[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20221013082913.0719721e@kernel.org>
Date: Thu, 13 Oct 2022 08:29:13 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: <edward.cree@....com>
Cc: <netdev@...r.kernel.org>, <linux-net-drivers@....com>,
<davem@...emloft.net>, <pabeni@...hat.com>, <edumazet@...gle.com>,
<habetsm.xilinx@...il.com>, <johannes@...solutions.net>,
<marcelo.leitner@...il.com>, Edward Cree <ecree.xilinx@...il.com>
Subject: Re: [RFC PATCH v2 net-next 1/3] netlink: add support for formatted
extack messages
On Thu, 13 Oct 2022 10:23:00 +0100 edward.cree@....com wrote:
> + if (snprintf(__extack->_msg_buf, NETLINK_MAX_FMTMSG_LEN, \
> + (fmt), ##args) >= NETLINK_MAX_FMTMSG_LEN) \
> + net_warn_ratelimited("truncated extack: " fmt "\n", \
> + ##args); \
> + \
Some "take it or leave it" comments:
- Jiri's idea of always printing may be worth exploring
- my preference would also be to produce a warning on overflow,
rather than the exact print, because I always worry about people
starting to depend on the print.
And WARN_ON() is really heavy and may trigger remediations even
tho truncated extack is just a minor nuisance.
I'd do:
pr(extack formatting overflow $__FILE__:$__func__:$__LINE__ $needed_len)
(I think splicing the "trunced extack:" with fmt will result
in the format string getting stored in .ro twice?)
Powered by blists - more mailing lists