Messages in this thread Patch in this message | | | Subject | Re: [PATCH 22/22] perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support | From | "Liang, Kan" <> | Date | Thu, 9 May 2019 11:04:01 -0400 |
| |
On 5/6/2019 5:26 PM, Len Brown wrote: > From: Len Brown <len.brown@intel.com> > > Syntax update only -- no logical or functional change. > > In response to the new multi-die/package changes, update variable names > to use the more generic "pmuid" terminology, instead of "pkgid", > as the pmu can refer to either packages or die.
The perf rapl "pmu" in the code is cross the pkg/die. We only register one rapl pmu for whole system for now. I think it may be better use "die" to replace the "pkg" as well. How about the patch as below?
Thanks, Kan
From a898867b24ed8ea3582b0c8b3218838701249065 Mon Sep 17 00:00:00 2001 From: Kan Liang <kan.liang@linux.intel.com> Date: Thu, 9 May 2019 07:21:35 -0700 Subject: [PATCH 2/2] perf/x86/intel/rapl: rename internal variables in response to multi-die/pkg support
Syntax update only -- no logical or functional change.
In response to the new multi-die/package changes, update variable names to use "die" terminology, instead of "pkg". For previous platforms which doesn't have multi-die, "die" is identical as "pkg".
Originally-by: Len Brown <len.brown@intel.com> Signed-off-by: Kan Liang <kan.liang@linux.intel.com> --- arch/x86/events/intel/rapl.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/arch/x86/events/intel/rapl.c b/arch/x86/events/intel/rapl.c index e49f69c..f7fc94d 100644 --- a/arch/x86/events/intel/rapl.c +++ b/arch/x86/events/intel/rapl.c @@ -148,7 +148,7 @@ struct rapl_pmu {
struct rapl_pmus { struct pmu pmu; - unsigned int maxpkg; + unsigned int maxdie; struct rapl_pmu *pmus[]; };
@@ -161,13 +161,13 @@ static u64 rapl_timer_ms;
static inline struct rapl_pmu *cpu_to_rapl_pmu(unsigned int cpu) { - unsigned int pkgid = topology_logical_die_id(cpu); + unsigned int dieid = topology_logical_die_id(cpu);
/* * The unsigned check also catches the '-1' return value for non * existent mappings in the topology map. */ - return pkgid < rapl_pmus->maxpkg ? rapl_pmus->pmus[pkgid] : NULL; + return dieid < rapl_pmus->maxdie ? rapl_pmus->pmus[dieid] : NULL; }
static inline u64 rapl_read_counter(struct perf_event *event) @@ -668,22 +668,22 @@ static void cleanup_rapl_pmus(void) { int i;
- for (i = 0; i < rapl_pmus->maxpkg; i++) + for (i = 0; i < rapl_pmus->maxdie; i++) kfree(rapl_pmus->pmus[i]); kfree(rapl_pmus); }
static int __init init_rapl_pmus(void) { - int maxpkg = topology_max_packages() * topology_max_die_per_package(); + int maxdie = topology_max_packages() * topology_max_die_per_package(); size_t size;
- size = sizeof(*rapl_pmus) + maxpkg * sizeof(struct rapl_pmu *); + size = sizeof(*rapl_pmus) + maxdie * sizeof(struct rapl_pmu *); rapl_pmus = kzalloc(size, GFP_KERNEL); if (!rapl_pmus) return -ENOMEM;
- rapl_pmus->maxpkg = maxpkg; + rapl_pmus->maxdie = maxdie; rapl_pmus->pmu.attr_groups = rapl_attr_groups; rapl_pmus->pmu.task_ctx_nr = perf_invalid_context; rapl_pmus->pmu.event_init = rapl_pmu_event_init; -- 2.7.4
| |