From: Borislav Petkov <bp@suse.de>
To: Huang Rui <ray.huang@amd.com>
Cc: linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org,
Alex Deucher <alexander.deucher@amd.com>,
Jason Bagavatsingham <jason.bagavatsingham@gmail.com>,
"Pierre-Loup A . Griffais" <pgriffais@valvesoftware.com>,
Nathan Fontenot <nathan.fontenot@amd.com>,
"Rafael J . Wysocki" <rafael.j.wysocki@intel.com>,
x86@kernel.org
Subject: Re: [PATCH] x86, sched: Fix the AMD CPPC maximum perf on some specific generations
Date: Tue, 20 Apr 2021 10:22:31 +0200 [thread overview]
Message-ID: <20210420082231.GE2326@zn.tnic> (raw)
In-Reply-To: <20210420080943.1045886-1-ray.huang@amd.com>
On Tue, Apr 20, 2021 at 04:09:43PM +0800, Huang Rui wrote:
> Some AMD Ryzen generations has different calculation method on maximum
> perf. 255 is not for all asics, some specific generations used 166 as
> the maximum perf. This patch is to fix the different maximum perf value
Avoid having "This patch" or "This commit" in the commit message. It is
tautologically useless.
Also, do
$ git grep 'This patch' Documentation/process
for more details.
> of AMD CPPC.
>
> Fixes: 41ea667227ba ("x86, sched: Calculate frequency invariance for AMD systems")
> Fixes: 3c55e94c0ade ("cpufreq: ACPI: Extend frequency tables to cover boost frequencies")
>
> Reported-by: Jason Bagavatsingham <jason.bagavatsingham@gmail.com>
> Tested-by: Jason Bagavatsingham <jason.bagavatsingham@gmail.com>
> Signed-off-by: Huang Rui <ray.huang@amd.com>
> Cc: Alex Deucher <alexander.deucher@amd.com>
> Cc: Nathan Fontenot <nathan.fontenot@amd.com>
> Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Cc: x86@kernel.org
> ---
> arch/x86/kernel/smpboot.c | 33 ++++++++++++++++++++++++++-
> drivers/cpufreq/acpi-cpufreq.c | 41 ++++++++++++++++++++++++++++++++++
> 2 files changed, 73 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
> index 02813a7f3a7c..705bc5ceb1ea 100644
> --- a/arch/x86/kernel/smpboot.c
> +++ b/arch/x86/kernel/smpboot.c
> @@ -2033,6 +2033,37 @@ static bool intel_set_max_freq_ratio(void)
> }
>
> #ifdef CONFIG_ACPI_CPPC_LIB
> +static u64 amd_get_highest_perf(void)
> +{
struct cpuinfo_x86 *c = &boot_cpu_data;
and then you can use "c" everywhere.
> + u64 cppc_max_perf;
u64 for something which fits in a byte?
Also,
max_perf = 255;
and you can remove the else and default branches below.
> +
> + switch (boot_cpu_data.x86) {
> + case 0x17:
> + if ((boot_cpu_data.x86_model >= 0x30 &&
> + boot_cpu_data.x86_model < 0x40) ||
> + (boot_cpu_data.x86_model >= 0x70 &&
> + boot_cpu_data.x86_model < 0x80))
> + cppc_max_perf = 166;
> + else
> + cppc_max_perf = 255;
> + break;
> + case 0x19:
> + if ((boot_cpu_data.x86_model >= 0x20 &&
> + boot_cpu_data.x86_model < 0x30) ||
> + (boot_cpu_data.x86_model >= 0x40 &&
> + boot_cpu_data.x86_model < 0x70))
> + cppc_max_perf = 166;
> + else
> + cppc_max_perf = 255;
> + break;
> + default:
> + cppc_max_perf = 255;
> + break;
> + }
> +
> + return cppc_max_perf;
> +}
Why is this here and not in arch/x86/kernel/cpu/amd.c?
> +
> static bool amd_set_max_freq_ratio(void)
> {
> struct cppc_perf_caps perf_caps;
> @@ -2046,8 +2077,8 @@ static bool amd_set_max_freq_ratio(void)
> return false;
> }
>
> - highest_perf = perf_caps.highest_perf;
> nominal_perf = perf_caps.nominal_perf;
> + highest_perf = amd_get_highest_perf();
>
> if (!highest_perf || !nominal_perf) {
> pr_debug("Could not retrieve highest or nominal performance\n");
> diff --git a/drivers/cpufreq/acpi-cpufreq.c b/drivers/cpufreq/acpi-cpufreq.c
> index d1bbc16fba4b..e5c03360db20 100644
> --- a/drivers/cpufreq/acpi-cpufreq.c
> +++ b/drivers/cpufreq/acpi-cpufreq.c
> @@ -630,6 +630,44 @@ static int acpi_cpufreq_blacklist(struct cpuinfo_x86 *c)
> #endif
>
> #ifdef CONFIG_ACPI_CPPC_LIB
> +
> +static u64 get_amd_max_boost_ratio(unsigned int cpu, u64 nominal_perf)
> +{
> + u64 boost_ratio, cppc_max_perf;
> +
> + if (!nominal_perf)
> + return 0;
> +
> + switch (boot_cpu_data.x86) {
> + case 0x17:
> + if ((boot_cpu_data.x86_model >= 0x30 &&
> + boot_cpu_data.x86_model < 0x40) ||
> + (boot_cpu_data.x86_model >= 0x70 &&
> + boot_cpu_data.x86_model < 0x80))
> + cppc_max_perf = 166;
> + else
> + cppc_max_perf = 255;
> + break;
> + case 0x19:
> + if ((boot_cpu_data.x86_model >= 0x20 &&
> + boot_cpu_data.x86_model < 0x30) ||
> + (boot_cpu_data.x86_model >= 0x40 &&
> + boot_cpu_data.x86_model < 0x70))
> + cppc_max_perf = 166;
> + else
> + cppc_max_perf = 255;
> + break;
> + default:
> + cppc_max_perf = 255;
> + break;
This chunk is repeated here. Why?
--
Regards/Gruss,
Boris.
SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer, HRB 36809, AG Nürnberg
next prev parent reply other threads:[~2021-04-20 8:22 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-20 8:09 [PATCH] x86, sched: Fix the AMD CPPC maximum perf on some specific generations Huang Rui
2021-04-20 8:22 ` Borislav Petkov [this message]
2021-04-20 10:48 ` Huang Rui
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210420082231.GE2326@zn.tnic \
--to=bp@suse.de \
--cc=alexander.deucher@amd.com \
--cc=jason.bagavatsingham@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=nathan.fontenot@amd.com \
--cc=pgriffais@valvesoftware.com \
--cc=rafael.j.wysocki@intel.com \
--cc=ray.huang@amd.com \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).