seq_buf: Introduce DECLARE_SEQ_BUF and seq_buf_cstr()

Message ID 20231026170722.work.638-kees@kernel.org
State New
Headers
Series seq_buf: Introduce DECLARE_SEQ_BUF and seq_buf_cstr() |

Commit Message

Kees Cook Oct. 26, 2023, 5:07 p.m. UTC
  Solve two ergonomic issues with struct seq_buf:

1) Too much boilerplate is required to initialize:

	struct seq_buf s;
	char buf[32];

	seq_buf_init(s, buf, sizeof(buf));

Instead, we can build this directly on the stack. Provide
DECLARE_SEQ_BUF() macro to do this:

	DECLARE_SEQ_BUF(s, 32);

2) %NUL termination is fragile and requires 2 steps to get a valid
   C String (and is a layering violation exposing the "internals" of
   seq_buf):

	seq_buf_terminate(s);
	do_something(s->buffer);

Instead, we can just return s->buffer direction after terminating it
in refactored seq_buf_terminate(), now known as seq_buf_cstr():

	do_soemthing(seq_buf_cstr(s));

Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Justin Stitt <justinstitt@google.com>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Yun Zhou <yun.zhou@windriver.com>
Cc: Jacob Keller <jacob.e.keller@intel.com>
Cc: Zhen Lei <thunder.leizhen@huawei.com>
Cc: linux-trace-kernel@vger.kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
---
 include/linux/seq_buf.h | 19 +++++++++++++++----
 kernel/trace/trace.c    | 11 +----------
 lib/seq_buf.c           |  4 +---
 3 files changed, 17 insertions(+), 17 deletions(-)
  

Comments

Steven Rostedt Oct. 26, 2023, 5:38 p.m. UTC | #1
On Thu, 26 Oct 2023 10:07:28 -0700
Kees Cook <keescook@chromium.org> wrote:

> Solve two ergonomic issues with struct seq_buf:

"ergonomic"? Does it cause carpal tunnel? ;-)

> 
> 1) Too much boilerplate is required to initialize:
> 
> 	struct seq_buf s;
> 	char buf[32];
> 
> 	seq_buf_init(s, buf, sizeof(buf));
> 
> Instead, we can build this directly on the stack. Provide
> DECLARE_SEQ_BUF() macro to do this:
> 
> 	DECLARE_SEQ_BUF(s, 32);
> 
> 2) %NUL termination is fragile and requires 2 steps to get a valid
>    C String (and is a layering violation exposing the "internals" of
>    seq_buf):
> 
> 	seq_buf_terminate(s);
> 	do_something(s->buffer);
> 
> Instead, we can just return s->buffer direction after terminating it
> in refactored seq_buf_terminate(), now known as seq_buf_cstr():
> 
> 	do_soemthing(seq_buf_cstr(s));

Do we really need to call it _cstr? Why not just have seq_buf_str() ?

I mean, this is C, do we need to state that in the name too?

BTW, I'm perfectly fine with this change, just the naming I have issues
with.

-- Steve
  
Kees Cook Oct. 26, 2023, 5:54 p.m. UTC | #2
On Thu, Oct 26, 2023 at 01:38:50PM -0400, Steven Rostedt wrote:
> On Thu, 26 Oct 2023 10:07:28 -0700
> Kees Cook <keescook@chromium.org> wrote:
> 
> > Solve two ergonomic issues with struct seq_buf:
> 
> "ergonomic"? Does it cause carpal tunnel? ;-)
> 
> > 
> > 1) Too much boilerplate is required to initialize:
> > 
> > 	struct seq_buf s;
> > 	char buf[32];
> > 
> > 	seq_buf_init(s, buf, sizeof(buf));
> > 
> > Instead, we can build this directly on the stack. Provide
> > DECLARE_SEQ_BUF() macro to do this:
> > 
> > 	DECLARE_SEQ_BUF(s, 32);
> > 
> > 2) %NUL termination is fragile and requires 2 steps to get a valid
> >    C String (and is a layering violation exposing the "internals" of
> >    seq_buf):
> > 
> > 	seq_buf_terminate(s);
> > 	do_something(s->buffer);
> > 
> > Instead, we can just return s->buffer direction after terminating it
> > in refactored seq_buf_terminate(), now known as seq_buf_cstr():
> > 
> > 	do_soemthing(seq_buf_cstr(s));
> 
> Do we really need to call it _cstr? Why not just have seq_buf_str() ?
> 
> I mean, this is C, do we need to state that in the name too?

I'm fine either way. I did that just to make the distinction between our
length-managed string of characters interface (seq_buf), and the
%NUL-terminated string of characters (traditionally called "C String" in
other languages). And it was still shorter than "seq_buf_terminate(s);
s->buffer" ;)

> BTW, I'm perfectly fine with this change, just the naming I have issues
> with.

Cool; thanks for looking at it!
  
Steven Rostedt Oct. 26, 2023, 6:02 p.m. UTC | #3
On Thu, 26 Oct 2023 10:54:26 -0700
Kees Cook <keescook@chromium.org> wrote:

> > Do we really need to call it _cstr? Why not just have seq_buf_str() ?
> > 
> > I mean, this is C, do we need to state that in the name too?  
> 
> I'm fine either way. I did that just to make the distinction between our
> length-managed string of characters interface (seq_buf), and the
> %NUL-terminated string of characters (traditionally called "C String" in
> other languages). And it was still shorter than "seq_buf_terminate(s);
> s->buffer" ;)

Do you believe that people might get confused with it as seq_buf_str()?

Can you envision that we would want a seq_buf_str() and seq_buf_cstr() that
do something different?

-- Steve
  
Kees Cook Oct. 26, 2023, 7:35 p.m. UTC | #4
On Thu, Oct 26, 2023 at 02:02:47PM -0400, Steven Rostedt wrote:
> On Thu, 26 Oct 2023 10:54:26 -0700
> Kees Cook <keescook@chromium.org> wrote:
> 
> > > Do we really need to call it _cstr? Why not just have seq_buf_str() ?
> > > 
> > > I mean, this is C, do we need to state that in the name too?  
> > 
> > I'm fine either way. I did that just to make the distinction between our
> > length-managed string of characters interface (seq_buf), and the
> > %NUL-terminated string of characters (traditionally called "C String" in
> > other languages). And it was still shorter than "seq_buf_terminate(s);
> > s->buffer" ;)
> 
> Do you believe that people might get confused with it as seq_buf_str()?
> 
> Can you envision that we would want a seq_buf_str() and seq_buf_cstr() that
> do something different?

No, I see your point. Like I said, I don't care either way. I was just
explaining why I did it that way. "string" means a lot of things to
different people. "C String" is unambiguous, and I try to be unambiguous
whenever possible. :)

I'll send a v2 as seq_buf_str()...
  
Steven Rostedt Oct. 26, 2023, 7:37 p.m. UTC | #5
On Thu, 26 Oct 2023 12:35:43 -0700
Kees Cook <keescook@chromium.org> wrote:

> I'll send a v2 as seq_buf_str()...

Thanks.

-- Steve
  
Matthew Wilcox Oct. 27, 2023, 10:43 a.m. UTC | #6
On Thu, Oct 26, 2023 at 10:54:26AM -0700, Kees Cook wrote:
> > > 	do_soemthing(seq_buf_cstr(s));
> > 
> > Do we really need to call it _cstr? Why not just have seq_buf_str() ?
> > 
> > I mean, this is C, do we need to state that in the name too?
> 
> I'm fine either way. I did that just to make the distinction between our
> length-managed string of characters interface (seq_buf), and the
> %NUL-terminated string of characters (traditionally called "C String" in
> other languages). And it was still shorter than "seq_buf_terminate(s);
> s->buffer" ;)

'cstr' might be short for 'counted string' ...
  

Patch

diff --git a/include/linux/seq_buf.h b/include/linux/seq_buf.h
index 8483e4b2d0d2..8896b830eb3d 100644
--- a/include/linux/seq_buf.h
+++ b/include/linux/seq_buf.h
@@ -21,9 +21,16 @@  struct seq_buf {
 	size_t			len;
 };
 
+#define DECLARE_SEQ_BUF(NAME, SIZE)					\
+	char __ ## NAME ## _buffer[SIZE] = "";				\
+	struct seq_buf NAME = { .buffer = &__ ## NAME ## _buffer,	\
+				.size = SIZE }
+
 static inline void seq_buf_clear(struct seq_buf *s)
 {
 	s->len = 0;
+	if (s->size)
+		s->buffer[0] = '\0';
 }
 
 static inline void
@@ -69,8 +76,8 @@  static inline unsigned int seq_buf_used(struct seq_buf *s)
 }
 
 /**
- * seq_buf_terminate - Make sure buffer is nul terminated
- * @s: the seq_buf descriptor to terminate.
+ * seq_buf_cstr - get %NUL-terminated C string from seq_buf
+ * @s: the seq_buf handle
  *
  * This makes sure that the buffer in @s is nul terminated and
  * safe to read as a string.
@@ -81,16 +88,20 @@  static inline unsigned int seq_buf_used(struct seq_buf *s)
  *
  * After this function is called, s->buffer is safe to use
  * in string operations.
+ *
+ * Returns @s->buf after making sure it is terminated.
  */
-static inline void seq_buf_terminate(struct seq_buf *s)
+static inline char *seq_buf_cstr(struct seq_buf *s)
 {
 	if (WARN_ON(s->size == 0))
-		return;
+		return "";
 
 	if (seq_buf_buffer_left(s))
 		s->buffer[s->len] = 0;
 	else
 		s->buffer[s->size - 1] = 0;
+
+	return s->buffer;
 }
 
 /**
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index d629065c2383..d83f36dc4bf8 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -3828,15 +3828,6 @@  static bool trace_safe_str(struct trace_iterator *iter, const char *str,
 	return false;
 }
 
-static const char *show_buffer(struct trace_seq *s)
-{
-	struct seq_buf *seq = &s->seq;
-
-	seq_buf_terminate(seq);
-
-	return seq->buffer;
-}
-
 static DEFINE_STATIC_KEY_FALSE(trace_no_verify);
 
 static int test_can_verify_check(const char *fmt, ...)
@@ -3976,7 +3967,7 @@  void trace_check_vprintf(struct trace_iterator *iter, const char *fmt,
 		 */
 		if (WARN_ONCE(!trace_safe_str(iter, str, star, len),
 			      "fmt: '%s' current_buffer: '%s'",
-			      fmt, show_buffer(&iter->seq))) {
+			      fmt, seq_buf_cstr(&iter->seq.seq))) {
 			int ret;
 
 			/* Try to safely read the string */
diff --git a/lib/seq_buf.c b/lib/seq_buf.c
index b7477aefff53..165caed5a74e 100644
--- a/lib/seq_buf.c
+++ b/lib/seq_buf.c
@@ -109,9 +109,7 @@  void seq_buf_do_printk(struct seq_buf *s, const char *lvl)
 	if (s->size == 0 || s->len == 0)
 		return;
 
-	seq_buf_terminate(s);
-
-	start = s->buffer;
+	start = seq_buf_cstr(s);
 	while ((lf = strchr(start, '\n'))) {
 		int len = lf - start + 1;