Re: [linus:master] BUILD REGRESSION a2e5790d841658485d642196dbb0927303d6c22f

[Borislav Petkov]

On Wed, Feb 07, 2018 at 11:43:37AM -0800, Linus Torvalds wrote:
> We could just also decide that the only thing that the modrm bytes of
> UD0 actually *affect* is how the CPU might act for a page-crossing
> instruction.
> 
> Because I think that's the only semantic difference: if it's a
> page-crosser, the instruction could take a page fault before raising
> the #UD.
> 
> Is there any other decode issue we might want to look out for?

Yes, AMD doesn't do UD0 with a ModRM:

"UD0 0F FF Raise an invalid opcode exception"

And I don't know about the other x86 vendors; what fun they've had with
the UD*.

I guess we should make sure the following bytes are a valid insn.

Btw Rudolf had experimented with this, CCed, and he showed that on Intel
you get a SIGSEGV for a page crosser with UD1 and SIGILL on AMD.

I guess you can get the same with UD0 apparently.

> Anyway, then we'd make the rule be:
> 
>  - we promise to always pad up the following bytes (our extra warning
> information etc) so that we never have a missing page afterwards (this
> is presumably practically speaking already the case). It might still
> be a page-crossing instruction, but we won't take a page fault in
> kernel space (due to it being at the end of some text sectoin or
> whatever that changes the NX bit or due to DEBUG_PAGEALLOC having
> unmapped the next page).

Yeah, I guess we can control the bytes there.

>  - we special-case the decoder so that we don't get this warning

That's easy - the decoder should simply say to upgrade objdump.

> and then we just ignore the issue entirely.
> 
> Hmm?

Sounds like a plan but with those things the devil's in the detail.
Nothing sounds too nasty now, though.

-- 
Regards/Gruss,
    Boris.

Good mailing practices for 400: avoid top-posting and trim the reply.