[for-next][PATCH 00/10] ring-buffer: Updates for v6.15

[for-next][PATCH 00/10] ring-buffer: Updates for v6.15

[Steven Rostedt]

  git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace.git
ring-buffer/for-next

Head SHA1: c6a1b7c2e58d048ed8b7b2f331fef42b52376336


Feng Yang (1):
      ring-buffer: Fix bytes_dropped calculation issue

Steven Rostedt (9):
      ring-buffer: Allow persistent ring buffers to be mmapped
      ring-buffer: Use kaslr address instead of text delta
      ring-buffer: Add buffer meta data for persistent ring buffer
      ring-buffer: Add ring_buffer_meta_scratch()
      tracing: Have persistent trace instances save KASLR offset
      module: Add module_for_each_mod() function
      tracing: Have persistent trace instances save module addresses
      tracing: Show module names and addresses of last boot
      tracing: Update modules to persistent instances when loaded

----
 include/linux/module.h      |   6 +
 include/linux/ring_buffer.h |   8 +-
 kernel/module/main.c        |  13 ++
 kernel/trace/ring_buffer.c  | 299 ++++++++++++++++++++++++++++----------------
 kernel/trace/trace.c        | 268 +++++++++++++++++++++++++++++++++++----
 kernel/trace/trace.h        |  15 ++-
 kernel/trace/trace_events.c |  40 ++++--
 7 files changed, 493 insertions(+), 156 deletions(-)

[for-next][PATCH 01/10] ring-buffer: Allow persistent ring buffers to be mmapped

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

The persistent ring buffer uses vmap()'d memory to map the reserved memory
from boot. But the user space mmap() to the ring buffer requires
virt_to_page() to return a valid page. But that only works for core kernel
addresses and not for vmap() addresses.

If virt_addr_valid() fails on the page to be mapped, use vmalloc_to_page()
instead.

Also, the persistent memory uses the page->id for its own purpose where as
the user mmap buffer currently uses that for the subbuf array mapped to
user space. If the buffer is a persistent buffer, use the page index into
that buffer as the identifier instead of the page->id.

That is, the page->id for a persistent buffer, represents the order of the
buffer is in the link list. ->id == 0 means it is the reader page.
When a reader page is swapped, the new reader page's ->id gets zero, and
the old reader page gets the ->id of the page that it swapped with.

The user space mapping has the ->id is the index of where it was mapped in
user space and does not change while it is mapped.

Since the persistent buffer is fixed in its location, the index of where
a page is in the memory range can be used as the "id" to put in the meta
page array, and it can be mapped in the same order to user space as it is
in the persistent memory.

Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Vincent Donnefort <vdonnefort@google.com>
Link: https://lore.kernel.org/20250214130442.3f4fee5d@gandalf.local.home
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Tested-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
---
 kernel/trace/ring_buffer.c | 49 ++++++++++++++++++++++++++++++++------
 kernel/trace/trace.c       |  4 ----
 2 files changed, 42 insertions(+), 11 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index bb6089c2951e..87caf9d48edb 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -5950,12 +5950,18 @@ static void rb_clear_buffer_page(struct buffer_page *page)
 static void rb_update_meta_page(struct ring_buffer_per_cpu *cpu_buffer)
 {
 	struct trace_buffer_meta *meta = cpu_buffer->meta_page;
+	struct page *page;
 
 	if (!meta)
 		return;
 
 	meta->reader.read = cpu_buffer->reader_page->read;
-	meta->reader.id = cpu_buffer->reader_page->id;
+	/* For boot buffers, the id is the index */
+	if (cpu_buffer->ring_meta)
+		meta->reader.id = rb_meta_subbuf_idx(cpu_buffer->ring_meta,
+						     cpu_buffer->reader_page->page);
+	else
+		meta->reader.id = cpu_buffer->reader_page->id;
 	meta->reader.lost_events = cpu_buffer->lost_events;
 
 	meta->entries = local_read(&cpu_buffer->entries);
@@ -5963,7 +5969,12 @@ static void rb_update_meta_page(struct ring_buffer_per_cpu *cpu_buffer)
 	meta->read = cpu_buffer->read;
 
 	/* Some archs do not have data cache coherency between kernel and user-space */
-	flush_dcache_folio(virt_to_folio(cpu_buffer->meta_page));
+	if (virt_addr_valid(cpu_buffer->meta_page))
+		page = virt_to_page(cpu_buffer->meta_page);
+	else
+		page = vmalloc_to_page(cpu_buffer->meta_page);
+
+	flush_dcache_folio(page_folio(page));
 }
 
 static void
@@ -6883,23 +6894,38 @@ static void rb_setup_ids_meta_page(struct ring_buffer_per_cpu *cpu_buffer,
 	struct trace_buffer_meta *meta = cpu_buffer->meta_page;
 	unsigned int nr_subbufs = cpu_buffer->nr_pages + 1;
 	struct buffer_page *first_subbuf, *subbuf;
+	int cnt = 0;
 	int id = 0;
 
-	subbuf_ids[id] = (unsigned long)cpu_buffer->reader_page->page;
-	cpu_buffer->reader_page->id = id++;
+	if (cpu_buffer->ring_meta)
+		id = rb_meta_subbuf_idx(cpu_buffer->ring_meta,
+					cpu_buffer->reader_page->page);
+	else
+		cpu_buffer->reader_page->id = id;
+
+	subbuf_ids[id++] = (unsigned long)cpu_buffer->reader_page->page;
+	cnt++;
 
 	first_subbuf = subbuf = rb_set_head_page(cpu_buffer);
 	do {
+		if (cpu_buffer->ring_meta)
+			id = rb_meta_subbuf_idx(cpu_buffer->ring_meta,
+						subbuf->page);
+		else
+			subbuf->id = id;
+
 		if (WARN_ON(id >= nr_subbufs))
 			break;
 
 		subbuf_ids[id] = (unsigned long)subbuf->page;
-		subbuf->id = id;
 
 		rb_inc_page(&subbuf);
 		id++;
+		cnt++;
 	} while (subbuf != first_subbuf);
 
+	WARN_ON(cnt != nr_subbufs);
+
 	/* install subbuf ID to kern VA translation */
 	cpu_buffer->subbuf_ids = subbuf_ids;
 
@@ -7064,7 +7090,10 @@ static int __rb_map_vma(struct ring_buffer_per_cpu *cpu_buffer,
 			goto out;
 		}
 
-		page = virt_to_page((void *)cpu_buffer->subbuf_ids[s]);
+		if (virt_addr_valid(cpu_buffer->subbuf_ids[s]))
+			page = virt_to_page((void *)cpu_buffer->subbuf_ids[s]);
+		else
+			page = vmalloc_to_page((void *)cpu_buffer->subbuf_ids[s]);
 
 		for (; off < (1 << (subbuf_order)); off++, page++) {
 			if (p >= nr_pages)
@@ -7210,6 +7239,7 @@ int ring_buffer_map_get_reader(struct trace_buffer *buffer, int cpu)
 	unsigned long missed_events;
 	unsigned long reader_size;
 	unsigned long flags;
+	struct page *page;
 
 	cpu_buffer = rb_get_mapped_buffer(buffer, cpu);
 	if (IS_ERR(cpu_buffer))
@@ -7278,7 +7308,12 @@ int ring_buffer_map_get_reader(struct trace_buffer *buffer, int cpu)
 
 out:
 	/* Some archs do not have data cache coherency between kernel and user-space */
-	flush_dcache_folio(virt_to_folio(cpu_buffer->reader_page->page));
+	if (virt_addr_valid(cpu_buffer->meta_page))
+		page = virt_to_page(cpu_buffer->meta_page);
+	else
+		page = vmalloc_to_page(cpu_buffer->meta_page);
+
+	flush_dcache_folio(page_folio(page));
 
 	rb_update_meta_page(cpu_buffer);
 
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 0e6d517e74e0..25ff37aab00f 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -8279,10 +8279,6 @@ static int tracing_buffers_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct trace_iterator *iter = &info->iter;
 	int ret = 0;
 
-	/* Currently the boot mapped buffer is not supported for mmap */
-	if (iter->tr->flags & TRACE_ARRAY_FL_BOOT)
-		return -ENODEV;
-
 	ret = get_snapshot_map(iter->tr);
 	if (ret)
 		return ret;
-- 
2.47.2

[for-next][PATCH 02/10] ring-buffer: Fix bytes_dropped calculation issue

[Steven Rostedt]

From: Feng Yang <yangfeng@kylinos.cn>

The calculation of bytes-dropped and bytes_dropped_nested is reversed.
Although it does not affect the final calculation of total_dropped,
it should still be modified.

Link: https://lore.kernel.org/20250223070106.6781-1-yangfeng59949@163.com
Fixes: 6c43e554a2a5 ("ring-buffer: Add ring buffer startup selftest")
Signed-off-by: Feng Yang <yangfeng@kylinos.cn>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 kernel/trace/ring_buffer.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 87caf9d48edb..7b6bd1f3daf4 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -7446,9 +7446,9 @@ static __init int rb_write_something(struct rb_test_data *data, bool nested)
 		/* Ignore dropped events before test starts. */
 		if (started) {
 			if (nested)
-				data->bytes_dropped += len;
-			else
 				data->bytes_dropped_nested += len;
+			else
+				data->bytes_dropped += len;
 		}
 		return len;
 	}
-- 
2.47.2

[for-next][PATCH 03/10] ring-buffer: Use kaslr address instead of text delta

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

Instead of saving off the text and data pointers and using them to compare
with the current boot's text and data pointers, just save off the KASLR
offset. Then that can be used to figure out how to read the previous boots
buffer.

The last_boot_info will now show this offset, but only if it is for a
previous boot:

  ~# cat instances/boot_mapped/last_boot_info
  39000000	[kernel]

  ~# echo function > instances/boot_mapped/current_tracer
  ~# cat instances/boot_mapped/last_boot_info
  # Current

If the KASLR offset saved is for the current boot, the last_boot_info will
show the value of "current".

Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Link: https://lore.kernel.org/20250305164608.274956504@goodmis.org
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 include/linux/ring_buffer.h |  3 +--
 kernel/trace/ring_buffer.c  | 31 ++++++++++++-------------------
 kernel/trace/trace.c        | 30 +++++++++++++++++++++---------
 kernel/trace/trace.h        |  9 +++++----
 4 files changed, 39 insertions(+), 34 deletions(-)

diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h
index 17fbb7855295..8de035f4f0d9 100644
--- a/include/linux/ring_buffer.h
+++ b/include/linux/ring_buffer.h
@@ -94,8 +94,7 @@ struct trace_buffer *__ring_buffer_alloc_range(unsigned long size, unsigned flag
 					       unsigned long range_size,
 					       struct lock_class_key *key);
 
-bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, long *text,
-				 long *data);
+bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, unsigned long *kaslr_addr);
 
 /*
  * Because the ring buffer is generic, if other users of the ring buffer get
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 7b6bd1f3daf4..49c3c46fe2f4 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -31,6 +31,7 @@
 
 #include <asm/local64.h>
 #include <asm/local.h>
+#include <asm/setup.h>
 
 #include "trace.h"
 
@@ -49,8 +50,7 @@ static void update_pages_handler(struct work_struct *work);
 struct ring_buffer_meta {
 	int		magic;
 	int		struct_size;
-	unsigned long	text_addr;
-	unsigned long	data_addr;
+	unsigned long	kaslr_addr;
 	unsigned long	first_buffer;
 	unsigned long	head_buffer;
 	unsigned long	commit_buffer;
@@ -550,8 +550,7 @@ struct trace_buffer {
 	unsigned long			range_addr_start;
 	unsigned long			range_addr_end;
 
-	long				last_text_delta;
-	long				last_data_delta;
+	unsigned long			kaslr_addr;
 
 	unsigned int			subbuf_size;
 	unsigned int			subbuf_order;
@@ -1891,16 +1890,13 @@ static void rb_meta_validate_events(struct ring_buffer_per_cpu *cpu_buffer)
 	}
 }
 
-/* Used to calculate data delta */
-static char rb_data_ptr[] = "";
-
-#define THIS_TEXT_PTR		((unsigned long)rb_meta_init_text_addr)
-#define THIS_DATA_PTR		((unsigned long)rb_data_ptr)
-
 static void rb_meta_init_text_addr(struct ring_buffer_meta *meta)
 {
-	meta->text_addr = THIS_TEXT_PTR;
-	meta->data_addr = THIS_DATA_PTR;
+#ifdef CONFIG_RANDOMIZE_BASE
+	meta->kaslr_addr = kaslr_offset();
+#else
+	meta->kaslr_addr = 0;
+#endif
 }
 
 static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages)
@@ -1928,8 +1924,7 @@ static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages)
 			meta->first_buffer += delta;
 			meta->head_buffer += delta;
 			meta->commit_buffer += delta;
-			buffer->last_text_delta = THIS_TEXT_PTR - meta->text_addr;
-			buffer->last_data_delta = THIS_DATA_PTR - meta->data_addr;
+			buffer->kaslr_addr = meta->kaslr_addr;
 			continue;
 		}
 
@@ -2482,17 +2477,15 @@ struct trace_buffer *__ring_buffer_alloc_range(unsigned long size, unsigned flag
  *
  * Returns: The true if the delta is non zero
  */
-bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, long *text,
-				 long *data)
+bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, unsigned long *kaslr_addr)
 {
 	if (!buffer)
 		return false;
 
-	if (!buffer->last_text_delta)
+	if (!buffer->kaslr_addr)
 		return false;
 
-	*text = buffer->last_text_delta;
-	*data = buffer->last_data_delta;
+	*kaslr_addr = buffer->kaslr_addr;
 
 	return true;
 }
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 25ff37aab00f..41d463e5327c 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -50,7 +50,7 @@
 #include <linux/irq_work.h>
 #include <linux/workqueue.h>
 
-#include <asm/setup.h> /* COMMAND_LINE_SIZE */
+#include <asm/setup.h> /* COMMAND_LINE_SIZE and kaslr_offset() */
 
 #include "trace.h"
 #include "trace_output.h"
@@ -4193,7 +4193,7 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter)
 		 * safe to use if the array has delta offsets
 		 * Force printing via the fields.
 		 */
-		if ((tr->text_delta || tr->data_delta) &&
+		if ((tr->text_delta) &&
 		    event->type > __TRACE_LAST_TYPE)
 			return print_event_fields(iter, event);
 
@@ -5990,7 +5990,7 @@ ssize_t tracing_resize_ring_buffer(struct trace_array *tr,
 
 static void update_last_data(struct trace_array *tr)
 {
-	if (!tr->text_delta && !tr->data_delta)
+	if (!(tr->flags & TRACE_ARRAY_FL_LAST_BOOT))
 		return;
 
 	/*
@@ -6003,7 +6003,8 @@ static void update_last_data(struct trace_array *tr)
 
 	/* Using current data now */
 	tr->text_delta = 0;
-	tr->data_delta = 0;
+
+	tr->flags &= ~TRACE_ARRAY_FL_LAST_BOOT;
 }
 
 /**
@@ -6821,8 +6822,17 @@ tracing_last_boot_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t
 
 	seq_buf_init(&seq, buf, 64);
 
-	seq_buf_printf(&seq, "text delta:\t%ld\n", tr->text_delta);
-	seq_buf_printf(&seq, "data delta:\t%ld\n", tr->data_delta);
+	/*
+	 * Do not leak KASLR address. This only shows the KASLR address of
+	 * the last boot. When the ring buffer is started, the LAST_BOOT
+	 * flag gets cleared, and this should only report "current".
+	 * Otherwise it shows the KASLR address from the previous boot which
+	 * should not be the same as the current boot.
+	 */
+	if (tr->flags & TRACE_ARRAY_FL_LAST_BOOT)
+		seq_buf_printf(&seq, "%lx\t[kernel]\n", tr->kaslr_addr);
+	else
+		seq_buf_puts(&seq, "# Current\n");
 
 	return simple_read_from_buffer(ubuf, cnt, ppos, buf, seq_buf_used(&seq));
 }
@@ -9206,8 +9216,10 @@ allocate_trace_buffer(struct trace_array *tr, struct array_buffer *buf, int size
 						      tr->range_addr_start,
 						      tr->range_addr_size);
 
-		ring_buffer_last_boot_delta(buf->buffer,
-					    &tr->text_delta, &tr->data_delta);
+#ifdef CONFIG_RANDOMIZE_BASE
+		if (ring_buffer_last_boot_delta(buf->buffer, &tr->kaslr_addr))
+			tr->text_delta = kaslr_offset() - tr->kaslr_addr;
+#endif
 		/*
 		 * This is basically the same as a mapped buffer,
 		 * with the same restrictions.
@@ -10455,7 +10467,7 @@ __init static void enable_instances(void)
 		 * to it.
 		 */
 		if (start) {
-			tr->flags |= TRACE_ARRAY_FL_BOOT;
+			tr->flags |= TRACE_ARRAY_FL_BOOT | TRACE_ARRAY_FL_LAST_BOOT;
 			tr->ref++;
 		}
 
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 9c21ba45b7af..abe8169c3e87 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -348,8 +348,8 @@ struct trace_array {
 	unsigned int		mapped;
 	unsigned long		range_addr_start;
 	unsigned long		range_addr_size;
+	unsigned long		kaslr_addr;
 	long			text_delta;
-	long			data_delta;
 
 	struct trace_pid_list	__rcu *filtered_pids;
 	struct trace_pid_list	__rcu *filtered_no_pids;
@@ -433,9 +433,10 @@ struct trace_array {
 };
 
 enum {
-	TRACE_ARRAY_FL_GLOBAL	= BIT(0),
-	TRACE_ARRAY_FL_BOOT	= BIT(1),
-	TRACE_ARRAY_FL_MOD_INIT	= BIT(2),
+	TRACE_ARRAY_FL_GLOBAL		= BIT(0),
+	TRACE_ARRAY_FL_BOOT		= BIT(1),
+	TRACE_ARRAY_FL_LAST_BOOT	= BIT(2),
+	TRACE_ARRAY_FL_MOD_INIT		= BIT(3),
 };
 
 #ifdef CONFIG_MODULES
-- 
2.47.2

[for-next][PATCH 04/10] ring-buffer: Add buffer meta data for persistent ring buffer

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

Instead of just having a meta data at the first page of each sub buffer
that has duplicate data, add a new meta page to the entire block of memory
that holds the duplicate data and remove it from the sub buffer meta data.

This will open up the extra memory in this first page to be used by the
tracer for its own persistent data.

Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Link: https://lore.kernel.org/20250305164608.446351513@goodmis.org
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 kernel/trace/ring_buffer.c | 171 ++++++++++++++++++++++++++-----------
 1 file changed, 119 insertions(+), 52 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 49c3c46fe2f4..ce78b57e1564 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -49,7 +49,12 @@ static void update_pages_handler(struct work_struct *work);
 
 struct ring_buffer_meta {
 	int		magic;
-	int		struct_size;
+	int		struct_sizes;
+	unsigned long	total_size;
+	unsigned long	buffers_offset;
+};
+
+struct ring_buffer_cpu_meta {
 	unsigned long	kaslr_addr;
 	unsigned long	first_buffer;
 	unsigned long	head_buffer;
@@ -517,7 +522,7 @@ struct ring_buffer_per_cpu {
 	struct mutex			mapping_lock;
 	unsigned long			*subbuf_ids;	/* ID to subbuf VA */
 	struct trace_buffer_meta	*meta_page;
-	struct ring_buffer_meta		*ring_meta;
+	struct ring_buffer_cpu_meta	*ring_meta;
 
 	/* ring buffer pages to update, > 0 to add, < 0 to remove */
 	long				nr_pages_to_update;
@@ -550,6 +555,8 @@ struct trace_buffer {
 	unsigned long			range_addr_start;
 	unsigned long			range_addr_end;
 
+	struct ring_buffer_meta		*meta;
+
 	unsigned long			kaslr_addr;
 
 	unsigned int			subbuf_size;
@@ -1270,7 +1277,7 @@ static void rb_head_page_activate(struct ring_buffer_per_cpu *cpu_buffer)
 	rb_set_list_to_head(head->list.prev);
 
 	if (cpu_buffer->ring_meta) {
-		struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+		struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 		meta->head_buffer = (unsigned long)head->page;
 	}
 }
@@ -1568,7 +1575,7 @@ static void rb_check_pages(struct ring_buffer_per_cpu *cpu_buffer)
 static unsigned long
 rb_range_align_subbuf(unsigned long addr, int subbuf_size, int nr_subbufs)
 {
-	addr += sizeof(struct ring_buffer_meta) +
+	addr += sizeof(struct ring_buffer_cpu_meta) +
 		sizeof(int) * nr_subbufs;
 	return ALIGN(addr, subbuf_size);
 }
@@ -1579,19 +1586,22 @@ rb_range_align_subbuf(unsigned long addr, int subbuf_size, int nr_subbufs)
 static void *rb_range_meta(struct trace_buffer *buffer, int nr_pages, int cpu)
 {
 	int subbuf_size = buffer->subbuf_size + BUF_PAGE_HDR_SIZE;
-	unsigned long ptr = buffer->range_addr_start;
-	struct ring_buffer_meta *meta;
+	struct ring_buffer_cpu_meta *meta;
+	struct ring_buffer_meta *bmeta;
+	unsigned long ptr;
 	int nr_subbufs;
 
-	if (!ptr)
+	bmeta = buffer->meta;
+	if (!bmeta)
 		return NULL;
 
+	ptr = (unsigned long)bmeta + bmeta->buffers_offset;
+	meta = (struct ring_buffer_cpu_meta *)ptr;
+
 	/* When nr_pages passed in is zero, the first meta has already been initialized */
 	if (!nr_pages) {
-		meta = (struct ring_buffer_meta *)ptr;
 		nr_subbufs = meta->nr_subbufs;
 	} else {
-		meta = NULL;
 		/* Include the reader page */
 		nr_subbufs = nr_pages + 1;
 	}
@@ -1623,7 +1633,7 @@ static void *rb_range_meta(struct trace_buffer *buffer, int nr_pages, int cpu)
 }
 
 /* Return the start of subbufs given the meta pointer */
-static void *rb_subbufs_from_meta(struct ring_buffer_meta *meta)
+static void *rb_subbufs_from_meta(struct ring_buffer_cpu_meta *meta)
 {
 	int subbuf_size = meta->subbuf_size;
 	unsigned long ptr;
@@ -1639,7 +1649,7 @@ static void *rb_subbufs_from_meta(struct ring_buffer_meta *meta)
  */
 static void *rb_range_buffer(struct ring_buffer_per_cpu *cpu_buffer, int idx)
 {
-	struct ring_buffer_meta *meta;
+	struct ring_buffer_cpu_meta *meta;
 	unsigned long ptr;
 	int subbuf_size;
 
@@ -1664,15 +1674,74 @@ static void *rb_range_buffer(struct ring_buffer_per_cpu *cpu_buffer, int idx)
 	return (void *)ptr;
 }
 
+/*
+ * See if the existing memory contains a valid meta section.
+ * if so, use that, otherwise initialize it.
+ */
+static bool rb_meta_init(struct trace_buffer *buffer)
+{
+	unsigned long ptr = buffer->range_addr_start;
+	struct ring_buffer_meta *bmeta;
+	unsigned long total_size;
+	int struct_sizes;
+
+	bmeta = (struct ring_buffer_meta *)ptr;
+	buffer->meta = bmeta;
+
+	total_size = buffer->range_addr_end - buffer->range_addr_start;
+
+	struct_sizes = sizeof(struct ring_buffer_cpu_meta);
+	struct_sizes |= sizeof(*bmeta) << 16;
+
+	/* The first buffer will start word size after the meta page */
+	ptr += sizeof(*bmeta);
+	ptr = ALIGN(ptr, sizeof(long));
+
+	if (bmeta->magic != RING_BUFFER_META_MAGIC) {
+		pr_info("Ring buffer boot meta mismatch of magic\n");
+		goto init;
+	}
+
+	if (bmeta->struct_sizes != struct_sizes) {
+		pr_info("Ring buffer boot meta mismatch of struct size\n");
+		goto init;
+	}
+
+	if (bmeta->total_size != total_size) {
+		pr_info("Ring buffer boot meta mismatch of total size\n");
+		goto init;
+	}
+
+	if (bmeta->buffers_offset > bmeta->total_size) {
+		pr_info("Ring buffer boot meta mismatch of offset outside of total size\n");
+		goto init;
+	}
+
+	if (bmeta->buffers_offset != (void *)ptr - (void *)bmeta) {
+		pr_info("Ring buffer boot meta mismatch of first buffer offset\n");
+		goto init;
+	}
+
+	return true;
+
+ init:
+	bmeta->magic = RING_BUFFER_META_MAGIC;
+	bmeta->struct_sizes = struct_sizes;
+	bmeta->total_size = total_size;
+	bmeta->buffers_offset = (void *)ptr - (void *)bmeta;
+
+	return false;
+}
+
 /*
  * See if the existing memory contains valid ring buffer data.
  * As the previous kernel must be the same as this kernel, all
  * the calculations (size of buffers and number of buffers)
  * must be the same.
  */
-static bool rb_meta_valid(struct ring_buffer_meta *meta, int cpu,
-			  struct trace_buffer *buffer, int nr_pages,
-			  unsigned long *subbuf_mask)
+static bool rb_cpu_meta_valid(struct ring_buffer_cpu_meta *meta, int cpu,
+			      struct trace_buffer *buffer, int nr_pages,
+			      unsigned long *subbuf_mask)
 {
 	int subbuf_size = PAGE_SIZE;
 	struct buffer_data_page *subbuf;
@@ -1683,20 +1752,6 @@ static bool rb_meta_valid(struct ring_buffer_meta *meta, int cpu,
 	if (!subbuf_mask)
 		return false;
 
-	/* Check the meta magic and meta struct size */
-	if (meta->magic != RING_BUFFER_META_MAGIC ||
-	    meta->struct_size != sizeof(*meta)) {
-		pr_info("Ring buffer boot meta[%d] mismatch of magic or struct size\n", cpu);
-		return false;
-	}
-
-	/* The subbuffer's size and number of subbuffers must match */
-	if (meta->subbuf_size != subbuf_size ||
-	    meta->nr_subbufs != nr_pages + 1) {
-		pr_info("Ring buffer boot meta [%d] mismatch of subbuf_size/nr_pages\n", cpu);
-		return false;
-	}
-
 	buffers_start = meta->first_buffer;
 	buffers_end = meta->first_buffer + (subbuf_size * meta->nr_subbufs);
 
@@ -1742,7 +1797,7 @@ static bool rb_meta_valid(struct ring_buffer_meta *meta, int cpu,
 	return true;
 }
 
-static int rb_meta_subbuf_idx(struct ring_buffer_meta *meta, void *subbuf);
+static int rb_meta_subbuf_idx(struct ring_buffer_cpu_meta *meta, void *subbuf);
 
 static int rb_read_data_buffer(struct buffer_data_page *dpage, int tail, int cpu,
 			       unsigned long long *timestamp, u64 *delta_ptr)
@@ -1809,7 +1864,7 @@ static int rb_validate_buffer(struct buffer_data_page *dpage, int cpu)
 /* If the meta data has been validated, now validate the events */
 static void rb_meta_validate_events(struct ring_buffer_per_cpu *cpu_buffer)
 {
-	struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+	struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 	struct buffer_page *head_page;
 	unsigned long entry_bytes = 0;
 	unsigned long entries = 0;
@@ -1890,7 +1945,7 @@ static void rb_meta_validate_events(struct ring_buffer_per_cpu *cpu_buffer)
 	}
 }
 
-static void rb_meta_init_text_addr(struct ring_buffer_meta *meta)
+static void rb_meta_init_text_addr(struct ring_buffer_cpu_meta *meta)
 {
 #ifdef CONFIG_RANDOMIZE_BASE
 	meta->kaslr_addr = kaslr_offset();
@@ -1901,10 +1956,12 @@ static void rb_meta_init_text_addr(struct ring_buffer_meta *meta)
 
 static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages)
 {
-	struct ring_buffer_meta *meta;
+	struct ring_buffer_cpu_meta *meta;
+	struct ring_buffer_meta *bmeta;
 	unsigned long *subbuf_mask;
 	unsigned long delta;
 	void *subbuf;
+	bool valid = false;
 	int cpu;
 	int i;
 
@@ -1912,12 +1969,17 @@ static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages)
 	subbuf_mask = bitmap_alloc(nr_pages + 1, GFP_KERNEL);
 	/* If subbuf_mask fails to allocate, then rb_meta_valid() will return false */
 
+	if (rb_meta_init(buffer))
+		valid = true;
+
+	bmeta = buffer->meta;
+
 	for (cpu = 0; cpu < nr_cpu_ids; cpu++) {
 		void *next_meta;
 
 		meta = rb_range_meta(buffer, nr_pages, cpu);
 
-		if (rb_meta_valid(meta, cpu, buffer, nr_pages, subbuf_mask)) {
+		if (valid && rb_cpu_meta_valid(meta, cpu, buffer, nr_pages, subbuf_mask)) {
 			/* Make the mappings match the current address */
 			subbuf = rb_subbufs_from_meta(meta);
 			delta = (unsigned long)subbuf - meta->first_buffer;
@@ -1935,9 +1997,6 @@ static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages)
 
 		memset(meta, 0, next_meta - (void *)meta);
 
-		meta->magic = RING_BUFFER_META_MAGIC;
-		meta->struct_size = sizeof(*meta);
-
 		meta->nr_subbufs = nr_pages + 1;
 		meta->subbuf_size = PAGE_SIZE;
 
@@ -1966,7 +2025,7 @@ static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages)
 static void *rbm_start(struct seq_file *m, loff_t *pos)
 {
 	struct ring_buffer_per_cpu *cpu_buffer = m->private;
-	struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+	struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 	unsigned long val;
 
 	if (!meta)
@@ -1991,7 +2050,7 @@ static void *rbm_next(struct seq_file *m, void *v, loff_t *pos)
 static int rbm_show(struct seq_file *m, void *v)
 {
 	struct ring_buffer_per_cpu *cpu_buffer = m->private;
-	struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+	struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 	unsigned long val = (unsigned long)v;
 
 	if (val == 1) {
@@ -2040,7 +2099,7 @@ int ring_buffer_meta_seq_init(struct file *file, struct trace_buffer *buffer, in
 static void rb_meta_buffer_update(struct ring_buffer_per_cpu *cpu_buffer,
 				  struct buffer_page *bpage)
 {
-	struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+	struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 
 	if (meta->head_buffer == (unsigned long)bpage->page)
 		cpu_buffer->head_page = bpage;
@@ -2055,7 +2114,7 @@ static int __rb_allocate_pages(struct ring_buffer_per_cpu *cpu_buffer,
 		long nr_pages, struct list_head *pages)
 {
 	struct trace_buffer *buffer = cpu_buffer->buffer;
-	struct ring_buffer_meta *meta = NULL;
+	struct ring_buffer_cpu_meta *meta = NULL;
 	struct buffer_page *bpage, *tmp;
 	bool user_thread = current->mm != NULL;
 	gfp_t mflags;
@@ -2179,7 +2238,7 @@ static struct ring_buffer_per_cpu *
 rb_allocate_cpu_buffer(struct trace_buffer *buffer, long nr_pages, int cpu)
 {
 	struct ring_buffer_per_cpu *cpu_buffer;
-	struct ring_buffer_meta *meta;
+	struct ring_buffer_cpu_meta *meta;
 	struct buffer_page *bpage;
 	struct page *page;
 	int ret;
@@ -2350,10 +2409,18 @@ static struct trace_buffer *alloc_buffer(unsigned long size, unsigned flags,
 
 	/* If start/end are specified, then that overrides size */
 	if (start && end) {
+		unsigned long buffers_start;
 		unsigned long ptr;
 		int n;
 
-		size = end - start;
+		/* Make sure that start is word aligned */
+		start = ALIGN(start, sizeof(long));
+
+		/* Subtract the buffer meta data and word aligned */
+		buffers_start = start + sizeof(struct ring_buffer_cpu_meta);
+		buffers_start = ALIGN(buffers_start, sizeof(long));
+
+		size = end - buffers_start;
 		size = size / nr_cpu_ids;
 
 		/*
@@ -2363,7 +2430,7 @@ static struct trace_buffer *alloc_buffer(unsigned long size, unsigned flags,
 		 * needed, plus account for the integer array index that
 		 * will be appended to the meta data.
 		 */
-		nr_pages = (size - sizeof(struct ring_buffer_meta)) /
+		nr_pages = (size - sizeof(struct ring_buffer_cpu_meta)) /
 			(subbuf_size + sizeof(int));
 		/* Need at least two pages plus the reader page */
 		if (nr_pages < 3)
@@ -2371,8 +2438,8 @@ static struct trace_buffer *alloc_buffer(unsigned long size, unsigned flags,
 
  again:
 		/* Make sure that the size fits aligned */
-		for (n = 0, ptr = start; n < nr_cpu_ids; n++) {
-			ptr += sizeof(struct ring_buffer_meta) +
+		for (n = 0, ptr = buffers_start; n < nr_cpu_ids; n++) {
+			ptr += sizeof(struct ring_buffer_cpu_meta) +
 				sizeof(int) * nr_pages;
 			ptr = ALIGN(ptr, subbuf_size);
 			ptr += subbuf_size * nr_pages;
@@ -3098,7 +3165,7 @@ static void rb_inc_iter(struct ring_buffer_iter *iter)
 }
 
 /* Return the index into the sub-buffers for a given sub-buffer */
-static int rb_meta_subbuf_idx(struct ring_buffer_meta *meta, void *subbuf)
+static int rb_meta_subbuf_idx(struct ring_buffer_cpu_meta *meta, void *subbuf)
 {
 	void *subbuf_array;
 
@@ -3110,7 +3177,7 @@ static int rb_meta_subbuf_idx(struct ring_buffer_meta *meta, void *subbuf)
 static void rb_update_meta_head(struct ring_buffer_per_cpu *cpu_buffer,
 				struct buffer_page *next_page)
 {
-	struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+	struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 	unsigned long old_head = (unsigned long)next_page->page;
 	unsigned long new_head;
 
@@ -3127,7 +3194,7 @@ static void rb_update_meta_head(struct ring_buffer_per_cpu *cpu_buffer,
 static void rb_update_meta_reader(struct ring_buffer_per_cpu *cpu_buffer,
 				  struct buffer_page *reader)
 {
-	struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+	struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 	void *old_reader = cpu_buffer->reader_page->page;
 	void *new_reader = reader->page;
 	int id;
@@ -3756,7 +3823,7 @@ rb_set_commit_to_write(struct ring_buffer_per_cpu *cpu_buffer)
 			  rb_page_write(cpu_buffer->commit_page));
 		rb_inc_page(&cpu_buffer->commit_page);
 		if (cpu_buffer->ring_meta) {
-			struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+			struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 			meta->commit_buffer = (unsigned long)cpu_buffer->commit_page->page;
 		}
 		/* add barrier to keep gcc from optimizing too much */
@@ -6020,7 +6087,7 @@ rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer)
 	if (cpu_buffer->mapped) {
 		rb_update_meta_page(cpu_buffer);
 		if (cpu_buffer->ring_meta) {
-			struct ring_buffer_meta *meta = cpu_buffer->ring_meta;
+			struct ring_buffer_cpu_meta *meta = cpu_buffer->ring_meta;
 			meta->commit_buffer = meta->head_buffer;
 		}
 	}
@@ -6054,7 +6121,7 @@ static void reset_disabled_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer)
 void ring_buffer_reset_cpu(struct trace_buffer *buffer, int cpu)
 {
 	struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu];
-	struct ring_buffer_meta *meta;
+	struct ring_buffer_cpu_meta *meta;
 
 	if (!cpumask_test_cpu(cpu, buffer->cpumask))
 		return;
@@ -6092,7 +6159,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu);
 void ring_buffer_reset_online_cpus(struct trace_buffer *buffer)
 {
 	struct ring_buffer_per_cpu *cpu_buffer;
-	struct ring_buffer_meta *meta;
+	struct ring_buffer_cpu_meta *meta;
 	int cpu;
 
 	/* prevent another thread from changing buffer sizes */
-- 
2.47.2

[for-next][PATCH 05/10] ring-buffer: Add ring_buffer_meta_scratch()

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

Now that there's one meta data at the start of the persistent memory used by
the ring buffer, allow the caller to request some memory right after that
data that it can use as its own persistent memory.

Also fix some white space issues with ring_buffer_alloc().

Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Link: https://lore.kernel.org/20250305164608.619631731@goodmis.org
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 include/linux/ring_buffer.h |  6 ++++--
 kernel/trace/ring_buffer.c  | 35 +++++++++++++++++++++++++++++------
 kernel/trace/trace.c        |  2 +-
 3 files changed, 34 insertions(+), 9 deletions(-)

diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h
index 8de035f4f0d9..019b59a0bbc9 100644
--- a/include/linux/ring_buffer.h
+++ b/include/linux/ring_buffer.h
@@ -92,9 +92,11 @@ __ring_buffer_alloc(unsigned long size, unsigned flags, struct lock_class_key *k
 struct trace_buffer *__ring_buffer_alloc_range(unsigned long size, unsigned flags,
 					       int order, unsigned long start,
 					       unsigned long range_size,
+					       unsigned long scratch_size,
 					       struct lock_class_key *key);
 
 bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, unsigned long *kaslr_addr);
+void *ring_buffer_meta_scratch(struct trace_buffer *buffer, unsigned int *size);
 
 /*
  * Because the ring buffer is generic, if other users of the ring buffer get
@@ -112,11 +114,11 @@ bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, unsigned long *kas
  * traced by ftrace, it can produce lockdep warnings. We need to keep each
  * ring buffer's lock class separate.
  */
-#define ring_buffer_alloc_range(size, flags, order, start, range_size)	\
+#define ring_buffer_alloc_range(size, flags, order, start, range_size, s_size)	\
 ({									\
 	static struct lock_class_key __key;				\
 	__ring_buffer_alloc_range((size), (flags), (order), (start),	\
-				  (range_size), &__key);		\
+				  (range_size), (s_size), &__key);	\
 })
 
 typedef bool (*ring_buffer_cond_fn)(void *data);
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index ce78b57e1564..bff199d31df3 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -1678,7 +1678,7 @@ static void *rb_range_buffer(struct ring_buffer_per_cpu *cpu_buffer, int idx)
  * See if the existing memory contains a valid meta section.
  * if so, use that, otherwise initialize it.
  */
-static bool rb_meta_init(struct trace_buffer *buffer)
+static bool rb_meta_init(struct trace_buffer *buffer, int scratch_size)
 {
 	unsigned long ptr = buffer->range_addr_start;
 	struct ring_buffer_meta *bmeta;
@@ -1696,6 +1696,7 @@ static bool rb_meta_init(struct trace_buffer *buffer)
 	/* The first buffer will start word size after the meta page */
 	ptr += sizeof(*bmeta);
 	ptr = ALIGN(ptr, sizeof(long));
+	ptr += scratch_size;
 
 	if (bmeta->magic != RING_BUFFER_META_MAGIC) {
 		pr_info("Ring buffer boot meta mismatch of magic\n");
@@ -1730,6 +1731,9 @@ static bool rb_meta_init(struct trace_buffer *buffer)
 	bmeta->total_size = total_size;
 	bmeta->buffers_offset = (void *)ptr - (void *)bmeta;
 
+	/* Zero out the scatch pad */
+	memset((void *)bmeta + sizeof(*bmeta), 0, bmeta->buffers_offset - sizeof(*bmeta));
+
 	return false;
 }
 
@@ -1954,7 +1958,7 @@ static void rb_meta_init_text_addr(struct ring_buffer_cpu_meta *meta)
 #endif
 }
 
-static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages)
+static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages, int scratch_size)
 {
 	struct ring_buffer_cpu_meta *meta;
 	struct ring_buffer_meta *bmeta;
@@ -1969,7 +1973,7 @@ static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages)
 	subbuf_mask = bitmap_alloc(nr_pages + 1, GFP_KERNEL);
 	/* If subbuf_mask fails to allocate, then rb_meta_valid() will return false */
 
-	if (rb_meta_init(buffer))
+	if (rb_meta_init(buffer, scratch_size))
 		valid = true;
 
 	bmeta = buffer->meta;
@@ -2367,6 +2371,7 @@ static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer)
 static struct trace_buffer *alloc_buffer(unsigned long size, unsigned flags,
 					 int order, unsigned long start,
 					 unsigned long end,
+					 unsigned long scratch_size,
 					 struct lock_class_key *key)
 {
 	struct trace_buffer *buffer;
@@ -2416,10 +2421,15 @@ static struct trace_buffer *alloc_buffer(unsigned long size, unsigned flags,
 		/* Make sure that start is word aligned */
 		start = ALIGN(start, sizeof(long));
 
+		/* scratch_size needs to be aligned too */
+		scratch_size = ALIGN(scratch_size, sizeof(long));
+
 		/* Subtract the buffer meta data and word aligned */
 		buffers_start = start + sizeof(struct ring_buffer_cpu_meta);
 		buffers_start = ALIGN(buffers_start, sizeof(long));
+		buffers_start += scratch_size;
 
+		/* Calculate the size for the per CPU data */
 		size = end - buffers_start;
 		size = size / nr_cpu_ids;
 
@@ -2456,7 +2466,7 @@ static struct trace_buffer *alloc_buffer(unsigned long size, unsigned flags,
 		buffer->range_addr_start = start;
 		buffer->range_addr_end = end;
 
-		rb_range_meta_init(buffer, nr_pages);
+		rb_range_meta_init(buffer, nr_pages, scratch_size);
 	} else {
 
 		/* need at least two pages */
@@ -2509,7 +2519,7 @@ struct trace_buffer *__ring_buffer_alloc(unsigned long size, unsigned flags,
 					struct lock_class_key *key)
 {
 	/* Default buffer page size - one system page */
-	return alloc_buffer(size, flags, 0, 0, 0,key);
+	return alloc_buffer(size, flags, 0, 0, 0, 0, key);
 
 }
 EXPORT_SYMBOL_GPL(__ring_buffer_alloc);
@@ -2521,6 +2531,7 @@ EXPORT_SYMBOL_GPL(__ring_buffer_alloc);
  * @order: sub-buffer order
  * @start: start of allocated range
  * @range_size: size of allocated range
+ * @scratch_size: size of scratch area (for preallocated memory buffers)
  * @key: ring buffer reader_lock_key.
  *
  * Currently the only flag that is available is the RB_FL_OVERWRITE
@@ -2531,9 +2542,11 @@ EXPORT_SYMBOL_GPL(__ring_buffer_alloc);
 struct trace_buffer *__ring_buffer_alloc_range(unsigned long size, unsigned flags,
 					       int order, unsigned long start,
 					       unsigned long range_size,
+					       unsigned long scratch_size,
 					       struct lock_class_key *key)
 {
-	return alloc_buffer(size, flags, order, start, start + range_size, key);
+	return alloc_buffer(size, flags, order, start, start + range_size,
+			    scratch_size, key);
 }
 
 /**
@@ -2557,6 +2570,16 @@ bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, unsigned long *kas
 	return true;
 }
 
+void *ring_buffer_meta_scratch(struct trace_buffer *buffer, unsigned int *size)
+{
+	if (!buffer || !buffer->meta)
+		return NULL;
+
+	*size = PAGE_SIZE - sizeof(*buffer->meta);
+
+	return (void *)buffer->meta + sizeof(*buffer->meta);
+}
+
 /**
  * ring_buffer_free - free a ring buffer.
  * @buffer: the buffer to free.
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 41d463e5327c..69dcd7edaf26 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -9214,7 +9214,7 @@ allocate_trace_buffer(struct trace_array *tr, struct array_buffer *buf, int size
 	if (tr->range_addr_start && tr->range_addr_size) {
 		buf->buffer = ring_buffer_alloc_range(size, rb_flags, 0,
 						      tr->range_addr_start,
-						      tr->range_addr_size);
+						      tr->range_addr_size, 0);
 
 #ifdef CONFIG_RANDOMIZE_BASE
 		if (ring_buffer_last_boot_delta(buf->buffer, &tr->kaslr_addr))
-- 
2.47.2

[for-next][PATCH 06/10] tracing: Have persistent trace instances save KASLR offset

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

There's no reason to save the KASLR offset for the ring buffer itself.
That is used by the tracer. Now that the tracer has a way to save data in
the persistent memory of the ring buffer, have the tracing infrastructure
take care of the saving of the KASLR offset.

Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Link: https://lore.kernel.org/20250305164608.792722274@goodmis.org
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 include/linux/ring_buffer.h |  1 -
 kernel/trace/ring_buffer.c  | 59 +++++++------------------------------
 kernel/trace/trace.c        | 37 +++++++++++++++++++----
 kernel/trace/trace.h        |  6 ++--
 4 files changed, 46 insertions(+), 57 deletions(-)

diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h
index 019b59a0bbc9..56e27263acf8 100644
--- a/include/linux/ring_buffer.h
+++ b/include/linux/ring_buffer.h
@@ -95,7 +95,6 @@ struct trace_buffer *__ring_buffer_alloc_range(unsigned long size, unsigned flag
 					       unsigned long scratch_size,
 					       struct lock_class_key *key);
 
-bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, unsigned long *kaslr_addr);
 void *ring_buffer_meta_scratch(struct trace_buffer *buffer, unsigned int *size);
 
 /*
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index bff199d31df3..649724efd1be 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -55,7 +55,6 @@ struct ring_buffer_meta {
 };
 
 struct ring_buffer_cpu_meta {
-	unsigned long	kaslr_addr;
 	unsigned long	first_buffer;
 	unsigned long	head_buffer;
 	unsigned long	commit_buffer;
@@ -557,8 +556,6 @@ struct trace_buffer {
 
 	struct ring_buffer_meta		*meta;
 
-	unsigned long			kaslr_addr;
-
 	unsigned int			subbuf_size;
 	unsigned int			subbuf_order;
 	unsigned int			max_data_size;
@@ -1949,15 +1946,6 @@ static void rb_meta_validate_events(struct ring_buffer_per_cpu *cpu_buffer)
 	}
 }
 
-static void rb_meta_init_text_addr(struct ring_buffer_cpu_meta *meta)
-{
-#ifdef CONFIG_RANDOMIZE_BASE
-	meta->kaslr_addr = kaslr_offset();
-#else
-	meta->kaslr_addr = 0;
-#endif
-}
-
 static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages, int scratch_size)
 {
 	struct ring_buffer_cpu_meta *meta;
@@ -1990,7 +1978,6 @@ static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages, int sc
 			meta->first_buffer += delta;
 			meta->head_buffer += delta;
 			meta->commit_buffer += delta;
-			buffer->kaslr_addr = meta->kaslr_addr;
 			continue;
 		}
 
@@ -2007,7 +1994,6 @@ static void rb_range_meta_init(struct trace_buffer *buffer, int nr_pages, int sc
 		subbuf = rb_subbufs_from_meta(meta);
 
 		meta->first_buffer = (unsigned long)subbuf;
-		rb_meta_init_text_addr(meta);
 
 		/*
 		 * The buffers[] array holds the order of the sub-buffers
@@ -2549,35 +2535,22 @@ struct trace_buffer *__ring_buffer_alloc_range(unsigned long size, unsigned flag
 			    scratch_size, key);
 }
 
-/**
- * ring_buffer_last_boot_delta - return the delta offset from last boot
- * @buffer: The buffer to return the delta from
- * @text: Return text delta
- * @data: Return data delta
- *
- * Returns: The true if the delta is non zero
- */
-bool ring_buffer_last_boot_delta(struct trace_buffer *buffer, unsigned long *kaslr_addr)
-{
-	if (!buffer)
-		return false;
-
-	if (!buffer->kaslr_addr)
-		return false;
-
-	*kaslr_addr = buffer->kaslr_addr;
-
-	return true;
-}
-
 void *ring_buffer_meta_scratch(struct trace_buffer *buffer, unsigned int *size)
 {
+	struct ring_buffer_meta *meta;
+	void *ptr;
+
 	if (!buffer || !buffer->meta)
 		return NULL;
 
-	*size = PAGE_SIZE - sizeof(*buffer->meta);
+	meta = buffer->meta;
 
-	return (void *)buffer->meta + sizeof(*buffer->meta);
+	ptr = (void *)ALIGN((unsigned long)meta + sizeof(*meta), sizeof(long));
+
+	if (size)
+		*size = (void *)meta + meta->buffers_offset - ptr;
+
+	return ptr;
 }
 
 /**
@@ -6144,7 +6117,6 @@ static void reset_disabled_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer)
 void ring_buffer_reset_cpu(struct trace_buffer *buffer, int cpu)
 {
 	struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu];
-	struct ring_buffer_cpu_meta *meta;
 
 	if (!cpumask_test_cpu(cpu, buffer->cpumask))
 		return;
@@ -6163,11 +6135,6 @@ void ring_buffer_reset_cpu(struct trace_buffer *buffer, int cpu)
 	atomic_dec(&cpu_buffer->record_disabled);
 	atomic_dec(&cpu_buffer->resize_disabled);
 
-	/* Make sure persistent meta now uses this buffer's addresses */
-	meta = rb_range_meta(buffer, 0, cpu_buffer->cpu);
-	if (meta)
-		rb_meta_init_text_addr(meta);
-
 	mutex_unlock(&buffer->mutex);
 }
 EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu);
@@ -6182,7 +6149,6 @@ EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu);
 void ring_buffer_reset_online_cpus(struct trace_buffer *buffer)
 {
 	struct ring_buffer_per_cpu *cpu_buffer;
-	struct ring_buffer_cpu_meta *meta;
 	int cpu;
 
 	/* prevent another thread from changing buffer sizes */
@@ -6210,11 +6176,6 @@ void ring_buffer_reset_online_cpus(struct trace_buffer *buffer)
 
 		reset_disabled_cpu_buffer(cpu_buffer);
 
-		/* Make sure persistent meta now uses this buffer's addresses */
-		meta = rb_range_meta(buffer, 0, cpu_buffer->cpu);
-		if (meta)
-			rb_meta_init_text_addr(meta);
-
 		atomic_dec(&cpu_buffer->record_disabled);
 		atomic_sub(RESET_BIT, &cpu_buffer->resize_disabled);
 	}
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 69dcd7edaf26..12a359c293c6 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5988,8 +5988,14 @@ ssize_t tracing_resize_ring_buffer(struct trace_array *tr,
 	return __tracing_resize_ring_buffer(tr, size, cpu_id);
 }
 
+struct trace_scratch {
+	unsigned long		kaslr_addr;
+};
+
 static void update_last_data(struct trace_array *tr)
 {
+	struct trace_scratch *tscratch;
+
 	if (!(tr->flags & TRACE_ARRAY_FL_LAST_BOOT))
 		return;
 
@@ -6004,6 +6010,17 @@ static void update_last_data(struct trace_array *tr)
 	/* Using current data now */
 	tr->text_delta = 0;
 
+	if (!tr->scratch)
+		return;
+
+	tscratch = tr->scratch;
+
+	/* Set the persistent ring buffer meta data to this address */
+#ifdef CONFIG_RANDOMIZE_BASE
+	tscratch->kaslr_addr = kaslr_offset();
+#else
+	tscratch->kaslr_addr = 0;
+#endif
 	tr->flags &= ~TRACE_ARRAY_FL_LAST_BOOT;
 }
 
@@ -6817,6 +6834,7 @@ static ssize_t
 tracing_last_boot_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos)
 {
 	struct trace_array *tr = filp->private_data;
+	struct trace_scratch *tscratch = tr->scratch;
 	struct seq_buf seq;
 	char buf[64];
 
@@ -6829,8 +6847,8 @@ tracing_last_boot_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t
 	 * Otherwise it shows the KASLR address from the previous boot which
 	 * should not be the same as the current boot.
 	 */
-	if (tr->flags & TRACE_ARRAY_FL_LAST_BOOT)
-		seq_buf_printf(&seq, "%lx\t[kernel]\n", tr->kaslr_addr);
+	if (tscratch && (tr->flags & TRACE_ARRAY_FL_LAST_BOOT))
+		seq_buf_printf(&seq, "%lx\t[kernel]\n", tscratch->kaslr_addr);
 	else
 		seq_buf_puts(&seq, "# Current\n");
 
@@ -9206,6 +9224,8 @@ static int
 allocate_trace_buffer(struct trace_array *tr, struct array_buffer *buf, int size)
 {
 	enum ring_buffer_flags rb_flags;
+	struct trace_scratch *tscratch;
+	unsigned int scratch_size;
 
 	rb_flags = tr->trace_flags & TRACE_ITER_OVERWRITE ? RB_FL_OVERWRITE : 0;
 
@@ -9214,12 +9234,19 @@ allocate_trace_buffer(struct trace_array *tr, struct array_buffer *buf, int size
 	if (tr->range_addr_start && tr->range_addr_size) {
 		buf->buffer = ring_buffer_alloc_range(size, rb_flags, 0,
 						      tr->range_addr_start,
-						      tr->range_addr_size, 0);
+						      tr->range_addr_size,
+						      sizeof(*tscratch));
+
+		tscratch = ring_buffer_meta_scratch(buf->buffer, &scratch_size);
+		if (tscratch) {
+			tr->scratch = tscratch;
+			tr->scratch_size = scratch_size;
 
 #ifdef CONFIG_RANDOMIZE_BASE
-		if (ring_buffer_last_boot_delta(buf->buffer, &tr->kaslr_addr))
-			tr->text_delta = kaslr_offset() - tr->kaslr_addr;
+			if (tscratch->kaslr_addr)
+				tr->text_delta = kaslr_offset() - tscratch->kaslr_addr;
 #endif
+		}
 		/*
 		 * This is basically the same as a mapped buffer,
 		 * with the same restrictions.
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index abe8169c3e87..3a020fb82a34 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -348,8 +348,11 @@ struct trace_array {
 	unsigned int		mapped;
 	unsigned long		range_addr_start;
 	unsigned long		range_addr_size;
-	unsigned long		kaslr_addr;
 	long			text_delta;
+	void			*scratch; /* pointer in persistent memory */
+	int			scratch_size;
+
+	int			buffer_disabled;
 
 	struct trace_pid_list	__rcu *filtered_pids;
 	struct trace_pid_list	__rcu *filtered_no_pids;
@@ -367,7 +370,6 @@ struct trace_array {
 	 * CONFIG_TRACER_MAX_TRACE.
 	 */
 	arch_spinlock_t		max_lock;
-	int			buffer_disabled;
 #ifdef CONFIG_FTRACE_SYSCALLS
 	int			sys_refcount_enter;
 	int			sys_refcount_exit;
-- 
2.47.2

[for-next][PATCH 07/10] module: Add module_for_each_mod() function

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

The tracing system needs a way to save all the currently loaded modules
and their addresses into persistent memory so that it can evaluate the
addresses on a reboot from a crash. When the persistent memory trace
starts, it will load the module addresses and names into the persistent
memory. To do so, it will call the module_for_each_mod() function and pass
it a function and data structure to get called on each loaded module. Then
it can record the memory.

This only implements that function.

Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Sami Tolvanen <samitolvanen@google.com>
Cc: Daniel Gomez <da.gomez@samsung.com>
Cc: linux-modules@vger.kernel.org
Link: https://lore.kernel.org/20250305164608.962615966@goodmis.org
Acked-by: Petr Pavlu <petr.pavlu@suse.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 include/linux/module.h |  6 ++++++
 kernel/module/main.c   | 13 +++++++++++++
 2 files changed, 19 insertions(+)

diff --git a/include/linux/module.h b/include/linux/module.h
index 30e5b19bafa9..9a71dd2cb11f 100644
--- a/include/linux/module.h
+++ b/include/linux/module.h
@@ -782,6 +782,8 @@ static inline void *module_writable_address(struct module *mod, void *loc)
 	return __module_writable_address(mod, loc);
 }
 
+void module_for_each_mod(int(*func)(struct module *mod, void *data), void *data);
+
 #else /* !CONFIG_MODULES... */
 
 static inline struct module *__module_address(unsigned long addr)
@@ -894,6 +896,10 @@ static inline void *module_writable_address(struct module *mod, void *loc)
 {
 	return loc;
 }
+
+static inline void module_for_each_mod(int(*func)(struct module *mod, void *data), void *data)
+{
+}
 #endif /* CONFIG_MODULES */
 
 #ifdef CONFIG_SYSFS
diff --git a/kernel/module/main.c b/kernel/module/main.c
index 1fb9ad289a6f..927a2e0ffd5f 100644
--- a/kernel/module/main.c
+++ b/kernel/module/main.c
@@ -3809,6 +3809,19 @@ bool is_module_text_address(unsigned long addr)
 	return ret;
 }
 
+void module_for_each_mod(int(*func)(struct module *mod, void *data), void *data)
+{
+	struct module *mod;
+
+	guard(rcu)();
+	list_for_each_entry_rcu(mod, &modules, list) {
+		if (mod->state == MODULE_STATE_UNFORMED)
+			continue;
+		if (func(mod, data))
+			break;
+	}
+}
+
 /**
  * __module_text_address() - get the module whose code contains an address.
  * @addr: the address.
-- 
2.47.2

[for-next][PATCH 08/10] tracing: Have persistent trace instances save module addresses

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

For trace instances that are mapped to persistent memory, have them use
the scratch area to save the currently loaded modules. This will allow
where the modules have been loaded on the next boot so that their
addresses can be deciphered by using where they were loaded previously.

Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Link: https://lore.kernel.org/20250305164609.129741650@goodmis.org
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 kernel/trace/trace.c | 98 ++++++++++++++++++++++++++++++++++++++++----
 1 file changed, 89 insertions(+), 9 deletions(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 12a359c293c6..6674dc985e52 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5988,14 +5988,60 @@ ssize_t tracing_resize_ring_buffer(struct trace_array *tr,
 	return __tracing_resize_ring_buffer(tr, size, cpu_id);
 }
 
+struct trace_mod_entry {
+	unsigned long	mod_addr;
+	char		mod_name[MODULE_NAME_LEN];
+};
+
 struct trace_scratch {
 	unsigned long		kaslr_addr;
+	unsigned long		nr_entries;
+	struct trace_mod_entry	entries[];
 };
 
+static int save_mod(struct module *mod, void *data)
+{
+	struct trace_array *tr = data;
+	struct trace_scratch *tscratch;
+	struct trace_mod_entry *entry;
+	unsigned int size;
+
+	tscratch = tr->scratch;
+	if (!tscratch)
+		return -1;
+	size = tr->scratch_size;
+
+	if (struct_size(tscratch, entries, tscratch->nr_entries + 1) > size)
+		return -1;
+
+	entry = &tscratch->entries[tscratch->nr_entries];
+
+	tscratch->nr_entries++;
+
+	entry->mod_addr = (unsigned long)mod->mem[MOD_TEXT].base;
+	strscpy(entry->mod_name, mod->name);
+
+	return 0;
+}
+
 static void update_last_data(struct trace_array *tr)
 {
 	struct trace_scratch *tscratch;
 
+	if (!(tr->flags & TRACE_ARRAY_FL_BOOT))
+		return;
+
+	/* Reset the module list and reload them */
+	if (tr->scratch) {
+		struct trace_scratch *tscratch = tr->scratch;
+
+		memset(tscratch->entries, 0,
+		       flex_array_size(tscratch, entries, tscratch->nr_entries));
+		tscratch->nr_entries = 0;
+
+		module_for_each_mod(save_mod, tr);
+	}
+
 	if (!(tr->flags & TRACE_ARRAY_FL_LAST_BOOT))
 		return;
 
@@ -9220,6 +9266,46 @@ static struct dentry *trace_instance_dir;
 static void
 init_tracer_tracefs(struct trace_array *tr, struct dentry *d_tracer);
 
+static void setup_trace_scratch(struct trace_array *tr,
+				struct trace_scratch *tscratch, unsigned int size)
+{
+	struct trace_mod_entry *entry;
+
+	if (!tscratch)
+		return;
+
+	tr->scratch = tscratch;
+	tr->scratch_size = size;
+
+#ifdef CONFIG_RANDOMIZE_BASE
+	if (tscratch->kaslr_addr)
+		tr->text_delta = kaslr_offset() - tscratch->kaslr_addr;
+#endif
+
+	if (struct_size(tscratch, entries, tscratch->nr_entries) > size)
+		goto reset;
+
+	/* Check if each module name is a valid string */
+	for (int i = 0; i < tscratch->nr_entries; i++) {
+		int n;
+
+		entry = &tscratch->entries[i];
+
+		for (n = 0; n < MODULE_NAME_LEN; n++) {
+			if (entry->mod_name[n] == '\0')
+				break;
+			if (!isprint(entry->mod_name[n]))
+				goto reset;
+		}
+		if (n == MODULE_NAME_LEN)
+			goto reset;
+	}
+	return;
+ reset:
+	/* Invalid trace modules */
+	memset(tscratch, 0, size);
+}
+
 static int
 allocate_trace_buffer(struct trace_array *tr, struct array_buffer *buf, int size)
 {
@@ -9232,21 +9318,15 @@ allocate_trace_buffer(struct trace_array *tr, struct array_buffer *buf, int size
 	buf->tr = tr;
 
 	if (tr->range_addr_start && tr->range_addr_size) {
+		/* Add scratch buffer to handle 128 modules */
 		buf->buffer = ring_buffer_alloc_range(size, rb_flags, 0,
 						      tr->range_addr_start,
 						      tr->range_addr_size,
-						      sizeof(*tscratch));
+						      struct_size(tscratch, entries, 128));
 
 		tscratch = ring_buffer_meta_scratch(buf->buffer, &scratch_size);
-		if (tscratch) {
-			tr->scratch = tscratch;
-			tr->scratch_size = scratch_size;
+		setup_trace_scratch(tr, tscratch, scratch_size);
 
-#ifdef CONFIG_RANDOMIZE_BASE
-			if (tscratch->kaslr_addr)
-				tr->text_delta = kaslr_offset() - tscratch->kaslr_addr;
-#endif
-		}
 		/*
 		 * This is basically the same as a mapped buffer,
 		 * with the same restrictions.
-- 
2.47.2

[for-next][PATCH 09/10] tracing: Show module names and addresses of last boot

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

Add the last boot module's names and addresses to the last_boot_info file.
This only shows the module information from a previous boot. If the buffer
is started and is recording the current boot, this file still will only
show "current".

  ~# cat instances/boot_mapped/last_boot_info
  10c00000		[kernel]
  ffffffffc00ca000	usb_serial_simple
  ffffffffc00ae000	usbserial
  ffffffffc008b000	bfq

  ~# echo function > instances/boot_mapped/current_tracer
  ~# cat instances/boot_mapped/last_boot_info
  # Current

Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Link: https://lore.kernel.org/20250305164609.299186021@goodmis.org
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 kernel/trace/trace.c | 102 +++++++++++++++++++++++++++++++++++++------
 1 file changed, 89 insertions(+), 13 deletions(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 6674dc985e52..e1f053ffe887 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5999,6 +5999,8 @@ struct trace_scratch {
 	struct trace_mod_entry	entries[];
 };
 
+static DEFINE_MUTEX(scratch_mutex);
+
 static int save_mod(struct module *mod, void *data)
 {
 	struct trace_array *tr = data;
@@ -6039,6 +6041,7 @@ static void update_last_data(struct trace_array *tr)
 		       flex_array_size(tscratch, entries, tscratch->nr_entries));
 		tscratch->nr_entries = 0;
 
+		guard(mutex)(&scratch_mutex);
 		module_for_each_mod(save_mod, tr);
 	}
 
@@ -6876,15 +6879,47 @@ tracing_total_entries_read(struct file *filp, char __user *ubuf,
 	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
 }
 
-static ssize_t
-tracing_last_boot_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos)
+#define LAST_BOOT_HEADER ((void *)1)
+
+static void *l_next(struct seq_file *m, void *v, loff_t *pos)
 {
-	struct trace_array *tr = filp->private_data;
+	struct trace_array *tr = m->private;
 	struct trace_scratch *tscratch = tr->scratch;
-	struct seq_buf seq;
-	char buf[64];
+	unsigned int index = *pos;
+
+	(*pos)++;
 
-	seq_buf_init(&seq, buf, 64);
+	if (*pos == 1)
+		return LAST_BOOT_HEADER;
+
+	/* Only show offsets of the last boot data */
+	if (!tscratch || !(tr->flags & TRACE_ARRAY_FL_LAST_BOOT))
+		return NULL;
+
+	/* *pos 0 is for the header, 1 is for the first module */
+	index--;
+
+	if (index >= tscratch->nr_entries)
+		return NULL;
+
+	return &tscratch->entries[index];
+}
+
+static void *l_start(struct seq_file *m, loff_t *pos)
+{
+	mutex_lock(&scratch_mutex);
+
+	return l_next(m, NULL, pos);
+}
+
+static void l_stop(struct seq_file *m, void *p)
+{
+	mutex_unlock(&scratch_mutex);
+}
+
+static void show_last_boot_header(struct seq_file *m, struct trace_array *tr)
+{
+	struct trace_scratch *tscratch = tr->scratch;
 
 	/*
 	 * Do not leak KASLR address. This only shows the KASLR address of
@@ -6894,11 +6929,52 @@ tracing_last_boot_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t
 	 * should not be the same as the current boot.
 	 */
 	if (tscratch && (tr->flags & TRACE_ARRAY_FL_LAST_BOOT))
-		seq_buf_printf(&seq, "%lx\t[kernel]\n", tscratch->kaslr_addr);
+		seq_printf(m, "%lx\t[kernel]\n", tscratch->kaslr_addr);
 	else
-		seq_buf_puts(&seq, "# Current\n");
+		seq_puts(m, "# Current\n");
+}
 
-	return simple_read_from_buffer(ubuf, cnt, ppos, buf, seq_buf_used(&seq));
+static int l_show(struct seq_file *m, void *v)
+{
+	struct trace_array *tr = m->private;
+	struct trace_mod_entry *entry = v;
+
+	if (v == LAST_BOOT_HEADER) {
+		show_last_boot_header(m, tr);
+		return 0;
+	}
+
+	seq_printf(m, "%lx\t%s\n", entry->mod_addr, entry->mod_name);
+	return 0;
+}
+
+static const struct seq_operations last_boot_seq_ops = {
+	.start		= l_start,
+	.next		= l_next,
+	.stop		= l_stop,
+	.show		= l_show,
+};
+
+static int tracing_last_boot_open(struct inode *inode, struct file *file)
+{
+	struct trace_array *tr = inode->i_private;
+	struct seq_file *m;
+	int ret;
+
+	ret = tracing_check_open_get_tr(tr);
+	if (ret)
+		return ret;
+
+	ret = seq_open(file, &last_boot_seq_ops);
+	if (ret) {
+		trace_array_put(tr);
+		return ret;
+	}
+
+	m = file->private_data;
+	m->private = tr;
+
+	return 0;
 }
 
 static int tracing_buffer_meta_open(struct inode *inode, struct file *filp)
@@ -7527,10 +7603,10 @@ static const struct file_operations trace_time_stamp_mode_fops = {
 };
 
 static const struct file_operations last_boot_fops = {
-	.open		= tracing_open_generic_tr,
-	.read		= tracing_last_boot_read,
-	.llseek		= generic_file_llseek,
-	.release	= tracing_release_generic_tr,
+	.open		= tracing_last_boot_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= tracing_seq_release,
 };
 
 #ifdef CONFIG_TRACER_SNAPSHOT
-- 
2.47.2

[for-next][PATCH 10/10] tracing: Update modules to persistent instances when loaded

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

When a module is loaded and a persistent buffer is actively tracing, add
it to the list of modules in the persistent memory.

Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Link: https://lore.kernel.org/20250305164609.469844721@goodmis.org
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
---
 kernel/trace/trace.c        | 27 +++++++++++++++++++++++++
 kernel/trace/trace.h        |  2 ++
 kernel/trace/trace_events.c | 40 ++++++++++++++++++++++++++-----------
 3 files changed, 57 insertions(+), 12 deletions(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index e1f053ffe887..177a7d921ff6 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -10083,6 +10083,32 @@ static void trace_module_remove_evals(struct module *mod)
 static inline void trace_module_remove_evals(struct module *mod) { }
 #endif /* CONFIG_TRACE_EVAL_MAP_FILE */
 
+static bool trace_array_active(struct trace_array *tr)
+{
+	if (tr->current_trace != &nop_trace)
+		return true;
+
+	/* 0 is no events, 1 is all disabled */
+	return trace_events_enabled(tr, NULL) > 1;
+}
+
+static void trace_module_record(struct module *mod)
+{
+	struct trace_array *tr;
+
+	list_for_each_entry(tr, &ftrace_trace_arrays, list) {
+		/* Update any persistent trace array that has already been started */
+		if ((tr->flags & (TRACE_ARRAY_FL_BOOT | TRACE_ARRAY_FL_LAST_BOOT)) ==
+		    TRACE_ARRAY_FL_BOOT) {
+			/* Only update if the trace array is active */
+			if (trace_array_active(tr)) {
+				guard(mutex)(&scratch_mutex);
+				save_mod(mod, tr);
+			}
+		}
+	}
+}
+
 static int trace_module_notify(struct notifier_block *self,
 			       unsigned long val, void *data)
 {
@@ -10091,6 +10117,7 @@ static int trace_module_notify(struct notifier_block *self,
 	switch (val) {
 	case MODULE_STATE_COMING:
 		trace_module_add_evals(mod);
+		trace_module_record(mod);
 		break;
 	case MODULE_STATE_GOING:
 		trace_module_remove_evals(mod);
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 3a020fb82a34..90493220c362 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -786,6 +786,8 @@ extern void trace_find_cmdline(int pid, char comm[]);
 extern int trace_find_tgid(int pid);
 extern void trace_event_follow_fork(struct trace_array *tr, bool enable);
 
+extern int trace_events_enabled(struct trace_array *tr, const char *system);
+
 #ifdef CONFIG_DYNAMIC_FTRACE
 extern unsigned long ftrace_update_tot_cnt;
 extern unsigned long ftrace_number_of_pages;
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 513de9ceb80e..7b3ef1d26167 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -1818,28 +1818,28 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt,
 	return cnt;
 }
 
-static ssize_t
-system_enable_read(struct file *filp, char __user *ubuf, size_t cnt,
-		   loff_t *ppos)
+/*
+ * Returns:
+ *   0 : no events exist?
+ *   1 : all events are disabled
+ *   2 : all events are enabled
+ *   3 : some events are enabled and some are enabled
+ */
+int trace_events_enabled(struct trace_array *tr, const char *system)
 {
-	const char set_to_char[4] = { '?', '0', '1', 'X' };
-	struct trace_subsystem_dir *dir = filp->private_data;
-	struct event_subsystem *system = dir->subsystem;
 	struct trace_event_call *call;
 	struct trace_event_file *file;
-	struct trace_array *tr = dir->tr;
-	char buf[2];
 	int set = 0;
-	int ret;
 
-	mutex_lock(&event_mutex);
+	guard(mutex)(&event_mutex);
+
 	list_for_each_entry(file, &tr->events, list) {
 		call = file->event_call;
 		if ((call->flags & TRACE_EVENT_FL_IGNORE_ENABLE) ||
 		    !trace_event_name(call) || !call->class || !call->class->reg)
 			continue;
 
-		if (system && strcmp(call->class->system, system->name) != 0)
+		if (system && strcmp(call->class->system, system) != 0)
 			continue;
 
 		/*
@@ -1855,7 +1855,23 @@ system_enable_read(struct file *filp, char __user *ubuf, size_t cnt,
 		if (set == 3)
 			break;
 	}
-	mutex_unlock(&event_mutex);
+
+	return set;
+}
+
+static ssize_t
+system_enable_read(struct file *filp, char __user *ubuf, size_t cnt,
+		   loff_t *ppos)
+{
+	const char set_to_char[4] = { '?', '0', '1', 'X' };
+	struct trace_subsystem_dir *dir = filp->private_data;
+	struct event_subsystem *system = dir->subsystem;
+	struct trace_array *tr = dir->tr;
+	char buf[2];
+	int set;
+	int ret;
+
+	set = trace_events_enabled(tr, system ? system->name : NULL);
 
 	buf[0] = set_to_char[set];
 	buf[1] = '\n';
-- 
2.47.2