Skip to contentSkip to conversation

bug in checkpatch (on pointers to typedefs?)

DISCUSSION Marcin Slusarz 4 participants 11 messages Feb 10, 2008 – Feb 14, 2008

lore
Share & tools
lore
#0 lore raw

Hi

Checkpatch in current mainline outputs following errors:

$ ./scripts/checkpatch.pl -q --file ./fs/udf/misc.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#205: FILE: fs/udf/misc.c:205:
+       tag *tag_p;
            ^
$ ./scripts/checkpatch.pl -q --file ./fs/udf/unicode.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#48: FILE: fs/udf/unicode.c:48:
+int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
                                               ^
(...)

I think the code is ok.

Marcin

#1 lore raw
Quoted 10 lines · Marcin Slusarz

Hi

Checkpatch in current mainline outputs following errors:

$ ./scripts/checkpatch.pl -q --file ./fs/udf/misc.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#205: FILE: fs/udf/misc.c:205:
+       tag *tag_p;
            ^
Quoted 8 lines
$ ./scripts/checkpatch.pl -q --file ./fs/udf/unicode.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#48: FILE: fs/udf/unicode.c:48:
+int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
                                               ^
(...)

I think the code is ok.

Yep the code is clearly correct. Can I have the whole patch fragment you got these against so I can figure out why we are unable to detect these two as types, I would expect us to have done so. Also what version of checkpatch is this? There is a version string at the top.

-apw

#2 lore raw

I saw this too with checkpatch.pl version 0.12
It seems like checkpatch.pl knows only about types derived from @typeList by build_types.

Example below...

Benny

$ cat <<EOF | scripts/checkpatch.pl -
Signed-off-by: [email protected]
---
diff a/f.c b/f.c
--- a/f.c
+++ b/f.c
@@ -1,0 +1,2 @@
+foo(int a, my_uint32 *);
+bar(int a, my_uint32_t *);
EOF
ERROR: need consistent spacing around '*' (ctx:WxB)
#7: FILE: f.c:1:
+foo(int a, my_uint32 *);
                      ^

total: 1 errors, 0 warnings, 2 lines checked

Quoted 31 lines · Andy Whitcroft
Quoted 10 lines · Marcin Slusarz

Hi

Checkpatch in current mainline outputs following errors:

$ ./scripts/checkpatch.pl -q --file ./fs/udf/misc.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#205: FILE: fs/udf/misc.c:205:
+       tag *tag_p;
            ^
Quoted 8 lines
$ ./scripts/checkpatch.pl -q --file ./fs/udf/unicode.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#48: FILE: fs/udf/unicode.c:48:
+int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
                                               ^
(...)

I think the code is ok.

Yep the code is clearly correct. Can I have the whole patch fragment you got these against so I can figure out why we are unable to detect these two as types, I would expect us to have done so. Also what version of checkpatch is this? There is a version string at the top.

-apw
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#3 lore raw
Quoted 15 lines · Benny Halevy

I saw this too with checkpatch.pl version 0.12
It seems like checkpatch.pl knows only about types derived from @typeList by build_types.

Example below...

Benny

$ cat <<EOF | scripts/checkpatch.pl -
Signed-off-by: [email protected]
---
diff a/f.c b/f.c

f.c line 2

Quoted 2 lines
f.c +2 -0
@@ -1,0 +1,2 @@
+foo(int a, my_uint32 *);
+bar(int a, my_uint32_t *);

But that isn't actually syntactically correct code is it? You have types as parameters like a function declaration, but no return type. So there is no hint to checkpatch that this is a function declaration and therefore the parameters are not expected to be types, nor are they checked as such.

The following diff is clean on the latest version of checkpatch:

Signed-off-by [email protected]

---

f.c +2 -0
diff a/f.c b/f.c
@@ -1,0 +1,2 @@
+void foo(int a, my_uint32 *);
+int bar(int a, my_uint32_t *);

EOF

Could you try out the version of checkpatch at the URL below on the real patch you are using to test, and let me know if it works. There are a number of improvements to type tracking in the face of ifdef's and the like. If it doesn't could I have the hunk which fails:

http://www.kernel.org/pub/linux/kernel/people/apw/checkpatch/checkpatch.pl-next

-apw

#4 lore raw
Quoted 34 lines · Andy Whitcroft
Quoted 15 lines · Benny Halevy

I saw this too with checkpatch.pl version 0.12
It seems like checkpatch.pl knows only about types derived from @typeList by build_types.

Example below...

Benny

$ cat <<EOF | scripts/checkpatch.pl -
Signed-off-by: [email protected]
---
diff a/f.c b/f.c

f.c line 2

Quoted 2 lines
f.c +2 -0
@@ -1,0 +1,2 @@
+foo(int a, my_uint32 *);
+bar(int a, my_uint32_t *);

But that isn't actually syntactically correct code is it? You have types as parameters like a function declaration, but no return type. So there is no hint to checkpatch that this is a function declaration and therefore the parameters are not expected to be types, nor are they checked as such.

The following diff is clean on the latest version of checkpatch:

Signed-off-by [email protected]

---
diff a/f.c b/f.c

f.c +2 -0
@@ -1,0 +1,2 @@
+void foo(int a, my_uint32 *);
+int bar(int a, my_uint32_t *);
Quoted 1 line

EOF

OK, but the return type doesn't have to be in the patched line, it could be in a synchronization line or even missing if the function has a long multi-line argument list.

Quoted 7 lines

Could you try out the version of checkpatch at the URL below on the real patch you are using to test, and let me know if it works. There are a number of improvements to type tracking in the face of ifdef's and the like. If it doesn't could I have the hunk which fails:

http://www.kernel.org/pub/linux/kernel/people/apw/checkpatch/checkpatch.pl-next

Your modified patch passes with version 0.12 as well as checkpatch.pl-next

However, the following patch that has the return type in the synchronization lines still produces the same error:

$ ./checkpatch.pl-next -
Signed-off-by: [email protected]
---
diff a/f.c b/f.c
--- a/f.c
+++ b/f.c
@@ -1,2 +1,4 @@
 int
+foo(int a, my_uint32 *);
 int
+bar(int a, my_uint32_t *);
ERROR: need consistent spacing around '*' (ctx:WxB)
#8: FILE: f.c:2:
+foo(int a, my_uint32 *);
                      ^

total: 1 errors, 0 warnings, 4 lines checked

Quoted 7 lines

-apw
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#5 lore raw
Quoted 3 lines · Benny Halevy

OK, but the return type doesn't have to be in the patched line, it could be in a synchronization line or even missing if the function has a long multi-line argument list.

Ok, I guess thats fair criticism. Could you check out the current checkpatch-next (0.14-8-g3737366 or later -9, -10 etc), and see if that works. It seems to on the simple examples you sent me :).

Thanks.

-apw

#6 lore raw
Quoted 8 lines · Andy Whitcroft
Quoted 3 lines · Benny Halevy

OK, but the return type doesn't have to be in the patched line, it could be in a synchronization line or even missing if the function has a long multi-line argument list.

Ok, I guess thats fair criticism. Could you check out the current checkpatch-next (0.14-8-g3737366 or later -9, -10 etc), and see if that works. It seems to on the simple examples you sent me :).

Confirmed with 0.14-8-g3737366.

Thanks!

Benny

Oh, and I really liked the fact that you print the patch file name in the summary line of each patch checked rather than "Your patch" :)

Quoted 4 lines

Thanks.

-apw

#7 lore raw
In reply to Andy Whitcroft
Quoted 18 lines · Andy Whitcroft
Quoted 17 lines · Marcin Slusarz

Hi

Checkpatch in current mainline outputs following errors:

$ ./scripts/checkpatch.pl -q --file ./fs/udf/misc.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#205: FILE: fs/udf/misc.c:205:
+       tag *tag_p;
            ^
$ ./scripts/checkpatch.pl -q --file ./fs/udf/unicode.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#48: FILE: fs/udf/unicode.c:48:
+int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
                                               ^
(...)
Quoted 6 lines
Quoted 1 line

I think the code is ok.

Yep the code is clearly correct. Can I have the whole patch fragment you got these against so I can figure out why we are unable to detect these two as types, I would expect us to have done so. Also what version of checkpatch is this? There is a version string at the top.

It's current Linus git tree (both checkpatch and "testcase").

Marcin

#8 lore raw
In reply to Marcin Slusarz
Quoted 7 lines · Marcin Slusarz

Checkpatch in current mainline outputs following errors:

$ ./scripts/checkpatch.pl -q --file ./fs/udf/misc.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#205: FILE: fs/udf/misc.c:205:
+       tag *tag_p;
           ^

I'd say "don't add any new typedefs" (and perhaps get rid of old ones).

#9 lore raw
Quoted 11 lines · Jan Engelhardt
Quoted 7 lines · Marcin Slusarz

Checkpatch in current mainline outputs following errors:

$ ./scripts/checkpatch.pl -q --file ./fs/udf/misc.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#205: FILE: fs/udf/misc.c:205:
+       tag *tag_p;
           ^

I'd say "don't add any new typedefs" (and perhaps get rid of old ones).

It should be doing that at the site of the new typedef :)

-apw

#10 lore raw
Quoted 11 lines · Andy Whitcroft
Quoted 10 lines · Jan Engelhardt
Quoted 7 lines · Marcin Slusarz

Checkpatch in current mainline outputs following errors:

$ ./scripts/checkpatch.pl -q --file ./fs/udf/misc.c
ERROR: need consistent spacing around '*' (ctx:WxV)
#205: FILE: fs/udf/misc.c:205:
+       tag *tag_p;
           ^
Quoted 4 lines
Quoted 1 line

I'd say "don't add any new typedefs" (and perhaps get rid of old ones).

It should be doing that at the site of the new typedef :)

Additionally, yes.
Given:

typedef struct {
	...
} tag_t;
void foo(void)
{
	tag_t name;
}

A user adding

+ tag_t newthing;

could at the same time give the struct a name if it already does not have one and instead use

+ struct tag newthing;

then.