[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v3 05/15] xen/x86: introduce "cpufreq=amd-cppc" xen cmdline


  • To: "Penny, Zheng" <penny.zheng@xxxxxxx>
  • From: Jan Beulich <jbeulich@xxxxxxxx>
  • Date: Thu, 27 Mar 2025 08:48:08 +0100
  • Autocrypt: addr=jbeulich@xxxxxxxx; keydata= xsDiBFk3nEQRBADAEaSw6zC/EJkiwGPXbWtPxl2xCdSoeepS07jW8UgcHNurfHvUzogEq5xk hu507c3BarVjyWCJOylMNR98Yd8VqD9UfmX0Hb8/BrA+Hl6/DB/eqGptrf4BSRwcZQM32aZK 7Pj2XbGWIUrZrd70x1eAP9QE3P79Y2oLrsCgbZJfEwCgvz9JjGmQqQkRiTVzlZVCJYcyGGsD /0tbFCzD2h20ahe8rC1gbb3K3qk+LpBtvjBu1RY9drYk0NymiGbJWZgab6t1jM7sk2vuf0Py O9Hf9XBmK0uE9IgMaiCpc32XV9oASz6UJebwkX+zF2jG5I1BfnO9g7KlotcA/v5ClMjgo6Gl MDY4HxoSRu3i1cqqSDtVlt+AOVBJBACrZcnHAUSuCXBPy0jOlBhxPqRWv6ND4c9PH1xjQ3NP nxJuMBS8rnNg22uyfAgmBKNLpLgAGVRMZGaGoJObGf72s6TeIqKJo/LtggAS9qAUiuKVnygo 3wjfkS9A3DRO+SpU7JqWdsveeIQyeyEJ/8PTowmSQLakF+3fote9ybzd880fSmFuIEJldWxp Y2ggPGpiZXVsaWNoQHN1c2UuY29tPsJgBBMRAgAgBQJZN5xEAhsDBgsJCAcDAgQVAggDBBYC AwECHgECF4AACgkQoDSui/t3IH4J+wCfQ5jHdEjCRHj23O/5ttg9r9OIruwAn3103WUITZee e7Sbg12UgcQ5lv7SzsFNBFk3nEQQCACCuTjCjFOUdi5Nm244F+78kLghRcin/awv+IrTcIWF hUpSs1Y91iQQ7KItirz5uwCPlwejSJDQJLIS+QtJHaXDXeV6NI0Uef1hP20+y8qydDiVkv6l IreXjTb7DvksRgJNvCkWtYnlS3mYvQ9NzS9PhyALWbXnH6sIJd2O9lKS1Mrfq+y0IXCP10eS FFGg+Av3IQeFatkJAyju0PPthyTqxSI4lZYuJVPknzgaeuJv/2NccrPvmeDg6Coe7ZIeQ8Yj t0ARxu2xytAkkLCel1Lz1WLmwLstV30g80nkgZf/wr+/BXJW/oIvRlonUkxv+IbBM3dX2OV8 AmRv1ySWPTP7AAMFB/9PQK/VtlNUJvg8GXj9ootzrteGfVZVVT4XBJkfwBcpC/XcPzldjv+3 HYudvpdNK3lLujXeA5fLOH+Z/G9WBc5pFVSMocI71I8bT8lIAzreg0WvkWg5V2WZsUMlnDL9 mpwIGFhlbM3gfDMs7MPMu8YQRFVdUvtSpaAs8OFfGQ0ia3LGZcjA6Ik2+xcqscEJzNH+qh8V m5jjp28yZgaqTaRbg3M/+MTbMpicpZuqF4rnB0AQD12/3BNWDR6bmh+EkYSMcEIpQmBM51qM EKYTQGybRCjpnKHGOxG0rfFY1085mBDZCH5Kx0cl0HVJuQKC+dV2ZY5AqjcKwAxpE75MLFkr wkkEGBECAAkFAlk3nEQCGwwACgkQoDSui/t3IH7nnwCfcJWUDUFKdCsBH/E5d+0ZnMQi+G0A nAuWpQkjM1ASeQwSHEeAWPgskBQL
  • Cc: "Huang, Ray" <Ray.Huang@xxxxxxx>, Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, Anthony PERARD <anthony.perard@xxxxxxxxxx>, "Orzel, Michal" <Michal.Orzel@xxxxxxx>, Julien Grall <julien@xxxxxxx>, Roger Pau Monné <roger.pau@xxxxxxxxxx>, Stefano Stabellini <sstabellini@xxxxxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxxx>
  • Delivery-date: Thu, 27 Mar 2025 07:48:17 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 27.03.2025 04:12, Penny, Zheng wrote:
>> -----Original Message-----
>> From: Jan Beulich <jbeulich@xxxxxxxx>
>> Sent: Wednesday, March 26, 2025 6:55 PM
>>
>> On 26.03.2025 09:35, Penny, Zheng wrote:
>>>> -----Original Message-----
>>>> From: Jan Beulich <jbeulich@xxxxxxxx>
>>>> Sent: Monday, March 24, 2025 11:26 PM
>>>>
>>>> On 06.03.2025 09:39, Penny Zheng wrote:
>>>>> --- a/xen/include/acpi/cpufreq/cpufreq.h
>>>>> +++ b/xen/include/acpi/cpufreq/cpufreq.h
>>>>> @@ -28,6 +28,7 @@ enum cpufreq_xen_opt {
>>>>>      CPUFREQ_none,
>>>>>      CPUFREQ_xen,
>>>>>      CPUFREQ_hwp,
>>>>> +    CPUFREQ_amd_cppc,
>>>>>  };
>>>>>  extern enum cpufreq_xen_opt cpufreq_xen_opts[2];
>>>>
>>>> I'm pretty sure I pointed out before that this array needs to grow,
>>>> now that you add a 3rd kind of handling.
>>>>
>>>
>>> Hmmm, but the CPUFREQ_hwp and CPUFREQ_amd_cppc are incompatible
>> options.
>>> I thought cpufreq_xen_opts[] shall reflect available choices on their 
>>> hardware.
>>> Even if users define "cpufreq=hwp;amd-cppc;xen", in Intel platform,
>>> cpufreq_xen_opts[] shall contain  CPUFREQ_hwp and CPUFREQ_xen, while
>>> in amd platform, cpufreq_xen_opts[] shall contain CPUFREQ_amd_cppc and
>>> CPUFREQ_xen
>>
>> Maybe I misread the code, but the impression I got was that "cpufreq=hwp;amd-
>> cppc;xen"
> 
> My bad. In my platform, I haven't enabled the CONFIG_INTEL. I previously 
> assumed that
> CONFIG_INTEL and CONFIG_AMD are incompatible options, which leads to the 
> following code
> ```
> else if ( IS_ENABLED(CONFIG_INTEL) && choice < 0 &&
>           !cmdline_strcmp(str, "hwp") )
> {
>     xen_processor_pmbits |= XEN_PROCESSOR_PM_PX;
>     cpufreq_controller = FREQCTL_xen;
> ```
> shall not be working in AMD platform...
> May I ask why not make them incompatible pair? I assumed it each wraps 
> vendor-specific feature, like vmx vs svm...

I'm sorry to say this, but that seems like a pretty odd question to ask. Distros
quite clearly want to build one single hypervisor which can be used on both
Intel and AMD hardware. CONFIG_* are build-time constants after all, not runtime
values. We use them in if() where possible (instead of in #if / #ifdef) simply
to expose as much code as possible to at least syntax and alike checking by the
compiler, irrespective of configuration used by a particular individual. This
way we limit the risk of bit-rotting and unexpected build failures at least 
some.

Jan



 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.