Skip to contentSkip to conversation

checkpatch: add check for snprintf to scnprintf

PATCH Superseded Justin Stitt 5 participants 7 messages Apr 8, 2024 – Apr 15, 2024 Checkpatch

v4 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

PATCH v4 1/1 checkpatch: add check for snprintf to scnprintf

I am going to quote Lee Jones who has been doing some snprintf -> scnprintf refactorings:

"There is a general misunderstanding amongst engineers that {v}snprintf() returns the length of the data *actually* encoded into the destination array. However, as per the C99 standard {v}snprintf() really returns the length of the data that *would have been* written if there were enough space for it. This misunderstanding has led to buffer-overruns in the past. It's generally considered safer to use the {v}scnprintf() variants in their place (or even sprintf() in simple cases). So let's do that."

To help prevent new instances of snprintf() from popping up, let's add a check to checkpatch.pl.

---
Changes in v4:
- also check for vsnprintf variant (thanks Bill)
- Link to v3: https://lore.kernel.org/r/[email protected]

Changes in v3:
- fix indentation
- add reference link (https://github.com/KSPP/linux/issues/105) (thanks Joe)
- Link to v2: https://lore.kernel.org/r/[email protected]

Changes in v2:
- Had a vim moment and deleted a character before sending the patch.
- Replaced the character :)
- Link to v1: https://lore.kernel.org/r/[email protected]
---
From a discussion here [1].

[1]: https://lore.kernel.org/all/[email protected]/

---
 scripts/checkpatch.pl | 6 ++++++
 1 file changed, 6 insertions(+)
scripts/checkpatch.pl +6 -0
index 9c4c4a61bc83..a0fd681ea837 100755
@@ -7012,6 +7012,12 @@ sub process {
 			     "Prefer strscpy, strscpy_pad, or __nonstring over strncpy - see: https://github.com/KSPP/linux/issues/90\n" . $herecurr);
 		}
 
+# {v}snprintf uses that should likely be {v}scnprintf
+		if ($line =~ /\b(v|)snprintf\s*\(\s*/) {
+			WARN("SNPRINTF",
+			     "Prefer {v}scnprintf over {v}snprintf - see: https://github.com/KSPP/linux/issues/105\n" . $herecurr);
+		}
+
 # ethtool_sprintf uses that should likely be ethtool_puts
 		if ($line =~ /\bethtool_sprintf\s*\(\s*$FuncArg\s*,\s*$FuncArg\s*\)/) {
 			if (WARN("PREFER_ETHTOOL_PUTS",

---
base-commit: b401b621758e46812da61fa58a67c3fd8d91de0d
change-id: 20240221-snprintf-checkpatch-a864ed67ebd0

Best regards,
--
Justin Stitt <[email protected]>

Reviewed-by on v4 1/1
#1 lore raw
Quoted 12 lines · Justin Stitt

I am going to quote Lee Jones who has been doing some snprintf -> scnprintf refactorings:

"There is a general misunderstanding amongst engineers that {v}snprintf() returns the length of the data *actually* encoded into the destination array. However, as per the C99 standard {v}snprintf() really returns the length of the data that *would have been* written if there were enough space for it. This misunderstanding has led to buffer-overruns in the past. It's generally considered safer to use the {v}scnprintf() variants in their place (or even sprintf() in simple cases). So let's do that."

Quoted 5 lines

To help prevent new instances of snprintf() from popping up, let's add a check to checkpatch.pl.

Signature · 1 line
Kees Cook
Reviewed-by on v4 1/1
#2 lore raw
In reply to Justin Stitt
Quoted 34 lines · Justin Stitt

I am going to quote Lee Jones who has been doing some snprintf -> scnprintf refactorings:

"There is a general misunderstanding amongst engineers that {v}snprintf() returns the length of the data *actually* encoded into the destination array. However, as per the C99 standard {v}snprintf() really returns the length of the data that *would have been* written if there were enough space for it. This misunderstanding has led to buffer-overruns in the past. It's generally considered safer to use the {v}scnprintf() variants in their place (or even sprintf() in simple cases). So let's do that."

To help prevent new instances of snprintf() from popping up, let's add a check to checkpatch.pl.

---
Changes in v4:
- also check for vsnprintf variant (thanks Bill)
- Link to v3: https://lore.kernel.org/r/[email protected]

Changes in v3:
- fix indentation
- add reference link (https://github.com/KSPP/linux/issues/105) (thanks Joe)
- Link to v2: https://lore.kernel.org/r/[email protected]

Changes in v2:
- Had a vim moment and deleted a character before sending the patch.
- Replaced the character :)
- Link to v1: https://lore.kernel.org/r/[email protected]
---
From a discussion here [1].

Quoted 4 lines

[1]: https://lore.kernel.org/all/[email protected]/

---
 scripts/checkpatch.pl | 6 ++++++
 1 file changed, 6 insertions(+)
Signature · 1 line
Lee Jones [李琼斯]
#3 lore raw
In reply to Justin Stitt

Le 08/04/2024 à 22:53, Justin Stitt a écrit :

Quoted 44 lines

I am going to quote Lee Jones who has been doing some snprintf -> scnprintf refactorings:

"There is a general misunderstanding amongst engineers that {v}snprintf() returns the length of the data *actually* encoded into the destination array. However, as per the C99 standard {v}snprintf() really returns the length of the data that *would have been* written if there were enough space for it. This misunderstanding has led to buffer-overruns in the past. It's generally considered safer to use the {v}scnprintf() variants in their place (or even sprintf() in simple cases). So let's do that."

To help prevent new instances of snprintf() from popping up, let's add a check to checkpatch.pl.

---
Changes in v4:
- also check for vsnprintf variant (thanks Bill)
- Link to v3: https://lore.kernel.org/r/[email protected]

Changes in v3:
- fix indentation
- add reference link (https://github.com/KSPP/linux/issues/105) (thanks Joe)
- Link to v2: https://lore.kernel.org/r/[email protected]

Changes in v2:
- Had a vim moment and deleted a character before sending the patch.
- Replaced the character :)
- Link to v1: https://lore.kernel.org/r/[email protected]
---
From a discussion here [1].

[1]: https://lore.kernel.org/all/[email protected]/

---
  scripts/checkpatch.pl | 6 ++++++
  1 file changed, 6 insertions(+)
scripts/checkpatch.pl +0 -0
index 9c4c4a61bc83..a0fd681ea837 100755

scripts/checkpatch.pl sub process { · line 7016

Quoted 5 lines
scripts/checkpatch.pl +2 -0
@@ -7012,6 +7012,12 @@ sub process {
  			     "Prefer strscpy, strscpy_pad, or __nonstring over strncpy - see: https://github.com/KSPP/linux/issues/90\n" . $herecurr);
  		}
 
+# {v}snprintf uses that should likely be {v}scnprintf
+		if ($line =~ /\b(v|)snprintf\s*\(\s*/) {

Hi,

for my understanding, what is the purpose of the 2nd "\s*"? IMHO, it could be just removed.

scripts/checkpatch.pl sub process { · line 7018

Quoted 2 lines
scripts/checkpatch.pl +2 -0
+			WARN("SNPRINTF",
+			     "Prefer {v}scnprintf over {v}snprintf - see: https://github.com/KSPP/linux/issues/105\n" . $herecurr);

Maybe $1 instead of {v} in both places, so that is displays the real function name that is and should be used?

CJ

Quoted 16 lines
scripts/checkpatch.pl +2 -1
+		}
+
  # ethtool_sprintf uses that should likely be ethtool_puts
  		if ($line =~ /\bethtool_sprintf\s*\(\s*$FuncArg\s*,\s*$FuncArg\s*\)/) {
  			if (WARN("PREFER_ETHTOOL_PUTS",
 
---

base-commit: b401b621758e46812da61fa58a67c3fd8d91de0d
change-id: 20240221-snprintf-checkpatch-a864ed67ebd0

Best regards,
--
Justin Stitt <[email protected]>

maintainer
#4 lore raw
Quoted 48 lines · Christophe JAILLET

Le 08/04/2024 à 22:53, Justin Stitt a écrit :

Quoted 47 lines

I am going to quote Lee Jones who has been doing some snprintf -> scnprintf refactorings:

"There is a general misunderstanding amongst engineers that {v}snprintf() returns the length of the data *actually* encoded into the destination array. However, as per the C99 standard {v}snprintf() really returns the length of the data that *would have been* written if there were enough space for it. This misunderstanding has led to buffer-overruns in the past. It's generally considered safer to use the {v}scnprintf() variants in their place (or even sprintf() in simple cases). So let's do that."

To help prevent new instances of snprintf() from popping up, let's add a check to checkpatch.pl.

---
Changes in v4:
- also check for vsnprintf variant (thanks Bill)
- Link to v3: https://lore.kernel.org/r/[email protected]

Changes in v3:
- fix indentation
- add reference link (https://github.com/KSPP/linux/issues/105) (thanks Joe)
- Link to v2: https://lore.kernel.org/r/[email protected]

Changes in v2:
- Had a vim moment and deleted a character before sending the patch.
- Replaced the character :)
- Link to v1: https://lore.kernel.org/r/[email protected]
---
From a discussion here [1].

[1]: https://lore.kernel.org/all/[email protected]/

---
  scripts/checkpatch.pl | 6 ++++++
  1 file changed, 6 insertions(+)
scripts/checkpatch.pl +0 -0
index 9c4c4a61bc83..a0fd681ea837 100755
@@ -7012,6 +7012,12 @@ sub process {
  			     "Prefer strscpy, strscpy_pad, or __nonstring over strncpy - see: https://github.com/KSPP/linux/issues/90\n" . $herecurr);
  		}
 
Quoted 7 lines

scripts/checkpatch.pl sub process { · line 7016

Quoted 2 lines
scripts/checkpatch.pl +2 -0
+# {v}snprintf uses that should likely be {v}scnprintf
+		if ($line =~ /\b(v|)snprintf\s*\(\s*/) {

Hi,

for my understanding, what is the purpose of the 2nd "\s*"? IMHO, it could be just removed.

It could.

# {v}snprintf uses that should likely be {v}scnprintf

		if ($line =~ /\b((v?)snprintf)\s*\(/) {
			WARN("SNPRINTF",
			     "Prefer ${2}scnprintf over $1 - see: https://github.com/KSPP/linux/issues/105\n" . $herecurr);
		}

Though I also think it's better to use lore rather than github

#5 lore raw
Quoted 11 lines · Joe Perches

It could.

# {v}snprintf uses that should likely be {v}scnprintf

                if ($line =~ /\b((v?)snprintf)\s*\(/) {
                        WARN("SNPRINTF",
                             "Prefer ${2}scnprintf over $1 - see: https://github.com/KSPP/linux/issues/105\n" . $herecurr);
                }

Though I also think it's better to use lore rather than github

I am fine with making the UX change in v5 regarding using ${2} and $1 but I wish someone could have said something about the Github links earlier, we already have a pattern going with these string api changes:

"Prefer strscpy over strcpy - see:
https://github.com/KSPP/linux/issues/88\n" . $herecurr); }

...

# strlcpy uses that should likely be strscpy
if ($line =~ /\bstrlcpy\s*\(/) {
WARN("STRLCPY",
     "Prefer strscpy over strlcpy - see:
https://github.com/KSPP/linux/issues/89\n" . $herecurr);
}
# strncpy uses that should likely be strscpy or strscpy_pad
if ($line =~ /\bstrncpy\s*\(/) {
WARN("STRNCPY",
     "Prefer strscpy, strscpy_pad, or __nonstring over strncpy - see:
https://github.com/KSPP/linux/issues/90\n" . $herecurr);
}

# {v}snprintf uses that should likely be {v}scnprintf if ($line =~ /\b(v|)snprintf\s*\(/) { WARN("SNPRINTF",

     "Prefer ${2}scnprintf over $1 - see:

https://github.com/KSPP/linux/issues/105\n" . $herecurr);
...

It should be noted that nowhere else is a lore link or github link provided during a warning, so there really is no precedence. Joe what should we do?

Quoted 1 line
#6 lore raw
Quoted 13 lines · Justin Stitt
Quoted 11 lines · Joe Perches

It could.

# {v}snprintf uses that should likely be {v}scnprintf

                if ($line =~ /\b((v?)snprintf)\s*\(/) {
                        WARN("SNPRINTF",
                             "Prefer ${2}scnprintf over $1 - see: https://github.com/KSPP/linux/issues/105\n" . $herecurr);
                }

Though I also think it's better to use lore rather than github

Quoted 8 lines

I am fine with making the UX change in v5 regarding using ${2} and $1 but I wish someone could have said something about the Github links earlier, we already have a pattern going with these string api changes:

"Prefer strscpy over strcpy - see:
https://github.com/KSPP/linux/issues/88\n" . $herecurr); }

KSPP isn't going anywhere -- we've used these links before and we can use them here too. I don't see any good reason to duplicate stuff into lore, etc.

Signature · 1 line
Kees Cook