I'm not positive on this one. Sorry for the noise if I'm wrong, but... I'm running profile hardened/x86/2.6. The current stable linux-headers, 2.6.27-r2, is newer than the current stable hardened-sources, 2.6.25-r11. I haven't merged these yet, but isn't this bad. If I understand correctly from previous LKML posts, even by Linus himself, it's imperative that the software is *not* built against kernel headers that are newer than the running kernel, as they would likely attempt to access an ABI that is not yet supported by the running kernel. Perhaps this situation has changed. If so, please enlighten me.
Don't worry it is safe. You can even rebuild glibc against linux-headers-2.6.27 and continue to run an older kernel.
Sorry for being paranoid, but the box I'm looking at currently is a remote headless server. A failure is incredibly inconvenient. Can you point me to some documentation on why this is safe or when things changed since posts like [ http://osdir.com/ml/linux.gentoo.server/2005-03/msg00097.html ]?
Sorry, I'm just reopening this bug hoping to get some attention to the followup question in comment #2.
Well, as there's been no further response, and the situation with the versions has now reversed anyway, I'll go ahead and switch this back to INVALID.