lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250814152707.6d16c342@kernel.org>
Date: Thu, 14 Aug 2025 15:27:07 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Stanislav Fomichev <sdf@...ichev.me>
Cc: netdev@...r.kernel.org, davem@...emloft.net, edumazet@...gle.com,
 pabeni@...hat.com, donald.hunter@...il.com, horms@...nel.org,
 jstancek@...hat.com, jacob.e.keller@...el.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next] tools: ynl: make ynl.c more c++ friendly

On Thu, 14 Aug 2025 09:44:13 -0700 Stanislav Fomichev wrote:
> Compiling ynl.c in a C++ code base requires invoking C compiler and
> using extern "C" for the headers. To make it easier, we can add
> small changes to the ynl.c file to make it palatable to the native
> C++ compiler. The changes are:
> - avoid using void* pointer arithmetic, use char* instead
> - avoid implicit void* type casts, add c-style explicit casts
> - avoid implicit int->enum type casts, add c-style explicit casts
> - avoid anonymous structs (for type casts)
> - namespacify cpp version, this should let us compile both ynl.c
>   as c and ynl.c as cpp in the same binary (YNL_CPP can be used
>   to enable/disable namespacing)
> 
> Also add test_cpp rule to make sure ynl.c won't break C++ in the future.

As I mentioned in person, ynl-cpp is a separate thing, and you'd all
benefit from making it more C++ than going the other way and massaging
YNL C.

With that said, commenting below on the few that I think would be okay.

> @@ -224,7 +228,7 @@ static inline void *ynl_attr_data_end(const struct nlattr *attr)
>  
>  #define ynl_attr_for_each_payload(start, len, attr)			\
>  	for ((attr) = ynl_attr_first(start, len, 0); attr;		\
> -	     (attr) = ynl_attr_next(start + len, attr))
> +	     (attr) = ynl_attr_next((char *)start + len, attr))

okay

> @@ -149,7 +153,7 @@ ynl_err_walk(struct ynl_sock *ys, void *start, void *end, unsigned int off,
>  		return n;
>  	}
>  
> -	data_len = end - start;
> +	data_len = (char *)end - (char *)start;

can we make the arguments char * instead of the casts?

>  static void ynl_err_reset(struct ynl_sock *ys)
>  {
> -	ys->err.code = 0;
> +	ys->err.code = YNL_ERROR_NONE;

sure

> @@ -56,6 +60,11 @@ struct ynl_family {
>  	unsigned int ntf_info_size;
>  };
>  
> +struct ynl_sock_mcast {

struct ynl_mcast_grp

> +	unsigned int id;
> +	char name[GENL_NAMSIZ];
> +};

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ