[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [RFC PATCH v2 05/25] x86: NUMA: Move generic dummy_numa_init to separate function



On Thu, Apr 20, 2017 at 9:42 PM, Julien Grall <julien.grall@xxxxxxx> wrote:
> Hi Vijay,
>
>
> On 28/03/17 16:53, vijay.kilari@xxxxxxxxx wrote:
>>
>> From: Vijaya Kumar K <Vijaya.Kumar@xxxxxxxxxx>
>>
>> Split numa_initmem_init() so that the numa fallback code is moved
>> as separate function which is generic.
>>
>> Signed-off-by: Vijaya Kumar K <Vijaya.Kumar@xxxxxxxxxx>
>> ---
>>  xen/arch/x86/numa.c | 29 +++++++++++++++++------------
>>  1 file changed, 17 insertions(+), 12 deletions(-)
>>
>> diff --git a/xen/arch/x86/numa.c b/xen/arch/x86/numa.c
>> index 6b794a7..0888d53 100644
>> --- a/xen/arch/x86/numa.c
>> +++ b/xen/arch/x86/numa.c
>> @@ -268,21 +268,10 @@ static int __init numa_emulation(uint64_t start_pfn,
>> uint64_t end_pfn)
>>  }
>>  #endif
>>
>> -void __init numa_initmem_init(unsigned long start_pfn, unsigned long
>> end_pfn)
>> +static void __init numa_dummy_init(unsigned long start_pfn, unsigned long
>> end_pfn)
>>  {
>>      int i;
>>
>> -#ifdef CONFIG_NUMA_EMU
>> -    if ( get_numa_fake() && !numa_emulation(start_pfn, end_pfn) )
>> -        return;
>> -#endif
>> -
>> -#ifdef CONFIG_ACPI_NUMA
>> -    if ( !is_numa_off() && !acpi_scan_nodes((uint64_t)start_pfn <<
>> PAGE_SHIFT,
>> -         (uint64_t)end_pfn << PAGE_SHIFT) )
>> -        return;
>> -#endif
>> -
>>      printk(KERN_INFO "%s\n",
>>             is_numa_off() ? "NUMA turned off" : "No NUMA configuration
>> found");
>>
>> @@ -301,6 +290,22 @@ void __init numa_initmem_init(unsigned long
>> start_pfn, unsigned long end_pfn)
>>                      (paddr_t)end_pfn << PAGE_SHIFT);
>>  }
>>
>> +void __init numa_initmem_init(unsigned long start_pfn, unsigned long
>> end_pfn)
>> +{
>> +#ifdef CONFIG_NUMA_EMU
>> +    if ( get_numa_fake() && !numa_emulation(start_pfn, end_pfn) )
>> +        return;
>> +#endif
>
>
> I am not sure where to comment about it in this series, so I will say it
> here.
>
> As asked on v1, why don't you consider fake NUMA? This would help to test
> the series on non-NUMA platform.

I have not tested non-NUMA case with this series. Agreed this two
lines should be added
to numa_initmem_init() of arm (xen/arch/arm/numa/numa.c)

>
>> +
>> +#ifdef CONFIG_ACPI_NUMA
>> +    if ( !is_numa_off() && !acpi_scan_nodes((uint64_t)start_pfn <<
>> PAGE_SHIFT,
>> +         (uint64_t)end_pfn << PAGE_SHIFT) )
>> +        return;
>> +#endif
>> +
>> +    numa_dummy_init(start_pfn, end_pfn);
>> +}
>> +
>>  void numa_add_cpu(int cpu)
>>  {
>>      cpumask_set_cpu(cpu, &node_to_cpumask[cpu_to_node(cpu)]);
>>
>
> Cheers,
>
> --
>  Julien Grall

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.