On Thu, Oct 24, 2024 at 07:57:40PM +0200, Pablo Neira Ayuso wrote:
> Hi,
>
> On Thu, Oct 24, 2024 at 06:56:43PM +0200, Matthieu Baerts wrote:
> > Hello,
> >
> > First, thank you for all the work you did and are still doing around
> > Netfilter!
> >
> > I'm writing you this email, because when I run the MPTCP test suite with
> > a VM running a kernel built with a debug config including
> > CONFIG_PROVE_RCU_LIST=y (and CONFIG_RCU_EXPERT=y), I get the following
> > warning:
> >
> >
> > > =============================
> > > WARNING: suspicious RCU usage
> > > 6.12.0-rc3+ #7 Not tainted
> > > -----------------------------
> > > net/netfilter/nf_tables_api.c:3420 RCU-list traversed in non-reader section!!
> > >
> > > other info that might help us debug this:
> > >
> > >
> > > rcu_scheduler_active = 2, debug_locks = 1
> > > 1 lock held by iptables/134:
> > > #0: ffff888008c4fcc8 (&nft_net->commit_mutex){+.+.}-{3:3}, at: nf_tables_valid_genid (include/linux/jiffies.h:101) nf_tables
> > >
> > > stack backtrace:
> > > CPU: 1 UID: 0 PID: 134 Comm: iptables Not tainted 6.12.0-rc3+ #7
> > > Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
> > > Call Trace:
> > > <TASK>
> > > dump_stack_lvl (lib/dump_stack.c:123)
> > > lockdep_rcu_suspicious (kernel/locking/lockdep.c:6822)
> > > __nft_rule_lookup (net/netfilter/nf_tables_api.c:3420 (discriminator 7)) nf_tables
>
> This is a _rcu notation which is not correct, while mutex is held, it
> was introduced here:
>
> d9adf22a2918 ("netfilter: nf_tables: use call_rcu in netlink dumps")
Hm. There is also:
3cb03edb4de3 ("netfilter: nf_tables: Add locking for NFT_MSG_GETRULE_RESET requests")
this comment below is also not valid anymore:
/* called with rcu_read_lock held */
static struct sk_buff *
nf_tables_getrule_single(u32 portid, const struct nfnl_info *info,
const struct nlattr * const nla[], bool reset)
This is not the only spot that can trigger rcu splats.
