Received-SPF: pass (google.com: domain of linux-kernel+bounces-133538-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Date: Fri, 5 Apr 2024 20:51:17 +0100
From: Simon Horman <horms@kernel.org>
To: Petr Tesarik <petr@tesarici.cz>
Cc: "David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	open list <netdev@vger.kernel.org>,
	Linux regressions mailing list <regressions@lists.linux.dev>,
	open list <linux-kernel@vger.kernel.org>,
	"Linux regression tracking (Thorsten Leemhuis)" <regressions@leemhuis.info>
Subject: Re: [PATCH net] u64_stats: fix u64_stats_init() for lockdep when
 used repeatedly in one file
Message-ID: <20240405195117.GV26556@kernel.org>
References: <20240404075740.30682-1-petr@tesarici.cz>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240404075740.30682-1-petr@tesarici.cz>

On Thu, Apr 04, 2024 at 09:57:40AM +0200, Petr Tesarik wrote:
> Fix bogus lockdep warnings if multiple u64_stats_sync variables are
> initialized in the same file.
> 
> With CONFIG_LOCKDEP, seqcount_init() is a macro which declares:
> 
> 	static struct lock_class_key __key;
> 
> Since u64_stats_init() is a function (albeit an inline one), all calls
> within the same file end up using the same instance, effectively treating
> them all as a single lock-class.
> 
> Fixes: 9464ca650008 ("net: make u64_stats_init() a function")
> Closes: https://lore.kernel.org/netdev/ea1567d9-ce66-45e6-8168-ac40a47d1821@roeck-us.net/
> Signed-off-by: Petr Tesarik <petr@tesarici.cz>

Reviewed-by: Simon Horman <horms@kernel.org>

Interesting bug. I'm wondering if you also looked over other users of
u64_stats_init() to see if any of them can result in unexpected aliasing of
lock keys too.

..