Hi Petr

-----Original Message-----
From: Petr Mladek <pmladek@suse.com>
Sent: Thursday, June 24, 2021 5:02 PM
To: Justin He <redacted>
Cc: Steven Rostedt <rostedt@goodmis.org>; Sergey Senozhatsky
[off-list ref]; Andy Shevchenko
[off-list ref]; Rasmus Villemoes
[off-list ref]; Jonathan Corbet [off-list ref]; Alexander
Viro [off-list ref]; Linus Torvalds <torvalds@linux-
foundation.org>; Peter Zijlstra (Intel) [off-list ref]; Eric
Biggers [off-list ref]; Ahmed S. Darwish [off-list ref];
linux-doc@vger.kernel.org; linux-kernel@vger.kernel.org; linux-
fsdevel@vger.kernel.org; Matthew Wilcox [off-list ref]; Christoph
Hellwig [off-list ref]; nd [off-list ref]
Subject: Re: [PATCH v2 2/4] lib/vsprintf.c: make '%pD' print the full path
of file

On Wed 2021-06-23 13:50:09, Jia He wrote:

quoted

Previously, the specifier '%pD' is for printing dentry name of struct
file. It may not be perfect (by default it only prints one component.)

As suggested by Linus [1]:

quoted

A dentry has a parent, but at the same time, a dentry really does
inherently have "one name" (and given just the dentry pointers, you
can't show mount-related parenthood, so in many ways the "show just
one name" makes sense for "%pd" in ways it doesn't necessarily for
"%pD"). But while a dentry arguably has that "one primary component",
a _file_ is certainly not exclusively about that last component.

Hence change the behavior of '%pD' to print the full path of that file.

If someone invokes snprintf() with small but positive space,
prepend_name_with_len() moves or truncates the string partially.

Does this comment belong to the 1st patch?
prepend_name_with_len() is not called in this patch.

Tend to remove this paragraph since the comments in do_path_unsafe (patch1/4)
was clear enough.

quoted

More
than that, kasprintf() will pass NULL @buf and @end as the parameters,
and @end - @buf can be negative in some case. Hence make it return at
the very beginning with false in these cases.

Same here. file_d_path_name() does not return bool.

Well, please mention in the commit message that %pD uses the entire
given buffer as a scratch space. It might write something behind
the trailing '\0'.

It would make sense to warn about this also in
Documentation/core-api/printk-formats.rst. It is a bit non-standard
behavior.

Okay

quoted

diff --git a/lib/vsprintf.c b/lib/vsprintf.c
index f0c35d9b65bf..f4494129081f 100644
--- a/lib/vsprintf.c
+++ b/lib/vsprintf.c

@@ -920,13 +921,44 @@ char *dentry_name(char *buf, char *end, const

struct dentry *d, struct printf_sp

quoted

 }

 static noinline_for_stack
-char *file_dentry_name(char *buf, char *end, const struct file *f,
+char *file_d_path_name(char *buf, char *end, const struct file *f,
 			struct printf_spec spec, const char *fmt)
 {
+	char *p;
+	const struct path *path;
+	int prepend_len, widen_len, dpath_len;
+
 	if (check_pointer(&buf, end, f, spec))
 		return buf;

-	return dentry_name(buf, end, f->f_path.dentry, spec, fmt);
+	path = &f->f_path;
+	if (check_pointer(&buf, end, path, spec))
+		return buf;
+
+	p = d_path_unsafe(path, buf, end - buf, &prepend_len);
+
+	/* Calculate the full d_path length, ignoring the tail '\0' */
+	dpath_len = end - buf - prepend_len - 1;
+
+	widen_len = max_t(int, dpath_len, spec.field_width);
+
+	/* Case 1: Already started past the buffer. Just forward @buf. */
+	if (buf >= end)
+		return buf + widen_len;
+
+	/*
+	 * Case 2: The entire remaining space of the buffer filled by
+	 * the truncated path. Still need to get moved right when
+	 * the filled width is greather than the full path length.

s/filled/field/ ?

Ah, sorry, I changed it. I would have thought it as a typo 😊


--
Cheers,
Justin (Jia He)