Re: [RFC PATCH glibc 4/8] glibc: Perform rseq(2) registration at C startup and thread creation (v15)

[Mathieu Desnoyers]

----- On Mar 19, 2020, at 2:16 PM, Florian Weimer fw@xxxxxxxxxxxxx wrote:

> * Mathieu Desnoyers:
> 
>>> You also need to add an assert that the compiler supports
>>> __attribute__ ((aligned)) because ignoring it produces an
>>> ABI-incompatible header.
>>
>> Are you aware of some helper macro I should use to do this, or
>> is it done elsewhere in glibc ?
> 
> I don't think we have any such GCC-only types yet.  max_align_t is
> provided by GCC itself.

I was thinking of adding the following to

sysdeps/unix/sysv/linux/rseq-internal.h: rseq_register_current_thread()

+  /* Ensure the compiler supports __attribute__ ((aligned)).  */
+  _Static_assert (__alignof__ (struct rseq_cs) >= 4 * sizeof(uint64_t),
+                 "alignment");
+  _Static_assert (__alignof__ (struct rseq) >= 4 * sizeof(uint64_t),
+                 "alignment");
+

>>> The struct rseq/struct rseq_cs definitions
>>> are broken, they should not try to change the alignment.
>>
>> AFAIU, this means we should ideally not have used __attribute__((aligned))
>> in the uapi headers in the first place. Why is it broken ?
> 
> Compilers which are not sufficiently GCC-compatible define
> __attribute__(X) as the empty expansion, so you silently get a
> different ABI.

It is worth noting that rseq.h is not the only Linux uapi header
which uses __attribute__ ((aligned)), so this ABI problem exists today
anyway for those compilers.

> 
> There is really no need to specify 32-byte alignment here.  Is not
> even the size of a standard cache line.  It can result in crashes if
> these structs are heap-allocated using malloc, when optimizing for
> AVX2.

Why would it be valid to allocate those with malloc ? Isn't it the
purpose of posix_memalign() ?

> 
> For example, clang turns
> 
> void
> clear (struct rseq *p)
> {
>  memset (p, 0, sizeof (*p));
> }
> 
> into:
> 
>	vxorps	%xmm0, %xmm0, %xmm0
>	vmovaps	%ymm0, (%rdi)
>	vzeroupper
>	retq
> 
> My understanding is that vmovaps will trap if the pointer is
> misaligned (âWhen the source or destination operand is a memory
> operand, the operand must be aligned on a 32-byte boundary or a
> general-protection exception (#GP) will be generated.â).
> 
>> However, now that it is in the wild, it's a bit late to change that.
> 
> I had forgotten about the alignment crashes.  I think we should
> seriously consider changing the types. 8-(

I don't think this is an option at this stage given that it is part
of the Linux kernel UAPI. I am not convinced that it is valid at all
to allocate struct rseq or struct rseq_cs with malloc(), because it
does not guarantee any alignment.

Thanks,

Mathieu

-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com