https://bugzilla.kernel.org/show_bug.cgi?id=205089 --- Comment #30 from David Nichols (david@xxxxxxxx) --- The amdgpu problems in my system were completely and definitively resolved with a memcpy() patch to glibc: https://gist.github.com/jnettlet/f6f8b49bb7c731255c46f541f875f436 The SoC I'm using (NXP LX2160A - SolidRun HoneyComb system) has a known bug regarding PCI device memory writes that can be completely addressed with a simple reordering of the assembly instructions in the arch-specific memcpy() implementation. In any case, this is not a kernel bug for me after all. I can't comment on the source of the problem for others who most likely are running an x86_64 kernel. -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
