Re: [PATCH v1] perf/core: Restore header fields in sideband output callbacks

[Ian Rogers <irogers@google.com>]

On Tue, Sep 29, 2026 at 5:04 AM Peter Zijlstra <peterz@infradead.org> wrote:
>
> On Mon, Sep 28, 2026 at 06:42:06PM -0700, Ian Rogers wrote:
> > perf_iterate_sb() invokes its callback for each matching perf_event on
> > the CPU and task context, passing a shared caller-allocated event
> > structure.
> >
> > perf_event_header__init_id() increments header->size by
> > event->id_header_size. Unlike perf_event_task_output(),
> > perf_event_comm_output(), perf_event_namespaces_output(),
> > perf_event_cgroup_output(), perf_event_mmap_output(), and
> > perf_callchain_deferred_output(), three sideband callbacks failed to
> > save and restore header.size around perf_event_header__init_id():
> > - perf_event_ksymbol_output()
> > - perf_event_bpf_output()
> > - perf_event_text_poke_output()
>
> I also found perf_event_switch_output(). Does something like so also
> work?

It does. I was considering a general pattern but in the case of
perf_event_switch_output the header size is computed per callback
event rather than in the caller. I wonder if the problem is really
with perf_event_header__init_id mutating and not just assigning the
header. It tangles the header size computation between the caller and
the callback. If we're assigning the header in every callback then
there is no need to save and restore it. It is inefficient to
reinitialize values in the header every time, but saving and restoring
the header is also inefficient. Below is the refactoring I mean and I
think it simplifies the code overall:

diff --git a/arch/powerpc/perf/imc-pmu.c b/arch/powerpc/perf/imc-pmu.c
index f401181d8669..ada2ebfe1fcf 100644
--- a/arch/powerpc/perf/imc-pmu.c
+++ b/arch/powerpc/perf/imc-pmu.c
@@ -1275,6 +1275,8 @@ static int trace_imc_prepare_sample(struct trace_imc_data *mem,
 				    struct perf_event_header *header,
 				    struct perf_event *event)
 {
+	__u16 misc = 0;
+
 	/* Sanity checks for a valid record */
 	if (be64_to_cpu(READ_ONCE(mem->tb1)) > *prev_tb)
 		*prev_tb = be64_to_cpu(READ_ONCE(mem->tb1));
@@ -1289,23 +1291,19 @@ static int trace_imc_prepare_sample(struct trace_imc_data *mem,
 	data->ip =  be64_to_cpu(READ_ONCE(mem->ip));
 	data->period = event->hw.last_period;
 
-	header->type = PERF_RECORD_SAMPLE;
-	header->size = sizeof(*header) + event->header_size;
-	header->misc = 0;
-
 	if (cpu_has_feature(CPU_FTR_ARCH_31)) {
 		switch (IMC_TRACE_RECORD_VAL_HVPR(be64_to_cpu(READ_ONCE(mem->val)))) {
 		case 0:/* when MSR HV and PR not set in the trace-record */
-			header->misc |= PERF_RECORD_MISC_GUEST_KERNEL;
+			misc |= PERF_RECORD_MISC_GUEST_KERNEL;
 			break;
 		case 1: /* MSR HV is 0 and PR is 1 */
-			header->misc |= PERF_RECORD_MISC_GUEST_USER;
+			misc |= PERF_RECORD_MISC_GUEST_USER;
 			break;
 		case 2: /* MSR HV is 1 and PR is 0 */
-			header->misc |= PERF_RECORD_MISC_KERNEL;
+			misc |= PERF_RECORD_MISC_KERNEL;
 			break;
 		case 3: /* MSR HV is 1 and PR is 1 */
-			header->misc |= PERF_RECORD_MISC_USER;
+			misc |= PERF_RECORD_MISC_USER;
 			break;
 		default:
 			pr_info("IMC: Unable to set the flag based on MSR bits\n");
@@ -1313,11 +1311,14 @@ static int trace_imc_prepare_sample(struct trace_imc_data *mem,
 		}
 	} else {
 		if (is_kernel_addr(data->ip))
-			header->misc |= PERF_RECORD_MISC_KERNEL;
+			misc |= PERF_RECORD_MISC_KERNEL;
 		else
-			header->misc |= PERF_RECORD_MISC_USER;
+			misc |= PERF_RECORD_MISC_USER;
 	}
-	perf_event_header__init_id(header, data, event);
+	perf_event_header__init_header_and_id(header, data,
+					      PERF_RECORD_SAMPLE, misc,
+					      sizeof(*header) + event->header_size,
+					      event);
 
 	return 0;
 }
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 5842552294c1..8912298e5d4d 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -1523,9 +1523,10 @@ is_default_overflow_handler(struct perf_event *event)
 }
 
 extern void
-perf_event_header__init_id(struct perf_event_header *header,
-			   struct perf_sample_data *data,
-			   struct perf_event *event);
+perf_event_header__init_header_and_id(struct perf_event_header *header,
+				      struct perf_sample_data *data,
+				      u32 type, u16 misc, u16 size,
+				      struct perf_event *event);
 extern void
 perf_event__output_id_sample(struct perf_event *event,
 			     struct perf_output_handle *handle,
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 33210aff3ee6..5a1bfffd46e6 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -8119,13 +8119,18 @@ static void __perf_event_header__init_id(struct perf_sample_data *data,
 	}
 }
 
-void perf_event_header__init_id(struct perf_event_header *header,
-				struct perf_sample_data *data,
-				struct perf_event *event)
+void perf_event_header__init_header_and_id(struct perf_event_header *header,
+					   struct perf_sample_data *data,
+					   u32 type, u16 misc, u16 size,
+					   struct perf_event *event)
 {
+	header->type = type;
+	header->misc = misc;
 	if (event->attr.sample_id_all) {
-		header->size += event->id_header_size;
+		header->size = size + event->id_header_size;
 		__perf_event_header__init_id(data, event, event->attr.sample_type);
+	} else {
+		header->size = size;
 	}
 }
 
@@ -8959,17 +8964,16 @@ perf_event_read_event(struct perf_event *event,
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
 	struct perf_read_event read_event = {
-		.header = {
-			.type = PERF_RECORD_READ,
-			.misc = 0,
-			.size = sizeof(read_event) + event->read_size,
-		},
 		.pid = perf_event_pid(event, task),
 		.tid = perf_event_tid(event, task),
 	};
 	int ret;
 
-	perf_event_header__init_id(&read_event.header, &sample, event);
+	perf_event_header__init_header_and_id(&read_event.header, &sample,
+					      PERF_RECORD_READ,
+					      /*misc=*/0,
+					      sizeof(read_event) + event->read_size,
+					      event);
 	ret = perf_output_begin(&handle, &sample, event, read_event.header.size);
 	if (ret)
 		return;
@@ -9210,6 +9214,7 @@ struct perf_task_event {
 		u32				ptid;
 		u64				time;
 	} event_id;
+	int new;
 };
 
 static int perf_event_task_match(struct perf_event *event)
@@ -9226,17 +9231,21 @@ static void perf_event_task_output(struct perf_event *event,
 	struct perf_output_handle handle;
 	struct perf_sample_data	sample;
 	struct task_struct *task = task_event->task;
-	int ret, size = task_event->event_id.header.size;
+	int ret;
 
 	if (!perf_event_task_match(event))
 		return;
 
-	perf_event_header__init_id(&task_event->event_id.header, &sample, event);
+	perf_event_header__init_header_and_id(&task_event->event_id.header, &sample,
+					      task_event->new ? PERF_RECORD_FORK : PERF_RECORD_EXIT,
+					      /*misc=*/0,
+					      sizeof(task_event->event_id),
+					      event);
 
 	ret = perf_output_begin(&handle, &sample, event,
 				task_event->event_id.header.size);
 	if (ret)
-		goto out;
+		return;
 
 	task_event->event_id.pid = perf_event_pid(event, task);
 	task_event->event_id.tid = perf_event_tid(event, task);
@@ -9258,8 +9267,6 @@ static void perf_event_task_output(struct perf_event *event,
 	perf_event__output_id_sample(event, &handle, &sample);
 
 	perf_output_end(&handle);
-out:
-	task_event->event_id.header.size = size;
 }
 
 static void perf_event_task(struct task_struct *task,
@@ -9276,18 +9283,7 @@ static void perf_event_task(struct task_struct *task,
 	task_event = (struct perf_task_event){
 		.task	  = task,
 		.task_ctx = task_ctx,
-		.event_id    = {
-			.header = {
-				.type = new ? PERF_RECORD_FORK : PERF_RECORD_EXIT,
-				.misc = 0,
-				.size = sizeof(task_event.event_id),
-			},
-			/* .pid  */
-			/* .ppid */
-			/* .tid  */
-			/* .ptid */
-			/* .time */
-		},
+		.new	  = new,
 	};
 
 	perf_iterate_sb(perf_event_task_output,
@@ -9364,6 +9360,7 @@ struct perf_comm_event {
 		u32				pid;
 		u32				tid;
 	} event_id;
+	bool			exec;
 };
 
 static int perf_event_comm_match(struct perf_event *event)
@@ -9377,18 +9374,21 @@ static void perf_event_comm_output(struct perf_event *event,
 	struct perf_comm_event *comm_event = data;
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
-	int size = comm_event->event_id.header.size;
 	int ret;
 
 	if (!perf_event_comm_match(event))
 		return;
 
-	perf_event_header__init_id(&comm_event->event_id.header, &sample, event);
+	perf_event_header__init_header_and_id(&comm_event->event_id.header, &sample,
+					      PERF_RECORD_COMM,
+					      comm_event->exec ? PERF_RECORD_MISC_COMM_EXEC : 0,
+					      sizeof(comm_event->event_id) + comm_event->comm_size,
+					      event);
 	ret = perf_output_begin(&handle, &sample, event,
 				comm_event->event_id.header.size);
 
 	if (ret)
-		goto out;
+		return;
 
 	comm_event->event_id.pid = perf_event_pid(event, comm_event->task);
 	comm_event->event_id.tid = perf_event_tid(event, comm_event->task);
@@ -9400,8 +9400,6 @@ static void perf_event_comm_output(struct perf_event *event,
 	perf_event__output_id_sample(event, &handle, &sample);
 
 	perf_output_end(&handle);
-out:
-	comm_event->event_id.header.size = size;
 }
 
 static void perf_event_comm_event(struct perf_comm_event *comm_event)
@@ -9416,8 +9414,6 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event)
 	comm_event->comm = comm;
 	comm_event->comm_size = size;
 
-	comm_event->event_id.header.size = sizeof(comm_event->event_id) + size;
-
 	perf_iterate_sb(perf_event_comm_output,
 		       comm_event,
 		       NULL);
@@ -9434,15 +9430,8 @@ void perf_event_comm(struct task_struct *task, bool exec)
 		.task	= task,
 		/* .comm      */
 		/* .comm_size */
-		.event_id  = {
-			.header = {
-				.type = PERF_RECORD_COMM,
-				.misc = exec ? PERF_RECORD_MISC_COMM_EXEC : 0,
-				/* .size */
-			},
-			/* .pid */
-			/* .tid */
-		},
+		/* .event_id */
+		.exec	= exec,
 	};
 
 	perf_event_comm_event(&comm_event);
@@ -9476,18 +9465,20 @@ static void perf_event_namespaces_output(struct perf_event *event,
 	struct perf_namespaces_event *namespaces_event = data;
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
-	u16 header_size = namespaces_event->event_id.header.size;
 	int ret;
 
 	if (!perf_event_namespaces_match(event))
 		return;
 
-	perf_event_header__init_id(&namespaces_event->event_id.header,
-				   &sample, event);
+	perf_event_header__init_header_and_id(&namespaces_event->event_id.header, &sample,
+					      PERF_RECORD_NAMESPACES,
+					      /*misc=*/0,
+					      sizeof(namespaces_event->event_id),
+					      event);
 	ret = perf_output_begin(&handle, &sample, event,
 				namespaces_event->event_id.header.size);
 	if (ret)
-		goto out;
+		return;
 
 	namespaces_event->event_id.pid = perf_event_pid(event,
 							namespaces_event->task);
@@ -9499,8 +9490,6 @@ static void perf_event_namespaces_output(struct perf_event *event,
 	perf_event__output_id_sample(event, &handle, &sample);
 
 	perf_output_end(&handle);
-out:
-	namespaces_event->event_id.header.size = header_size;
 }
 
 static void perf_fill_ns_link_info(struct perf_ns_link_info *ns_link_info,
@@ -9531,11 +9520,7 @@ void perf_event_namespaces(struct task_struct *task)
 	namespaces_event = (struct perf_namespaces_event){
 		.task	= task,
 		.event_id  = {
-			.header = {
-				.type = PERF_RECORD_NAMESPACES,
-				.misc = 0,
-				.size = sizeof(namespaces_event.event_id),
-			},
+			/* .header */
 			/* .pid */
 			/* .tid */
 			.nr_namespaces = NR_NAMESPACES,
@@ -9603,18 +9588,19 @@ static void perf_event_cgroup_output(struct perf_event *event, void *data)
 	struct perf_cgroup_event *cgroup_event = data;
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
-	u16 header_size = cgroup_event->event_id.header.size;
 	int ret;
+	__u16 size = sizeof(cgroup_event->event_id) + cgroup_event->path_size;
 
 	if (!perf_event_cgroup_match(event))
 		return;
 
-	perf_event_header__init_id(&cgroup_event->event_id.header,
-				   &sample, event);
+	perf_event_header__init_header_and_id(&cgroup_event->event_id.header, &sample,
+					      PERF_RECORD_CGROUP, /*misc=*/0, size,
+					      event);
 	ret = perf_output_begin(&handle, &sample, event,
 				cgroup_event->event_id.header.size);
 	if (ret)
-		goto out;
+		return;
 
 	perf_output_put(&handle, cgroup_event->event_id);
 	__output_copy(&handle, cgroup_event->path, cgroup_event->path_size);
@@ -9622,8 +9608,6 @@ static void perf_event_cgroup_output(struct perf_event *event, void *data)
 	perf_event__output_id_sample(event, &handle, &sample);
 
 	perf_output_end(&handle);
-out:
-	cgroup_event->event_id.header.size = header_size;
 }
 
 static void perf_event_cgroup(struct cgroup *cgrp)
@@ -9638,11 +9622,6 @@ static void perf_event_cgroup(struct cgroup *cgrp)
 
 	cgroup_event = (struct perf_cgroup_event){
 		.event_id  = {
-			.header = {
-				.type = PERF_RECORD_CGROUP,
-				.misc = 0,
-				.size = sizeof(cgroup_event.event_id),
-			},
 			.id = cgroup_id(cgrp),
 		},
 	};
@@ -9665,7 +9644,6 @@ static void perf_event_cgroup(struct cgroup *cgrp)
 	while (!IS_ALIGNED(size, sizeof(u64)))
 		cgroup_event.path[size++] = '\0';
 
-	cgroup_event.event_id.header.size += size;
 	cgroup_event.path_size = size;
 
 	perf_iterate_sb(perf_event_cgroup_output,
@@ -9721,8 +9699,9 @@ static void perf_event_mmap_output(struct perf_event *event,
 	struct perf_mmap_event *mmap_event = data;
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
-	int size = mmap_event->event_id.header.size;
-	u32 type = mmap_event->event_id.header.type;
+	int size = sizeof(mmap_event->event_id) + mmap_event->file_size;
+	u32 type = PERF_RECORD_MMAP;
+	u16 misc = PERF_RECORD_MISC_USER;
 	bool use_build_id;
 	int ret;
 
@@ -9730,20 +9709,23 @@ static void perf_event_mmap_output(struct perf_event *event,
 		return;
 
 	if (event->attr.mmap2) {
-		mmap_event->event_id.header.type = PERF_RECORD_MMAP2;
-		mmap_event->event_id.header.size += sizeof(mmap_event->maj);
-		mmap_event->event_id.header.size += sizeof(mmap_event->min);
-		mmap_event->event_id.header.size += sizeof(mmap_event->ino);
-		mmap_event->event_id.header.size += sizeof(mmap_event->ino_generation);
-		mmap_event->event_id.header.size += sizeof(mmap_event->prot);
-		mmap_event->event_id.header.size += sizeof(mmap_event->flags);
-	}
-
-	perf_event_header__init_id(&mmap_event->event_id.header, &sample, event);
+		type = PERF_RECORD_MMAP2;
+		size += sizeof(mmap_event->maj);
+		size += sizeof(mmap_event->min);
+		size += sizeof(mmap_event->ino);
+		size += sizeof(mmap_event->ino_generation);
+		size += sizeof(mmap_event->prot);
+		size += sizeof(mmap_event->flags);
+	}
+	if (!(mmap_event->vma->vm_flags & VM_EXEC))
+		misc |= PERF_RECORD_MISC_MMAP_DATA;
+
+	perf_event_header__init_header_and_id(&mmap_event->event_id.header, &sample,
+					      type, misc, size, event);
 	ret = perf_output_begin(&handle, &sample, event,
 				mmap_event->event_id.header.size);
 	if (ret)
-		goto out;
+		return;
 
 	mmap_event->event_id.pid = perf_event_pid(event, current);
 	mmap_event->event_id.tid = perf_event_tid(event, current);
@@ -9777,9 +9759,6 @@ static void perf_event_mmap_output(struct perf_event *event,
 	perf_event__output_id_sample(event, &handle, &sample);
 
 	perf_output_end(&handle);
-out:
-	mmap_event->event_id.header.size = size;
-	mmap_event->event_id.header.type = type;
 }
 
 static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
@@ -9875,11 +9854,6 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
 	mmap_event->prot = prot;
 	mmap_event->flags = flags;
 
-	if (!(vma->vm_flags & VM_EXEC))
-		mmap_event->event_id.header.misc |= PERF_RECORD_MISC_MMAP_DATA;
-
-	mmap_event->event_id.header.size = sizeof(mmap_event->event_id) + size;
-
 	if (atomic_read(&nr_build_id_events))
 		build_id_parse_nofault(vma, mmap_event->build_id, &mmap_event->build_id_size);
 
@@ -9999,11 +9973,7 @@ void perf_event_mmap(struct vm_area_struct *vma)
 		/* .file_name */
 		/* .file_size */
 		.event_id  = {
-			.header = {
-				.type = PERF_RECORD_MMAP,
-				.misc = PERF_RECORD_MISC_USER,
-				/* .size */
-			},
+			/* .header */
 			/* .pid */
 			/* .tid */
 			.start  = vma->vm_start,
@@ -10033,18 +10003,15 @@ void perf_event_aux_event(struct perf_event *event, unsigned long head,
 		u64				size;
 		u64				flags;
 	} rec = {
-		.header = {
-			.type = PERF_RECORD_AUX,
-			.misc = 0,
-			.size = sizeof(rec),
-		},
 		.offset		= head,
 		.size		= size,
 		.flags		= flags,
 	};
 	int ret;
 
-	perf_event_header__init_id(&rec.header, &sample, event);
+	perf_event_header__init_header_and_id(&rec.header, &sample,
+					      PERF_RECORD_AUX, /*misc=*/0, sizeof(rec),
+					      event);
 	ret = perf_output_begin(&handle, &sample, event, rec.header.size);
 
 	if (ret)
@@ -10069,15 +10036,14 @@ void perf_log_lost_samples(struct perf_event *event, u64 lost)
 		struct perf_event_header	header;
 		u64				lost;
 	} lost_samples_event = {
-		.header = {
-			.type = PERF_RECORD_LOST_SAMPLES,
-			.misc = 0,
-			.size = sizeof(lost_samples_event),
-		},
 		.lost		= lost,
 	};
 
-	perf_event_header__init_id(&lost_samples_event.header, &sample, event);
+	perf_event_header__init_header_and_id(&lost_samples_event.header, &sample,
+					      PERF_RECORD_LOST_SAMPLES,
+					      /*misc=*/0,
+					      sizeof(lost_samples_event),
+					      event);
 
 	ret = perf_output_begin(&handle, &sample, event,
 				lost_samples_event.header.size);
@@ -10102,6 +10068,8 @@ struct perf_switch_event {
 		u32				next_prev_pid;
 		u32				next_prev_tid;
 	} event_id;
+	bool			sched_in;
+	bool			preempt;
 };
 
 static int perf_event_switch_match(struct perf_event *event)
@@ -10114,6 +10082,9 @@ static void perf_event_switch_output(struct perf_event *event, void *data)
 	struct perf_switch_event *se = data;
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
+	__u32 type;
+	__u16 misc;
+	__u16 size;
 	int ret;
 
 	if (!perf_event_switch_match(event))
@@ -10121,18 +10092,22 @@ static void perf_event_switch_output(struct perf_event *event, void *data)
 
 	/* Only CPU-wide events are allowed to see next/prev pid/tid */
 	if (event->ctx->task) {
-		se->event_id.header.type = PERF_RECORD_SWITCH;
-		se->event_id.header.size = sizeof(se->event_id.header);
+		type = PERF_RECORD_SWITCH;
+		size = sizeof(se->event_id.header);
 	} else {
-		se->event_id.header.type = PERF_RECORD_SWITCH_CPU_WIDE;
-		se->event_id.header.size = sizeof(se->event_id);
+		type = PERF_RECORD_SWITCH_CPU_WIDE;
+		size = sizeof(se->event_id);
 		se->event_id.next_prev_pid =
 					perf_event_pid(event, se->next_prev);
 		se->event_id.next_prev_tid =
 					perf_event_tid(event, se->next_prev);
 	}
+	misc = se->sched_in ? 0 : PERF_RECORD_MISC_SWITCH_OUT;
+	if (se->preempt)
+		misc |= PERF_RECORD_MISC_SWITCH_OUT_PREEMPT;
 
-	perf_event_header__init_id(&se->event_id.header, &sample, event);
+	perf_event_header__init_header_and_id(&se->event_id.header, &sample,
+					      type, misc, size, event);
 
 	ret = perf_output_begin(&handle, &sample, event, se->event_id.header.size);
 	if (ret)
@@ -10158,15 +10133,9 @@ static void perf_event_switch(struct task_struct *task,
 	switch_event = (struct perf_switch_event){
 		.task		= task,
 		.next_prev	= next_prev,
-		.event_id	= {
-			.header = {
-				/* .type */
-				.misc = sched_in ? 0 : PERF_RECORD_MISC_SWITCH_OUT,
-				/* .size */
-			},
-			/* .next_prev_pid */
-			/* .next_prev_tid */
-		},
+		/* .event_id */
+		.sched_in	= sched_in,
+		.preempt	= !sched_in && task_is_runnable(task),
 	};
 
 	if (!sched_in && task_is_runnable(task)) {
@@ -10193,11 +10162,6 @@ static void perf_log_throttle(struct perf_event *event, int enable)
 		u64				id;
 		u64				stream_id;
 	} throttle_event = {
-		.header = {
-			.type = PERF_RECORD_THROTTLE,
-			.misc = 0,
-			.size = sizeof(throttle_event),
-		},
 		.time		= perf_event_clock(event),
 		.id		= primary_event_id(event),
 		.stream_id	= event->id,
@@ -10206,7 +10170,11 @@ static void perf_log_throttle(struct perf_event *event, int enable)
 	if (enable)
 		throttle_event.header.type = PERF_RECORD_UNTHROTTLE;
 
-	perf_event_header__init_id(&throttle_event.header, &sample, event);
+	perf_event_header__init_header_and_id(&throttle_event.header, &sample,
+					      PERF_RECORD_THROTTLE,
+					      /*misc=*/0,
+					      sizeof(throttle_event),
+					      event);
 
 	ret = perf_output_begin(&handle, &sample, event,
 				throttle_event.header.size);
@@ -10245,12 +10213,14 @@ static void perf_event_ksymbol_output(struct perf_event *event, void *data)
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
 	int ret;
+	__u16 size = sizeof(ksymbol_event->event_id) + ksymbol_event->name_len;
 
 	if (!perf_event_ksymbol_match(event))
 		return;
 
-	perf_event_header__init_id(&ksymbol_event->event_id.header,
-				   &sample, event);
+	perf_event_header__init_header_and_id(&ksymbol_event->event_id.header, &sample,
+					      PERF_RECORD_KSYMBOL, /*misc=*/0, size,
+					      event);
 	ret = perf_output_begin(&handle, &sample, event,
 				ksymbol_event->event_id.header.size);
 	if (ret)
@@ -10291,11 +10261,6 @@ void perf_event_ksymbol(u16 ksym_type, u64 addr, u32 len, bool unregister,
 		.name = name,
 		.name_len = name_len,
 		.event_id = {
-			.header = {
-				.type = PERF_RECORD_KSYMBOL,
-				.size = sizeof(ksymbol_event.event_id) +
-					name_len,
-			},
 			.addr = addr,
 			.len = len,
 			.ksym_type = ksym_type,
@@ -10339,8 +10304,11 @@ static void perf_event_bpf_output(struct perf_event *event, void *data)
 	if (!perf_event_bpf_match(event))
 		return;
 
-	perf_event_header__init_id(&bpf_event->event_id.header,
-				   &sample, event);
+	perf_event_header__init_header_and_id(&bpf_event->event_id.header, &sample,
+					      PERF_RECORD_BPF_EVENT,
+					      /*misc=*/0,
+					      sizeof(bpf_event->event_id),
+					      event);
 	ret = perf_output_begin(&handle, &sample, event,
 				bpf_event->event_id.header.size);
 	if (ret)
@@ -10396,10 +10364,6 @@ void perf_event_bpf_event(struct bpf_prog *prog,
 	bpf_event = (struct perf_bpf_event){
 		.prog = prog,
 		.event_id = {
-			.header = {
-				.type = PERF_RECORD_BPF_EVENT,
-				.size = sizeof(bpf_event.event_id),
-			},
 			.type = type,
 			.flags = flags,
 			.id = prog->aux->id,
@@ -10427,18 +10391,23 @@ static void perf_callchain_deferred_output(struct perf_event *event, void *data)
 	struct perf_callchain_deferred_event *deferred_event = data;
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
-	int ret, size = deferred_event->event.header.size;
+	int ret;
+	__u16 size = sizeof(deferred_event->event) + (deferred_event->trace->nr * sizeof(u64));
 
 	if (!event->attr.defer_output)
 		return;
 
 	/* XXX do we really need sample_id_all for this ??? */
-	perf_event_header__init_id(&deferred_event->event.header, &sample, event);
+	perf_event_header__init_header_and_id(&deferred_event->event.header, &sample,
+					      PERF_RECORD_CALLCHAIN_DEFERRED,
+					      PERF_RECORD_MISC_USER,
+					      size,
+					      event);
 
 	ret = perf_output_begin(&handle, &sample, event,
 				deferred_event->event.header.size);
 	if (ret)
-		goto out;
+		return;
 
 	perf_output_put(&handle, deferred_event->event);
 	for (int i = 0; i < deferred_event->trace->nr; i++) {
@@ -10448,8 +10417,6 @@ static void perf_callchain_deferred_output(struct perf_event *event, void *data)
 	perf_event__output_id_sample(event, &handle, &sample);
 
 	perf_output_end(&handle);
-out:
-	deferred_event->event.header.size = size;
 }
 
 static void perf_unwind_deferred_callback(struct unwind_work *work,
@@ -10458,12 +10425,6 @@ static void perf_unwind_deferred_callback(struct unwind_work *work,
 	struct perf_callchain_deferred_event deferred_event = {
 		.trace = trace,
 		.event = {
-			.header = {
-				.type = PERF_RECORD_CALLCHAIN_DEFERRED,
-				.misc = PERF_RECORD_MISC_USER,
-				.size = sizeof(deferred_event.event) +
-					(trace->nr * sizeof(u64)),
-			},
 			.cookie = cookie,
 			.nr = trace->nr,
 		},
@@ -10475,7 +10436,8 @@ static void perf_unwind_deferred_callback(struct unwind_work *work,
 struct perf_text_poke_event {
 	const void		*old_bytes;
 	const void		*new_bytes;
-	size_t			pad;
+	u16			tot;
+	u16			pad;
 	u16			old_len;
 	u16			new_len;
 
@@ -10496,13 +10458,18 @@ static void perf_event_text_poke_output(struct perf_event *event, void *data)
 	struct perf_text_poke_event *text_poke_event = data;
 	struct perf_output_handle handle;
 	struct perf_sample_data sample;
+	__u16 size = sizeof(text_poke_event->event_id) + ALIGN(text_poke_event->tot, sizeof(u64));
 	u64 padding = 0;
 	int ret;
 
 	if (!perf_event_text_poke_match(event))
 		return;
 
-	perf_event_header__init_id(&text_poke_event->event_id.header, &sample, event);
+	perf_event_header__init_header_and_id(&text_poke_event->event_id.header, &sample,
+					      PERF_RECORD_TEXT_POKE,
+					      PERF_RECORD_MISC_KERNEL,
+					      size,
+					      event);
 
 	ret = perf_output_begin(&handle, &sample, event,
 				text_poke_event->event_id.header.size);
@@ -10535,20 +10502,15 @@ void perf_event_text_poke(const void *addr, const void *old_bytes,
 
 	tot  = sizeof(text_poke_event.old_len) + old_len;
 	tot += sizeof(text_poke_event.new_len) + new_len;
-	pad  = ALIGN(tot, sizeof(u64)) - tot;
 
 	text_poke_event = (struct perf_text_poke_event){
 		.old_bytes    = old_bytes,
 		.new_bytes    = new_bytes,
+		.tot          = tot,
 		.pad          = pad,
 		.old_len      = old_len,
 		.new_len      = new_len,
 		.event_id  = {
-			.header = {
-				.type = PERF_RECORD_TEXT_POKE,
-				.misc = PERF_RECORD_MISC_KERNEL,
-				.size = sizeof(text_poke_event.event_id) + tot + pad,
-			},
 			.addr = (unsigned long)addr,
 		},
 	};
@@ -10579,13 +10541,12 @@ static void perf_log_itrace_start(struct perf_event *event)
 	    event->attach_state & PERF_ATTACH_ITRACE)
 		return;
 
-	rec.header.type	= PERF_RECORD_ITRACE_START;
-	rec.header.misc	= 0;
-	rec.header.size	= sizeof(rec);
 	rec.pid	= perf_event_pid(event, current);
 	rec.tid	= perf_event_tid(event, current);
 
-	perf_event_header__init_id(&rec.header, &sample, event);
+	perf_event_header__init_header_and_id(&rec.header, &sample,
+					      PERF_RECORD_ITRACE_START, /*misc=*/0, sizeof(rec),
+					      event);
 	ret = perf_output_begin(&handle, &sample, event, rec.header.size);
 
 	if (ret)
@@ -10610,12 +10571,10 @@ void perf_report_aux_output_id(struct perf_event *event, u64 hw_id)
 	if (event->parent)
 		event = event->parent;
 
-	rec.header.type	= PERF_RECORD_AUX_OUTPUT_HW_ID;
-	rec.header.misc	= 0;
-	rec.header.size	= sizeof(rec);
-	rec.hw_id	= hw_id;
-
-	perf_event_header__init_id(&rec.header, &sample, event);
+	rec.hw_id = hw_id;
+	perf_event_header__init_header_and_id(&rec.header, &sample,
+					      PERF_RECORD_AUX_OUTPUT_HW_ID, /*misc=*/0, sizeof(rec),
+					      event);
 	ret = perf_output_begin(&handle, &sample, event, rec.header.size);
 
 	if (ret)