On 08.10.2020 18:28, Roger Pau Monné wrote:
> On Mon, Sep 28, 2020 at 02:31:49PM +0200, Jan Beulich wrote:
>> Under certain conditions CPUs can speculate into the instruction stream
>> past a RET instruction. Guard against this just like 3b7dab93f240
>> ("x86/spec-ctrl: Protect against CALL/JMP straight-line speculation")
>> did - by inserting an "INT $3" insn. It's merely the mechanics of how to
>> achieve this that differ: A set of macros gets introduced to post-
>> process RET insns issued by the compiler (or living in assembly files).
>>
>> Unfortunately for clang this requires further features their built-in
>> assembler doesn't support: We need to be able to override insn mnemonics
>> produced by the compiler (which may be impossible, if internally
>> assembly mnemonics never get generated), and we want to use \(text)
>> escaping / quoting in the auxiliary macro.
>>
>> Signed-off-by: Jan Beulich <jbeul...@suse.com>
>
> Code LGTM.
>
> Acked-by: Roger Pau Monné <roger....@citrix.com>
Thanks.
>> ---
>> TBD: Should this depend on CONFIG_SPECULATIVE_HARDEN_BRANCH?
>
> I don't see the additions done in 3b7dab93f240 being guarded by
> CONFIG_SPECULATIVE_HARDEN_BRANCH, so in that regard I would say no.
> However those are already guarded by CONFIG_INDIRECT_THUNK so it's
> slightly weird that the addition of such protections cannot be turned
> off in any way.
>
> I would be fine with having the additions done in 3b7dab93f240
> protected by CONFIG_SPECULATIVE_HARDEN_BRANCH, and then the additions
> done here also.
Okay, perhaps I'll make a separate patch then to add the conditional
at all respective places.
Jan
