Re: [PATCH] powerpc/numa: Fix topology_physical_package_id() on pSeries

[Daniel Henrique Barboza <danielhb413@gmail.com>]

On 3/15/21 1:16 PM, Cédric Le Goater wrote:
> On 3/15/21 4:12 PM, Daniel Henrique Barboza wrote:
>>
>>
>> On 3/12/21 11:31 AM, Cédric Le Goater wrote:
>>> Initial commit 15863ff3b8da ("powerpc: Make chip-id information
>>> available to userspace") introduce a cpu_to_chip_id() routine for the
>>> PowerNV platform using the "ibm,chip-id" property to query the chip id
>>> of a CPU. But PAPR does not specify such a property and the node id
>>> query is broken.
>>>
>>> Use cpu_to_node() instead which guarantees to have a correct value on
>>> all platforms, PowerNV an pSeries.
>>
>> It is worth mentioning that that this patch will change how
>> topology_physical_package_id() represents in a QEMU guest. Right now, ibm,chip-id
>> in QEMU is matching the socket-id. After this patch, topology_physical_package_id()
>> will now match the NUMA id of the CPU.
> 
> yes. I should have added some more background.
> 
> LPARs are impacted by the use of ibm,chip-id because the property
> does not exist under PowerVM and the topology-id in sysfs is always
> -1 even if NUMA nodes are defined.
> 
> Under QEMU/KVM, ibm,chip-id is badly calculated when using uncommon
> SMT configuration. This leads to a bogus topology-id value being
> exported in sysfs.
> 
> The use of cpu_to_node() guarantees to have a correct NUMA node id
> under both environments QEMU/KVM and PowerVM.
> 
> On the PowerNV platform, the numa node id returned by cpu_to_node()
> is computed from the "ibm,associativity" property of the CPU. Its
> value is built from the OPAL chip id and is equivalent to ibm,chip-id.
> 
> May be I should rephrase the commit log in a v2 ?

It's a fine idea, given that apparently we don't have documentation explaining
these details (well, at least I didn't find any). We can reference the commit
message later on as explanation :)




Thanks,

DHB

> 
> C.
> 
>   
>> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
>> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
>>
>>>
>>> Cc: Nathan Lynch <nathanl@linux.ibm.com>
>>> Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
>>> Cc: Vasant Hegde <hegdevasant@linux.vnet.ibm.com>
>>> Signed-off-by: Cédric Le Goater <clg@kaod.org>
>>> ---
>>>    arch/powerpc/include/asm/topology.h | 2 +-
>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/arch/powerpc/include/asm/topology.h b/arch/powerpc/include/asm/topology.h
>>> index 3beeb030cd78..887c42a4e43d 100644
>>> --- a/arch/powerpc/include/asm/topology.h
>>> +++ b/arch/powerpc/include/asm/topology.h
>>> @@ -123,7 +123,7 @@ static inline int cpu_to_coregroup_id(int cpu)
>>>    #ifdef CONFIG_PPC64
>>>    #include <asm/smp.h>
>>>    -#define topology_physical_package_id(cpu)    (cpu_to_chip_id(cpu))
>>> +#define topology_physical_package_id(cpu)    (cpu_to_node(cpu))
>>>      #define topology_sibling_cpumask(cpu)    (per_cpu(cpu_sibling_map, cpu))
>>>    #define topology_core_cpumask(cpu)    (cpu_cpu_mask(cpu))
>>>
>