[PATCH v6 13/14] arm64, acpi, numa: NUMA support based on SRAT and SLIT

David Daney ddaney at caviumnetworks.com
Thu May 12 13:40:24 PDT 2016

On 05/12/2016 09:24 AM, Catalin Marinas wrote:
> On Thu, May 12, 2016 at 08:27:08AM -0700, David Daney wrote:
>> On 05/12/2016 02:49 AM, Catalin Marinas wrote:
>>> On Wed, May 11, 2016 at 05:06:13PM -0700, David Daney wrote:
>>>> On 05/11/2016 03:39 AM, Catalin Marinas wrote:
>> [...]
>>>>> I wonder whether you could replace the get_mpidr_in_madt() function with
>>>>> something like acpi_get_phys_id(). It looks like get_mpidr_in_madt()
>>>>> duplicates functionality already available elsewhere.
>>>> I just tried that, and it doesn't work.
>>>> The problem is that this code is being run very early in the boot, and
>>>> kmalloc cannot be used.  acpi_get_phys_id() and its ilk can only be used
>>>> once we have working kmalloc.  We need to extract the NUMA information early
>>>> like this precisely because it is needed to initializing the slab system
>>>> Notice that we are using early_acpi_os_unmap_memory() et al. in
>>>> get_mpidr_in_madt() explicitly for this reason.
>>>> In summary: I don't think we need another revision of this patch, it is like
>>>> this for a good reason.
>>> Slightly confusing, in another reply you said you are going to address
>>> my comment. So, is it doable?
>> I don't think so.
>> My previous reply, to the thread in 0/14, was prematurely made with the
>> incorrect assumption that it was a simple change.  Now, after really digging
>> in to the code, and attempting to do as you suggested, I have changed my
>> mind.
> Would the snippet below help with avoiding any kmalloc calls? At a quick
> look, it seems that it's only map_mat_entry() that ends up using
> kmalloc() calls. Alternatively, exporting map_madt_entry() may work as
> well.

I think making a non-static version of map_madt_entry() is the cleanest 
solution.  That works, so I will send a new version of the patch set 
that does that instead.


> diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c
> index 33a38d604630..77af0a7df914 100644
> --- a/drivers/acpi/processor_core.c
> +++ b/drivers/acpi/processor_core.c
> @@ -152,6 +152,9 @@ static phys_cpuid_t map_mat_entry(acpi_handle handle, int type, u32 acpi_id)
>   	struct acpi_subtable_header *header;
>   	phys_cpuid_t phys_id = PHYS_CPUID_INVALID;
> +	if (!acpi_gbl_permanent_mmap)
> +		return phys_id;
> +
>   	if (ACPI_FAILURE(acpi_evaluate_object(handle, "_MAT", NULL, &buffer)))
>   		goto exit;

More information about the linux-arm-kernel mailing list