Sorry about this, there was a file I forgot in my commit command line for a
change two weeks ago:
https://mail-index.netbsd.org/source-changes/2020/05/23/msg117587.html
The lack of call to elf_build_info() caused a sanity check in the prekern to
fail.
Thanks for reporting
Le 22/05/2020 à 12:42, Chavdar Ivanov a écrit :
After the latest updates to /usr/mdec/prekern, now I am getting
------------
FATAL
elf sym lookup: symbol beyond table
------------
when I try to load the KASLR kernel.
Chavdar
On Tue, 5 May 2020 at 21:26, Chavdar Ivanov <ci4...@gmail.com> wrote:
Ok, thanks.
On Tue, 5 May 2020 at 20:21, Maxime Villard <m...@m00nbsd.net> wrote:
Le 05/05/2020 à 21:01, Chavdar Ivanov a écrit :
On Mon, 4 May 2020 at 21:21, Chavdar Ivanov <ci4...@gmail.com> wrote:
Hi,
netbsd-GENERIC_KASLR from yesterday fails to pass the prekern stage
with a frightening message in red 'Page fault'. I've never seen this
before, no idea if I can get any further info.
This machine has been otherwise running KASLR kernel for some four
months now without a problem; it is a VirtualBox guest.
Same with just completed clean buid of 9.99.60.
After "[+] Rel relocations applied" I get:
****** Fault occurred ******
page fault
******************************
I switched th vm to the normal kernel, no problem otherwise.
I know, this is because of the Xen section in the binary, which is not
marked as allocatable. The prekern explicitly decides not to map the note
but then proceeds to relocate its RELAs regardless, which causes a fatal
page fault.
Trivial to fix but I'm not sure which policy to choose, may take a few
days.
--
----