When running the command: 'perf record -e "{instructions,instructions:p}" -j any,counter sleep 1', a "shift-out-of-bounds" warning is reported on CWF.
[ 5231.981423][ C17] UBSAN: shift-out-of-bounds in /kbuild/src/consumer/arch/x86/events/intel/lbr.c:970:15 [ 5231.981428][ C17] shift exponent 64 is too large for 64-bit type 'long long unsigned int' [ 5231.981436][ C17] CPU: 17 UID: 0 PID: 211871 Comm: sleep Tainted: G S W 6.18.0-2025-12-09-intel-next-48166-g6cf574943ba3 #1 PREEMPT(none) [ 5231.981445][ C17] Tainted: [S]=CPU_OUT_OF_SPEC, [W]=WARN [ 5231.981447][ C17] Hardware name: Intel Corporation AvenueCity/AvenueCity, BIOS BHSDCRB1.IPC.3544.P98.2508260307 08/26/2025 [ 5231.981449][ C17] Call Trace: [ 5231.981453][ C17] <NMI> [ 5231.981455][ C17] dump_stack_lvl+0x4b/0x70 [ 5231.981463][ C17] ubsan_epilogue+0x5/0x2b [ 5231.981468][ C17] __ubsan_handle_shift_out_of_bounds.cold+0x61/0xe6 [ 5231.981472][ C17] ? __entry_text_end+0x158b/0x102259 [ 5231.981475][ C17] intel_pmu_lbr_counters_reorder.isra.0.cold+0x2a/0xa7 [ 5231.981480][ C17] ? __task_pid_nr_ns+0x134/0x2a0 [ 5231.981483][ C17] ? __pfx_intel_pmu_lbr_counters_reorder.isra.0+0x10/0x10 [ 5231.981486][ C17] ? __pfx_perf_output_sample+0x10/0x10 [ 5231.981489][ C17] ? arch_perf_update_userpage+0x293/0x310 [ 5231.981491][ C17] ? __pfx_arch_perf_update_userpage+0x10/0x10 [ 5231.981494][ C17] ? local_clock_noinstr+0xd/0x100 [ 5231.981498][ C17] ? calc_timer_values+0x2cb/0x860 [ 5231.981501][ C17] ? perf_event_update_userpage+0x399/0x5b0 [ 5231.981505][ C17] ? __pfx_perf_event_update_userpage+0x10/0x10 [ 5231.981508][ C17] ? local_clock_noinstr+0xd/0x100 [ 5231.981511][ C17] ? __perf_event_account_interrupt+0x11c/0x540 [ 5231.981514][ C17] intel_pmu_lbr_save_brstack+0xc0/0x4c0 [ 5231.981518][ C17] setup_arch_pebs_sample_data+0x114b/0x2400 [ 5231.981522][ C17] ? __pfx_x86_perf_event_set_period+0x10/0x10 [ 5231.981526][ C17] intel_pmu_drain_arch_pebs+0x64d/0xcc0 [ 5231.981530][ C17] ? __pfx_intel_pmu_drain_arch_pebs+0x10/0x10 [ 5231.981534][ C17] ? unwind_next_frame+0x11c5/0x1df0 [ 5231.981541][ C17] ? intel_pmu_drain_bts_buffer+0xbf/0x6e0 [ 5231.981545][ C17] ? __pfx_intel_pmu_drain_bts_buffer+0x10/0x10 [ 5231.981550][ C17] handle_pmi_common+0x5c5/0xcb0 [ 5231.981553][ C17] ? __pfx_handle_pmi_common+0x10/0x10 [ 5231.981556][ C17] ? intel_idle+0x64/0xb0 [ 5231.981560][ C17] ? intel_bts_interrupt+0xe5/0x4c0 [ 5231.981562][ C17] ? __pfx_intel_bts_interrupt+0x10/0x10 [ 5231.981565][ C17] ? intel_pmu_lbr_filter+0x27f/0x910 [ 5231.981568][ C17] intel_pmu_handle_irq+0x2ed/0x600 [ 5231.981571][ C17] perf_event_nmi_handler+0x219/0x280 [ 5231.981575][ C17] ? __pfx_perf_event_nmi_handler+0x10/0x10 [ 5231.981579][ C17] ? unwind_next_frame+0x11c5/0x1df0 [ 5231.981582][ C17] nmi_handle.part.0+0x11b/0x3a0 [ 5231.981585][ C17] ? unwind_next_frame+0x11c5/0x1df0 [ 5231.981588][ C17] default_do_nmi+0x6b/0x180 [ 5231.981591][ C17] fred_exc_nmi+0x3e/0x80 [ 5231.981594][ C17] asm_fred_entrypoint_kernel+0x41/0x60 [ 5231.981596][ C17] RIP: 0010:unwind_next_frame+0x11c5/0x1df0 ......
The warning occurs because the second "instructions:p" event, which involves branch counters sampling, is incorrectly programmed to fixed counter 0 instead of the general-purpose (GP) counters 0-3 that support branch counters sampling. Currently only GP counters 0~3 support branch counters sampling on CWF, any event involving branch counters sampling should be programed on GP counters 0~3. Since the counter index of fixed counter 0 is 32, it leads to the "src" value in below code is right shifted 64 bits and trigger the "shift-out-of-bounds" warning.
cnt = (src >> (order[j] * LBR_INFO_BR_CNTR_BITS)) & LBR_INFO_BR_CNTR_MASK;
The root cause is the loss of the branch counters constraint for the last event in the branch counters sampling event group. This results in the second "instructions:p" event being programmed on fixed counter 0 incorrectly instead of the appropriate GP counters 0~3.
To address this, we apply the missing branch counters constraint for the last event in the group. Additionally, we introduce a new function, `intel_set_branch_counter_constr()`, to apply the branch counters constraint and avoid code duplication.
Cc: stable@vger.kernel.org Reported-by: Xudong Hao xudong.hao@intel.com Fixes: 33744916196b ("perf/x86/intel: Support branch counters logging") Signed-off-by: Dapeng Mi dapeng1.mi@linux.intel.com --- arch/x86/events/intel/core.c | 30 ++++++++++++++++++++---------- 1 file changed, 20 insertions(+), 10 deletions(-)
diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c index aad89c9d9514..7c6a0001c8e4 100644 --- a/arch/x86/events/intel/core.c +++ b/arch/x86/events/intel/core.c @@ -4364,6 +4364,19 @@ static inline void intel_pmu_set_acr_caused_constr(struct perf_event *event, event->hw.dyn_constraint &= hybrid(event->pmu, acr_cause_mask64); }
+static inline int intel_set_branch_counter_constr(struct perf_event *event, + int *num) +{ + if (branch_sample_call_stack(event)) + return -EINVAL; + if (branch_sample_counters(event)) { + (*num)++; + event->hw.dyn_constraint &= x86_pmu.lbr_counters; + } + + return 0; +} + static int intel_pmu_hw_config(struct perf_event *event) { int ret = x86_pmu_hw_config(event); @@ -4434,21 +4447,18 @@ static int intel_pmu_hw_config(struct perf_event *event) * group, which requires the extra space to store the counters. */ leader = event->group_leader; - if (branch_sample_call_stack(leader)) + if (intel_set_branch_counter_constr(leader, &num)) return -EINVAL; - if (branch_sample_counters(leader)) { - num++; - leader->hw.dyn_constraint &= x86_pmu.lbr_counters; - } leader->hw.flags |= PERF_X86_EVENT_BRANCH_COUNTERS;
for_each_sibling_event(sibling, leader) { - if (branch_sample_call_stack(sibling)) + if (intel_set_branch_counter_constr(sibling, &num)) + return -EINVAL; + } + + if (event != leader) { + if (intel_set_branch_counter_constr(event, &num)) return -EINVAL; - if (branch_sample_counters(sibling)) { - num++; - sibling->hw.dyn_constraint &= x86_pmu.lbr_counters; - } }
if (num > fls(x86_pmu.lbr_counters))
base-commit: 9929dffce5ed7e2988e0274f4db98035508b16d9