08:03 <jenneron[m]> digetx: hi, do you mind of having my configs in grate? minimal defconfig (supports 8 devices) and config fragment with options needed by pmOS for contributors who test kernel in pmOS systems. both configs make 3.1M zImage

09:16 <digetx> jenneron[m]: hi, I don't mind

11:20 <tagr_> pgwipeout[m]: yeah, I suppose U-Boot could be taught to transcribe the downstream DTB into an upstream-compatible DTB

11:20 <DavidHeidelberg> cool, after merge I'll add CI also for grate_defconfig

11:21 <tagr_> my preference would be for that not to be necessary, not least because the upstream-compatible DTB content is basically just forwarding the binary structure that nvtboot writes to memory

11:23 <tagr_> I've got patches to make nvtboot write both versions and a patch to U-Boot to basically copy the upstream-compatible bits from the downstream DTB to the upstream DTB (usually loaded externally, via extlinux.conf or BOOTP or whatever)

11:23 <tagr_> at this point it's mostly a matter of figuring out how/when to release it

11:26 <kwizart> tagr_, nvtboot is closed source from our POV , right ? same as cboot (for tegra210 at least) any chances that this change ?

11:28 <kwizart> I was wondering quite a similar question on forum https://forums.developer.nvidia.com/t/efi-support-for-jetson-tx1-and-running-upstream-kernel/187945/3 but unfortunately no "clean solution for EFI on TX1"

11:29 <kwizart> (exept building upstream u-boot, but that broke switching with downstream userspace)

11:32 <kwizart> ha, nvtboot is the bootrom, so basically it means too late for TX1 and will be available on newer SOC ?

11:33 <cyndis> nvtboot is not the bootrom

11:34 <cyndis> it's one of the bootloader stages

11:35 <kwizart> ok, I've miss-read the appnote. got it

12:04 <tagr_> yeah, nvtboot and cboot are both closed source on Tegra210 and I don't know if that's ever going to change, hence why I'm pushing to get these changes in before it becomes EOL

12:09 <tagr_> kwizart: what exactly is it that keeps downstream userspace from working with an upstream kernel? I mean the absence of the downstream kernel drivers, obviously, but that's going to be a problem whether you boot with upstream U-Boot or downstream U-Boot)

12:10 <tagr_> to my knowledge, using upstream U-Boot only has an influence on the downstream kernel

12:11 <tagr_> or are you trying to boot a downstream kernel in EFI mode using an upstream U-Boot?

14:43 <kwizart> tagr_, I'm trying to keep downstream uboot for all, which "barelly works" to boot upstream kernel with EFI (and only a little hack) on external MMC; and downstream kernel in eMMC (using downstream rootfs partition layout, etc)

14:45 <kwizart> So I know I will have the memory training issue as you've describes, so I really hopes pushing that for nvtboot/cboot/uboot can still be done as I expect Jetson Nano will benefit from it. (so it could also be done on JTX1)

14:46 <kwizart> but I wonder if others issue may arise from using downstream uboot with upstream kernel (like iommu setup, gpu enablement, etc ?)