Re: [PATCH v10] perf: add qcom l2 cache perf events driver
From: Mark Rutland
Date: Tue Feb 07 2017 - 13:46:05 EST
Hi Neil,
On Tue, Feb 07, 2017 at 01:14:04PM -0500, Neil Leeder wrote:
> Adds perf events support for L2 cache PMU.
>
> The L2 cache PMU driver is named 'l2cache_0' and can be used
> with perf events to profile L2 events such as cache hits
> and misses on Qualcomm Technologies processors.
>
> Signed-off-by: Neil Leeder <nleeder@xxxxxxxxxxxxxx>
Thanks for respinning this. This looks good to me now:
Reviewed-by: Mark Rutland <mark.rutland@xxxxxxx>
Will and I should be able to pick this up shortly.
There's one minor thing I'd like to clean up below, but we can sort that
out when applying -- there's no need to respin.
> +static struct cluster_pmu *l2_cache_associate_cpu_with_cluster(
> + struct l2cache_pmu *l2cache_pmu, int cpu)
> +{
> + u64 mpidr;
> + int cpu_cluster_id;
> + struct cluster_pmu *cluster;
> +
> + /*
> + * This assumes that the cluster_id is in MPIDR[aff1] for
> + * single-threaded cores, and MPIDR[aff2] for multi-threaded
> + * cores. This logic will have to be updated if this changes.
> + */
> + mpidr = read_cpuid_mpidr();
> + if (mpidr & MPIDR_MT_BITMASK)
> + cpu_cluster_id = MPIDR_AFFINITY_LEVEL(mpidr, 2);
> + else
> + cpu_cluster_id = MPIDR_AFFINITY_LEVEL(mpidr, 1);
> +
> + list_for_each_entry(cluster, &l2cache_pmu->clusters, next) {
> + if (cluster->cluster_id == cpu_cluster_id) {
> + dev_info(&l2cache_pmu->pdev->dev,
> + "CPU%d associated with cluster %d\n", cpu,
> + cluster->cluster_id);
> + cpumask_set_cpu(cpu, &cluster->cluster_cpus);
> + *per_cpu_ptr(l2cache_pmu->pmu_cluster, cpu) = cluster;
> + return cluster;
> + }
> + }
To minimise nesting, I'd like to fix this up as:
list_for_each_entry(cluster, &l2cache_pmu->clusters, next) {
if (cluster->cluster_id != cpu_cluster_id)
continue;
dev_info(&l2cache_pmu->pdev->dev,
"CPU%d associated with cluster %d\n", cpu,
cluster->cluster_id);
cpumask_set_cpu(cpu, &cluster->cluster_cpus);
*per_cpu_ptr(l2cache_pmu->pmu_cluster, cpu) = cluster;
return cluster;
}
Regardless, this is fine by me.
Thanks,
Mark.