Re: [PATCH] Input: mouse: vsxxxaa: Replace strlcat() with strscpy()

From
Alejandro Colomar <alx+linux-hardening@kernel.org>
Date
2026-11-05 11:50:03
Also in
linux-hardening, lkml
Reply to
David Laight, 2026-11-05
Hi David,
Date: 2026-11-05 11:28:13+0000
From: David Laight <redacted>

On Tue,  3 Nov 2026 22:23:18 -0800
Mariia Nikitash [off-list ref] wrote:
quoted
In preparation for removing the deprecated strlcat() API[1], use
strscpy() to initialize the mouse device name.

The destination is zero-initialized, and the constant string fits
within the buffer. Since the return value is unused, this preserves
the existing behavior.

Link: https://github.com/KSPP/linux/issues/370 [1]
Signed-off-by: Mariia Nikitash <redacted>
---
 drivers/input/mouse/vsxxxaa.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/input/mouse/vsxxxaa.c b/drivers/input/mouse/vsxxxaa.c
index 042a17fe81db..b3a0d7ff5738 100644
--- a/drivers/input/mouse/vsxxxaa.c
+++ b/drivers/input/mouse/vsxxxaa.c
@@ -463,8 +463,8 @@ static int vsxxxaa_connect(struct serio *serio, struct serio_driver *drv)
 
 	mouse->dev = input_dev;
 	mouse->serio = serio;
-	strlcat(mouse->name, "DEC VSXXX-AA/-GA mouse or VSXXX-AB digitizer",
-		 sizeof(mouse->name));
+	strscpy(mouse->name, "DEC VSXXX-AA/-GA mouse or VSXXX-AB digitizer",
+		sizeof(mouse->name));
You might as well use the 2 argument form.
Moreover, if you use strcpy() you'll get a compile-time error if the
string doesn't fit.
strscpy() will truncate and return an errors.

The other question is whether mouse->name even has to be char[].

Personally I think that while there are issues with strlcat()
(mostly to do with its return value), and that there are often
much better (and faster - so sprintf is out) ways of getting
the same result, the deprecation really needs to propose a
sensible alternative (not inlined strlen() + strscpy()).
One alternative would be to add strscat(), with the obvious internal
strlen(3).  This would be a drop-in replacement, and wouldn't have the
issues of strlcat(3) --mainly, the DoS vulnerability which is
consequence of the return value, and also the more difficult error
handling of doing '>=sizeof(...)' than '<0'--.

There's an implementation of strscat() in shadow-utils, where we call it
strtcat() --because what it does is Truncating the string, so a T feels
more appropriate--:
<https://github.com/shadow-maint/shadow/blob/214a2e6e0a5f81fa281b272be7cf9e9d3aff8f34/lib/string/strcpy/strtcat.h#L28>

Another alternative which would avoid the internal strlen(3) call, and
thus would be faster, would be to add a stpecpy() function, similar to
stpcpy(3), but which would take a pointer one-past-the-last element in
the buffer, to perform truncation.  This is fundamentally Plan9's
strecpy(2), except that Plan9's API has a bug.

There's an implementation of stpecpy() in shadow-utils:
<https://github.com/shadow-maint/shadow/blob/214a2e6e0a5f81fa281b272be7cf9e9d3aff8f34/lib/string/strcpy/stpecpy.h#L28>

We've been using stpecpy() quite extensively in shadow-utils, and it's
a great API.  Of course, there's also the decades-long experience in
Plan9, which apart from their bug --which I found, by comparing their
implementation to mine--, shows that it's a quite useful API.

And stpecpy() produces the simplest code of all truncating functions.


Have a lovely day!
Alex
This case is, of course one of the 'better ways' cases.

David
quoted
 	snprintf(mouse->phys, sizeof(mouse->phys), "%s/input0", serio->phys);
 
 	input_dev->name = mouse->name;
-- 
<https://www.alejandro-colomar.es>

Attachments

Thread (9 messages) 9 messages, 5 authors, last reply 5d ago
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help