Re: [PATCH v2 13/16] cpufreq/amd-pstate: Check if CPPC request has changed before writing to the MSR or shared memory

[Mario Limonciello]

On 12/9/2024 02:56, Gautham R. Shenoy wrote:
> Hello Mario,
>
>
> On Sun, Dec 08, 2024 at 12:30:28AM -0600, Mario Limonciello wrote:
> > Move the common MSR field formatting code to msr_update_perf() from
> > its callers.
> >
> > Ensure that the MSR write is necessary before flushing a write out.
> > Also drop the comparison from the passive flow tracing.
> >
> > Reviewed-and-tested-by: Dhananjay Ugwekar <dhananjay.ugwekar@xxxxxxx>
> > Signed-off-by: Mario Limonciello <mario.limonciello@xxxxxxx>
> > ---
>
> [..snip..]
>
> > diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c
> > index dd11ba6c00cc3..2178931fbf87b 100644
> > --- a/drivers/cpufreq/amd-pstate.c
> > +++ b/drivers/cpufreq/amd-pstate.c
> > @@ -224,15 +224,26 @@ static s16 shmem_get_epp(struct amd_cpudata *cpudata)
> >   static int msr_update_perf(struct amd_cpudata *cpudata, u32 min_perf,
> >   			   u32 des_perf, u32 max_perf, u32 epp, bool fast_switch)
> >   {
> > -	u64 value;
> > +	u64 value, prev;
> > +
> > +	value = prev = READ_ONCE(cpudata->cppc_req_cached);
> > +
> > +	value &= ~(AMD_CPPC_MAX_PERF_MASK | AMD_CPPC_MIN_PERF_MASK |
> > +		   AMD_CPPC_DES_PERF_MASK | AMD_CPPC_EPP_PERF_MASK);
> > +	value |= FIELD_PREP(AMD_CPPC_MAX_PERF_MASK, max_perf);
> > +	value |= FIELD_PREP(AMD_CPPC_DES_PERF_MASK, des_perf);
> > +	value |= FIELD_PREP(AMD_CPPC_MIN_PERF_MASK, min_perf);
> > +	value |= FIELD_PREP(AMD_CPPC_EPP_PERF_MASK, epp);
> > +
> > +	if (value == prev)
> > +		return 0;
> >   
> > -	value = READ_ONCE(cpudata->cppc_req_cached);
> >   	if (fast_switch) {
> > -		wrmsrl(MSR_AMD_CPPC_REQ, READ_ONCE(cpudata->cppc_req_cached));
> > +		wrmsrl(MSR_AMD_CPPC_REQ, value);
> >   		return 0;
> >   	} else {
> > -		int ret = wrmsrl_on_cpu(cpudata->cpu, MSR_AMD_CPPC_REQ,
> > -					READ_ONCE(cpudata->cppc_req_cached));
> > +		int ret = wrmsrl_on_cpu(cpudata->cpu, MSR_AMD_CPPC_REQ, value);
> > +
> >   		if (ret)
> >   			return ret;
>
> Ok, so you are recomputing the value in this patch. Does it also make
> sense to move trace_amd_pstate_perf() call to this place?

You mean so that essentially it only logs events that REALLY write to 
the MSR/shared memory region?

I guess with an ftrace function call graph + that event you can 
effectively determine that is what happened.

An alternative is to move it right above the (value == prev) check and 
include "prev" in the trace event.

I think that's my preference at least - you capture whether it changed 
and what the intent was.

> --
> Thanks and Regards
> gautham.
>
> >   	}
> > @@ -528,9 +539,7 @@ static void amd_pstate_update(struct amd_cpudata *cpudata, u32 min_perf,
> >   {
> >   	unsigned long max_freq;
> >   	struct cpufreq_policy *policy = cpufreq_cpu_get(cpudata->cpu);
> > -	u64 prev = READ_ONCE(cpudata->cppc_req_cached);
> >   	u32 nominal_perf = READ_ONCE(cpudata->nominal_perf);
> > -	u64 value = prev;
> >   
> >   	des_perf = clamp_t(unsigned long, des_perf, min_perf, max_perf);
> >   
> > @@ -546,27 +555,14 @@ static void amd_pstate_update(struct amd_cpudata *cpudata, u32 min_perf,
> >   	if (!cpudata->boost_supported)
> >   		max_perf = min_t(unsigned long, nominal_perf, max_perf);
> >   
> > -	value &= ~(AMD_CPPC_MAX_PERF_MASK | AMD_CPPC_MIN_PERF_MASK |
> > -		   AMD_CPPC_DES_PERF_MASK);
> > -	value |= FIELD_PREP(AMD_CPPC_MAX_PERF_MASK, max_perf);
> > -	value |= FIELD_PREP(AMD_CPPC_DES_PERF_MASK, des_perf);
> > -	value |= FIELD_PREP(AMD_CPPC_MIN_PERF_MASK, min_perf);
> > -
> >   	if (trace_amd_pstate_perf_enabled() && amd_pstate_sample(cpudata)) {
> >   		trace_amd_pstate_perf(min_perf, des_perf, max_perf, cpudata->freq,
> >   			cpudata->cur.mperf, cpudata->cur.aperf, cpudata->cur.tsc,
> > -				cpudata->cpu, (value != prev), fast_switch);
> > +				cpudata->cpu, fast_switch);
> >   	}
> >   
> > -	if (value == prev)
> > -		goto cpufreq_policy_put;
> > -
> > -	WRITE_ONCE(cpudata->cppc_req_cached, value);
> > -
> >   	amd_pstate_update_perf(cpudata, min_perf, des_perf, max_perf, 0, fast_switch);
> >   
> > -cpufreq_policy_put:
> > -
> >   	cpufreq_cpu_put(policy);
> >   }
> >   
> > @@ -1562,19 +1558,10 @@ static void amd_pstate_epp_cpu_exit(struct cpufreq_policy *policy)
> >   static int amd_pstate_epp_update_limit(struct cpufreq_policy *policy)
> >   {
> >   	struct amd_cpudata *cpudata = policy->driver_data;
> > -	u64 value;
> >   	u32 epp;
> >   
> >   	amd_pstate_update_min_max_limit(policy);
> >   
> > -	value = READ_ONCE(cpudata->cppc_req_cached);
> > -
> > -	value &= ~(AMD_CPPC_MAX_PERF_MASK | AMD_CPPC_MIN_PERF_MASK |
> > -		   AMD_CPPC_DES_PERF_MASK | AMD_CPPC_EPP_PERF_MASK);
> > -	value |= FIELD_PREP(AMD_CPPC_MAX_PERF_MASK, cpudata->max_limit_perf);
> > -	value |= FIELD_PREP(AMD_CPPC_DES_PERF_MASK, 0);
> > -	value |= FIELD_PREP(AMD_CPPC_MIN_PERF_MASK, cpudata->min_limit_perf);
> > -
> >   	if (cpudata->policy == CPUFREQ_POLICY_PERFORMANCE)
> >   		epp = 0;
> >   	else
> > --
> > 2.43.0