On Tue, Jan 06, 2026 at 08:47:19AM -0700, Theo de Raadt wrote: > Mark Kettenis <[email protected]> wrote: > > > To me, it really feels like the issue is a kernel layout thing. Some > > kernels consistently work, some consistently don't. Even if the only > > difference is a relink. > > random relinking does one surprising thing which has never bit us, but > I've always wondered when it would. > > the sub-sections of each .o end up with alignment requirements based > upon the .o file randomly placed before them. If an alignment requirement > is not correctly defined, there could be problems > > But this hasn't happened yet. > > We've had problems on riscv64 which we thought were related to random > relinking, That is at boot, though. Not sure anymore, the failure mode > is extremely rare. > > I think unhibernate relies upon the page-table mapping of the unhib kernel. > After all, it is the same image. >
sure, all this is correct but the issue is that the checksums aren't mathcing. that should be 100% independent of layout. (see other mail for debugging questions/ideas)
