| 00:01 | _rmk | 00:01 * _rmk_ puts some music on... this hotel's wonderful because the TV audio also comes out in the bathroom |
| 00:04 | flixh | cbxbiker61: hm, it does not look like dm-mod is there |
| 00:10 | cbxbiker61 | flixh, i just looked in the Modules archive, it's in there |
| 00:11 | cbxbiker61 | make sure it pulls from release/2, also wouldn't hurt to rm -rf /lib/modules/3.11.6 before running UPDATE-KERNEL.sh |
| 00:12 | flixh | strange, not after running UPDATE-KERNEL.sh in /lib/modules, trying |
| 00:13 | cbxbiker61 | you could also "tar tf" the Modules archive to make sure there are no errors |
| 00:15 | flixh | having rm'd the dir and running UPDATE-KERNEL.sh again it's still missing. |
| 00:16 | cbxbiker61 | manually download the Modules archive and md5 from release/2 directory, maybe it's a minor glitch in my web server strategy |
| 00:16 | cbxbiker61 | i have a cdn and that may be the problem |
| 00:18 | cbxbiker61 | did you make sure to rm the *3.11.6* files, if you don't...it wont download the newer version |
| 00:18 | flixh | i guess that was the problem, it's redownloading now |
| 00:27 | flixh | thanks a lot, that seems to work now. need to sleep, bye |
| 09:47 | jnettlet | found the offending patch causing the mmc instability |
| 10:52 | jnettlet | rabeeh, pushed a fix for the sdhc card detection problems. the solidworks-imx6-3.10 kernel should now boot. |
| 10:53 | jnettlet | DRI still seems flaky |
| 10:53 | jnettle | 10:53 * jnettlet needs to work on other things for a while now |
| 13:37 | jnettlet | dv_, otavio, trying to boot up yocto and am losing my serial console when X starts up. |
| 13:37 | jnettlet | any idea what could be up with that? |
| 14:45 | dv__ | jnettlet: huh? never saw that |
| 14:46 | jnettlet | dv__, damn. was hoping it was some yocto thing |
| 14:46 | jnettlet | I do have the yocto splash screen coming up on hdmi |
| 14:46 | jnettlet | although ethernet is only sending packets out :-{ |
| 14:52 | jnettlet | Oh interesting it is definitely something with X |
| 14:52 | jnettlet | booting to init3 gives me a console with working hdmi and keyboard |
| 15:08 | _rmk_ | robclark: looks like armada drm does need a depends on ARM |
| 15:08 | _rmk_ | _relaxed accessors |
| 15:08 | robclark | ahh |
| 15:08 | robclark | well, even if it depends on ARM, if it is able to not depend on ARMADA for compile tests that would be good enough |
| 15:08 | _rmk_ | just generating a patch for that |
| 15:09 | robclark | ie. if I can do one x86 and one arm build to test core changes, then I am a happy camper :-P |
| 15:09 | robclark | (as opposed to N arm builds) |
| 15:10 | _rmk_ | I'm thinking about just "depends on ARM" |
| 15:10 | _rmk_ | since that's what the writel_relaxed needs |
| 15:10 | robclark | yeah, that sounds like the right thing |
| 15:13 | _rmk_ | top most commit of: http://ftp.arm.linux.org.uk/cgit/linux-cubox.git/log/?h=drm-tda998x-3.12-fixes |
| 15:30 | _rmk_ | robclark: happy with that ^^^ ? |
| 15:31 | robclark | _rmk_, yup |
| 15:32 | _rmk_ | thanks, I'll send airlied either the patch or another pull for it |
| 15:32 | _rmk_ | with your ack? |
| 15:32 | robclark | _rmk_, sounds good |
| 15:33 | _rmk_ | thanks |
| 15:33 | _rmk_ | Acked-by: Rob Clark |
| 15:33 | _rmk_ | ? |
| 15:33 | _rmk_ | (that's from memory) |
| 15:33 | robclark | _rmk_, d in there: [email protected] |
| 15:34 | _rmk_ | ta. |
| 15:34 | robclark | np |
| 15:36 | _rmk_ | sent. |
| 16:08 | jnettlet | okay unaccelerated HDMI via the framebuffer is running on the 3.10 kernel |
| 16:08 | jnettlet | just need to figure out what is up with ethernet |
| 16:13 | dv_ | damnit |
| 16:13 | dv_ | what the hell is going on with my freenode connection |
| 16:24 | jnettlet | what the hell is going on with the stupid ethernet driver on this stupid kernel |
| 16:24 | jnettle | 16:24 * jnettlet thought we had this all worked out. |
| 16:29 | jnettlet | dv_, have you tested gpu acceleration on the carrier-1 yet? |
| 16:29 | jnettlet | wumpus, have you tested the gpu on the C1 yet? |
| 16:30 | jnettle | 16:30 * jnettlet has to leave. just leave a message will be back in 45 |
| 16:31 | dv_ | jnettlet: I tried it with the EGL sink, which worked |
| 16:34 | wumpus | jnettlet: yes |
| 16:34 | wumpus | I've run the etnaviv demos, they worked |
| 16:37 | wumpus | this was with the linux kernel from solidrun/linux-imx6 |
| 16:37 | dv_ | (my modified gstreamer EGL sink) |
| 17:37 | jnettlet | dv_, wumpus_, okay thanks. with the 3.10 kernel I am just getting a hardware hang. Need to sort that out. |
| 17:48 | jnettlet | okay have that sorted out. Now just need ethernet |
| 17:58 | wumpus | jnettlet: what was the problem? |
| 17:58 | jnettlet | wumpus, device-tree config. my fault. |
| 17:58 | wumpus | okay |
| 17:59 | jnettlet | although the Xorg driver doesn't draw properly. |
| 17:59 | jnettlet | but that is fine. I can fix that. |
| 17:59 | wumpus | hmm but that's a sw issue? |
| 17:59 | wumpus | or the clock is wrong :) |
| 18:00 | jnettlet | most likely a software issue as it isn't refreshing at all. Just random pieces. |
| 21:45 | jnettlet | okay changes pushed for the 3.10 alpha kernel repo. still missing ethernet. It comes up and detects link properly but won't pass packets to save itself. |
| 21:45 | jnettlet | other than that I think everything else is working. |
| 21:48 | dv__ | well, the VPU and IPU stuff is missing. but thats obvious. |
| 21:48 | dv__ | what about galcore and/or the DRM patches though? |
| 21:50 | jnettlet | dv_, root@solidrunc1:~# dmesg | grep -i vpu |
| 21:50 | jnettlet | mxc_vpu 2040000.vpu: VPU initialized |
| 21:50 | jnettlet | root@solidrunc1:~# dmesg | grep -i ipu |
| 21:50 | jnettlet | imx-ipuv3 2400000.ipu: IPU DMFC NORMAL mode: 1(0~1), 5B(4,5), 5F(6,7) |
| 21:50 | jnettlet | imx-ipuv3 2400000.ipu: IPU DMFC DP HIGH RESOLUTION: 1(0,1), 5B(2~5), 5F(6,7) |
| 21:51 | dv_ | oh? then where are the headers? |
| 21:51 | dv_ | try to build imx-lib with your kernel headers |
| 21:51 | jnettlet | dv_, all on github. |
| 21:51 | dv_ | hmm .. right, I forget that this is based on the boundarydevices work |
| 21:51 | dv_ | still, imx-lib didnt build with it |
| 21:51 | jnettlet | dv_, nope I ported everything over to freescale's 3.10 kernel |
| 21:52 | dv_ | okay. can you try to build imx-lib with it? |
| 21:52 | dv_ | perhaps I missed something |
| 21:53 | jnettlet | dv_, probably not right away. in the middle of making dinner :-) |
| 21:53 | dv_ | k :) |
| 22:01 | jnettlet | dv_, https://github.com/linux4kix/linux-imx6/tree/solidrun-imx_3.10.9_1.0.0_alpha can be built if you want to test yourself. |
| 22:03 | dv_ | I already did |
| 22:03 | dv_ | I built the kernel. but the imx-lib did not build with its headers. |
| 22:08 | dv_ | otavio: was it you who told me that it is a better idea to use libfslvpuwrap instead of imx-lib ? |
| 22:09 | dv_ | because the VPU parts of imx-lib will eventually get factored out into their own library? |
| 22:09 | otavio | dv_: it has been; in 3.10.9 |
| 22:20 | jnettlet | well that might explain things |
| 22:25 | dv_ | otavio: where is this new library ? |
| 22:27 | otavio | dv_: imx-vpu |
| 22:27 | dv_ | otavio: ah |
| 22:27 | otavio | dv_: check master-next |
| 22:27 | dv_ | I am tempted to rebase my plugins on top of that |
| 22:28 | dv_ | the VPU wrapper has critical design mistakes |
| 22:28 | otavio | dv_: The split has been also done for mx53; the 11.09.02 imx-vpu and imx-lib packages did the same. |
| 22:28 | dv_ | can you explain me this strange versioning system? |
| 22:28 | dv_ | 3.10.9 vs 11.09.02 ? |
| 22:29 | otavio | dv_: I am not against it but as I said it'd be good to have a layer on top of this mess (using macros or so) |
| 22:29 | otavio | dv_: it is a mess |
| 22:29 | otavio | dv_: no logic |
| 22:29 | dv_ | reminds me of a certain company called "vivante" |
| 22:29 | otavio | dv_: they used the name it as release date |
| 22:29 | dv_ | why is it so damn difficult to come up with a consistent versioning system? |
| 22:29 | otavio | now as kernel version |
| 22:29 | dv_ | its not rocket science |
| 22:29 | otavio | agreed |
| 22:30 | dv_ | otavio: the problem is that the layer on top of this mess is also a mess :) |
| 22:31 | otavio | dv_: but it will be your own controlled mess |
| 22:31 | otavio | dv_: which makes it doable |
| 22:31 | dv_ | I mean the VPU wrapper |
| 22:31 | otavio | dv_: you can define well defined api |
| 22:31 | otavio | and make it to vpu internalls |
| 22:32 | dv_ | the VPU wrapper's code is chaotic, full of #ifdefs , zero explanations.. |
| 22:32 | dv_ | which is why I am eyeing the imx-vpu library |
| 22:33 | dv_ | using the VPU wrapper code as a reference for imx version specific details |
| 22:34 | dv_ | I still wish pengutronix doesnt make my work obsolete though ... :/ although it does sound like their patches wont see the light of day in the near or midterm future. your opinion? |