linux-kernel - Re: [patch V4 01/32] mm/slab: Fix broken stack trace storage

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: <20190415132339.wiqyzygqklliyml7@treble>
Date:   Mon, 15 Apr 2019 08:23:39 -0500
From:   Josh Poimboeuf <jpoimboe@...hat.com>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     Andy Lutomirski <luto@...nel.org>,
        LKML <linux-kernel@...r.kernel.org>, X86 ML <x86@...nel.org>,
        Sean Christopherson <sean.j.christopherson@...el.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Pekka Enberg <penberg@...nel.org>,
        Linux-MM <linux-mm@...ck.org>
Subject: Re: [patch V4 01/32] mm/slab: Fix broken stack trace storage

On Mon, Apr 15, 2019 at 11:02:58AM +0200, Thomas Gleixner wrote:
> kstack_end() is broken on interrupt stacks as they are not guaranteed to be
> sized THREAD_SIZE and THREAD_SIZE aligned.
> 
> Use the stack tracer instead. Remove the pointless pointer increment at the
> end of the function while at it.
> 
> Fixes: 98eb235b7feb ("[PATCH] page unmapping debug") - History tree
> Signed-off-by: Thomas Gleixner <tglx@...utronix.de>
> Cc: Andrew Morton <akpm@...ux-foundation.org>
> Cc: Pekka Enberg <penberg@...nel.org>
> Cc: linux-mm@...ck.org
> ---
> V4: Made the code simpler to understand (Andy) and make it actually compile
> ---
>  mm/slab.c |   30 ++++++++++++++----------------
>  1 file changed, 14 insertions(+), 16 deletions(-)
> 
> --- a/mm/slab.c
> +++ b/mm/slab.c
> @@ -1470,33 +1470,31 @@ static bool is_debug_pagealloc_cache(str
>  static void store_stackinfo(struct kmem_cache *cachep, unsigned long *addr,
>  			    unsigned long caller)
>  {
> -	int size = cachep->object_size;
> +	int size = cachep->object_size / sizeof(unsigned long);
>  
>  	addr = (unsigned long *)&((char *)addr)[obj_offset(cachep)];
>  
> -	if (size < 5 * sizeof(unsigned long))
> +	if (size < 5)
>  		return;
>  
>  	*addr++ = 0x12345678;
>  	*addr++ = caller;
>  	*addr++ = smp_processor_id();
> -	size -= 3 * sizeof(unsigned long);
> +	size -= 3;
> +#ifdef CONFIG_STACKTRACE
>  	{
> -		unsigned long *sptr = &caller;
> -		unsigned long svalue;
> -
> -		while (!kstack_end(sptr)) {
> -			svalue = *sptr++;
> -			if (kernel_text_address(svalue)) {
> -				*addr++ = svalue;
> -				size -= sizeof(unsigned long);
> -				if (size <= sizeof(unsigned long))
> -					break;
> -			}
> -		}
> +		struct stack_trace trace = {
> +			/* Leave one for the end marker below */
> +			.max_entries	= size - 1,
> +			.entries	= addr,
> +			.skip		= 3,
> +		};
>  
> +		save_stack_trace(&trace);
> +		addr += trace.nr_entries;
>  	}
> -	*addr++ = 0x87654321;
> +#endif
> +	*addr = 0x87654321;

Looks like stack_trace.nr_entries isn't initialized?  (though this code
gets eventually replaced by a later patch)

Who actually reads this stack trace?  I couldn't find a consumer.

-- 
Josh