[RFC PATCH 05/38] arm64: mpam: Add helpers to change a task or cpu's MPAM PARTID/PMG values

[Jonathan Cameron]

On Fri, 5 Dec 2025 21:58:28 +0000
James Morse <james.morse at arm.com> wrote:

> Care must be taken when modifying the PARTID and PMG of a task in any
> per-task structure as writing these values may race with the task being
> scheduled in, and reading the modified values.
> 
> Add helpers to set the task properties, and the CPU default value.
> These use WRITE_ONCE() that pairs with the READ_ONCE() in mpam_get_regval()
> to avoid causing torn values.
> 
> CC: Dave Martin <Dave.Martin at arm.com>
> CC: Ben Horgan <ben.horgan at arm.com>
> Signed-off-by: James Morse <james.morse at arm.com>
> ---
>  arch/arm64/include/asm/mpam.h | 30 ++++++++++++++++++++++++++++++
>  1 file changed, 30 insertions(+)
> 
> diff --git a/arch/arm64/include/asm/mpam.h b/arch/arm64/include/asm/mpam.h
> index 86a55176f884..2960ffaf6574 100644
> --- a/arch/arm64/include/asm/mpam.h
> +++ b/arch/arm64/include/asm/mpam.h
> @@ -5,6 +5,7 @@
>  #define __ASM__MPAM_H
>  
>  #include <linux/bitops.h>
> +#include <linux/bitfield.h>
>  #include <linux/init.h>
>  #include <linux/jump_label.h>
>  #include <linux/percpu.h>
> @@ -37,6 +38,35 @@ extern u64 arm64_mpam_global_default;
>   * A value in struct thread_info is used instead of struct task_struct as the
>   * cpu's u64 register format is used, but struct task_struct has two u32'.
>   */
I'd be tempted to reorder this so the comment ends up near at least
one use of task_thread_info(tsk)->mpam_partid_pmg. Otherwise it is even
less obvious what it is referring to.

Easiest might be to put the mpam_get_regval() before the mpam_set_task_partid_pmg()


> +static inline void mpam_set_cpu_defaults(int cpu, u16 partid_d, u16 partid_i,
> +					 u8 pmg_d, u8 pmg_i)
> +{
> +	u64 default_val;
> +
> +	default_val = FIELD_PREP(MPAM0_EL1_PARTID_D, partid_d);
> +	default_val |= FIELD_PREP(MPAM0_EL1_PARTID_I, partid_i);
> +	default_val |= FIELD_PREP(MPAM0_EL1_PMG_D, pmg_d);
> +	default_val |= FIELD_PREP(MPAM0_EL1_PMG_I, pmg_i);
> +
> +	WRITE_ONCE(per_cpu(arm64_mpam_default, cpu), default_val);
> +}
> +
> +static inline void mpam_set_task_partid_pmg(struct task_struct *tsk,
> +					    u16 partid_d, u16 partid_i,
> +					    u8 pmg_d, u8 pmg_i)
> +{
> +#ifdef CONFIG_ARM64_MPAM
> +	u64 regval;
> +
> +	regval = FIELD_PREP(MPAM0_EL1_PARTID_D, partid_d);
> +	regval |= FIELD_PREP(MPAM0_EL1_PARTID_I, partid_i);
> +	regval |= FIELD_PREP(MPAM0_EL1_PMG_D, pmg_d);
> +	regval |= FIELD_PREP(MPAM0_EL1_PMG_I, pmg_i);

Maybe a macro or helper to build regval given this replicates the block in
mpam_set_cpu_defaults.  Obviously ignore if this changes in later patches!

> +
> +	WRITE_ONCE(task_thread_info(tsk)->mpam_partid_pmg, regval);
> +#endif
> +}
> +
>  static inline u64 mpam_get_regval(struct task_struct *tsk)
>  {
>  #ifdef CONFIG_ARM64_MPAM