[PATCH] init/main.c: Initialize early LSMs after arch code

[Nathan Chancellor]

On Tue, Aug 06, 2024 at 01:29:37AM +0200, KP Singh wrote:
> On Mon, Aug 5, 2024 at 9:58 PM Paul Moore <paul at paul-moore.com> wrote:
> >
> > On Thu, Aug 1, 2024 at 1:17 PM KP Singh <kpsingh at kernel.org> wrote:
> > >
> > > With LSMs using static calls, early_lsm_init needs to wait for setup_arch
> > > for architecture specific functionality which includes jump tables and
> > > static calls to be initialized.
> > >
> > > This only affects "early LSMs" i.e. only lockdown when
> > > CONFIG_SECURITY_LOCKDOWN_LSM_EARLY is set.
> > >
> > > Fixes: 2732ad5ecd5b ("lsm: replace indirect LSM hook calls with static calls")
> > > Signed-off-by: KP Singh <kpsingh at kernel.org>
> > > ---
> > >  init/main.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > Considering the problems we've had, I'd like to hear more about how
...
> I guess it would not harm Boris, Nathan and others to look at it as
> well and see if it breaks any of their tests.

For what it's worth, I have not noticed any issues in my -next testing
with this patch applied but I only build architectures that build with
LLVM due to the nature of my work. If exposure to more architectures is
desirable, perhaps Guenter Roeck would not mind testing it with his
matrix?

Cheers,
Nathan

> > > diff --git a/init/main.c b/init/main.c
> > > index 206acdde51f5..a0e3f3c720e6 100644
> > > --- a/init/main.c
> > > +++ b/init/main.c
> > > @@ -922,8 +922,8 @@ void start_kernel(void)
> > >         boot_cpu_init();
> > >         page_address_init();
> > >         pr_notice("%s", linux_banner);
> > > -       early_security_init();
> > >         setup_arch(&command_line);
> > > +       early_security_init();
> > >         setup_boot_config();
> > >         setup_command_line(command_line);
> > >         setup_nr_cpu_ids();
> > > --
> > > 2.46.0.rc2.264.g509ed76dc8-goog
> >
> > --
> > paul-moore.com