15:28 <mwalle> xypron: I see an error on layerscape, if efi_allocate_pages() is called (its actually the spin table, which is relocated in arch/arm/cpu/armv8/fsl-layerscape/mp.c) it might be the case that the memory is allocated twice. in my case its also allocated by the lmb for the initrd

15:30 <mwalle> xypron: how is this supposed to work? should efi_allocate_pages() call lmb_reserve()? what about the other way around. Or is this up to the caller?

16:15 <mwalle> xypron: mhh the lmb isn't global, so disregard the lmb_reserve() comment

20:01 <ac_slater> hey guys, anyone here use u-boot on zynqmp? I hate the xilinx forums so I'm going to try here first

20:01 <ac_slater> (maybe my question isn't really zynq specific)

20:14 <ac_slater> In the Xilinx fork of u-boot, they have some hardcoded load addresses https://github.com/Xilinx/u-boot-xlnx/blob/xilinx-v2019.2/include/configs/xilinx_zynqmp.h#L107

20:15 <ac_slater> anyone know WHY these are there? I'm digging in the ARM and Zynq TRMs but can't seem to find those. Just curious if someone knew

20:20 <ac_slater> ah I think I found the answer - no u-boot specific

20:25 <ac_slater> I guess a general question: How does Uboot feed the device-tree to a Linux kernel?