05:35 <digetx> jonasschwoebel[m: I don't understand the question, nvflash has miniloader blob embedded into it

09:23 <digetx> kwizart: usb fix is pushed to grate kernel, please give it a test

09:27 <kwizart> in progress, thanks

11:12 <kwizart> digetx, okay xusb is back again on jetson-tk1 with current grate, worth to scheudle the patch for 5.15 as the regresion first appeared there IIRC

11:15 <kwizart> digetx, Testing Xorg on grate tk1 is working, (fedora 35, with updated xorg-server 1.21) but I have the tearing issue

11:15 <kwizart> wayland is discarded, but I have no reason why yet...

11:15 <kwizart> I might have blacklisted nouveau on earlier attempt

11:17 <digetx> do you have a working opengl? are you using compositor that supports vsync?

11:17 <digetx> I'll send the usb patch soon

11:19 <digetx> could you please also test that usb works after re-plug and after suspend?

11:19 <digetx> and that there are no error messages in kmsg

11:22 <kwizart> digetx, I'm using gnome 41, so I expect mutter support vsync

11:22 <jenneron[m]> digetx: i'm sorry for the test taking so long, but i'm getting this with vdpau on T114 https://dpaste.com/CH45AP5DZ

11:23 <jenneron[m]> i use this file https://download.blender.org/peach/trailer/trailer_720p.mov

11:24 <digetx> jenneron[m]: what's in the kernel log?

11:24 <jenneron[m]> https://dpaste.com/FNHZNY7ZP

11:25 <jenneron[m]> this is my patch https://dpaste.com/A5UXV3TWK

11:25 <digetx> could you try to remove "iommus" from DT?

11:25 <digetx> ah

11:26 <digetx> yes, it shouldn't work with iommu

11:26 <digetx> need an extra patch for that

11:26 <kwizart> digetx, serial output: https://paste.centos.org/view/2baebc2f I can unplug/replug usb, it still works, but I have an issue when entering suspend (then I don't know how to leave suspend, keyboard stroke doesn't work)

11:28 <digetx> kwizart: use `rtcwake -s10 -mmem`, it will suspend and resume after 10 seconds

11:29 <digetx> jenneron[m]: for now please try to remove iommus, I'll try to add the extra patches and then ping you to test

11:30 <jenneron[m]> ok

11:31 <kwizart> digetx, okay, using power button on the board seems to have the needed effect to leave suspend. Then usb still works...

11:31 <digetx> alright, thanks

11:31 <kwizart> digetx, interrupt 150 is related to PCIe enp1s0, so unrelated to anything USB

11:32 <jenneron[m]> still present without iommus

11:33 <digetx> show me new kernel log

11:37 <jenneron[m]> without iommus

11:37 <jenneron[m]> mpv https://dpaste.com/BKMN8TU57

11:37 <jenneron[m]> kmsg https://dpaste.com/32PXDZ5Q2

11:37 <jenneron[m]> although video seems to play fine

11:39 <digetx> jenneron[m]: disable CONFIG_TEGRA_IOMMU_SMMU, it falls back to software decoding

11:39 <digetx> display iommu also needs to be disabled

11:39 <jenneron[m]> oh well T114 is powerfull enough :D

11:39 <jenneron[m]> will disable and test

11:40 <digetx> powerful until cpu will get very hot

11:44 <jenneron[m]> without CONFIG_TEGRA_IOMMU_SMMU https://dpaste.com/79KK6CEFA

11:45 <digetx> right, memory doesn't start with 0x0 on t114

11:46 <digetx> need extra patches then