x86, perf: Add Top Down events to Intel Goldmont
Subsystems (2)
PATCH 1/1 x86, perf: Add Top Down events to Intel Goldmont
From: Kan Liang <[email protected]>
Goldmont supports full Top Down level 1 metrics (FrontendBound, Retiring, Backend Bound and Bad Speculation).
It has 3 wide pipeline.
--- arch/x86/events/intel/core.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+)
arch/x86/events/intel/core.c +26 -0
@@ -1553,6 +1553,31 @@ static __initconst const u64 slm_hw_cache_event_ids }, }; +EVENT_ATTR_STR (topdown-total-slots, td_total_slots_glm,"event=0x3c" ); +EVENT_ATTR_STR (topdown-total-slots.scale, td_total_slots_scale_glm,"3" ); +/* UOPS_NOT_DELIVERED.ANY */ +EVENT_ATTR_STR (topdown-fetch-bubbles, td_fetch_bubbles_glm, +"event=0x9c" ); +/* ISSUE_SLOTS_NOT_CONSUMED.RECOVERY */ +EVENT_ATTR_STR (topdown-recovery-bubbles, td_recovery_bubbles_glm, +"event=0xca,umask=0x02" ); +/* UOPS_RETIRED.ANY */ +EVENT_ATTR_STR (topdown-slots-retired, td_slots_retired_glm, +"event=0xc2" ); +/* UOPS_ISSUED.ANY */ +EVENT_ATTR_STR (topdown-slots-issued, td_slots_issued_glm, +"event=0x0e" ); + +static struct attribute *glm_events_attrs[] = { +EVENT_PTR (td_total_slots_glm), +EVENT_PTR (td_total_slots_scale_glm), +EVENT_PTR (td_fetch_bubbles_glm), +EVENT_PTR (td_recovery_bubbles_glm), +EVENT_PTR (td_slots_issued_glm), +EVENT_PTR (td_slots_retired_glm), +NULL +}; +static struct extra_reg intel_glm_extra_regs[]__read_mostly = {/* must define OFFCORE_RSP_X first, see intel_fixup_er() */ INTEL_UEVENT_EXTRA_REG (0x01b7 ,MSR_OFFCORE_RSP_0 ,0x760005ffbfull ,RSP_0 ),
@@ -3750,6 +3775,7 @@ __init int intel_pmu_init(void) x86_pmu.pebs_prec_dist = true; x86_pmu.lbr_pt_coexist = true; x86_pmu.flags |=PMU_FL_HAS_RSP_1 ; + x86_pmu.cpu_events = glm_events_attrs;pr_cont ("Goldmont events, " );break ;
Signature · 1 line
2.4.3
Quoted 23 lines · * [email protected] <[email protected]> wrote:
From: Kan Liang <[email protected]>
Goldmont supports full Top Down level 1 metrics (FrontendBound, Retiring, Backend Bound and Bad Speculation).
It has 3 wide pipeline.Signed-off-by Kan Liang <[email protected]>--- arch/x86/events/intel/core.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+)arch/x86/events/intel/core.c +4
-0@@ -1553,6 +1553,31 @@ static __initconst const u64 slm_hw_cache_event_ids }, }; +EVENT_ATTR_STR (topdown-total-slots, td_total_slots_glm,"event=0x3c" ); +EVENT_ATTR_STR (topdown-total-slots.scale, td_total_slots_scale_glm,"3" ); +/* UOPS_NOT_DELIVERED.ANY */ +EVENT_ATTR_STR(topdown-fetch-bubbles, td_fetch_bubbles_glm,
arch/x86/events/intel/core.c static __initconst const u64 slm_hw_cache_event_ids · line 1569
Quoted 10 lines
arch/x86/events/intel/core.c +10
-0+ "event=0x9c"); +/* ISSUE_SLOTS_NOT_CONSUMED.RECOVERY */ +EVENT_ATTR_STR (topdown-recovery-bubbles, td_recovery_bubbles_glm, +"event=0xca,umask=0x02" ); +/* UOPS_RETIRED.ANY */ +EVENT_ATTR_STR (topdown-slots-retired, td_slots_retired_glm, +"event=0xc2" ); +/* UOPS_ISSUED.ANY */ +EVENT_ATTR_STR (topdown-slots-issued, td_slots_issued_glm, +"event=0x0e" );
Please don't
break lines
in funny places and make code
unreadable.
Instead have a look at the existing event listing style in arch/x86/events/intel/core.c and match that!
Ok?
Thanks,
Ingo