[nfs-utils PATCH 0/7] silence some warning in rpcgen

[nfs-utils PATCH 0/7] silence some warning in rpcgen

[Giulio Benetti]

Since I'm trying to bump version of nfs-utils to latest in Buildroot, I've
noticed some warning in rpcgen, so I've decided to clean them up by fixing
code or #pragma ignoring them. Hope this is useful. Other warnings are
still there waiting to be fixed and if you find these patches useful I'm
going to complete all warning correction.

Giulio Benetti (7):
  rpcgen: rpc_cout: silence unused def parameter
  rpcgen: rpc_util: add storeval args to prototype
  rpcgen: rpc_util: add findval args to prototype
  rpcgen: rpc_parse: add get_definition() void argument
  rpcgen: rpc_cout: fix potential -Wformat-nonliteral warning
  rpcgen: rpc_hout: fix potential -Wformat-security warning
  rpcgen: rpc_hout: fix indentation on f_print() argument separator

 tools/rpcgen/rpc_cout.c  | 8 ++++----
 tools/rpcgen/rpc_hout.c  | 4 +++-
 tools/rpcgen/rpc_parse.h | 2 +-
 tools/rpcgen/rpc_util.h  | 4 ++--
 4 files changed, 10 insertions(+), 8 deletions(-)

-- 
2.20.1

[nfs-utils PATCH 1/7] rpcgen: rpc_cout: silence unused def parameter

[Giulio Benetti]

In emit_enum() argument def is not used and can cause a warning. So
let's mark it with __attribute__((unused)).

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
---
 tools/rpcgen/rpc_cout.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/rpcgen/rpc_cout.c b/tools/rpcgen/rpc_cout.c
index a61214fd..f806a86a 100644
--- a/tools/rpcgen/rpc_cout.c
+++ b/tools/rpcgen/rpc_cout.c
@@ -53,7 +53,7 @@ static void	print_ifsizeof(char *prefix, char *type);
 static void	print_ifclose(int indent);
 static void	print_ifstat(int indent, char *prefix, char *type, relation rel,
 			char *amax, char *objname, char *name);
-static void	emit_enum(definition *def);
+static void	emit_enum(__attribute__((unused)) definition *def);
 static void	emit_program(definition *def);
 static void	emit_union(definition *def);
 static void	emit_struct(definition *def);
@@ -286,7 +286,7 @@ print_ifstat(int indent, char *prefix, char *type, relation rel,
 }
 
 static void
-emit_enum(definition *def)
+emit_enum(__attribute__((unused)) definition *def)
 {
 	print_ifopen(1, "enum");
 	print_ifarg("(enum_t *)objp");
-- 
2.20.1

[nfs-utils PATCH 2/7] rpcgen: rpc_util: add storeval args to prototype

[Giulio Benetti]

storeval() prototype has no arguments and this can cause warnings during
building. Let's add its arguments to prototype according to its
implementation.

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
---
 tools/rpcgen/rpc_util.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/rpcgen/rpc_util.h b/tools/rpcgen/rpc_util.h
index fa115bee..bd7b15ca 100644
--- a/tools/rpcgen/rpc_util.h
+++ b/tools/rpcgen/rpc_util.h
@@ -91,7 +91,7 @@ extern int nonfatalerrors;
 /*
  * rpc_util routines 
  */
-void storeval();
+void storeval(list **, definition *);
 
 #define STOREVAL(list,item)	\
 	storeval(list,item)
-- 
2.20.1

[nfs-utils PATCH 3/7] rpcgen: rpc_util: add findval args to prototype

[Giulio Benetti]

findval() prototype has no arguments and this can cause warnings during
building. Let's add its arguments to prototype according to its
implementation.

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
---
 tools/rpcgen/rpc_util.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/rpcgen/rpc_util.h b/tools/rpcgen/rpc_util.h
index bd7b15ca..97b87f2b 100644
--- a/tools/rpcgen/rpc_util.h
+++ b/tools/rpcgen/rpc_util.h
@@ -96,7 +96,7 @@ void storeval(list **, definition *);
 #define STOREVAL(list,item)	\
 	storeval(list,item)
 
-definition *findval();
+definition *findval(list *, char *, int (*)(definition *, char *));
 
 #define FINDVAL(list,item,finder) \
 	findval(list, item, finder)
-- 
2.20.1

[nfs-utils PATCH 4/7] rpcgen: rpc_parse: add get_definition() void argument

[Giulio Benetti]

get_definition() prototype has no arguments and this can cause warnings
during building. Let's add void argument to prototype according to its
implementation.

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
---
 tools/rpcgen/rpc_parse.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/rpcgen/rpc_parse.h b/tools/rpcgen/rpc_parse.h
index 2afae104..6c134dd8 100644
--- a/tools/rpcgen/rpc_parse.h
+++ b/tools/rpcgen/rpc_parse.h
@@ -153,7 +153,7 @@ struct definition {
 };
 typedef struct definition definition;
 
-definition *get_definition();
+definition *get_definition(void);
 
 
 struct bas_type
-- 
2.20.1

[nfs-utils PATCH 5/7] rpcgen: rpc_cout: fix potential -Wformat-nonliteral warning

[Giulio Benetti]

format and vecformat must be declared as "char * const" to be really
treated as constant when building with -Werror=format-nonliteral,
otherwise compiler will consider them subject to change throughout the
function.

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
---
 tools/rpcgen/rpc_cout.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/rpcgen/rpc_cout.c b/tools/rpcgen/rpc_cout.c
index f806a86a..df2609c4 100644
--- a/tools/rpcgen/rpc_cout.c
+++ b/tools/rpcgen/rpc_cout.c
@@ -319,8 +319,8 @@ emit_union(definition *def)
   case_list *cl;
   declaration *cs;
   char *object;
-  char *vecformat = "objp->%s_u.%s";
-  char *format = "&objp->%s_u.%s";
+  char * const vecformat = "objp->%s_u.%s";
+  char * const format = "&objp->%s_u.%s";
 
   print_stat(1,&def->def.un.enum_decl);
   f_print(fout, "\tswitch (objp->%s) {\n", def->def.un.enum_decl.name);
-- 
2.20.1

[nfs-utils PATCH 6/7] rpcgen: rpc_hout: fix potential -Wformat-security warning

[Giulio Benetti]

f_print()'s argument "separator" is not known because it's passed as an
argument and with -Wformat-security will cause a useless warning. Let's
ignore by adding "#pragma GCC diagnostic ignored/warning" before and
after f_print().

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
---
 tools/rpcgen/rpc_hout.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/tools/rpcgen/rpc_hout.c b/tools/rpcgen/rpc_hout.c
index ea1cb24f..999c061f 100644
--- a/tools/rpcgen/rpc_hout.c
+++ b/tools/rpcgen/rpc_hout.c
@@ -467,7 +467,9 @@ pdeclaration(char *name, declaration *dec, int tab, char *separator)
 			break;
 		}
 	}
+#pragma GCC diagnostic ignored "-Wformat-security"
 	f_print(fout, separator );
+#pragma GCC diagnostic warning "-Wformat-security"
 }
 
 static int
-- 
2.20.1

[nfs-utils PATCH 7/7] rpcgen: rpc_hout: fix indentation on f_print() argument separator

[Giulio Benetti]

Remove useless space before closing parenthesys.

Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
---
 tools/rpcgen/rpc_hout.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/rpcgen/rpc_hout.c b/tools/rpcgen/rpc_hout.c
index 999c061f..ed668778 100644
--- a/tools/rpcgen/rpc_hout.c
+++ b/tools/rpcgen/rpc_hout.c
@@ -468,7 +468,7 @@ pdeclaration(char *name, declaration *dec, int tab, char *separator)
 		}
 	}
 #pragma GCC diagnostic ignored "-Wformat-security"
-	f_print(fout, separator );
+	f_print(fout, separator);
 #pragma GCC diagnostic warning "-Wformat-security"
 }
 
-- 
2.20.1

Re: [nfs-utils PATCH 0/7] silence some warning in rpcgen

[Steve Dickson]

On 1/3/20 4:50 PM, Giulio Benetti wrote:
> Since I'm trying to bump version of nfs-utils to latest in Buildroot, I've
> noticed some warning in rpcgen, so I've decided to clean them up by fixing
> code or #pragma ignoring them. Hope this is useful. Other warnings are
> still there waiting to be fixed and if you find these patches useful I'm
> going to complete all warning correction.
> 
> Giulio Benetti (7):
>   rpcgen: rpc_cout: silence unused def parameter
>   rpcgen: rpc_util: add storeval args to prototype
>   rpcgen: rpc_util: add findval args to prototype
>   rpcgen: rpc_parse: add get_definition() void argument
>   rpcgen: rpc_cout: fix potential -Wformat-nonliteral warning
>   rpcgen: rpc_hout: fix potential -Wformat-security warning
>   rpcgen: rpc_hout: fix indentation on f_print() argument separator
> 
>  tools/rpcgen/rpc_cout.c  | 8 ++++----
>  tools/rpcgen/rpc_hout.c  | 4 +++-
>  tools/rpcgen/rpc_parse.h | 2 +-
>  tools/rpcgen/rpc_util.h  | 4 ++--
>  4 files changed, 10 insertions(+), 8 deletions(-)
> 
Committed (tag: nfs-utils-2-4-3-rc5)

I must admit this code is actually being used... I assume they do the right thing...

The rpcgen we been using is the old one that came out 
of the glibc code at https://github.com/thkukuk/rpcsvc-proto

I wonder what the difference is.... 

steved.

Re: [nfs-utils PATCH 0/7] silence some warning in rpcgen

[Giulio Benetti]

On 1/7/20 8:06 PM, Steve Dickson wrote:
> 
> 
> On 1/3/20 4:50 PM, Giulio Benetti wrote:
>> Since I'm trying to bump version of nfs-utils to latest in Buildroot, I've
>> noticed some warning in rpcgen, so I've decided to clean them up by fixing
>> code or #pragma ignoring them. Hope this is useful. Other warnings are
>> still there waiting to be fixed and if you find these patches useful I'm
>> going to complete all warning correction.
>>
>> Giulio Benetti (7):
>>    rpcgen: rpc_cout: silence unused def parameter
>>    rpcgen: rpc_util: add storeval args to prototype
>>    rpcgen: rpc_util: add findval args to prototype
>>    rpcgen: rpc_parse: add get_definition() void argument
>>    rpcgen: rpc_cout: fix potential -Wformat-nonliteral warning
>>    rpcgen: rpc_hout: fix potential -Wformat-security warning
>>    rpcgen: rpc_hout: fix indentation on f_print() argument separator
>>
>>   tools/rpcgen/rpc_cout.c  | 8 ++++----
>>   tools/rpcgen/rpc_hout.c  | 4 +++-
>>   tools/rpcgen/rpc_parse.h | 2 +-
>>   tools/rpcgen/rpc_util.h  | 4 ++--
>>   4 files changed, 10 insertions(+), 8 deletions(-)
>>
> Committed (tag: nfs-utils-2-4-3-rc5)
> 
> I must admit this code is actually being used... I assume they do the right thing...
> 
> The rpcgen we been using is the old one that came out
> of the glibc code at https://github.com/thkukuk/rpcsvc-proto
> 
> I wonder what the difference is....

I can check it and use that one as upstream maybe and update it here in 
nfs-utils if you see that it makes sense.

Best regards
-- 
Giulio Benetti
Benetti Engineering sas

> steved.
> 

Re: [nfs-utils PATCH 0/7] silence some warning in rpcgen

[Steve Dickson]

On 1/9/20 9:00 AM, Giulio Benetti wrote:
> On 1/7/20 8:06 PM, Steve Dickson wrote:
>>
>>
>> On 1/3/20 4:50 PM, Giulio Benetti wrote:
>>> Since I'm trying to bump version of nfs-utils to latest in Buildroot, I've
>>> noticed some warning in rpcgen, so I've decided to clean them up by fixing
>>> code or #pragma ignoring them. Hope this is useful. Other warnings are
>>> still there waiting to be fixed and if you find these patches useful I'm
>>> going to complete all warning correction.
>>>
>>> Giulio Benetti (7):
>>>    rpcgen: rpc_cout: silence unused def parameter
>>>    rpcgen: rpc_util: add storeval args to prototype
>>>    rpcgen: rpc_util: add findval args to prototype
>>>    rpcgen: rpc_parse: add get_definition() void argument
>>>    rpcgen: rpc_cout: fix potential -Wformat-nonliteral warning
>>>    rpcgen: rpc_hout: fix potential -Wformat-security warning
>>>    rpcgen: rpc_hout: fix indentation on f_print() argument separator
>>>
>>>   tools/rpcgen/rpc_cout.c  | 8 ++++----
>>>   tools/rpcgen/rpc_hout.c  | 4 +++-
>>>   tools/rpcgen/rpc_parse.h | 2 +-
>>>   tools/rpcgen/rpc_util.h  | 4 ++--
>>>   4 files changed, 10 insertions(+), 8 deletions(-)
>>>
>> Committed (tag: nfs-utils-2-4-3-rc5)
>>
>> I must admit this code is actually being used... I assume they do the right thing...
>>
>> The rpcgen we been using is the old one that came out
>> of the glibc code at https://github.com/thkukuk/rpcsvc-proto
>>
>> I wonder what the difference is....
> 
> I can check it and use that one as upstream maybe and update it here in nfs-utils if you see that it makes sense.
That would be interest... If they both generate the same
code... two are probably not needed... 

but I bet either code base as not changed in 40 yrs ;-) 

steved.

> 
> Best regards

Re: [nfs-utils PATCH 5/7] rpcgen: rpc_cout: fix potential -Wformat-nonliteral warning

[Giulio Benetti]

Hi Steve,

you've missed this patch while applying the series. Can you please 
commit it?

Thank you
Kind regards
-- 
Giulio Benetti
Benetti Engineering sas

On 1/3/20 10:50 PM, Giulio Benetti wrote:
> format and vecformat must be declared as "char * const" to be really
> treated as constant when building with -Werror=format-nonliteral,
> otherwise compiler will consider them subject to change throughout the
> function.
> 
> Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
> ---
>   tools/rpcgen/rpc_cout.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/tools/rpcgen/rpc_cout.c b/tools/rpcgen/rpc_cout.c
> index f806a86a..df2609c4 100644
> --- a/tools/rpcgen/rpc_cout.c
> +++ b/tools/rpcgen/rpc_cout.c
> @@ -319,8 +319,8 @@ emit_union(definition *def)
>     case_list *cl;
>     declaration *cs;
>     char *object;
> -  char *vecformat = "objp->%s_u.%s";
> -  char *format = "&objp->%s_u.%s";
> +  char * const vecformat = "objp->%s_u.%s";
> +  char * const format = "&objp->%s_u.%s";
>   
>     print_stat(1,&def->def.un.enum_decl);
>     f_print(fout, "\tswitch (objp->%s) {\n", def->def.un.enum_decl.name);
> 

Re: [nfs-utils PATCH 5/7] rpcgen: rpc_cout: fix potential -Wformat-nonliteral warning

[Steve Dickson]

On 1/15/20 11:29 AM, Giulio Benetti wrote:
> Hi Steve,
> 
> you've missed this patch while applying the series. Can you please commit it?
It is in...

commit 6f4568f1f7395f967cc03995dcfb79a1ac5c11cd
Author: Giulio Benetti <giulio.benetti@benettiengineering.com>
Date:   Mon Jan 6 14:23:04 2020 -0500

    rpcgen: rpc_hout: fix potential -Wformat-security warning
    
    f_print()'s argument "separator" is not known because it's passed as an
    argument and with -Wformat-security will cause a useless warning. Let's
    ignore by adding "#pragma GCC diagnostic ignored/warning" before and
    after f_print().
    
    Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
    Signed-off-by: Steve Dickson <steved@redhat.com>

what am I missing?

steved.

Re: [nfs-utils PATCH 5/7] rpcgen: rpc_cout: fix potential -Wformat-nonliteral warning

[Giulio Benetti]

On 1/16/20 9:07 PM, Steve Dickson wrote:
> 
> 
> On 1/15/20 11:29 AM, Giulio Benetti wrote:
>> Hi Steve,
>>
>> you've missed this patch while applying the series. Can you please commit it?
> It is in...
> 
> commit 6f4568f1f7395f967cc03995dcfb79a1ac5c11cd
> Author: Giulio Benetti <giulio.benetti@benettiengineering.com>
> Date:   Mon Jan 6 14:23:04 2020 -0500
> 
>      rpcgen: rpc_hout: fix potential -Wformat-security warning
>      
>      f_print()'s argument "separator" is not known because it's passed as an
>      argument and with -Wformat-security will cause a useless warning. Let's
>      ignore by adding "#pragma GCC diagnostic ignored/warning" before and
>      after f_print().
>      
>      Signed-off-by: Giulio Benetti <giulio.benetti@benettiengineering.com>
>      Signed-off-by: Steve Dickson <steved@redhat.com>
> 
> what am I missing?

Ah you've merged them together, now I see:
[5/7] https://patchwork.kernel.org/patch/11317493/
[6/7] https://patchwork.kernel.org/patch/11317489/

I didn't notice it while pull rebasing from upstream, because I have 
another patch similar to that after bumping to latest version of rpcgen.

Sorry for the noise and thank you!
Best regards
-- 
Giulio Benetti
Benetti Engineering sas