lkml.org 
[lkml]   [2021]   [Nov]   [30]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH v3 2/2] x86/tsc: skip tsc watchdog checking for qualified platforms
On Tue, Nov 30, 2021 at 10:55:45PM +0100, Thomas Gleixner wrote:
> On Tue, Nov 30 2021 at 12:47, Paul E. McKenney wrote:
> > On Tue, Nov 30, 2021 at 09:39:32PM +0100, Thomas Gleixner wrote:
> >> Seriously. Jiffies is not usable as watchdog simply because lost ticks
> >> cannot be compensated and you cannot use TSC to bridge them because you
> >> are not trusting TSC. This is simply a circulus vitiosus.
> >
> > OK, HPET or nothing, then.
>
> Older machines also have pm_timer. But those beasts seem to have lost
> that too.

I suppose that one way of avoiding clock-skew messages is to have only
one clock.

> >> We really need to remove the watchdog requirement for modern hardware.
> >> Let me stare at those patches and get them merged.
> >
> > You are more trusting of modern hardware than I am, but for all I know,
> > maybe rightfully so. ;-)
>
> Well, I rather put a bet on the hardware, which has become reasonable
> over the last decade, than on trying to solve a circular dependency
> problem with tons of heuristics which won't ever work correctly.

Use of HPET to check the interval length would not be circular, right?

> TSC_ADJUST is a reasonable safety net and since its invention the amount
> of BIOS wreckage has been massively reduced. Seems the nastigram in
> dmesg when detecting a change in TSC_ADJUST had an effect or maybe
> Microsoft enforces a tinkerfree TSC by now and we get the benefit. :)
>
> I still wish to have a knob to lock down TSC to read only, but that's
> probably for christmas 2030 or later. :)

Indeed. How would BIOS writers hide their SMI handlers? :-/

Thanx, Paul

\
 
 \ /
  Last update: 2021-11-30 23:48    [W:0.056 / U:0.336 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site