linux-kernel - Re: [PATCH] atomic: Fix bugs in 'fetch_or()' and rename it to 'xchg

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: <20160315120147.GA9742@gmail.com>
Date:	Tue, 15 Mar 2016 13:01:47 +0100
From:	Ingo Molnar <mingo@...nel.org>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Frédéric Weisbecker <fweisbec@...il.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH] atomic: Fix bugs in 'fetch_or()' and rename it to
 'xchg_or()'


* Ingo Molnar <mingo@...nel.org> wrote:

> Subject: [PATCH] atomic: Fix bugs in 'fetch_or()' and rename it to 'xchg_or()'
> 
> Linus noticed a couple of problems with the fetch_or() implementation introduced 
> by 5fd7a09cfb8c ("atomic: Export fetch_or()"):
> 
>  - Sloppy macro implementation: 'mask' and 'ptr' is evaluated multiple times, 
>    which will break if arguments have side effects.

So this shiny new patch manages to crash the x86 kernel with a NULL pointer 
dereference:

  [    0.143027] BUG: unable to handle kernel NULL pointer dereference at           (null)
  [    0.144000] IP: [<ffffffff8107c64c>] resched_curr+0x3c/0xc0

GCC manages to turn this:

static bool set_nr_and_not_polling(struct task_struct *p)
{
        struct thread_info *ti = task_thread_info(p);
        return !(xchg_or(&ti->flags, _TIF_NEED_RESCHED) & _TIF_POLLING_NRFLAG);
}

and this:

>  /**
> + * xchg_or - perform *ptr |= mask atomically and return old value of *ptr
> + * @ptr: pointer to value (cmpxchg() compatible integer pointer type)
>   * @mask: mask to OR on the value
>   *
> + * cmpxchg() based, it's a macro so it works for different integer types.
>   */
> +#ifndef xchg_or
> +# define xchg_or(ptr, mask)						\
> +({									\
> +	typeof(ptr)  __ptr  = (ptr);					\
> +	typeof(mask) __mask = (mask);					\
> +									\
> +	typeof(*(__ptr)) __old, __val = *__ptr;				\
> +									\
>  	for (;;) {							\
> +		__old = cmpxchg(__ptr, __val, __val | __mask);		\
>  		if (__old == __val)					\
>  			break;						\
>  		__val = __old;						\
>  	}								\
> +									\
>  	__old;								\
>  })

into:

    41c1:       89 c2                   mov    %eax,%edx
    41c3:       89 d6                   mov    %edx,%esi
    41c5:       31 c9                   xor    %ecx,%ecx
    41c7:       89 d0                   mov    %edx,%eax
    41c9:       83 ce 08                or     $0x8,%esi
    41cc:       f0 0f b1 31             lock cmpxchg %esi,(%rcx)

note the RCX zeroing via XOR...

The original, working sequence is:

    41c4:       89 c2                   mov    %eax,%edx
    41c6:       89 d6                   mov    %edx,%esi
    41c8:       89 d0                   mov    %edx,%eax
    41ca:       83 ce 08                or     $0x8,%esi
    41cd:       f0 0f b1 31             lock cmpxchg %esi,(%rcx)

The change that makes the difference is the 'ptr' part of:

> +		__old = cmpxchg(__ptr, __val, __val | __mask);		\

This variant works:

> +		__old = cmpxchg((ptr), __val, __val | __mask);		\

After a lot of staring PeterZ realized that __ptr aliases with the x86 cmpxchg() 
macro-jungle's __ptr name!!

So if I do a s/__ptr/_ptr it all works...

But IMHO this really highlights a fundamental weakness of all this macro magic, 
it's all way too fragile.

Why don't we introduce a boring family of APIs:

	cmpxchg_8()
	cmpxchg_16()
	cmpxchg_32()
	cmpxchg_64()

	xchg_or_32()
	xchg_or_64()
	...

... with none of this pesky auto-typing property and none of the 
macro-inside-a-macro crap? We could do clean types and would write them all in 
proper C, not fragile CPP.

It's not like we migrate between the types all that frequently - and even if we 
do, it's trivial.

hm?

Thanks,

	Ingo