linux-kernel - Re: [PATCH 1/3] x86/debug: Fix BTF handling

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: <20201028182025.4bb6d633719d7ce76300aafa@kernel.org>
Date:   Wed, 28 Oct 2020 18:20:25 +0900
From:   Masami Hiramatsu <mhiramat@...nel.org>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     tglx@...utronix.de, luto@...nel.org, me@...ehuey.com,
        x86@...nel.org, linux-kernel@...r.kernel.org,
        torvalds@...ux-foundation.org, rocallahan@...il.com,
        alexandre.chartre@...cle.com, paulmck@...nel.org,
        frederic@...nel.org, pbonzini@...hat.com,
        sean.j.christopherson@...el.com, mhiramat@...nel.org,
        pmladek@...e.com, joel@...lfernandes.org, rostedt@...dmis.org,
        boris.ostrovsky@...cle.com, jgross@...e.com, brgerst@...il.com,
        jpoimboe@...hat.com, daniel.thompson@...aro.org,
        julliard@...ehq.org, pgofman@...eweavers.com
Subject: Re: [PATCH 1/3] x86/debug: Fix BTF handling

On Tue, 27 Oct 2020 20:41:26 +0100
Peter Zijlstra <peterz@...radead.org> wrote:

> On Tue, Oct 27, 2020 at 10:15:05AM +0100, Peter Zijlstra wrote:
> 
> > @@ -873,6 +866,20 @@ static __always_inline void exc_debug_ke
> >  	 */
> >  	WARN_ON_ONCE(user_mode(regs));
> >  
> > +	if (test_thread_flag(TIF_BLOCKSTEP)) {
> > +		/*
> > +		 * The SDM says "The processor clears the BTF flag when it
> > +		 * generates a debug exception." but PTRACE_BLOCKSTEP requested
> > +		 * it for userspace, but we just took a kernel #DB, so re-set
> > +		 * BTF.
> > +		 */
> > +		unsigned long debugctl;
> > +
> > +		rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctl);
> > +		debugctl |= DEBUGCTLMSR_BTF;
> > +		wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctl);
> > +	}
> > +
> >  	/*
> >  	 * Catch SYSENTER with TF set and clear DR_STEP. If this hit a
> >  	 * watchpoint at the same time then that will still be handled.
> 
> Masami, how does BTF interact with !optimized kprobes that single-step?

Good question, BTF is cleared right before single-stepping and restored
after single-stepping. It will be done accoding to TIF_BLOCKSTEP bit as below.

(in arch/x86/kernel/kprobes/core.c)

static nokprobe_inline void clear_btf(void)
{
        if (test_thread_flag(TIF_BLOCKSTEP)) {
                unsigned long debugctl = get_debugctlmsr();

                debugctl &= ~DEBUGCTLMSR_BTF;
                update_debugctlmsr(debugctl);
        }
}

static nokprobe_inline void restore_btf(void)
{
        if (test_thread_flag(TIF_BLOCKSTEP)) {
                unsigned long debugctl = get_debugctlmsr();

                debugctl |= DEBUGCTLMSR_BTF;
                update_debugctlmsr(debugctl);
        }
}

Hrm, so it seems that we do same ... maybe we don't need clear_btf() too?

> 
> The best answer I can come up with is 'poorly' :/

Is this what you expected? :)

Thank you,


-- 
Masami Hiramatsu <mhiramat@...nel.org>