[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <83a6596b-d9c4-4f2f-9eae-fd35cae561dc@linux.alibaba.com>
Date: Mon, 3 Jun 2024 10:57:55 +0800
From: "D. Wythe" <alibuda@...ux.alibaba.com>
To: Simon Horman <horms@...nel.org>
Cc: kgraul@...ux.ibm.com, wenjia@...ux.ibm.com, jaka@...ux.ibm.com,
wintera@...ux.ibm.com, guwen@...ux.alibaba.com, kuba@...nel.org,
davem@...emloft.net, netdev@...r.kernel.org, linux-s390@...r.kernel.org,
linux-rdma@...r.kernel.org, tonylu@...ux.alibaba.com, pabeni@...hat.com,
edumazet@...gle.com
Subject: Re: [PATCH net-next v4 3/3] net/smc: Introduce IPPROTO_SMC
On 6/1/24 9:06 PM, Simon Horman wrote:
> On Wed, May 29, 2024 at 11:59:07AM +0800, D. Wythe wrote:
>> From: "D. Wythe" <alibuda@...ux.alibaba.com>
>>
>> This patch allows to create smc socket via AF_INET,
>> similar to the following code,
>>
>> /* create v4 smc sock */
>> v4 = socket(AF_INET, SOCK_STREAM, IPPROTO_SMC);
>>
>> /* create v6 smc sock */
>> v6 = socket(AF_INET6, SOCK_STREAM, IPPROTO_SMC);
>>
>> There are several reasons why we believe it is appropriate here:
>>
>> 1. For smc sockets, it actually use IPv4 (AF-INET) or IPv6 (AF-INET6)
>> address. There is no AF_SMC address at all.
>>
>> 2. Create smc socket in the AF_INET(6) path, which allows us to reuse
>> the infrastructure of AF_INET(6) path, such as common ebpf hooks.
>> Otherwise, smc have to implement it again in AF_SMC path.
>>
>> Signed-off-by: D. Wythe <alibuda@...ux.alibaba.com>
> ...
>
>> diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c
> ...
>
>> @@ -3594,9 +3595,31 @@ static int __init smc_init(void)
>> goto out_lo;
>> }
>>
>> + rc = proto_register(&smc_inet_prot, 1);
>> + if (rc) {
>> + pr_err("%s: proto_register smc_inet_prot fails with %d\n", __func__, rc);
> Hi,
>
> FWIIW, my feeling is that if a log message includes __func__ then it should
> be a debug level message, and even then I'm dubious about the value of
> __func__: we do have many tools including dynamic tracing or pinpointing
> problems.
>
> So I would suggest rephrasing this message and dropping __func__.
> Or maybe removing it entirely.
> Or if not, lowering the priority of this message to debug.
>
> If for some reason __func__ remains, please do consider wrapping
> the line to 80c columns or less, which can be trivially done here
> (please don't split the format string in any case).
>
> Flagged by checkpatch.pl --max-line-length=80
Hi Simon,
Thank you very much for your feedback.
Allow me to briefly explain the reasons for using pr_err and __func__ here.
Regarding pr_err, the failure here leads to the failure of the module
loading, which is definitely an error-level message rather than a
debug-level one.
As for __func__, I must admit that the purpose here is simply to align
with the format of other error messages in smc_init(). In fact, I also
feel that the presence of
__func__ doesn't hold significant value because this error will only
occur within this function. It's meaningless information for both users
and kernel developers.
Perhaps a more suitable format would be “smc: xxx: %d”.
However, if changes are needed, I think they should be made across the
board in order to maintain a consistent style. Maybe this can be
addressed by
submitting a new patch after this patch. @Wenjia, what do you think?
Therefore, for now, I would like to wrap this line to not exceed 80
characters, to ensure it can pass the checkpatch.pl.
What do you think?
Best wishes,
D. Wythe
>
>> + goto out_ulp;
>> + }
>> + inet_register_protosw(&smc_inet_protosw);
>> +#if IS_ENABLED(CONFIG_IPV6)
>> + rc = proto_register(&smc_inet6_prot, 1);
>> + if (rc) {
>> + pr_err("%s: proto_register smc_inet6_prot fails with %d\n", __func__, rc);
> Here too.
>
>> + goto out_inet_prot;
>> + }
>> + inet6_register_protosw(&smc_inet6_protosw);
>> +#endif
> ...
Powered by blists - more mailing lists