All of lore.kernel.org
 help / color / mirror / Atom feed
* x86 fpu: command-line parameters broken post-FPU-rewrite
@ 2015-12-08 20:43 Dave Hansen
  2015-12-09  8:32 ` Ingo Molnar
  0 siblings, 1 reply; 3+ messages in thread
From: Dave Hansen @ 2015-12-08 20:43 UTC (permalink / raw)
  To: Ingo Molnar, the arch/x86 maintainers, LKML, Yu, Yu-cheng, Yu, Fenghua

Hey Ingo,

We were starting to look at reenabling XSAVES support and tried to use
the 'noxsave' and 'noxsaves' kernel command-line options.  The rewrite
moved the FPU initialization to before we even are parsing command-line
options, even the early_param()s.

Do you have any preferences on how it gets fixed?

The most obvious thing would be to just defer as much of the FPU setup
as we can until after parse_early_param() has happened.  The only other
think I can think of doing would be to try to do some *really* early,
simple, command-line parsing to look for 'noxsave' and friends.

Thoughts?

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: x86 fpu: command-line parameters broken post-FPU-rewrite
  2015-12-08 20:43 x86 fpu: command-line parameters broken post-FPU-rewrite Dave Hansen
@ 2015-12-09  8:32 ` Ingo Molnar
  2015-12-09 19:14   ` Borislav Petkov
  0 siblings, 1 reply; 3+ messages in thread
From: Ingo Molnar @ 2015-12-09  8:32 UTC (permalink / raw)
  To: Dave Hansen
  Cc: the arch/x86 maintainers, LKML, Yu, Yu-cheng, Yu, Fenghua,
	Thomas Gleixner, H. Peter Anvin, Borislav Petkov


* Dave Hansen <dave.hansen@intel.com> wrote:

> Hey Ingo,
> 
> We were starting to look at reenabling XSAVES support and tried to use the 
> 'noxsave' and 'noxsaves' kernel command-line options.  The rewrite moved the FPU 
> initialization to before we even are parsing command-line options, even the 
> early_param()s.
> 
> Do you have any preferences on how it gets fixed?
> 
> The most obvious thing would be to just defer as much of the FPU setup as we can 
> until after parse_early_param() has happened.  The only other think I can think 
> of doing would be to try to do some *really* early, simple, command-line parsing 
> to look for 'noxsave' and friends.
> 
> Thoughts?

Hm, so given that having the FPU operational is key to a functioning kernel (and 
other kernel subsystems may make use of FPU functionality), I'd rather bring 
option parsing earlier than change the FPU setup sequence. FPU setup should be 
done when we identify the CPU - not 'very late' as we used to.

Thanks,

	Ingo

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: x86 fpu: command-line parameters broken post-FPU-rewrite
  2015-12-09  8:32 ` Ingo Molnar
@ 2015-12-09 19:14   ` Borislav Petkov
  0 siblings, 0 replies; 3+ messages in thread
From: Borislav Petkov @ 2015-12-09 19:14 UTC (permalink / raw)
  To: Ingo Molnar
  Cc: Dave Hansen, the arch/x86 maintainers, LKML, Yu, Yu-cheng, Yu,
	Fenghua, Thomas Gleixner, H. Peter Anvin

On Wed, Dec 09, 2015 at 09:32:10AM +0100, Ingo Molnar wrote:
> 
> * Dave Hansen <dave.hansen@intel.com> wrote:
> 
> > Hey Ingo,
> > 
> > We were starting to look at reenabling XSAVES support and tried to use the 
> > 'noxsave' and 'noxsaves' kernel command-line options.  The rewrite moved the FPU 
> > initialization to before we even are parsing command-line options, even the 
> > early_param()s.
> > 
> > Do you have any preferences on how it gets fixed?
> > 
> > The most obvious thing would be to just defer as much of the FPU setup as we can 
> > until after parse_early_param() has happened.  The only other think I can think 
> > of doing would be to try to do some *really* early, simple, command-line parsing 
> > to look for 'noxsave' and friends.
> > 
> > Thoughts?
> 
> Hm, so given that having the FPU operational is key to a functioning kernel (and 
> other kernel subsystems may make use of FPU functionality), I'd rather bring 
> option parsing earlier than change the FPU setup sequence. FPU setup should be 
> done when we identify the CPU - not 'very late' as we used to.

We already have early cmdline parsing: arch/x86/lib/cmdline.c

See users for examples.

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2015-12-09 19:14 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-12-08 20:43 x86 fpu: command-line parameters broken post-FPU-rewrite Dave Hansen
2015-12-09  8:32 ` Ingo Molnar
2015-12-09 19:14   ` Borislav Petkov

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.