threaded programs crash.
With this fix the new weak symbol's st_other is not merged in, i.e. NOPV
is not copied from the libc __libc_thr_init.
* elflink.c (_bfd_elf_merge_symbol): If merging a new weak
symbol that will be skipped, we don't have a new definition.
means that zero is returned, and the kernel keeps mounting (and it
probably ends up in a deadlock/memory corruption somewhere).
2) 'nentries' and 'gnentries' are int and user-controlled, and there's no
check to ensure they are greater than zero. Since they are used to
compute the size of two copyin's, a user can control the copied size
by giving a negative value (like 128-2^29), and thus overwrite kernel
memory.
Both triggerable from root only.
at the top of CHANGES:
* Start with <tab><heading><colon><one space><capital letter>...
* Indent with exactly two tabs.
* End with <full stop><one or two spaces, or newline and two tabs>
<left square bracket><initials><space><date><right square bracket>
I edited only in the "Changes from NetBSD 6.0 to NetBSD 7.0" section,
and I left one entry starting with a lowercase letter because it was a
literal variable name.
possibly, we want to build the i386 fenv.c for compat build, but they
seem largely compatible and for now, just go back to what we had before
that wasn't breaking the build or generating bug reports.
This caused recursion on rndpool_mtx when called from an rndsource's
callback.
XXX The old and new code here is a minefield demanding another
rototill.
XXX rnd_source_anonymous was not actually used before, and still is
not used now, to be cleaned up later.
the moment, this can only load very simple modules due to missing
symbols. It is being add at this time to make pullups to the
netbsd-7 branch easier. It is not enabled by default in any kernels.
they are included on all boot media.
Some of the previous method of setting the console (and conspeed, when
using serial console) is retained, but it should probably be reworked.
Also, adapt to the fact that we may be installing to a disk wedge.
With this change, it is now possible (but still ugly) for sysinst to
create a bootable GPT-labeled disk.
matt