linux-kernel - Re: objtool warning for next-20221118

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20221124163928.dof5ldfer6rswlnx@treble>
Date:   Thu, 24 Nov 2022 08:39:28 -0800
From:   Josh Poimboeuf <jpoimboe@...nel.org>
To:     Juergen Gross <jgross@...e.com>
Cc:     Andrew Cooper <Andrew.Cooper3@...rix.com>,
        Peter Zijlstra <peterz@...radead.org>,
        "Paul E. McKenney" <paulmck@...nel.org>,
        "sfr@...b.auug.org.au" <sfr@...b.auug.org.au>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "sstabellini@...nel.org" <sstabellini@...nel.org>,
        "boris.ostrovsky@...cle.com" <boris.ostrovsky@...cle.com>,
        "xen-devel@...ts.xenproject.org" <xen-devel@...ts.xenproject.org>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: objtool warning for next-20221118

On Thu, Nov 24, 2022 at 08:47:47AM +0100, Juergen Gross wrote:
> > > +++ b/arch/x86/xen/smp_pv.c
> > > @@ -385,17 +385,9 @@ static void xen_pv_play_dead(void) /* used only
> > > with HOTPLUG_CPU */
> > >   {
> > >       play_dead_common();
> > >       HYPERVISOR_vcpu_op(VCPUOP_down, xen_vcpu_nr(smp_processor_id()), NULL);
> > > -    cpu_bringup();
> > > -    /*
> > > -     * commit 4b0c0f294 (tick: Cleanup NOHZ per cpu data on cpu down)
> > > -     * clears certain data that the cpu_idle loop (which called us
> > > -     * and that we return from) expects. The only way to get that
> > > -     * data back is to call:
> > > -     */
> > > -    tick_nohz_idle_enter();
> > > -    tick_nohz_idle_stop_tick_protected();
> > > -    cpuhp_online_idle(CPUHP_AP_ONLINE_IDLE);
> > > +    /* FIXME: converge cpu_bringup_and_idle() and start_secondary() */
> > > +    cpu_bringup_and_idle();
> > 
> > I think this will leak stack memory. Multiple cpu offline/online cycles of
> > the same cpu will finally exhaust the idle stack.

Doh!  Of course...

I was actually thinking ahead, to where eventually xen_pv_play_dead()
can call start_cpu0(), which can be changed to automatically reset the
stack pointer like this:

SYM_CODE_START(start_cpu0)
	ANNOTATE_NOENDBR
	UNWIND_HINT_EMPTY
	movq	PER_CPU_VAR(pcpu_hot + X86_top_of_stack), %rax
	leaq	-PTREGS_SIZE(%rax), %rsp
	jmp	.Ljump_to_C_code
SYM_CODE_END(start_cpu0)

but that would only be possible be after more cleanups which converge
cpu_bringup_and_idle() with start_secondary().

> The attached patch seems to work fine.

The patch looks good to me.

It doesn't solve Paul's original issue where arch_cpu_idle_dead() needs
to be __noreturn.  But that should probably be a separate patch anyway.

> The __noreturn annotation seems to trigger an objtool warning, though, in
> spite of the added BUG() at the end of xen_pv_play_dead():
>
> arch/x86/xen/smp_pv.o: warning: objtool: xen_pv_play_dead() falls through to
> next function xen_pv_cpu_die()

You'll need to tell objtool that xen_cpu_bringup_again() is noreturn by
adding "xen_cpu_bringup_again" to global_noreturns[] in
tools/objtool/check.c.

(Yes it's a pain, I'll be working an improved solution to the noreturn
thing...)

-- 
Josh