Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH net-next v2 2/2] bonding: combine netlink and console
 error messages
To:     Joe Perches <joe@perches.com>, Jakub Kicinski <kuba@kernel.org>
Cc:     netdev@vger.kernel.org, leon@kernel.org,
        Jay Vosburgh <j.vosburgh@gmail.com>,
        Veaceslav Falico <vfalico@gmail.com>,
        Andy Gospodarek <andy@greyhouse.net>,
        "David S. Miller" <davem@davemloft.net>,
        linux-kernel@vger.kernel.org
References: <cover.1628650079.git.jtoppins@redhat.com>
 <e6b78ce8f5904a5411a809cf4205d745f8af98cb.1628650079.git.jtoppins@redhat.com>
 <d5e1aada694465fd62f57695e264259815e60746.camel@perches.com>
 <20210811054917.722bd988@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
 <c296dd2f66e97ad38d5456c0fab4e0ff99b14634.camel@perches.com>
From:   Jonathan Toppins <jtoppins@redhat.com>
Message-ID: <5100eeab-dd01-d739-95f0-0a0267652bae@redhat.com>
Date:   Wed, 11 Aug 2021 10:31:20 -0400
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.12.0
MIME-Version: 1.0
In-Reply-To: <c296dd2f66e97ad38d5456c0fab4e0ff99b14634.camel@perches.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 8/11/21 9:23 AM, Joe Perches wrote:
> On Wed, 2021-08-11 at 05:49 -0700, Jakub Kicinski wrote:
>> On Tue, 10 Aug 2021 20:27:01 -0700 Joe Perches wrote:
>>>> +#define BOND_NL_ERR(bond_dev, extack, errmsg) do {		\
>>>> +	if (extack)						\
>>>> +		NL_SET_ERR_MSG(extack, errmsg);			\
>>>> +	else							\
>>>> +		netdev_err(bond_dev, "Error: %s\n", errmsg);	\
>>>> +} while (0)
>>>> +
>>>> +#define SLAVE_NL_ERR(bond_dev, slave_dev, extack, errmsg) do {		\
>>>> +	if (extack)							\
>>>> +		NL_SET_ERR_MSG(extack, errmsg);				\
>>>> +	else								\
>>>> +		slave_err(bond_dev, slave_dev, "Error: %s\n", errmsg);	\
>>>> +} while (0)
>>>
>>> Ideally both of these would be static functions and not macros.
>>
>> That may break our ability for NL_SET_ERR_MSG to place strings
>> back in a static buffer, no?
> 
> Not really.
> 
> The most common way to place things in a particular section is to
> use __section("whatever")
> 
> It's pretty trivial to mark these errmsg strings as above.

I am unable to convert these to functions at this time, due to how 
NL_SET_ERR_MSG is expanded. This is with either a param list prototype 
of, "const char *errmsg" or "const char errmsg[]".

$ make C=1 drivers/net/bonding/bonding.ko
   CALL    scripts/checksyscalls.sh
   CALL    scripts/atomic/check-atomics.sh
   DESCEND objtool
   DESCEND bpf/resolve_btfids
   CC [M]  drivers/net/bonding/bond_main.o
In file included from ./include/uapi/linux/neighbour.h:6,
                  from ./include/linux/netdevice.h:45,
                  from ./include/net/inet_sock.h:19,
                  from ./include/net/ip.h:28,
                  from drivers/net/bonding/bond_main.c:42:
drivers/net/bonding/bond_main.c: In function ‘bond_nl_err’:
drivers/net/bonding/bond_main.c:1733:26: error: invalid initializer
    NL_SET_ERR_MSG(extack, errmsg);
                           ^~~~~~
./include/linux/netlink.h:92:30: note: in definition of macro 
‘NL_SET_ERR_MSG’
   static const char __msg[] = msg;  \
                               ^~~
drivers/net/bonding/bond_main.c: In function ‘slave_nl_err’:
drivers/net/bonding/bond_main.c:1744:26: error: invalid initializer
    NL_SET_ERR_MSG(extack, errmsg);
                           ^~~~~~
./include/linux/netlink.h:92:30: note: in definition of macro 
‘NL_SET_ERR_MSG’
   static const char __msg[] = msg;  \
                               ^~~
make[3]: *** [scripts/Makefile.build:271: 
drivers/net/bonding/bond_main.o] Error 1
make[2]: *** [scripts/Makefile.build:514: drivers/net/bonding] Error 2
make[1]: *** [scripts/Makefile.build:514: drivers/net] Error 2
make: *** [Makefile:1841: drivers] Error 2