Re: [PATCH 1/4] x86, amd: Normalize compute unit IDs on multi-node processors

[Andrew Morton <akpm@linux-foundation.org>]

On Mon, 24 Jan 2011 16:05:40 +0100
Hans Rosenfeld <hans.rosenfeld@amd.com> wrote:

> From: Andreas Herrmann <andreas.herrmann3@amd.com>
> 
> On multi-node CPUs we don't need the socket wide compute unit ID but
> the node-wide compute unit ID. Thus we need to normalize the value.
> This is similar to what we do with cpu_core_id.
> 
> A compute unit is then identified by physical_package_id, node_id, and
> compute_unit_id.
> 
> ...
>
> --- a/arch/x86/kernel/cpu/amd.c
> +++ b/arch/x86/kernel/cpu/amd.c
> @@ -261,7 +261,7 @@ static int __cpuinit nearby_node(int apicid)
>  #ifdef CONFIG_X86_HT
>  static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c)
>  {
> -	u32 nodes;
> +	u32 nodes, cores_per_cu;
>  	u8 node_id;
>  	int cpu = smp_processor_id();
>  
> @@ -276,6 +276,7 @@ static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c)
>  		/* get compute unit information */
>  		smp_num_siblings = ((ebx >> 8) & 3) + 1;
>  		c->compute_unit_id = ebx & 0xff;
> +		cores_per_cu = ((ebx >> 8) & 3) + 1;
>  	} else if (cpu_has(c, X86_FEATURE_NODEID_MSR)) {
>  		u64 value;
>  
> @@ -288,15 +289,18 @@ static void __cpuinit amd_get_topology(struct cpuinfo_x86 *c)
>  	/* fixup multi-node processor information */
>  	if (nodes > 1) {
>  		u32 cores_per_node;
> +		u32 cus_per_node;
>  
>  		set_cpu_cap(c, X86_FEATURE_AMD_DCM);
>  		cores_per_node = c->x86_max_cores / nodes;
> +		cus_per_node = cores_per_node / cores_per_cu;
>  
>  		/* store NodeID, use llc_shared_map to store sibling info */
>  		per_cpu(cpu_llc_id, cpu) = node_id;
>  
>  		/* core id to be in range from 0 to (cores_per_node - 1) */
> -		c->cpu_core_id = c->cpu_core_id % cores_per_node;
> +		c->cpu_core_id %= cores_per_node;
> +		c->compute_unit_id %= cus_per_node;
>  	}
>  }
>  #endif

arch/x86/kernel/cpu/amd.c: In function 'init_amd':
arch/x86/kernel/cpu/amd.c:268: warning: 'cores_per_cu' may be used uninitialized in this function

The code looks buggy to me.