On 07/10, Steven Rostedt wrote:

On Wed, 28 Jun 2023 18:02:43 +0300
Andy Shevchenko [off-list ref] wrote:

quoted

Since strreplace() returns the pointer to the string itself,
we may use it directly in the code.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 include/trace/events/f2fs.h | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h
index 793f82cc1515..f5994515290c 100644
--- a/include/trace/events/f2fs.h
+++ b/include/trace/events/f2fs.h

@@ -2234,13 +2234,11 @@ DECLARE_EVENT_CLASS(f2fs__rw_start,
 		 * because this screws up the tooling that parses
 		 * the traces.
 		 */
-		__assign_str(pathbuf, pathname);
-		(void)strreplace(__get_str(pathbuf), ' ', '_');
+		__assign_str(pathbuf, strreplace(pathname, ' ', '_'));

But this modifies the pathname that is passed into the trace event, which
is something that a trace point should never do! In fact, the char
*pathname, really should be a const char * (for which this would fail to
build).

Note, I went to look for these events and I can not find where they are
used. Should these events just be deleted?

Hmm, this was a part of upstream effort to replace the previous android_fs
tracepoints like:
https://android-review.git.corp.google.com/c/platform/system/extras/+/2223339

-- Steve

quoted

 		__entry->offset = offset;
 		__entry->bytes = bytes;
 		__entry->i_size = i_size_read(inode);
-		__assign_str(cmdline, command);
-		(void)strreplace(__get_str(cmdline), ' ', '_');
+		__assign_str(cmdline, strreplace(command, ' ', '_'));
 		__entry->pid = pid;
 		__entry->ino = inode->i_ino;
 	),