On 12/13/24 01:54, Yan Zhao wrote: > Accept the destination addresses during the kexec load, immediately after > they pass sanity checks. This ensures the code is located in a common place > shared by both the kexec_load and kexec_file_load system calls. So, we've got an end-user-visible bug. Eric raised some good concerns about the hardware and firmware design, but I think they've all been addressed. The only other even solution I've seen proposed is to not do unaccepted memory in the first place. I don't think that's viable or justified, especially since we've got at least one end user[1] that seems to think unaccepted memory fits their needs. This bug can _probably_ be fixed in arch/x86 as well, but having the solution in general code seems like the right place to me: Acked-by: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx> Andrew, it seems like a lot of kexec work flows through you. Are you the right one to pick this up? 1. https://lore.kernel.org/all/CAMGD6P3r-S-Va-TRvVjZ808on9+-wFJ_VeTpQ+FEN1jBbhmnXw@xxxxxxxxxxxxxx/
