Re: [PATCH v10 07/14] unwind_user/deferred: Make unwind deferral requests NMI-safe

[Peter Zijlstra]

On Thu, Jun 19, 2025 at 11:32:26AM +0200, Peter Zijlstra wrote:
> On Thu, Jun 19, 2025 at 05:07:10AM -0400, Steven Rostedt wrote:
> 
> > Does #DB make in_nmi() true? If that's the case then we do need to handle that.
> 
> Yes: #DF, #MC, #BP (int3), #DB and NMI all have in_nmi() true.

Note: these are all the from-kernel parts of those exceptions. The
from-user side is significantly different.

> Ignoring #DF because that's mostly game over, you can get them all
> nested for up to 4 (you're well aware of the normal NMI recursion
> crap).
> 
> Then there is the SEV #VC stuff, which is also NMI like. So if you're a
> CoCo-nut, you can perhaps get it up to 5.