linux-kernel - Re: [PATCH] powerpc/smp: Expose die_id and die

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <53fd537b-5549-427f-ad7d-28d58daaa5bd@linux.ibm.com>
Date: Mon, 10 Nov 2025 22:32:48 +0530
From: Shrikanth Hegde <sshegde@...ux.ibm.com>
To: Srikar Dronamraju <srikar@...ux.ibm.com>, linuxppc-dev@...ts.ozlabs.org
Cc: Christophe Leroy <christophe.leroy@...roup.eu>,
        linux-kernel@...r.kernel.org,
        Madhavan Srinivasan <maddy@...ux.ibm.com>,
        Michael Ellerman <mpe@...erman.id.au>,
        Nicholas Piggin <npiggin@...il.com>,
        Peter Zijlstra <peterz@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH] powerpc/smp: Expose die_id and die_cpumask



> 
> Without this change.
> $ grep . /sys/devices/system/cpu/cpu8/topology/{die*,package*} 2> /dev/null
> /sys/devices/system/cpu/cpu8/topology/package_cpus:0000,ffffff00
> /sys/devices/system/cpu/cpu8/topology/package_cpus_list:8-31
> 
> With this change.
> $ grep . /sys/devices/system/cpu/cpu8/topology/{die*,package*} 2> /dev/null
> /sys/devices/system/cpu/cpu8/topology/die_cpus:0000,0000ff00
> /sys/devices/system/cpu/cpu8/topology/die_cpus_list:8-15
> /sys/devices/system/cpu/cpu8/topology/die_id:4
> /sys/devices/system/cpu/cpu8/topology/package_cpus:0000,ffffff00
> /sys/devices/system/cpu/cpu8/topology/package_cpus_list:8-31
> 
> snipped lstopo-no-graphics o/p
> Group0 L#0
>    Package L#1
>      NUMANode L#1 (P#2 8135MB)
>      Die L#0 + Core L#1

Die L#0 -> Here L# indicates logical die number.
One should run --verbose  to see theactual die_id.

For example:
Die L#8 (P#25) --> die_id is 25.


Maybe you could the same CPU's (CPU8) info of lstopo here?


>        L3 L#2 (4096KB) + L2 L#2 (1024KB) + L1d L#2 (32KB) + L1i L#2 (48KB)
>          PU L#8 (P#8)

> 
> Signed-off-by: Srikar Dronamraju <srikar@...ux.ibm.com>
> ---
>   arch/powerpc/include/asm/topology.h |  4 ++++
>   arch/powerpc/kernel/smp.c           | 18 ++++++++++++++++++
>   2 files changed, 22 insertions(+)
> 
> diff --git a/arch/powerpc/include/asm/topology.h b/arch/powerpc/include/asm/topology.h
> index f19ca44512d1..c6ad1eb7e44a 100644
> --- a/arch/powerpc/include/asm/topology.h
> +++ b/arch/powerpc/include/asm/topology.h
> @@ -132,6 +132,8 @@ static inline int cpu_to_coregroup_id(int cpu)
>   #include <asm/cputable.h>
>   
>   struct cpumask *cpu_coregroup_mask(int cpu);
> +const struct cpumask *cpu_die_mask(int cpu);
> +int cpu_die_id(int cpu);
>   
>   #ifdef CONFIG_PPC64
>   #include <asm/smp.h>
> @@ -141,6 +143,8 @@ struct cpumask *cpu_coregroup_mask(int cpu);
>   #define topology_sibling_cpumask(cpu)	(per_cpu(cpu_sibling_map, cpu))
>   #define topology_core_cpumask(cpu)	(per_cpu(cpu_core_map, cpu))
>   #define topology_core_id(cpu)		(cpu_to_core_id(cpu))
> +#define topology_die_id(cpu) (cpu_die_id(cpu))
> +#define topology_die_cpumask(cpu) (cpu_die_mask(cpu))
>   

nit:

can we fix the tab spacing here please? Maybe like below?

+#define topology_sibling_cpumask(cpu)          (per_cpu(cpu_sibling_map, cpu))
+#define topology_core_cpumask(cpu)             (per_cpu(cpu_core_map, cpu))
+#define topology_core_id(cpu)                  (cpu_to_core_id(cpu))
+#define topology_die_id(cpu)                   (cpu_die_id(cpu))
+#define topology_die_cpumask(cpu)              (cpu_die_mask(cpu))

>   #endif
>   #endif
> diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
> index 68edb66c2964..a0b0b46b78e3 100644
> --- a/arch/powerpc/kernel/smp.c
> +++ b/arch/powerpc/kernel/smp.c
> @@ -1085,6 +1085,24 @@ static int __init init_big_cores(void)
>   	return 0;
>   }
>

As you have mentioned in the changelog, could you add a comment here
explaining on which systems we see this information correctly.   
> +const struct cpumask *cpu_die_mask(int cpu)
> +{
> +	if (has_coregroup_support())
> +		return per_cpu(cpu_coregroup_map, cpu);
> +	else
> +		return cpu_node_mask(cpu);
> +}
> +EXPORT_SYMBOL_GPL(cpu_die_mask);
> +
> +int cpu_die_id(int cpu)
> +{
> +	if (has_coregroup_support())
> +		return cpu_to_coregroup_id(cpu);
> +	else
> +		return -1;
> +}
> +EXPORT_SYMBOL_GPL(cpu_die_id);
> +
>   void __init smp_prepare_cpus(unsigned int max_cpus)
>   {
>   	unsigned int cpu, num_threads;


Other than above:

Reviewed-by: Shrikanth Hegde <sshegde@...ux.ibm.com>