Re: patch 'telemetry: lower log level on socket error' has been queued to stable release 23.11.2

[Xueming Li <xuemingl@nvidia.com>]

[-- Attachment #1: Type: text/plain, Size: 2542 bytes --]

Hi David,

Thanks for pointing this out, fixed both patches.

Regards,
Xueming
________________________________
From: David Marchand <david.marchand@redhat.com>
Sent: Saturday, July 13, 2024 9:38 PM
To: Xueming Li <xuemingl@nvidia.com>; Luca Boccassi <bluca@debian.org>; Kevin Traynor <ktraynor@redhat.com>
Cc: Christian Ehrhardt <christian.ehrhardt@canonical.com>; Bruce Richardson <bruce.richardson@intel.com>; dpdk stable <stable@dpdk.org>
Subject: Re: patch 'telemetry: lower log level on socket error' has been queued to stable release 23.11.2

Hello LTS maintainers,

On Fri, Jul 12, 2024 at 1:04 PM Xueming Li <xuemingl@nvidia.com> wrote:
> diff --git a/lib/telemetry/telemetry.c b/lib/telemetry/telemetry.c
> index 88d6410980..551bc92e9d 100644
> --- a/lib/telemetry/telemetry.c
> +++ b/lib/telemetry/telemetry.c
> @@ -381,7 +381,7 @@ client_handler(void *sock_id)
>                         "{\"version\":\"%s\",\"pid\":%d,\"max_output_len\":%d}",
>                         telemetry_version, getpid(), MAX_OUTPUT_LEN);
>         if (write(s, info_str, strlen(info_str)) < 0) {
> -               TMTY_LOG(ERR, "Socket write base info to client failed\n");
> +               TMTY_LOG(DEBUG, "Socket write base info to client failed");

Xueming:

There is a similar issue to what I reported for EAL, for this backport.

In the main branch, we have TMTY_LOG_LINE() that appends a \n to the
format string.
But TMTY_LOG() in older branches (and as you can see in the context of
this hunk) will not do this.


To all LTS maintainers:

Many libraries have been updated with similar changes (making the \n
added by the log macros) in the main branch.
I introduced new macros every time possible when doing those changes.
The intent was that backports would trigger small conflicts and make
the person backporting raise an eyebrow :-).

For such small conflicts, some simple rules:
- if the commit from main branch uses <MACRO>_LINE(.*, format, args),
then it becomes <MACRO>(.*, format "\n", args) in the LTS branches,
- if the commit from main branch uses another macro that does not
exist in the LTS branch, there is a good chance a "\n" is missing too,
And in doubt, just look at logs in the same file. This gives a quick
hint in general.

Note: backporting those macros could be an option too, but this could
be some work as I suspect there will be a lot of conflicts.

I went and double checked the rest of the 23.11-staging branch and it
seems okay.


--
David Marchand


[-- Attachment #2: Type: text/html, Size: 5004 bytes --]