| 08:18 | jnettlet | rabeeh, are you happy with the u-boot 2013 carrier-1 support? The yocto guys want a patch submitted upstream so they can pull from it. |
| 10:10 | rabeeh | jnettlet / otavio: one big item missing in u-boot is SPL support where the idea is having a single u-boot that supports all different SoC and memory configs |
| 10:34 | jnettlet | rabeeh, I don't think the patches to support SPL multi-boot are upstream yet. |
| 10:34 | rabeeh | they are not |
| 10:34 | jnettlet | I have seen some discussion on the mailing list, but didn't think a "correct" way had been agreed upon yet. |
| 10:35 | rabeeh | http://lists.denx.de/pipermail/u-boot/2013-August/161452.html |
| 10:36 | jnettlet | I haven't seen his patches yet.d |
| 10:36 | rabeeh | neither i |
| 10:36 | rabeeh | i'v sent him an email though (offering C1 too) |
| 10:37 | jnettlet | I guess if we push the carrier-1 code mainline he may add support to it when he pushes his patches. |
| 10:37 | rabeeh | that's a good option |
| 10:37 | rabeeh | when do you think this will happen? i have minor typos to fix |
| 10:38 | jnettlet | there has been some more recent discussions in Sept. I think otavio would know best as they are working on the wandboard support. |
| 10:39 | jnettlet | the dual and quad configs still need to be added to the 2013 codebase. |
| 10:39 | jnettlet | If I add that in can you test? |
| 10:41 | rabeeh | yes |
| 10:45 | jnettlet | yeah are the ram configs for the other microsoms? 800mhz 4x256? and 1066mhz 4x512? |
| 10:46 | rabeeh | yes |
| 10:47 | rabeeh | solo - 800mhz, 2 devices by x16 256Mb |
| 10:47 | rabeeh | dual lite - 800mhz, 4 devices by x16 256Mb |
| 10:47 | rabeeh | dual - 1066mhz, 4 devices by x16 256Mb |
| 10:47 | rabeeh | quad - 1066mhz, 4 devices by x16 512Mb |
| 10:48 | jnettlet | and everything but the solo is 64-bit |
| 10:48 | rabeeh | yes |
| 10:49 | jnettlet | got it |
| 11:09 | rabeeh | jnettlet: just sent you two patches for small typos |
| 11:38 | jnettlet | rabeeh, in your code I just see three ddr configurations. 32-bit and 64-bit mx6dl and then one for mx6q |
| 11:39 | rabeeh | yes |
| 11:39 | jnettlet | I guess the imx6d and imx6q use the same ram config |
| 11:39 | jnettlet | I don't see anywhere to distinguish between 1GB and 2GB's |
| 11:42 | rabeeh | jnettlet: true; that one was missed |
| 11:42 | rabeeh | the most important configuration for the memory is the write leveling values which are 100% dependent on how to boards are configured |
| 12:28 | jnettlet | rabeeh, I have pushed all the changes for the other boards to github. you can run make mx6_c1dl_config/mx6_c1d_config/mx6_c1q_config to switch between the different board configs |
| 12:28 | rabeeh | great. i will test them. |
| 12:28 | rabeeh | thanks |
| 12:28 | jnettlet | let me know how it works out. I need to walk the dogs. |
| 15:22 | jnettlet | Has anyone released official distro ports for the Carrier-1 yet? |
| 15:24 | _rmk | 15:24 * _rmk_ is back |
| 15:24 | jnettlet | *and the crowd goes wild* |
| 15:24 | _rmk_ | asoc is now sorted for cubox :) |
| 15:25 | jnettle | 15:25 * jnettlet can't wait to see the final solution |
| 15:25 | jnettlet | as I probably need to do the same to properly fix hdmi audio on the XO-4 |
| 15:25 | _rmk_ | however, there may be disruption for me in the next few days: the met office are forecasting a storm worse than that of October '87 which brought a lot of destruction to the southern UK for the next couple of days. |
| 15:26 | jnettlet | yeah I was reading about that earlier. Supposed to be a real doozy |
| 15:26 | _rmk_ | and we have very similar conditions: the ground is saturated and the trees still have most of their leaves. |
| 15:26 | _rmk_ | perfect for lots of trees to be uprooted |
| 15:26 | jnettlet | hopefully you don't have lots of trees near your house. |
| 15:27 | jnettlet | or at least a good chainsaw |
| 15:27 | _rmk_ | we have some. we have one which the top came out of in '87 |
| 15:27 | _rmk_ | blue atlas fir tree |
| 15:28 | _rmk_ | anyway, I need to unpack, but more importantly, unpack the cubox :) |
| 15:28 | jnettlet | enjoy |
| 15:29 | _rmk_ | and drink coffee :) |
| 15:32 | jnettlet_ | and the dog knocked my computers powercord out of the wall :-) |
| 16:27 | _rmk_ | UK 13 amp plugs don't suffer from that problem :) |
| 16:30 | jnettlet | one of the most annoying parts about the EU are the outlets |
| 16:31 | _rmk_ | ours may be bulky but one thing they generally don't do is fall out even with assistance from pets |
| 16:33 | Bluerise | jnettlet: Why are those annoying? |
| 16:34 | _rmk_ | the different types coupled by the ease of accidental disconnection? |
| 16:34 | jnettlet | Bluerise, because they are enormous yet easily fall out of the socket. |
| 16:35 | Bluerise | Really? |
| 16:35 | Bluerise | It's rather hard here to get mine out of the socket |
| 16:35 | _rmk_ | here = ? |
| 16:35 | Bluerise | Germany |
| 16:36 | Bluerise | We're talking about those, right? http://en.wikipedia.org/wiki/File:Stikkontakt.jpg |
| 16:36 | Bluerise | http://en.wikipedia.org/wiki/File:Schuko_plug_and_socket.png |
| 16:37 | Bluerise | those are a bit easier to get out: http://en.wikipedia.org/wiki/File:Euro-Flachstecker_2.jpg |
| 16:37 | jnettlet | The ones in Denmark generally aren't that deep and don't have the notches in the side of the plug |
| 16:38 | _rmk | 16:38 * _rmk_ gatecrashed (well, not really) the ELCE end game last night, along with many ARM Ltd folk and a few others. |
| 16:39 | jnettlet | _rmk_, yeah so did you solve all of the problems faced by ARM Linux or what? |
| 16:39 | _rmk | 16:39 * _rmk_ won... nothing :( |
| 16:39 | _rmk_ | jnettlet: oh, that'll be DT then :) |
| 16:39 | _rmk_ | There were discussions about DT from day 1 through til the end. |
| 16:39 | _rmk_ | imx-drm is going to get a rewrite and I think I've probably gained that job. |
| 16:40 | _rmk_ | we've got an idea about how to approach sorting out DRM vs DT |
| 16:41 | jnettle | 16:41 * jnettlet can't wait for the writeup. |
| 16:42 | jnettlet | Bluerise, these are pretty common in DK. https://lh6.googleusercontent.com/-Ir17PK_P8tc/UmvT-kl-8BI/AAAAAAAAFik/fJFUHCvR2X8/w644-h858-no/IMG_20131026_163810.jpg |
| 16:42 | _rmk_ | I don't think we elected for anyone in particularly to do any kind of writeup |
| 16:42 | _rmk_ | s/ly// |
| 16:43 | jnettlet | will keep an eye on the mailing lists then. |
| 16:44 | jnettle | 16:44 * jnettlet made some good progress on backporting _rmk_'s drm patches and needed supporting patches to the imx-3.10 kernel |
| 16:44 | jnettlet | HDMI on the KMS driver is not working for some reason but haven't tackled that yet. |
| 16:45 | jnettlet | also been beating the crap out the galcore driver bringing it in line with IMX latest and MRVL latest. Think I am just about there. |
| 16:46 | _rmk_ | I mentioned that to the pengutronix people |
| 16:46 | jnettlet | although now I have found another series of patches in the boundary tree that I need to evaluate and merge in. |
| 16:47 | _rmk_ | does this mean we have later galcore libs for dove? |
| 16:47 | jnettlet | not yet. I opened a ticket with Marvell, but given the state of OLPC I am not sure if I can get any traction. |
| 16:48 | jnettlet | But I have sifted through the differences and with only minor ifdef's can support both drivers in a single codebase. |
| 16:49 | _rmk_ | sounds like what I did with my version when I used the pre-armhf libs |
| 16:49 | _rmk_ | or... post armhf libs with rabeeh's kernel space |
| 16:49 | _rmk_ | btw, I met shesselba :) |
| 16:49 | jnettlet | yeah I saw you guys finally found each other. |
| 16:50 | jnettle | 16:50 * jnettlet wishes he could have made it. |
| 16:51 | _rmk_ | I need to check that my cubox has survived properly: yesterday it stopped driving the hotel's tv at 1080p @60 or @50Hz but would @25Hz |
| 16:51 | _rmk_ | @50Hz, it would occasionally show a picture but with green spots all over and blank |
| 16:51 | _rmk_ | I suspect crappy stuff on the hotels part rather than the cubox, but I want to be sure |
| 16:52 | _rmk_ | and... I need to sort out my cubox git tree again: it's had quite an amount of stuff piled into it over the last four days, especially for NX kernel mappings |
| 16:53 | jnettlet | should I wait for you to clean that out before doing a pull? |
| 16:53 | _rmk_ | not pushed anything out yet |
| 16:53 | jnettlet | ah okay great. |
| 16:53 | _rmk_ | and my old phone lasted the entire conference, being on 24/5 :) |
| 16:53 | _rmk_ | including getting a call or two |
| 16:54 | jnettlet | that is still the benefit of "dumb" mobile phones |
| 16:56 | jnettlet | If you have any thoughts on the strange CPU_IDLE behavior I am seeing with the MX6 I would love to hear them. I figured if anyone would know about those interactions within the kernel it would be you. |
| 16:56 | _rmk_ | its the only time the battery goes below 100% on the display |
| 16:56 | _rmk_ | I couldn't get cpuidle to work on 3.12-rc: it seemed to be missing some stuff |
| 16:57 | jnettlet | _rmk_, there are a bunch of patches in the IMX tree for it. |
| 16:57 | jnettlet | I can probably give you a short list of needed patches if you want. |
| 16:57 | _rmk_ | btw, from the sounds of it, pengutronix does no upstreaming of anything unless there's a customer asking for it to happen |
| 16:58 | Coolgeek | I saw that you are talking to each other some stuff. What is it ? Hardware decoding audio and video ? |
| 16:58 | jnettlet | that is kind of what I gathered. |
| 16:59 | _rmk_ | I think Grant Likely, airlied, myself and a few others knocked the CDF people on the head and basically said no to them too. |
| 16:59 | _rmk_ | essentially, Framework = bad. library = good. |
| 17:00 | jnettlet | oh great! That had clusterf**K written all over it. |
| 17:00 | _rmk_ | even airlied wishes that DRM was a library rather than being a layer between drivers and userspace |
| 17:03 | jnettlet | yeah like the difference between an idea and a belief. You can always change an idea, a belief is a little bit harder. |
| 17:03 | jnettlet | Once it is in the kernel you can't break userspace, and that can make progress a little slow |
| 17:05 | _rmk_ | the big problem with frameworks its it is very easy to misdesign them, and then not be able to fix them later |
| 17:08 | jnettle | 17:08 * jnettlet hopes that rabeeh's silence is him doing non-computer related things and not pouring through broken u-boot support for the other C1 board types. |
| 19:26 | rabeeh | jnettlet: just went to lunch with family; i'v put this aside since i need to spin the C1 too |
| 22:01 | jnettlet | rabeeh, no worries. we are getting pretty caught up with features at this point that there is no need to rush. |
| 23:48 | Bluerise | Why the do you license device tree sources as GPL? |
| 23:48 | Bluerise | -do+heck |
| 23:48 | Bluerise | +do |
| 23:49 | rabeeh | Bluerise? |
| 23:49 | rabeeh | do you want to reuse for openbsd? |
| 23:49 | Bluerise | Oh, I mean "Linux" in general :) |
| 23:49 | Bluerise | not "you" as in SolidRun |
| 23:49 | Bluerise | It'd be much easier if I could just pull the device tree sources and put it into OpenBSD/Bitrig |
| 23:49 | Bluerise | FDT is the way to go. |
| 23:50 | Bluerise | Looks like I will have to manually craft the device trees. and in the end it will look like Linux's, or very similar... |
| 23:50 | Bluerise | imho something like that should be public domain |
| 23:54 | Bluerise | Not sure how compatible it is to OpenBSD. if it's only a single file which isn't compiled into (as it's not really code), it might not be that problematic. |
| 23:54 | Bluerise | or some files. binutils (up to 2.17) is also in the repo, which is GPLv2 |
| 23:54 | Bluerise | but it's still a little bit of a headache. |
| 23:56 | _rmk_ | Bluerise: how about trying to talk to people about it sensibly? |
| 23:57 | _rmk_ | like "can we add license X to the DT files?" |
| 23:58 | rabeeh | is there a restriction in the kernel having an additional license besides GPLv2? |
| 23:58 | Bluerise | I'm sorry. :( It's just often frustrating how much is GPL and incompatible, although we have the goals. |
| 23:58 | rabeeh | if there isn't then i don't think there is a problem |