19:04
prpplaguekoen: i did,it burned19:05
mranostayprpplague: only you would get hurt eating :)19:10
* mranostay steps away from prpplague's bad luck19:10
prpplaguemranostay: indee19:10
dvhartsweet, platform code working without nasty hacks in gpio-sch.c19:17
dvharterm.... fewer nasty hacks19:17
prpplaguedvhart: watchout for dangerous tortilla chips19:23
prpplaguekoen: did you see that extruder device that chews up old plastic for 3dprinting?19:26
zenlinuxthanks prpplague, will share19:29
prpplaguezenlinux / dvhart either of you guys got admin for the channel? we need to get chanserv in the channel and get a few folks setup for op19:35
prpplagueas the news of minnow gets around, we are going to attract some folks that will need attention, and the weekends and late nights are important19:35
mranostayprpplague: so with multiple types of lures on one board it is going to look weird :)19:37
prpplaguemranostay: possibly19:37
mranostayI'm thinking Jenga for some reason now19:37
prpplaguemranostay: i really don't think we'll see that many stacks19:41
prpplaguemranostay: at least not with the 100 expansion header19:41
mranostayprpplague: how pins of that header are grounds?19:42
zenlinuxprpplague, I'll get that set up for you later today19:42
prpplaguemranostay: grounds between all the differential pairs with 100ohm matched impedance, and grounds to offset the matching power rails19:43
*** yocti sets mode: +o zenlinux20:03
*** yocti sets mode: +o prpplague20:05
*** zenlinux <zenlinux!> has quit IRC20:08
*** zenlinux <zenlinux!> has joined #minnowboard20:08
*** yocti sets mode: +o zenlinux20:09
* dvhart returns20:09
prpplague     /msg Chanserv set #minnowboard guard on20:18
prpplaguezenlinux / dvhart looks like you have me for ops, but not as a founder, so i can't register or maintain the channel, which is fine, but someone will need to setup chanserv registration and put chanserv on guard20:19
zenlinuxprpplague, I'm working w/ Michael Halstead on that, cc'ed you in an email20:20
prpplaguezenlinux: ahh dandy20:20
* prpplague waits for email20:20
*** panto <panto!~panto@> has quit IRC20:31
mranostayprpplague: so are there power usage numbers up?20:33
* mranostay is wondering what it is with and without the PCI-E20:33
dvhartcurrent dev gpio mapping:
dvhartkoen, I have pch[0:7] in sysfs currently, the rest have specific drivers20:37
dvhartseems like the best out-of-the-box setup20:37
zenlinuxI've hooked up a Kill-A-Watt while running a game demo, seems to top out around 11-12W max21:11
zenlinuxidle is more like 8-9W21:11
zenlinuxwe haven't looked into any power optimizations yet21:11
dvhartI'd be interested in the powertop numbers21:12
dvhartI suspect we have quite a bit we can do just in software21:12
zenlinuxyeah, those are really preliminary numbers, I did it just out of curiosity21:13
prpplaguethe X3/X4 designs are going to be a little more power hungry, the plan is to have a X5 with power consumption cleanups, cost reductions, and board size shrink21:21
*** davest <davest!Adium@nat/intel/x-mvckiutfajiptxft> has joined #minnowboard21:22
prpplagueit's super davest21:27
*** jayabharath <jayabharath!> has quit IRC22:04
mranostaysuper davest?22:10
mranostaywhat is his superpower?22:10
prpplaguemranostay: to open wine bottles with a single pull, and to consume wine faster than a speeding bullet22:12
mranostayprpplague: and forget about it even22:12
*** jayabharath <jayabharath!> has joined #minnowboard22:32
*** jayabharath <jayabharath!> has quit IRC22:39
davestmranostay: hi there22:44
davestprpplague: there is a method for opening a wine bottle without a cork screw22:45
prpplaguedavest: hehe22:45
davestprpplague: it uses a shoe22:46
*** ChanServ sets mode: -o yocti22:46
*** ChanServ sets mode: +o yocti22:46
prpplaguedavest: not going to ask... not going to ask22:46
mranostayi will however22:47
* mranostay noticed his Swiss Army knife doesn't even have a corkscrew now22:48
*** jayabharath <jayabharath!> has joined #minnowboard23:11
*** yocti sets mode: +o prpplague23:16
*** prpplague <prpplague!> has quit IRC23:17
*** jayabharath <jayabharath!> has quit IRC23:19
*** dvhart <dvhart!~dvhart@> has quit IRC23:58

Generated by 2.11.0 by Marius Gedminas - find it at!