Re: [PATCH v4 05/15] xen/x86: introduce "cpufreq=amd-cppc" xen cmdline

[Jan Beulich <jbeulich@xxxxxxxx>]

On 14.04.2025 09:40, Penny Zheng wrote:
> @@ -514,5 +515,16 @@ acpi_cpufreq_driver = {
>  
>  int __init acpi_cpufreq_register(void)
>  {
> -    return cpufreq_register_driver(&acpi_cpufreq_driver);
> +    int ret;
> +
> +    ret = cpufreq_register_driver(&acpi_cpufreq_driver);
> +    if ( ret )
> +        return ret;
> +    /*
> +     * After cpufreq driver registeration, XEN_PROCESSOR_PM_CPPC
> +     * and XEN_PROCESSOR_PM_PX shall become exclusive flags
> +     */
> +    xen_processor_pmbits &= ~XEN_PROCESSOR_PM_CPPC;
> +
> +    return ret;
>  }

Why is no similar adjustment needed in powernow_register_driver()? In principle
I would have expected that it's not each individual driver which needs to care
about this aspect, but that the framework is taking care of this.

> --- /dev/null
> +++ b/xen/arch/x86/acpi/cpufreq/amd-cppc.c
> @@ -0,0 +1,81 @@
> +/* SPDX-License-Identifier: GPL-2.0-only */
> +/*
> + * amd-cppc.c - AMD Processor CPPC Frequency Driver
> + *
> + * Copyright (C) 2025 Advanced Micro Devices, Inc. All Rights Reserved.
> + *
> + * Author: Penny Zheng <penny.zheng@xxxxxxx>
> + *
> + * AMD CPPC cpufreq driver introduces a new CPU performance scaling design
> + * for AMD processors using the ACPI Collaborative Performance and Power
> + * Control (CPPC) feature which provides finer grained frequency control 
> range.
> + */
> +
> +#include <xen/domain.h>
> +#include <xen/init.h>
> +#include <xen/param.h>
> +#include <acpi/cpufreq/cpufreq.h>
> +
> +static bool __init amd_cppc_handle_option(const char *s, const char *end)
> +{
> +    int ret;
> +
> +    ret = parse_boolean("verbose", s, end);
> +    if ( ret >= 0 )
> +    {
> +        cpufreq_verbose = ret;
> +        return true;
> +    }
> +
> +    return false;
> +}
> +
> +int __init amd_cppc_cmdline_parse(const char *s, const char *e)
> +{
> +    do
> +    {

Nit (style): Brace placement is special here, just like it is ...

> +        const char *end = strpbrk(s, ",;");
> +
> +        if ( !amd_cppc_handle_option(s, end) )
> +        {
> +            printk(XENLOG_WARNING
> +                   "cpufreq/amd-cppc: option '%.*s' not recognized\n",
> +                   (int)((end ?: e) - s), s);
> +
> +            return -EINVAL;
> +        }
> +
> +        s = end ? end + 1 : NULL;
> +    } while ( s && s < e );

... here.

> --- a/xen/arch/x86/platform_hypercall.c
> +++ b/xen/arch/x86/platform_hypercall.c
> @@ -542,6 +542,9 @@ ret_t do_platform_op(
>                  ret = -ENOSYS;
>                  break;
>              }
> +            /* Xen doesn't support mixed mode */
> +            ASSERT((xen_processor_pmbits & XEN_PROCESSOR_PM_CPPC) == 0);

Please prefer ! over "== 0" in such purely boolean contexts.

> @@ -573,6 +576,14 @@ ret_t do_platform_op(
>          }
>  
>          case XEN_PM_CPPC:
> +            if ( !(xen_processor_pmbits & XEN_PROCESSOR_PM_CPPC) )
> +            {
> +                ret = -EOPNOTSUPP;
> +                break;
> +            }

While at least you no longer use -ENOSYS here, I question this behavior,
including that for the pre-existing cases: How is the caller supposed to know
whether to invoke this sub-op? Ignoring errors is generally not a good idea,
so it would be better if the caller could blindly issue this request, getting
back success unless there really was an issue with the data provided.

> @@ -102,6 +103,9 @@ static int __init handle_cpufreq_cmdline(enum 
> cpufreq_xen_opt option)
>      cpufreq_xen_opts[cpufreq_xen_cnt++] = option;
>      switch ( option )
>      {
> +    case CPUFREQ_amd_cppc:
> +        xen_processor_pmbits |= XEN_PROCESSOR_PM_CPPC;
> +        break;
>      case CPUFREQ_hwp:
>      case CPUFREQ_xen:
>          xen_processor_pmbits |= XEN_PROCESSOR_PM_PX;

Here and (about) everywhere else: Blank line please between non-fall-through
case blocks. I guess I'm not going to repeat this any further. There are
very tiny switch() statements where it is okay to violate this principle, but
as a rule of thumb - if in doubt, put a blank line there.

> --- a/xen/include/acpi/cpufreq/processor_perf.h
> +++ b/xen/include/acpi/cpufreq/processor_perf.h
> @@ -5,6 +5,9 @@
>  #include <public/sysctl.h>
>  #include <xen/acpi.h>
>  
> +/* ability bits */
> +#define XEN_PROCESSOR_PM_CPPC   8

This needs correlating (at least via commentary, better by build-time checking)
with the other XEN_PROCESSOR_PM_* values. Otherwise someone adding a new #define
in the public header may not (easily) notice a possible conflict. With that in
mind I also question whether 8 is actually a good choice: That's the obvious
next value to use in the public interface. SIF_PM_MASK is 8 bits wide, so a
sensible value to use here would by e.g. 0x100.

Jan