Re: NMI Reported with console_blast.sh

["John B. Wyatt IV" <jwyatt@redhat.com>]

On Thu, Feb 29, 2024 at 12:15:28PM +0100, Petr Mladek wrote:
> On Thu 2024-02-22 00:21:19, John B. Wyatt IV wrote:
> > On Mon, Feb 12, 2024 at 04:23:04PM -0500, John B. Wyatt IV wrote:
> > 
> > I found 2 additional NMIs for a total of 3. Number 2 is very
> > large-please feel free to let me know what specific information you
> > wanted if it was unnecessary.
> > 
> > This first one (the original) is with the real-time profile (no preemption).
> > The next two are with the performance-througput (no preemption).
> > 
> > This is what real-time NMI reports with the caller information enabled. It
> > looks like it is lacking some information compared to the other two further below.
> > 
> > [ T2481] Call Trace:
> > [ T2477] Kernel panic - not syncing: sysrq triggered crash
> > [    C0] NMI backtrace for cpu 0
> 
> This message seems to be printed by nmi_cpu_backtrace().
> 
> I am surprised. I would expect to see the backtrace printed from panic().
> It calls dump_stack() directly on the panic-CPU. And this panic() should
> be called from sysrq_handle_crash(). IMHO, it should be (normal)
> interrupt context.
> 
> Is it different on RT?

Yes. There are no NMIs on this machine in my tests for when I run an RT
enabled kernel.

> 
> Or have you somehow modified the kernel to use NMI for sysrq?
> 

I have not modified the kernel source code from 6.7.0-rt6.

-- 
Sincerly,
John Wyatt
Software Engineer, Core Kernel