[PATCH v5 11/17] irqchip/riscv-imsic: Add S-mode MSI address list

Anup Patel anup at brainfault.org
Wed Sep 23 04:55:11 PDT 2026


On Mon, Aug 31, 2026 at 8:30 PM Andrew Jones
<andrew.jones at oss.qualcomm.com> wrote:
>
> Upcoming RISC-V IOMMU interrupt remapping needs to map every possible
> host IMSIC target into one contiguous MSI IOVA range. MSI message
> composition then uses the target CPU's position within that range.
>
> Build an S-mode IMSIC physical address array indexed by logical CPU.
> Require every possible CPU to have an initialized IMSIC page before
> publishing the array. This preserves physical address zero as a valid
> target and guarantees the array has num_possible_cpus() entries.
>
> The array size and each target's index can therefore be derived
> directly, without storing duplicate count or per-CPU index metadata.
>
> Signed-off-by: Andrew Jones <andrew.jones at oss.qualcomm.com>
> ---
>  drivers/irqchip/irq-riscv-imsic-state.c | 32 +++++++++++++++++++++++++
>  drivers/irqchip/irq-riscv-imsic-state.h |  1 +
>  2 files changed, 33 insertions(+)
>
> diff --git a/drivers/irqchip/irq-riscv-imsic-state.c b/drivers/irqchip/irq-riscv-imsic-state.c
> index df38a7670a89..c7ebddd308ec 100644
> --- a/drivers/irqchip/irq-riscv-imsic-state.c
> +++ b/drivers/irqchip/irq-riscv-imsic-state.c
> @@ -709,6 +709,31 @@ static int __init imsic_get_mmio_resource(struct fwnode_handle *fwnode,
>         return of_address_to_resource(to_of_node(fwnode), index, res);
>  }
>
> +static int __init imsic_init_smode_msi_pa(void)
> +{
> +       struct imsic_global_config *global = &imsic->global;
> +       phys_addr_t *smode_msi_pa;
> +       unsigned int cpu;
> +
> +       smode_msi_pa = kcalloc(num_possible_cpus(), sizeof(*smode_msi_pa), GFP_KERNEL);
> +       if (!smode_msi_pa)
> +               return -ENOMEM;
> +
> +       for_each_possible_cpu(cpu) {
> +               struct imsic_local_config *local = per_cpu_ptr(global->local, cpu);
> +
> +               if (!local->msi_va) {
> +                       kfree(smode_msi_pa);
> +                       return -ENODEV;
> +               }
> +
> +               smode_msi_pa[cpu] = local->msi_pa;
> +       }
> +
> +       imsic->smode_msi_pa = smode_msi_pa;
> +       return 0;
> +}
> +
>  static int __init imsic_parse_fwnode(struct fwnode_handle *fwnode,
>                                      struct imsic_global_config *global,
>                                      u32 *nr_parent_irqs,
> @@ -959,6 +984,12 @@ int __init imsic_setup_state(struct fwnode_handle *fwnode, void *opaque)
>                 goto out_local_cleanup;
>         }
>
> +       rc = imsic_init_smode_msi_pa();
> +       if (rc) {
> +               pr_err("%pfwP: failed to initialize S-mode MSI addresses\n", fwnode);
> +               goto out_local_cleanup;
> +       }
> +
>         /* Initialize matrix allocator */
>         rc = imsic_matrix_init();
>         if (rc) {
> @@ -984,6 +1015,7 @@ int __init imsic_setup_state(struct fwnode_handle *fwnode, void *opaque)
>  out_free_local:
>         free_percpu(imsic->global.local);
>  out_free_priv:
> +       kfree(imsic->smode_msi_pa);
>         kfree(imsic);
>         imsic = NULL;
>         return rc;
> diff --git a/drivers/irqchip/irq-riscv-imsic-state.h b/drivers/irqchip/irq-riscv-imsic-state.h
> index c42ee180b305..11352f14e32d 100644
> --- a/drivers/irqchip/irq-riscv-imsic-state.h
> +++ b/drivers/irqchip/irq-riscv-imsic-state.h
> @@ -49,6 +49,7 @@ struct imsic_priv {
>
>         /* Global configuration common for all HARTs */
>         struct imsic_global_config              global;
> +       phys_addr_t                             *smode_msi_pa;

Please a comment over here about what this is for.

I also suggest renaming "smode_msi_pa" to "msi_pa_list" because
theoritically IMSIC driver can work for M-mode as well although
we have not tried it for M-mode.

Regards,
Anup



More information about the linux-riscv mailing list