Re: [RFC PATCH 2/8] jump label v4 - x86: Intr oduce generic jump patching without stop_machine
From: H. Peter Anvin
Date: Sun Jan 17 2010 - 18:02:53 EST
If single-byte updates weren't atomic, then the int3 scheme would not be possible in the first place. Of course, if you want there to be a synchronization point beyond which the modification is guaranteed to have affected all CPUs, you need an IPI-IRET on all CPUs.
The other thing to watch out for is that the CPU itself is subject to text modification through a different alias, which means some of the hardware SMC protections are ineffective.
"Mathieu Desnoyers" <mathieu.desnoyers@xxxxxxxxxx> wrote:
>* H. Peter Anvin (hpa@xxxxxxxxx) wrote:
>> On 01/14/2010 07:32 AM, Steven Rostedt wrote:
>> >> +
>> >> + /* Replacing 1 byte can be done atomically. */
>> >> + if (unlikely(len <= 1))
>> >> + return text_poke(addr, opcode, len);
>> > This part bothers me. The text_poke just writes over the text directly
>> > (using a separate mapping). But if that memory is in the pipeline of
>> > another CPU, I think this could cause a GPF.
>> Could you clarify why you think that?
>Basically, what Steven and I were concerned about in this particular
>patch version is the fact that this code took a "shortcut" for
>single-byte text modification, thus bypassing the int3-bypass scheme
>As mere atomicity of the modification is not the only concern here
>(because we also have to deal with instruction trace cache coherency and
>so forth), then the int3 breakpoint scheme is, I think, also needed for
>> H. Peter Anvin, Intel Open Source Technology Center
>> I work for Intel. I don't speak on their behalf.
>OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F BA06 3F25 A8FE 3BAE 9A68
Sent from my mobile phone, pardon any lack of formatting.