Bugzilla 217447 points out that ftrace bitmask fields still use the legacy dynamic-array format, which makes trace consumers treat them as unsigned long arrays instead of bitmaps.
This is visible in the ipi events today: ipi_send_cpumask already reports its CPU mask as '__data_loc cpumask_t', but ipi_raise still exposes target_cpus as '__data_loc unsigned long[]'. Switch ipi_raise to __cpumask() and the matching helpers so its tracefs format matches the existing cpumask representation used by the other ipi event. The underlying storage size stays the same, but trace data consumers can now recognize the field as a cpumask directly. Link: https://bugzilla.kernel.org/show_bug.cgi?id=217447 Signed-off-by: CaoRuichuang <[email protected]> --- include/trace/events/ipi.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/include/trace/events/ipi.h b/include/trace/events/ipi.h index 9912f0ded..fae4f8eac 100644 --- a/include/trace/events/ipi.h +++ b/include/trace/events/ipi.h @@ -68,16 +68,16 @@ TRACE_EVENT(ipi_raise, TP_ARGS(mask, reason), TP_STRUCT__entry( - __bitmask(target_cpus, nr_cpumask_bits) + __cpumask(target_cpus) __field(const char *, reason) ), TP_fast_assign( - __assign_bitmask(target_cpus, cpumask_bits(mask), nr_cpumask_bits); + __assign_cpumask(target_cpus, cpumask_bits(mask)); __entry->reason = reason; ), - TP_printk("target_mask=%s (%s)", __get_bitmask(target_cpus), __entry->reason) + TP_printk("target_mask=%s (%s)", __get_cpumask(target_cpus), __entry->reason) ); DECLARE_EVENT_CLASS(ipi_handler, -- 2.39.5 (Apple Git-154)
