Re: [PATCH] Append 'nohpet' in dom0 cmdline to prevent dom0 from using HPET

[Jeremy Fitzhardinge <jeremy@goop.org>]

On 10/15/09 19:19, Wei, Gang wrote:
> Wei, Gang wrote:
>   
>> Jeremy Fitzhardinge wrote:
>>     
>>> Well, making disable_hpet() non-static and calling it from Xen setup
>>> is the same as appending nohpet to the command line, but less hacky.
>>>       
>> Yes, you are right. I mixed disable_hpet() with hpet_disable(). I
>> just try to avoid changes in normal kernel code. If such slight
>> change is acceptable, please just make it this way.  
>>     
> Here is the new patch.
>
> diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
> index 9b32c88..366e1e5 100644
> --- a/arch/x86/kernel/hpet.c
> +++ b/arch/x86/kernel/hpet.c
> @@ -96,7 +96,7 @@ static int __init hpet_setup(char *str)
>  }
>  __setup("hpet=", hpet_setup);
>  
> -static int __init disable_hpet(char *str)
> +int __init disable_hpet(char *str)
>  {
>  	boot_hpet_disable = 1;
>  	return 1;
> diff --git a/arch/x86/xen/setup.c b/arch/x86/xen/setup.c
> index 5eeeedb..3774578 100644
> --- a/arch/x86/xen/setup.c
> +++ b/arch/x86/xen/setup.c
> @@ -34,6 +34,10 @@ extern void xen_sysenter_target(void);
>  extern void xen_syscall_target(void);
>  extern void xen_syscall32_target(void);
>  
> +#ifdef CONFIG_HPET_TIMER
> +extern int __init disable_hpet(char *str);
> +#endif
>   
It would be better to add

#else
static inline int disable_hpet(char *str) { return 0; }
#endif


> +
>  static unsigned long __init xen_release_chunk(phys_addr_t start_addr, phys_addr_t end_addr)
>  {
>  	struct xen_memory_reservation reservation = {
> @@ -280,6 +284,10 @@ void __init xen_arch_setup(void)
>  		printk(KERN_INFO "ACPI in unprivileged domain disabled\n");
>  		disable_acpi();
>  	}
> +#endif 
> +
> +#ifdef CONFIG_HPET_TIMER
> +	disable_hpet(NULL);
>  #endif
>   
and do without the #ifdef here.  Also, it wouldn't hurt to put a small
explanatory message here.

    J