07:42 <Myy> にゃ！

07:42 <Myy> Assertion p_hal->vcodec_type & ((0x00000200) | (0x00000001) | (0x00000002)) failed at hal_h264d_init:187

07:42 <Myy> Is that a known issue ?

07:43 <Myy> I guess that this refers to : mpp_assert(p_hal->vcodec_type & (HAVE_RKVDEC | HAVE_VPU1 | HAVE_VPU2));

07:44 <Myy> So something obviously went wrong with the VPU driver I guess

07:45 <Myy> I got the same error, whether I use the mpp test suite or LongChair's patched mpv with MPP support.

07:50 <Myy> Well, adding some mpp_log_f to log p_hal->vcodec_type value, it seems that its value is 0 at that point...

07:50 <rperier> mhhhhh... starting linux 4.13-rc2 from NAND (via RK bootloader), then loading the same kernel via kexec, ethernet is working

07:50 <rperier> LEDs blue and LEDs green are fixed

07:50 <rperier> that's not the case when loading kernel via uboot, it's blicking all the time

07:51 <rperier> interesting, perhaps a regulator issue...

07:58 <Myy> Are special compilation options required for the release version of MPP ?

08:24 <Myy> Turns out that every call to mpp_get_vcodec_type seems to return 0

08:58 <Myy> Alright, turns out that -DHAVE_DRM=ON and -DRKPLATFORM=ON were needed

08:58 <Myy> Now I got VPU driver issues with nice oops messages in the logs

09:00 <Myy> https://ghostbin.com/paste/kuum8

09:50 <mrjay> nand is working ... mkimage makes rk3066 nand boot image ... i can also flash this boot image with nand driver ugly hack

09:54 <mmind00> mrjay: you are a hero for making this work :-)

09:55 <mrjay> mmind00: hehe not a hero ... just stuborn

09:55 <naobsd> awesome!

09:56 <mrjay> all related to nand in rk3066 has weir format

09:56 <mrjay> weird*

14:56 <rperier> did you notice in "regulator_summary" (debugfs) that 10204000.ethernet (subnode of VCC_RMII) is set to 0mV ? that's normal

14:56 <rperier> ?

14:56 <rperier> (compared to the other regulators I mean)

15:29 <ayaka> rperier, rk3036?

15:34 <rperier> ayaka: no rk3188

15:35 <rperier> that's with the rock

15:36 <ayaka> ok, I don't have that board

16:34 <Myy> I've added some debug messages to the VPU driver, in order to understand where things went wrong. It seems that vcodec_iommu_import returns -ENOMEM for some reason. I'll try to investigate why, but if anyone has any idea about why this could happen, I'm all ears.

16:34 <Myy> vcodec_iommu_import(ed150000, eebde000, 21) → -12

17:20 <Myy> The issue seems to be triggered by this code : https://github.com/Miouyouyou/rockchip-vcodec/blob/master/vcodec_iommu_drm.c#L329

18:24 <Myy> ayaka, wzyy2, when you got time, I'll be trying to pinpoint my issue with the VPU here : https://github.com/Miouyouyou/rockchip-vcodec/issues/3 . If you have any idea about what's going on, feel free to let a message there

18:24 <Myy> *I'm