[PATCH] scripts/kernel-doc: handle object-like macros

[PATCH] scripts/kernel-doc: handle object-like macros

[Horia Geanta]

Object-like macros are different than function-like macros:
https://gcc.gnu.org/onlinedocs/cpp/Object-like-Macros.html
https://gcc.gnu.org/onlinedocs/cpp/Function-like-Macros.html

They are not parsed correctly, generating invalid intermediate
files (xmls) for cases like:
    #define BIT_MASK    (0xFF << BIT_SHIFT)
where "OxFF <<" is considered to be parameter type.

When parsing, we can differentiate beween these two types of macros by
checking whether there is at least one whitespace b/w "#define" and
first opening paranthesis.

Signed-off-by: Horia Geanta <horia.geanta@freescale.com>
---
 scripts/kernel-doc | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/scripts/kernel-doc b/scripts/kernel-doc
index da058da413e7..ffd35c9d616a 100755
--- a/scripts/kernel-doc
+++ b/scripts/kernel-doc
@@ -2073,6 +2073,7 @@ sub check_return_section {
 sub dump_function($$) {
     my $prototype = shift;
     my $file = shift;
+    my $noret = 0;
 
     $prototype =~ s/^static +//;
     $prototype =~ s/^extern +//;
@@ -2086,7 +2087,7 @@ sub dump_function($$) {
     $prototype =~ s/__init_or_module +//;
     $prototype =~ s/__must_check +//;
     $prototype =~ s/__weak +//;
-    $prototype =~ s/^#\s*define\s+//; #ak added
+    my $define = $prototype =~ s/^#\s*define\s+//; #ak added
     $prototype =~ s/__attribute__\s*\(\([a-z,]*\)\)//;
 
     # Yes, this truly is vile.  We are looking for:
@@ -2105,7 +2106,15 @@ sub dump_function($$) {
     # - atomic_set (macro)
     # - pci_match_device, __copy_to_user (long return type)
 
-    if ($prototype =~ m/^()([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
+    if ($define && $prototype =~ m/^()([a-zA-Z0-9_~:]+)\s+/) {
+        # This is an object-like macro, it has no return type and no parameter
+        # list.
+        # Function-like macros are not allowed to have spaces between
+        # declaration_name and opening paranthesis (notice the \s+).
+        $return_type = $1;
+        $declaration_name = $2;
+        $noret = 1;
+    } elsif ($prototype =~ m/^()([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
 	$prototype =~ m/^(\w+)\s+([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
 	$prototype =~ m/^(\w+\s*\*)\s*([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
 	$prototype =~ m/^(\w+\s+\w+)\s+([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
@@ -2140,7 +2149,7 @@ sub dump_function($$) {
         # of warnings goes sufficiently down, the check is only performed in
         # verbose mode.
         # TODO: always perform the check.
-        if ($verbose) {
+        if ($verbose && !$noret) {
                 check_return_section($file, $declaration_name, $return_type);
         }
 
-- 
1.8.3.1

Re: [PATCH] scripts/kernel-doc: handle object-like macros

[Randy Dunlap]

On 06/23/14 05:15, Horia Geanta wrote:
> Object-like macros are different than function-like macros:
> https://gcc.gnu.org/onlinedocs/cpp/Object-like-Macros.html
> https://gcc.gnu.org/onlinedocs/cpp/Function-like-Macros.html
> 
> They are not parsed correctly, generating invalid intermediate
> files (xmls) for cases like:
>     #define BIT_MASK    (0xFF << BIT_SHIFT)
> where "OxFF <<" is considered to be parameter type.
> 
> When parsing, we can differentiate beween these two types of macros by
> checking whether there is at least one whitespace b/w "#define" and
> first opening paranthesis.

                parenthesis.

Where did you see a problem?  I tested the patch and all of the
Documentation/DocBook/*.xml files are the same without or with the patch.


> 
> Signed-off-by: Horia Geanta <horia.geanta@freescale.com>
> ---
>  scripts/kernel-doc | 15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
> 
> diff --git a/scripts/kernel-doc b/scripts/kernel-doc
> index da058da413e7..ffd35c9d616a 100755
> --- a/scripts/kernel-doc
> +++ b/scripts/kernel-doc
> @@ -2073,6 +2073,7 @@ sub check_return_section {
>  sub dump_function($$) {
>      my $prototype = shift;
>      my $file = shift;
> +    my $noret = 0;
>  
>      $prototype =~ s/^static +//;
>      $prototype =~ s/^extern +//;
> @@ -2086,7 +2087,7 @@ sub dump_function($$) {
>      $prototype =~ s/__init_or_module +//;
>      $prototype =~ s/__must_check +//;
>      $prototype =~ s/__weak +//;
> -    $prototype =~ s/^#\s*define\s+//; #ak added
> +    my $define = $prototype =~ s/^#\s*define\s+//; #ak added
>      $prototype =~ s/__attribute__\s*\(\([a-z,]*\)\)//;
>  
>      # Yes, this truly is vile.  We are looking for:
> @@ -2105,7 +2106,15 @@ sub dump_function($$) {
>      # - atomic_set (macro)
>      # - pci_match_device, __copy_to_user (long return type)
>  
> -    if ($prototype =~ m/^()([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
> +    if ($define && $prototype =~ m/^()([a-zA-Z0-9_~:]+)\s+/) {
> +        # This is an object-like macro, it has no return type and no parameter
> +        # list.
> +        # Function-like macros are not allowed to have spaces between
> +        # declaration_name and opening paranthesis (notice the \s+).
> +        $return_type = $1;
> +        $declaration_name = $2;
> +        $noret = 1;
> +    } elsif ($prototype =~ m/^()([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
>  	$prototype =~ m/^(\w+)\s+([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
>  	$prototype =~ m/^(\w+\s*\*)\s*([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
>  	$prototype =~ m/^(\w+\s+\w+)\s+([a-zA-Z0-9_~:]+)\s*\(([^\(]*)\)/ ||
> @@ -2140,7 +2149,7 @@ sub dump_function($$) {
>          # of warnings goes sufficiently down, the check is only performed in
>          # verbose mode.
>          # TODO: always perform the check.
> -        if ($verbose) {
> +        if ($verbose && !$noret) {
>                  check_return_section($file, $declaration_name, $return_type);
>          }
>  
> 


-- 
~Randy

Re: [PATCH] scripts/kernel-doc: handle object-like macros

[Horia Geantă]

On 6/29/2014 5:52 AM, Randy Dunlap wrote:
> On 06/23/14 05:15, Horia Geanta wrote:
>> Object-like macros are different than function-like macros:
>> https://gcc.gnu.org/onlinedocs/cpp/Object-like-Macros.html
>> https://gcc.gnu.org/onlinedocs/cpp/Function-like-Macros.html
>>
>> They are not parsed correctly, generating invalid intermediate
>> files (xmls) for cases like:
>>      #define BIT_MASK    (0xFF << BIT_SHIFT)
>> where "OxFF <<" is considered to be parameter type.
>>
>> When parsing, we can differentiate beween these two types of macros by

                                      between.

>> checking whether there is at least one whitespace b/w "#define" and
>> first opening paranthesis.
>
>                  parenthesis.

Thanks, I'll fix both typos in v2.

>
> Where did you see a problem?  I tested the patch and all of the
> Documentation/DocBook/*.xml files are the same without or with the patch.
>

Indeed, right now there is no kernel-doc for "object-like" macros in 
Docbook.

I am preparing documentation for not-yet-upstreamed code in 
drivers/crypto/caam and I thought it would be useful to use kernel-doc 
for some of the #defines.

Doc would look like this (no @param list, no "Return" section, only a 
description):

/**
  * DEFAULT_SEC_ERA - the default value for the SEC era in case the user
  * provides an unsupported value.
  */
#define DEFAULT_SEC_ERA 6

Regards,
Horia

Re: [PATCH] scripts/kernel-doc: handle object-like macros

[Randy Dunlap]

On 06/30/14 06:51, Horia Geantă wrote:
> On 6/29/2014 5:52 AM, Randy Dunlap wrote:
>> On 06/23/14 05:15, Horia Geanta wrote:
>>> Object-like macros are different than function-like macros:
>>> https://gcc.gnu.org/onlinedocs/cpp/Object-like-Macros.html
>>> https://gcc.gnu.org/onlinedocs/cpp/Function-like-Macros.html
>>>
>>> They are not parsed correctly, generating invalid intermediate
>>> files (xmls) for cases like:
>>>      #define BIT_MASK    (0xFF << BIT_SHIFT)
>>> where "OxFF <<" is considered to be parameter type.
>>>
>>> When parsing, we can differentiate beween these two types of macros by
> 
>                                      between.
> 
>>> checking whether there is at least one whitespace b/w "#define" and
>>> first opening paranthesis.
>>
>>                  parenthesis.
> 
> Thanks, I'll fix both typos in v2.
> 
>>
>> Where did you see a problem?  I tested the patch and all of the
>> Documentation/DocBook/*.xml files are the same without or with the patch.
>>
> 
> Indeed, right now there is no kernel-doc for "object-like" macros in Docbook.
> 
> I am preparing documentation for not-yet-upstreamed code in drivers/crypto/caam and I thought it would be useful to use kernel-doc for some of the #defines.

OK, thanks for the explanation.  I am applying your patch.

> Doc would look like this (no @param list, no "Return" section, only a description):
> 
> /**
>  * DEFAULT_SEC_ERA - the default value for the SEC era in case the user
>  * provides an unsupported value.
>  */
> #define DEFAULT_SEC_ERA 6


-- 
~Randy