Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 3 Mar 2022 11:09:34 -0800
From:   Kees Cook <keescook@chromium.org>
To:     Dan Li <ashimida@linux.alibaba.com>
Cc:     akpm@linux-foundation.org, arnd@arndb.de, catalin.marinas@arm.com,
        gregkh@linuxfoundation.org, linux@roeck-us.net,
        luc.vanoostenryck@gmail.com, elver@google.com,
        mark.rutland@arm.com, masahiroy@kernel.org, ojeda@kernel.org,
        nathan@kernel.org, npiggin@gmail.com, ndesaulniers@google.com,
        samitolvanen@google.com, shuah@kernel.org, tglx@linutronix.de,
        will@kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
        llvm@lists.linux.dev, linux-hardening@vger.kernel.org
Subject: Re: [PATCH v3 2/2] lkdtm: Add Shadow Call Stack tests
Message-ID: <202203031105.A1B4CAE6@keescook>
References: <20220303073340.86008-1-ashimida@linux.alibaba.com>
 <20220303074339.86337-1-ashimida@linux.alibaba.com>
 <202203031010.0A492D114@keescook>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <202203031010.0A492D114@keescook>
Precedence: bulk

On Thu, Mar 03, 2022 at 10:42:45AM -0800, Kees Cook wrote:
> Though, having the IS_ENABLED in there makes me wonder if this test
> should instead be made _survivable_ on failure. Something like this,
> completely untested:
> 
> 
> #ifdef CONFIG_ARM64
> static noinline void lkdtm_scs_set_lr(unsigned long *addr)
> {
> 	unsigned long **lr = (unsigned long **)__builtin_frame_address(0) + 1;
> 	*lr = addr;
> }
> 
> /* Function with __noscs attribute clears its return address. */
> static noinline void __noscs lkdtm_noscs_set_lr(unsigned long *addr)
> {
> 	unsigned long **lr = (unsigned long **)__builtin_frame_address(0) + 1;
> 	*lr = addr;
> }
> #endif
> 
> 
> void lkdtm_CFI_BACKWARD_SHADOW(void)
> {
> #ifdef CONFIG_ARM64
> 
> 	/* Verify the "normal" condition of LR corruption working. */
> 	do {
> 		/* Keep label in scope to avoid compiler warning. */
> 		if ((volatile int)0)
> 			goto unexpected;
> 
> 		pr_info("Trying to corrupt lr in a function without scs protection ...\n");
> 		lkdtm_noscs_set_lr(&&expected);
> 
> unexpected:
> 		pr_err("XPASS: Unexpectedly survived lr corruption without scs?!\n");
> 		break;
> 
> expected:
> 		pr_err("ok: lr corruption redirected without scs.\n");
> 	} while (0);
> 
> 
> 	do {
> 		/* Keep labe in scope to avoid compiler warning. */
> 		if ((volatile int)0)
> 			goto good_scs;
> 
> 		pr_info("Trying to corrupt lr in a function with scs protection ...\n");
> 		lkdtm_scs_set_lr(&&bad_scs);
> 
> good_scs:
> 		pr_info("ok: scs takes effect.\n");
> 		break;
> 
> bad_scs:
> 		pr_err("FAIL: return address rewritten!\n");
> 		pr_expected_config(CONFIG_SHADOW_CALL_STACK);
> 	} while (0);
> #else
> 	pr_err("XFAIL: this test is arm64-only\n");
> #endif
> }
> 
> And we should, actually, be able to make the "set_lr" functions be
> arch-specific, leaving the test itself arch-agnostic....

Yeah, as a tested example, this works for x86_64, and based on what you
had, I'd expect it to work on arm64 too:

#include <stdio.h>

static __attribute__((noinline))
void set_return_addr(unsigned long *expected, unsigned long *addr)
{
    /* Use of volatile is to make sure final write isn't seen as a dead store. */
    unsigned long * volatile *ret_addr = (unsigned long **)__builtin_frame_address(0) + 1;

    /* Make sure we've found the right place on the stack before writing it. */
    if (*ret_addr == expected)
        *ret_addr = addr;
}

volatile int force_label;
int main(void)
{
    do {
        /* Keep labels in scope. */
        if (force_label)
            goto normal;
        if (force_label)
            goto redirected;

        set_return_addr(&&normal, &&redirected);
normal:
        printf("I should be skipped\n");
        break;
redirected:
        printf("Redirected\n");
    } while (0);

    return 0;
}


It does _not_ work under Clang, though, which I'm still looking at.

-- 
Kees Cook