Re: [RFC PATCH] x86/mce/inject: Add sanity check in inject_mce()

[Ingo Molnar <mingo@kernel.org>]

* Borislav Petkov <bp@alien8.de> wrote:

> On Sat, May 31, 2025 at 10:14:41AM +0200, Ingo Molnar wrote:
> > Uhm, avoiding a hard kernel crash in case of a tooling or human error:
> 
> Do you know what this tool is about?
> 
> It is supposed to cause crashes, among others.

What tool? If you mean syzkaller, then it is very much not supposed to 
crash on a correctly functioning kernel. If it causes a crash, then the 
proper response is to fix the crash.

Or if you mean the MCE-injection interface, that's not supposed to 
trigger avoidable crashes either, it's an injection facility for 
testing purposes:

          Provide support for injecting machine checks for testing purposes.

It's really simple really:

- If the kernel unnecessarily locks up on the receipt of a 
  hardware-generated MCE then that's a kernel bug that
  should be fixed.

- If the kernel unnecessarily locks up on the receipt of a 
  software-generated MCE then that's a kernel bug that
  should be fixed.

TL;DR, this is not an acceptable kernel response:

> > [  306.335489][ T3298] mce: CPUs not responding to MCE broadcast (may include false positives): 1-3
> > [  306.336332][ T3298] Kernel panic - not syncing: Timeout: Not all CPUs entered broadcast exception handler
> > [  306.337786][ T3298] Kernel Offset: 0x17400000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)

Thanks,

	Ingo