| 00:04 | kasvain | Artox: yeah just some good recent version that works with i2-ex, so can install jessie. my current bootloader have the bug it can't find devices to boot anything |
| 00:06 | kasvain | unable to get the official installer work in other words |
| 00:07 | Artox | kasvain: try this one please: https://images.solid-build.xyz/IMX6/Debian/sr-imx6-debian-jessie-cli-20151019.img.xz |
| 00:08 | Artox | but decompress it first |
| 00:08 | Artox | I cant help you debug it though, as I need to go to bed approximately 2 hours ago |
| 00:08 | Artox | ttyl |
| 00:09 | Artox | user=debian, password=debian |
| 00:10 | Artox | and, mind that this is still a prerelease |
| 00:16 | kasvain | thanks, just main thing i get the bootloader to work |
| 07:52 | jnettlet | mk01, I fixed that error message. It is a stupid annoying upstream mistake that has been fixed. |
| 10:11 | mk01 | jnettlet: I can't find any discussion in docs about kernel's global workqueue and when this (the queue itself) is started. I have an INIT_WORK and latyer |
| 10:12 | mk01 | later schedule_work(). although the schedule_work() is executed at about +1s, the work only at +2s |
| 10:12 | mk01 | if I change the work to delayed work and put delay to 1,5s, it is executed at +2,5s - so this works. |
| 10:13 | mk01 | just it looks like the global queue would be only 'launched' around +2s ? |
| 10:24 | jnettlet | mk01, the work queues are brought up during early_init that should happen right after the cpus are initialized basically |
| 10:24 | mk01 | hm |
| 10:25 | jnettlet | mk01, where are you scheduling the work? |
| 10:27 | mk01 | basically this started by observing mxc_hdmi -> (intr) -> jitter_timer -> hotplug_work_launch() -> schedule work of hotplug_worker() |
| 10:28 | mk01 | so created dumb driver and the timing is quite the same |
| 10:29 | mk01 | (module_init happens early, then init_work(), at end of init schedule_work and until +2s nothing) |
| 10:30 | mk01 | jnettlet: not crucial for today's work so let's put it into morning's cafe mysteries |
| 10:31 | mk01 | will take 10 more minutes to retest this in 3.14 |
| 12:11 | sanjeevs | Hello all |
| 12:11 | sanjeevs | is HummingBoard Base is compatiable with Quad |
| 12:47 | jnettlet | sanjeevs, yes. All the mx6 carriers are compatible with all the mx6 microsoms |
| 18:23 | vpeter | Any idea what could be the cause: EGL error in CreateSurface: 3003 |
| 18:25 | Artox | vpeter: wild guess: wrong graphics system |
| 18:25 | Artox | (i. e. x11, fb) |
| 18:26 | mk01 | vpeter: one shot or repetitive ? |
| 18:27 | vpeter | Maybe Artox is correct. Investigating. |
| 18:27 | mk01 | but 3003 is failed allocation on start - mostly because of non compatible kernel driver version <> user space libs |
| 18:27 | mk01 | (of galcore) |
| 18:28 | Artox | good point mk01 |
| 18:29 | Artox | I think EGLInitialize will already bail out for windowsys mismatch |
| 18:35 | vpeter | Yes, application crashes too. Missed that part. Let's clean and rebuild. |
| 18:53 | vpeter | Ok, it's not that. Seems I broke something lately. |
| 19:00 | jnettlet | vpeter, it can be there way things are initialized as well. |
| 19:01 | jnettlet | I just had to fix glmark2 for testing |
| 19:07 | vpeter | I changed back to old kernel but seems I lost something from last time it worked. |
| 19:08 | vpeter | Now I have small display working but big one does not :-) |
| 19:08 | vpeter | Last time it was opposite. |
| 19:10 | vpeter | splash in fb I see but kodi doesn't start. |