Re: [PATCH v2 2/4] x86, mwaitt: make delay method configurable

[Huang Rui]

On Tue, Jun 09, 2015 at 05:05:43PM +0800, Borislav Petkov wrote:
> On Tue, Jun 09, 2015 at 11:13:39AM +0800, Huang Rui wrote:
> > This patch introduces a kernel parameter (delay), user is able to configure
> > it at boot loader to choose different types of delay method.
> > 
> > Default schema is to use TSC delay. This update can be more flexiable to
> > add new delay method.
> > 
> > Suggested-by: Suravee Suthikulanit <suravee.suthikulpanit@xxxxxxx>
> > Signed-off-by: Huang Rui <ray.huang@xxxxxxx>
> > ---
> >  arch/x86/include/asm/delay.h |    7 +++++++
> >  arch/x86/kernel/setup.c      |   19 +++++++++++++++++++
> >  arch/x86/lib/delay.c         |   12 +++++++++++-
> >  3 files changed, 37 insertions(+), 1 deletion(-)
> > 
> > diff --git a/arch/x86/include/asm/delay.h b/arch/x86/include/asm/delay.h
> > index 9b3b4f2..99873ec 100644
> > --- a/arch/x86/include/asm/delay.h
> > +++ b/arch/x86/include/asm/delay.h
> > @@ -5,4 +5,11 @@
> >  
> >  void use_tsc_delay(void);
> >  
> > +extern unsigned long boot_option_delay;
> > +
> > +enum delay_type {
> > +	DELAY_LOOP=0,
> > +	DELAY_TSC,
> > +};
> > +
> >  #endif /* _ASM_X86_DELAY_H */
> > diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
> > index 0b10698..cc2886d 100644
> > --- a/arch/x86/kernel/setup.c
> > +++ b/arch/x86/kernel/setup.c
> > @@ -111,6 +111,7 @@
> >  #include <asm/mce.h>
> >  #include <asm/alternative.h>
> >  #include <asm/prom.h>
> > +#include <asm/delay.h>
> >  
> >  /*
> >   * max_low_pfn_mapped: highest direct mapped pfn under 4GB
> > @@ -844,6 +845,24 @@ dump_kernel_offset(struct notifier_block *self, unsigned long v, void *p)
> >  	return 0;
> >  }
> >  
> > +static int __init delay_setup(char *str)
> > +{
> > +	if (!str)
> > +		return -EINVAL;
> > +
> > +	if (!strcmp(str, "tsc")) {
> > +		pr_info("using tsc delay\n");
> > +		boot_option_delay = DELAY_TSC;
> > +	} else if (!strcmp(str, "loop")) {
> > +		pr_info("using loop delay\n");
> > +		boot_option_delay = DELAY_LOOP;
> > +	} else
> > +		return -1;
> 
> What Peter said.
> 
> And we did talk about this already in the last review. You guys want to
> drop all those kernel parameters and use MWAITX delay by default when:
> 
> 1. HW supports MWAITX
> 
> *and*
> 
> 2. delay fits in u32.
> 
> Kernel parameters is a bad bad idea.
> 

Sorry, I am suggested to make the mwaitx delay configurable firstly by
our guys...
Why do you think that method is bad?

Thanks,
Rui
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/