On Tue, 20 Aug 2024, Edgar E. Iglesias wrote:
> From: "Edgar E. Iglesias" <edgar.igles...@amd.com>
> 
> Add SMP support for Xen PVH ARM guests.
> Create ms->smp.max_cpus ioreq servers to handle hotplug.
> 
> Note that ms->smp.max_cpus will be passed to us by the
> user (Xen tools) set to the guests maxvcpus.
> 
> The value in mc->max_cpus is an absolute maximum for the
> -smp option and won't be used to setup ioreq servers unless
> the user explicitly specifies it with -smp.
> 
> If the user doesn't pass -smp on the command-line, smp.cpus
> and smp.max_cpus will default to 1.
> 
> Signed-off-by: Edgar E. Iglesias <edgar.igles...@amd.com>


Reviewed-by: Stefano Stabellini <sstabell...@kernel.org>


> ---
>  hw/arm/xen_arm.c | 23 +++++++++++++++++++++--
>  1 file changed, 21 insertions(+), 2 deletions(-)
> 
> diff --git a/hw/arm/xen_arm.c b/hw/arm/xen_arm.c
> index 5f75cc3779..fda65d0d8d 100644
> --- a/hw/arm/xen_arm.c
> +++ b/hw/arm/xen_arm.c
> @@ -173,7 +173,7 @@ static void xen_arm_init(MachineState *machine)
>  
>      xen_init_ram(machine);
>  
> -    xen_register_ioreq(xam->state, machine->smp.cpus, &xen_memory_listener);
> +    xen_register_ioreq(xam->state, machine->smp.max_cpus, 
> &xen_memory_listener);
>  
>      xen_create_virtio_mmio_devices(xam);
>  
> @@ -218,7 +218,26 @@ static void xen_arm_machine_class_init(ObjectClass *oc, 
> void *data)
>      MachineClass *mc = MACHINE_CLASS(oc);
>      mc->desc = "Xen PVH ARM machine";
>      mc->init = xen_arm_init;
> -    mc->max_cpus = 1;
> +
> +    /*
> +     * mc->max_cpus holds the MAX value allowed in the -smp command-line 
> opts.
> +     *
> +     * 1. If users don't pass any -smp option:
> +     *   ms->smp.cpus will default to 1.
> +     *   ms->smp.max_cpus will default to 1.
> +     *
> +     * 2. If users pass -smp X:
> +     *   ms->smp.cpus will be set to X.
> +     *   ms->smp.max_cpus will also be set to X.
> +     *
> +     * 3. If users pass -smp X,maxcpus=Y:
> +     *   ms->smp.cpus will be set to X.
> +     *   ms->smp.max_cpus will be set to Y.
> +     *
> +     * In scenarios 2 and 3, if X or Y are set to something larger than
> +     * mc->max_cpus, QEMU will bail out with an error message.
> +     */
> +    mc->max_cpus = GUEST_MAX_VCPUS;
>      mc->default_machine_opts = "accel=xen";
>      /* Set explicitly here to make sure that real ram_size is passed */
>      mc->default_ram_size = 0;
> -- 
> 2.43.0
> 

Reply via email to