On Thu, Mar 31, 2011 at 11:07:36AM +0400, Pan Tsu wrote:
> Konstantin Belousov <kib@FreeBSD.org> writes:
> > Author: kib
> > Date: Wed Mar 30 14:46:12 2011
> > New Revision: 220158
> > URL: http://svn.freebsd.org/changeset/base/220158
> > Log:
> > Provide compat32 shims for kldstat(2).
> > Requested and tested by: jpaetzel
> > MFC after: 1 week
> It breaks loading modules from loader(8) for me. The box panics early
> #13 0xffffffff804f2f2a in sysctl_find_oidname (name=0xffffffff80eec156 "kstat", list=0x657a69735f796c)
> at /usr/src/sys/kern/kern_sysctl.c:106
> #14 0xffffffff804f3e69 in sysctl_register_oid (oidp=0xffffffff80ef0a80)
> at /usr/src/sys/kern/kern_sysctl.c:149
> #15 0xffffffff804cb83c in linker_file_register_sysctls (lf=Variable "lf" is not available.
> ) at /usr/src/sys/kern/kern_linker.c:301
> #16 0xffffffff804cd112 in linker_preload (arg=Variable "arg" is not available.
> ) at /usr/src/sys/kern/kern_linker.c:1624
> #17 0xffffffff80499ad7 in mi_startup () at /usr/src/sys/kern/init_main.c:258
> #18 0xffffffff80299a3c in btext () at /usr/src/sys/amd64/amd64/locore.S:81
What supports your claim that the referenced commit breaks early
module loading ? You even did not provided the panic message, btw.
Commit only changed the wrapper around kldstat(2) syscall, that cannot
be used during bootstrap, and the changed code does not reference kldstat(9).
The later cannot prove that the change did not break something, but
due to its nature, it is unlikely.