Re: Self-XORing BPF registers is undefined behavior

On 12/13/2018 01:24 PM, Alexander Potapenko wrote:

On Thu, Dec 13, 2018 at 1:20 PM Michal Kubecek [off-list ref] wrote:

quoted

On Thu, Dec 13, 2018 at 12:59:36PM +0100, Michal Kubecek wrote:

quoted

On Thu, Dec 13, 2018 at 12:00:59PM +0100, Alexander Potapenko wrote:

quoted

Hi BPF maintainers,

some time ago KMSAN found an issue in BPF code which we decided to
suppress at that point, but now I'd like to bring it to your
attention.
Namely, some BPF programs may contain instructions that XOR a register
with itself.
This effectively results in the following C code:
  regs[BPF_REG_A] = regs[BPF_REG_A] ^ regs[BPF_REG_A];
or
  regs[BPF_REG_X] = regs[BPF_REG_X] ^ regs[BPF_REG_X];
being executed.

According to the C11 standard this is undefined behavior, so KMSAN
reports an error in this case.

Can you elaborate / quote the exact bit from C11 that KMSAN is referring
to? (The below that Michal was quoting or something else?)

Does that only refer to C11 standard? Note that kernel's Makefile +430
explicitly states 'std=gnu89' and not 'std=c11' [0].

  [0] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=51b97e354ba9fce1890cf38ecc754aa49677fc89

quoted

quoted

Can you quote the part of the standard saying this is undefined
behavior? I couldn't find anything else than

  If the value being stored in an object is read from another object
  that overlaps in any way the storage of the first object, then the
  overlap shall be exact and the two objects shall have qualified or
  unqualified versions of a compatible type; otherwise, the behavior
  is undefined.

(but I only have a draft for obvious reasons). I'm not sure what exactly
they mean by "exact overlap" and the standard doesn't seem to define
the term but if the two objects are actually the same, they certainly
have compatible types.

Here is an example for the overlap quoted above; I don't think this
applies to our case since it would be "exact". Quote [1]:

  struct S { int x; int y; };
  struct T { int z; struct S s; };
  union U { struct S f ; struct T g; } u;

  main(){
    u.f = u.g.s;
    return 0;
  }

  [1] https://bts.frama-c.com/print_bug_page.php?bug_id=945

quoted

I think I understand now. You didn't want to say that the statement

  regs[BPF_REG_A] = regs[BPF_REG_A] ^ regs[BPF_REG_A];

as such is undefined behavior but that it's UB when regs[BPF_REG_A] is
uninitialized. Right?

Yes. Sorry for being unclear.
By default regs[] is uninitialized, so we need to initialize it before
using the register values.
I am also wondering if it's possible to simply copy the uninitialized
register values from regs[] to the userspace via maps.

quoted

Michal Kubecek