Tuesday, 2013-02-26

*** yocti sets mode: +o zenlinux00:37
*** prpplague <prpplague!~prpplague@107-206-64-184.lightspeed.rcsntx.sbcglobal.net> has joined #minnowboard00:43
prpplaguegreetings all! hows the fishing this evening?00:43
*** dvhart <dvhart!~dvhart@static-50-53-98-238.bvtn.or.frontiernet.net> has joined #minnowboard01:05
prpplaguedvhart: greetings01:06
dvhartHey prpplague01:08
prpplaguedvhart / koenkooi what is the line of reasoning on the usage of the 0.1" gpio header with 4 GPIOS?02:23
dvhartwhat do you mean?02:25
dvhartas opposed to what?02:25
prpplaguedvhart: just wondering if there was a specific use case that you had in mind it was added to the design02:26
dvhartAh. The idea is just to give people some buttons, some LEDs, and some lines they can do whatever with.02:26
dvhartThese are all just the GPIO from the CPU02:26
dvhartthe PCH GPIOs go over the expansion connector02:26
prpplagueright just trying to follow the line of thinking for it02:27
* dvhart breaks for SIGFOOD02:31
dvhartprpplague, did I answer your question?02:31
prpplaguedvhart: back later?02:31
prpplaguedvhart: yep thanks, just going over a bunch of ME details now02:31
*** davest <davest!Adium@nat/intel/x-dstvqcnqiyjjfjlr> has joined #minnowboard02:49
dvhartyeah, back later (prolly after 8:30 PT)02:49
dvhartheya davest02:49
davestdvhart: ho!02:50
davestfrom 20,000 feet!02:50
dvhartnice :-)02:51
prpplaguedavest: isn't technology grand?02:51
* dvhart struggles to get the LEDs to match what the tool says it is02:51
prpplaguedvhart: when you get back, i'd like to pick your brain on how your envision the lures to be formated02:55
dvhartprpplague, OK, that might be a longer discussion - but we can start. Catch you back here at 8:30 PT?02:56
koenkooioh, food03:06
koenkooigood idea03:06
* koenkooi runs across the parking lot to Jimmy Johns for a sandwich03:07
koenkooidvhart: can you email me your 3.8 bzImage (provided it has an EFI_STUB)? I think I now what's tripping up gummiboot03:12
prpplaguekoenkooi: already back with food03:14
prpplaguekoenkooi: ?03:14
koenkooiit's a 100 yards and there slogan is "subs so fast you'll freak"03:54
*** davest <davest!Adium@nat/intel/x-dstvqcnqiyjjfjlr> has quit IRC03:57
prpplaguekoenkooi: hehe03:57
prpplagueguess dave's plane is out of range03:57
prpplaguekoenkooi: how are we planning to do board revision control? i don't see any gpio pins used for board id, or any write protect eeprom other than the spi flash03:59
koenkooithere's 4MiB of flash for the BIOS, so let's use that :)04:02
prpplaguethats what i figured, but no other fixed hardware method?04:02
koenkooinot that I know off04:02
koenkooiwe might need it, but I try to trust the fw people to come up with something04:02
* prpplague doesn't trust anyone04:03
* koenkooi has figured out the ratio of iced-tea:sweet-tea:ice04:04
* koenkooi afk04:04
dvhartkoenkooi, I'll kick off a build with EFI_STUB and send you the binary (no EMGD, no modules, just boot testing)04:19
dvhartkoenkooi, I have 4 email addresses for you!04:20
* dvhart sends to circuitco04:20
prpplaguedvhart: do we have a meeting scheduled to review the pinout of the expansion header?04:22
prpplaguedvhart: i think koenkooi implied that there was one scheduled04:22
dvhartyes, tomorrow and thursday04:22
prpplagueok, Dave Albert let me know about one tomorrow at 11am cst04:23
prpplaguedvhart: have you gotten an FPDlink display working with this type of design before?04:23
dvhartsorry, need that last 6 minutes - getting kid out of tub and into bed, back in 604:24
prpplaguedvhart: no worries :)04:24
dvhartOK, I"m here again04:32
prpplagueno worries, i have all night ,hehe04:32
dvhartSo, prpplague I think you had a number of questions - I have about 30 minutes.04:32
dvhartwhere did you want to start?04:32
prpplaguedvhart: i think i can hold off with the expansion header questions until tomorrow04:32
dvhartI might be able to go an hour, but after that it's the dog house for me04:33
prpplaguedvhart: for the lure boards04:33
prpplaguedvhart: how did you foresee them mechanically fitting on the top of the minnow?04:33
dvhartWe have a couple of form-factors planned04:34
dvharthalf length, full length, and full-size04:34
dvhartyou'll see the support holds by the usb/eth stack, next to the second RAM chip, and for full-size, the corner holes04:34
dvhartI would prefer to have two at each length, but we'll have to discuss mounting options for X3 as we refactor some of the other components04:35
dvhartJ20 is in the way, for example04:35
dvhartSATA might be as well04:35
prpplagueindeed, plus the mating height for the current connector is some what limited04:35
dvharteach lure would need to pass through the mounting locations and keep the connector in the same spot04:36
dvhartI think there are multiple connector heights04:36
dvhartthat was one reason we chose that connector iirc04:36
dvhartI am a bit concerned with the caps already though - those things are tall!04:36
prpplaguedvhart: yea i just a did a mockup board and have a bunch of issues04:37
dvhartprpplague, great! glad to see that being done already04:37
prpplaguedvhart: we can go over this in more detail tomorrow, i just wanted to make sure i wasn't missing something obvious04:37
dvhartusing the corner mounts is made difficult by eth and usb for example04:37
dvhartnope, the layout poses a variety of issues for the lures04:38
prpplaguedvhart: yea, on the pandaboard5 we had some similar issues that we resolved04:38
dvhartI would move things to the bottom, but that makes it harder to use an aluminum case which replaces the heat sync04:38
dvhartwell, not harder, but it increases the spacing between the board and the bottom heat spreader04:39
prpplaguedvhart: one last item for this evening....  normally on my most of my ARM board designs i select 4 or 5 gpios that i can configure with pullup/pulldowns as a board revision id04:39
prpplaguedvhart: koenkooi was planning to handle board revisions in uefi flash04:39
dvhartYeah, this is x86, so we have SMBIOS04:40
dvhartI'd agree with him on that, save the GPIOS04:40
dvhartthat does mean the FW has to match the board, or you could wreck things I guess04:40
dvhartSomething to ask Lee about04:40
dvhartwe have.... 12 GPIOS from the PCH04:41
dvhartare all of them exported over the expansion header?04:41
prpplaguewe've done this in the past to compensate for things like different boards with various memory configurations04:41
* dvhart checks the spec04:41
prpplaguenot all, but we have the option to work with an i2c gpio expander if needed04:41
prpplaguethis will fall in with the discussions tomorrow on the expansion header04:41
dvhartyeah, we have to lose some signals over that connector anyway it sounds like04:42
dvhartI wouldn't be opposed to losing some buttons or LEDs either04:42
prpplaguedvhart: ahh right that was the other item, do we have anyone looking at the fpdlink support?04:42
dvhart2 of each, leaving four for ID wouldn't be too bad04:42
prpplagueyea that would be plenty04:42
dvhartbut GPIO is pretty popular, hate to tie it up that way.... hrm, will think on it04:43
prpplagueindeed, something to think about04:43
dvhartSo by FPDLink, you just mean LVDS right?04:43
prpplaguehehe, yea lvds (incorrectly named), but yea, hehe04:43
dvhartah, OK - just what I'm used to an dhow it's labeled in the datasheet04:44
dvhartso, we have tested that on crownbay (very similar platform)04:44
prpplagueyea, one of those items that people repeat incorrectly, hehe04:44
dvhartOne issue with it is this CPU uses various GPIO lines for some of the LVDS signals04:44
prpplaguedirectly from uefi, or with a linux build?04:44
prpplagueyea i noticed that as well04:44
prpplaguei have solutions for that, which we can discuss04:45
prpplaguedvhart: ahh ok good04:45
dvhartalthough, that said, a previous system I worked on (also e6xx)04:45
dvharthad the firmware use the LVDS output by default04:45
dvhartand you only got SDVO with X04:45
dvhart(very obnoxious)04:45
dvhartso I'm glad this one defaults to SDVO by default, but I suspect some work could be done to make it work over LVDS if necessary04:45
dvhartnote that LVDS on this board is limited to 1366x76804:46
dvhartwhile the SDVO is 1920x108004:46
prpplaguedvhart: ok, i'll get that on my research list, yea FPDlink-I with single channel04:46
dvhart(despite the EFI FB being 800x600 by default)04:46
dvhartwe need to look at the GPIOs as well and make sure we have the right ones going over the expander to support LVDS04:46
dvhartand discuss how we mux those04:47
dvhartwhat are your thoughts on the buttons/leds04:48
dvhartI wanted plenty to experiment with out of the box04:48
dvhartbut I wonder if 4 of each is excessive04:48
prpplaguedvhart: it varies with the target group, but generally i've found two and two are a good combination04:48
prpplaguedvhart: we can do a wide range of other options with i2c gpio expanders and such so there is a lot of ways to add expansion04:49
dvhartOK, that's what I'm thinking as well. That gives us 4 back for either ID, or a wider GPIO header, etc.04:49
prpplaguedvhart: i'll have some example stuff for the meeting tomorrow to review04:49
dvhartlooking forward to it04:49
dvhartdo you happen to have a current "Minnow Rev x?.pdf" ?04:50
dvhartwith all the netlists, schematics, etc?04:50
dvhartmine is X404:50
dvhartrather old04:50
prpplagueit reports to be the latest. as i got it from dave albert this afternoon for review04:51
dvhartoh, there are two GPIOs from the PCH not being used at all it seems04:51
prpplaguealong with the gerbers04:51
prpplaguedvhart: no worries, i just wanted to put a bug in your ear as to some of the items we need to discuss04:52
dvhartah crap, the lines we need for LVDS are currently used for the switches04:53
dvhartwe're going to have to switch that around and use the PCH GPIOS for the mainboard GPIO04:53
dvhartand route the E6xx GPIOs over the expansion header04:53
dvhartyuk, folks aren't going to like that change now04:53
dvhart(nor am I since I am having to tweak this driver now)04:53
prpplaguedvhart: the lvds stuff isn't a major item, i can show you some basic examples for that04:54
* prpplague has done a lot of lvds design work04:54
dvhartbut we still need the GPIO lines right?04:54
dvhartGPIO_SUS[4:1] ?04:55
prpplaguedvhart: really as long as we have i2c we are good for the lvds04:55
dvhartOK, interested to hear more about that as the CPU datasheet has all kinds of caveats about GPIO_SUS[4:1] hinging on whether LVDS is disabled or not04:56
prpplaguedvhart: ahh there may be cpu related issues on that then, i'll have to review the cpu datasheet for those items as well04:57
dvhart7.5.2.2, Figure 804:58
dvhartGPIO is used for: DDC_CLK, DDC_DATA, VDDEN, BKLTEN, BKLCTL04:58
prpplaguei don't think i got the datasheet today, let me check04:58
dvhartif not, we'll discuss tomorrow and get it to you04:58
dvhartso that's [4:0] actually04:59
prpplaguedvhart: yea i didn't get the cpu datasheet, just the design guides04:59
dvharthere's what you need to know:04:59
dvhartTo support LVDS Backlight Control, the IntelĀ® AtomTM Processor E6xx Series will04:59
dvhartgenerate five types of messages to the display cluster. The display cluster will in turn04:59
dvhartdrive VDDEN and BKLTEN and modulate the duty cycle before driving it out through04:59
dvhartBKLTCTL. These three signals are multiplexed with the normal GPIO pins under the04:59
dvhartLVDS_CTL_MODE (address to be defined). The DDC_CLK and DDC_DATA is emulated04:59
dvhartthrough software.04:59
dvhartthat sums it up nicely04:59
prpplaguedvhart: so it sounds like the cpu has a dedicated ip controller for lvds mode then05:00
dvhartand it muxes with the GPIO signals05:00
prpplagueahh interesting05:00
dvhartI didn't see anything about that in the lpc_sch or the gpio-sch drivers05:00
dvhartthere may be some dragons there05:00
prpplagueneed to read more then05:01
prpplaguelots of work to do :)05:01
dvhartalways :-)05:01
prpplaguedvhart: thanks for the info, i'll get with you at the meeting to discuss these items in more detail tomorrow!05:01
* prpplague wanders off in search of a cup of tea05:01
dvhartsounds good, have a good night.05:01
* dvhart signs off05:02
*** dvhart is now known as dvhart|afk05:02
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has quit IRC09:41
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has joined #minnowboard09:41
* koenkooi yawns13:42
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has quit IRC14:22
*** DevBot <DevBot!~supybot@2001:6f8:12e0::7> has joined #minnowboard14:22
*** davest <davest!Adium@nat/intel/x-qyykcszklhdphqlb> has joined #minnowboard14:35
*** prpplague <prpplague!~prpplague@107-206-64-184.lightspeed.rcsntx.sbcglobal.net> has quit IRC14:39
*** dvhart|afk <dvhart|afk!~dvhart@static-50-53-98-238.bvtn.or.frontiernet.net> has quit IRC15:04
*** dvhart <dvhart!~dvhart@> has joined #minnowboard15:59
dvhartmorning all16:01
davestdvhart: GM16:06
*** zenlinux_ <zenlinux_!~sgarman@> has joined #minnowboard16:32
*** zenlinux_ <zenlinux_!~sgarman@> has quit IRC16:49
*** zenlinux_ <zenlinux_!~sgarman@> has joined #minnowboard16:51
*** davest <davest!Adium@nat/intel/x-qyykcszklhdphqlb> has quit IRC17:15
*** zenlinux_ <zenlinux_!~sgarman@> has quit IRC17:33
*** davest <davest!~Adium@> has joined #minnowboard18:45
*** prpplague <prpplague!~danders@> has joined #minnowboard19:58
prpplaguedvhart: hey sorry we had to leave the meeting19:58
prpplaguedvhart: davealbert gave us an update on the discussion19:59
prpplaguedoh, blocked20:05
prpplaguekoenkooi: looks like we can get 10/12/14mm stacked height connectors20:07
prpplaguemore expensive of course for the taller20:09
prpplagueso we still need to have an effect to work around the caps and the 0.1" headers20:10
prpplagueideal would be the 8mm configuration20:10
prpplaguedvhart: one of the boards i am looking at to do for quick and easy demo Lure boards is a variation of this -  http://www.elinux.org/BeaconBoard20:24
prpplaguedvhart: 7 seg, pwm rgb led, i2c eeprom20:24
prpplaguedvhart: we'll have a lot more room on a lure, but this can be easily used for diagnostics as well as educational20:24
*** prpplague <prpplague!~danders@> has quit IRC20:45
* dvhart looks up21:22
dvhartpwm and eeprom are good21:23
koenkooidvhart: no luck loading your bzImage or booting it gummiboot21:35
* koenkooi is back to grub-efi 21:35
koenkooidvhart: do you have RTC working?21:36
dvharthaven't tried21:36
dvharthaven't even installed the battery21:36
dvhartstill hashing out GPIO gunk21:36
dvhartwhat a pain!21:36
koenkooiroot@minnow:~# hwclock21:36
koenkooihwclock: can't open '/dev/misc/rtc': No such file or directory21:36
koenkooiroot@minnow:~# dmesg | grep -i rtc21:36
dvhartpossibly just a missing driver?21:37
dvhartwill have a look when I get a chance - 5 hours of meetings today21:37
koenkooiroot@minnow:~# zcat /proc/config.gz | grep -i rtc21:37
koenkooi# CONFIG_PM_TRACE_RTC is not set21:37
koenkooi# CONFIG_RTC is not set21:37
koenkooi# CONFIG_GEN_RTC is not set21:37
koenkooi# CONFIG_RTC_CLASS is not set21:37
koenkooithat explains it21:37
dvhartyou can just toss CONFIG_RTC=y into your user.cfg21:37
koenkooiI know21:37
koenkooibut I need to change a lot more21:38
koenkooichange a lot of PCI stuff like ivtv to M21:38
dvhartI haven't looked closely at that at all21:38
* koenkooi gets back to running Lee's test programs in EFI shell21:38
dvhartstill trying to get the core platform stuff going honestly21:38
dvhartso if you have CONFIG changes you want to see, please just send them along21:38
dvharteven just a list is fine21:38
dvhartand I'll incorporate21:38
dvhartotherwise, I'll be getting to it soonish21:39
dvhartkoenkooi, did the 3.8 bzImage help?21:43
dvhartkoenkooi, how eager are you to have a 3.8 recipe?21:43
koenkooivery eager21:48
koenkooicouldn't get it to boot from efi shell21:48
* koenkooi afk for a bit21:48
*** prpplague <prpplague!~prpplague@107-206-64-184.lightspeed.rcsntx.sbcglobal.net> has joined #minnowboard22:14
*** dvhart <dvhart!~dvhart@> has quit IRC22:15
prpplaguedavest: how important is it that we support Intel DPST for an lvds backlight control?22:22
*** davest <davest!~Adium@> has quit IRC22:29
*** davest <davest!~Adium@> has joined #minnowboard22:39
prpplagueany idea where i can purchase a shell bay kit?22:50
koenkooishe sells sea shells at the sea shore22:54
prpplaguekoenkooi: that too22:55
* prpplague needs a board to work oon22:55
koenkooimy board is hooked up to the LA to debug ethernet22:55
koenkooiefi test app says "25MHz"22:55
koenkooiLA says 1.563MHz22:56
koenkooiefi test app says "125MHz"22:56
koenkooiLA still says 1.563MHz22:56
prpplaguehmm, i think someone isn't telling the truth22:57
*** davest <davest!~Adium@> has quit IRC23:24
*** davest <davest!~Adium@> has joined #minnowboard23:29

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!