perf/amd/uncore: Set all slices and threads to restore perf stat -a behaviour
Subsystems (2)
PATCH 1/1 perf/amd/uncore: Set all slices and threads to restore perf stat -a behaviour
Commit 2f217d58a8a0 ("perf/x86/amd/uncore: Set the thread mask for F17h L3 PMCs") inadvertently changed the uncore driver's behaviour wrt perf tool invocations with or without a CPU list, specified with -C / --cpu=.
Change the behaviour of the driver to assume the former all-cpu (-a) case, which is the more commonly desired default. This fixes '-a -A' invocations without explicit cpu lists (-C) to not count L3 events only on behalf of the first thread of the first core in the L3 domain.
BEFORE:
Activity performed by the first thread of the last core (CPU#43) in CPU#40's L3 domain is not reported by CPU#40:
sudo perf stat -a -A -e l3_request_g1.caching_l3_cache_accesses taskset -c 43 perf bench mem memcpy -s 32mb -l 100 -f default
...
CPU36 21,835 l3_request_g1.caching_l3_cache_accesses CPU40 87,066 l3_request_g1.caching_l3_cache_accesses CPU44 17,360 l3_request_g1.caching_l3_cache_accesses
...
AFTER:
The L3 domain activity is now reported by CPU#40:
sudo perf stat -a -A -e l3_request_g1.caching_l3_cache_accesses taskset -c 43 perf bench mem memcpy -s 32mb -l 100 -f default
...
CPU36 354,891 l3_request_g1.caching_l3_cache_accesses CPU40 1,780,870 l3_request_g1.caching_l3_cache_accesses CPU44 315,062 l3_request_g1.caching_l3_cache_accesses
...
---
Stephane, if you want to replace this with the manual thread specification patch you have, that's fine with me.
arch/x86/events/amd/uncore.c | 28 ++++++++-------------------- 1 file changed, 8 insertions(+), 20 deletions(-)
arch/x86/events/amd/uncore.c +8 -20
@@ -181,28 +181,16 @@ static void amd_uncore_del(struct perf_event *event, int flags) }/* -* Convert logical CPU number to L3 PMC Config ThreadMask format +* Return a full thread and slice mask until per-CPU is +* properly supported. */ -static u64 l3_thread_slice_mask (int cpu) +static u64 l3_thread_slice_mask (void ) { -u64 thread_mask, core =topology_core_id (cpu); -unsigned int shift, thread =0 ; +if (boot_cpu_data.x86 <=0x18 ) +return AMD64_L3_SLICE_MASK |AMD64_L3_THREAD_MASK ; -if (topology_smt_supported () && !topology_is_primary_thread (cpu)) - thread =1 ; - -if (boot_cpu_data.x86 <=0x18 ) { - shift =AMD64_L3_THREAD_SHIFT +2 * (core %4 ) + thread; - thread_mask =BIT_ULL (shift); - -return AMD64_L3_SLICE_MASK | thread_mask; - } - - core = (core <<AMD64_L3_COREID_SHIFT ) &AMD64_L3_COREID_MASK ; - shift =AMD64_L3_THREAD_SHIFT + thread; - thread_mask =BIT_ULL (shift); - -return AMD64_L3_EN_ALL_SLICES | core | thread_mask; +return AMD64_L3_EN_ALL_SLICES |AMD64_L3_EN_ALL_CORES | +AMD64_L3_F19H_THREAD_MASK ; }static int amd_uncore_event_init (struct perf_event *event)
@@ -232,7 +220,7 @@ static int amd_uncore_event_init(struct perf_event *event)* For other events, the two fields do not affect the count. */ if (l3_mask &&is_llc_event (event)) - hwc->config |=l3_thread_slice_mask (event->cpu); + hwc->config |=l3_thread_slice_mask (); uncore =event_to_amd_uncore (event);if (!uncore)
Signature · 1 line
2.20.1