Re: [Qemu-devel] [PATCH v2 1/1] target-ppc: Implement rtas_get_sysparm(PROCESSOR_MODULE_INFO)

[Alexey Kardashevskiy]

On 11/10/2015 03:46 PM, Sukadev Bhattiprolu wrote:
> Alexey Kardashevskiy address@hidden wrote:
> <snip>
>
> | >| When exactly does a socket become a module? The SPAPR spec uses "sockets" 
> here.
> | >
> | >I am trying to get the terminology too :-) Is socket a slot where a
> | >module is attached?
> |
> | Sorry, no idea.
>
> Ok.
>
> |
> |
> | >
> | >I will change the variable name 'modules' to 'sockets'.
> | >|
> | >|
> | >| >+    modinfo->si[0].chips = chips;
> | >| >+    modinfo->si[0].cores_per_chip = cores / chips;
> | >|
> | >|
> | >| What if no "ibm,chip-id" was found and chips == 0?
> | >
> | >If we fail to readdir(xscom) or fail to read the 'ibm,chip-id',
> | >we return an error which we check above.
> |
> |
> | You assume that if there is /proc/device-tree, then there is always
> | "xscom@" but this might not be always the case, like PR KVM on
> | embedded PPC64.
>
> For ibm,chip-id, we do try to read the file (and eliminate duplicates


No, you do not try reading "ibm,chip-id" if there is no "xscom@" under 
/proc/device-tree, there is "continue":

+        if (strncmp(xscom_prefix, ent->d_name, strlen(xscom_prefix))) {
+            continue;
+        }



> chip ids) If we can't read the file (hash_table_add_contents()) we
> return an error, but will check again.



--
Alexey