Messages in this thread Patch in this message |  | | From | Jiri Olsa <> | Subject | [PATCH] perf/x86: Fix overlap counter scheduling bug | Date | Tue, 1 Nov 2016 16:44:28 +0100 |
| |
My fuzzer testing hits following warning in the counter scheduling code:
WARNING: CPU: 0 PID: 0 at arch/x86/events/core.c:718 perf_assign_events+0x2ae/0x2c0 Call Trace: <IRQ> dump_stack+0x68/0x93 __warn+0xcb/0xf0 warn_slowpath_null+0x1d/0x20 perf_assign_events+0x2ae/0x2c0 uncore_assign_events+0x1a7/0x250 [intel_uncore] uncore_pmu_event_add+0x7a/0x3c0 [intel_uncore] event_sched_in.isra.104+0xf6/0x2e0 group_sched_in+0x6e/0x190 ...
The reason is that the counter scheduling code assumes overlap constraints with mask weight < SCHED_STATES_MAX.
This assumption is broken with uncore cbox constraint added for snbep in: 3b19e4c98c03 perf/x86: Fix event constraint for SandyBridge-EP C-Box
It's also easily triggered by running following perf command on snbep box: # perf stat -e 'uncore_cbox_0/event=0x1f/,uncore_cbox_0/event=0x1f/,uncore_cbox_0/event=0x1f/' -a
Fixing this by increasing the SCHED_STATES_MAX to 3 and adding build check for EVENT_CONSTRAINT_OVERLAP macro.
Cc: Vince Weaver <vince@deater.net> Cc: Robert Richter <rric@kernel.org> Cc: Yan Zheng <zheng.z.yan@intel.com> Signed-off-by: Jiri Olsa <jolsa@kernel.org> --- arch/x86/events/core.c | 3 --- arch/x86/events/perf_event.h | 10 +++++++++- 2 files changed, 9 insertions(+), 4 deletions(-)
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c index d31735f37ed7..c725f8854216 100644 --- a/arch/x86/events/core.c +++ b/arch/x86/events/core.c @@ -676,9 +676,6 @@ struct sched_state { unsigned long used[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; }; -/* Total max is X86_PMC_IDX_MAX, but we are O(n!) limited */ -#define SCHED_STATES_MAX 2 - struct perf_sched { int max_weight; int max_events; diff --git a/arch/x86/events/perf_event.h b/arch/x86/events/perf_event.h index 5874d8de1f8d..4553275b37d4 100644 --- a/arch/x86/events/perf_event.h +++ b/arch/x86/events/perf_event.h @@ -277,8 +277,16 @@ struct cpu_hw_events { * dramatically. The number of such EVENT_CONSTRAINT_OVERLAP() macros * and its counter masks must be kept at a minimum. */ + +/* Total max is X86_PMC_IDX_MAX, but we are O(n!) limited */ +#define SCHED_STATES_MAX 3 + +/* Check we dont overlap beyond the states max. */ +#define OVERLAP_CHECK(n) (!!sizeof(char[1 - 2*!!(HWEIGHT(n) > SCHED_STATES_MAX)])) +#define OVERLAP_HWEIGHT(n) (OVERLAP_CHECK(n)*HWEIGHT(n)) + #define EVENT_CONSTRAINT_OVERLAP(c, n, m) \ - __EVENT_CONSTRAINT(c, n, m, HWEIGHT(n), 1, 0) + __EVENT_CONSTRAINT(c, n, m, OVERLAP_HWEIGHT(n), 1, 0) /* * Constraint on the Event code. -- 2.7.4
|  |