Re: [RFC PATCH 02/16] x86/split_lock: Handle #AC exception for split lock in kernel mode

[Thomas Gleixner]

On Fri, 22 Jun 2018, Fenghua Yu wrote:
> On Fri, Jun 22, 2018 at 12:49:00PM +0200, Thomas Gleixner wrote:
> > On Sun, 27 May 2018, Fenghua Yu wrote:
> > > +static void wait_for_reexecution(void)
> > > +{
> > > +	while (time_before(jiffies, disable_split_lock_jiffies +
> > > +			   reenable_split_lock_delay))
> > > +		cpu_relax();
> > > +}
> > > +
> > > +/*
> > > + * TEST_CTL MSR is shared among threads on the same core. To simplify
> > > + * situation, disable_split_lock_jiffies is global instead of per core.
> > 
> > This patch surely earns extra points in the trainwreck engineering contest,
> > but that's not taking place on LKML.
> > 
> > The whole thing is simply:
> > 
> > handle_ac()
> > {
> > 	if (user_mode(regs)) {
> > 		 do_trap(AC, SIGBUS, ...);
> > 	} else {
> > 		disable_ac_on_local_cpu();
> > 		WARN_ONCE(1);
> > 	}
> > }
> 
> Should I add kernel parameter or control knob to opt-out the feature?

A simple command line option 'acoff' or something more sensible should be
ok. No sysfs knobs or whatever please. The Kconfig option is not required
either.

> I'm afraid firmware may hang system after handling split lock if the
> feature is enabled by kernel, e.g. "reboot" hits split lock in firmware
> and firmware hangs the system after handling #AC.

Have you observed the problem in reality? I mean why would 'reboot' be the
critical path? I'd rather expect that EFI callbacks or SMM 'value add'
would trip over it.

Vs. reboot. If that is the only problem then we might just have to clear
#AC enable before issuing it, but that does not need to be part of the
initial patch set. Its an orthogonal issue.

Thanks,

	tglx