From: Florian Westphal <fw@strlen.de>
To: <netfilter-devel@vger.kernel.org>
Cc: Florian Westphal <fw@strlen.de>
Subject: [PATCH nf-next 6/7] netfilter: nf_tables: must hold rcu read lock while iterating expression type list
Date: Fri, 25 Oct 2024 15:32:23 +0200 [thread overview]
Message-ID: <20241025133230.22491-7-fw@strlen.de> (raw)
In-Reply-To: <20241025133230.22491-1-fw@strlen.de>
nft shell tests trigger:
WARNING: suspicious RCU usage
net/netfilter/nf_tables_api.c:3125 RCU-list traversed in non-reader section!!
1 lock held by nft/2068:
#0: ffff888106c6f8c8 (&nft_net->commit_mutex){+.+.}-{4:4}, at: nf_tables_valid_genid+0x3c/0xf0
But the transaction mutex doesn't protect this list, the nfnl subsystem
mutex would, but we can't acquire it here without risk of ABBA
deadlocks.
Acquire the rcu read lock and also reject any types that reside in
modules, we'd need to extend this to hold a module reference.
Fixes: 3a07327d10a0 ("netfilter: nft_inner: support for inner tunnel header matching")
Signed-off-by: Florian Westphal <fw@strlen.de>
---
net/netfilter/nf_tables_api.c | 24 ++++++++++++++++++------
1 file changed, 18 insertions(+), 6 deletions(-)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 3b5154f2dd79..3db4457f46a2 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -3296,25 +3296,37 @@ int nft_expr_inner_parse(const struct nft_ctx *ctx, const struct nlattr *nla,
if (!tb[NFTA_EXPR_DATA] || !tb[NFTA_EXPR_NAME])
return -EINVAL;
+ rcu_read_lock();
+
type = __nft_expr_type_get(ctx->family, tb[NFTA_EXPR_NAME]);
- if (!type)
- return -ENOENT;
+ if (!type) {
+ err = -ENOENT;
+ goto out_unlock;
+ }
- if (!type->inner_ops)
- return -EOPNOTSUPP;
+ if (!type->inner_ops || !type->owner) {
+ err = -EOPNOTSUPP;
+ goto out_unlock;
+ }
err = nla_parse_nested_deprecated(info->tb, type->maxattr,
tb[NFTA_EXPR_DATA],
type->policy, NULL);
if (err < 0)
- goto err_nla_parse;
+ goto out_unlock;
info->attr = nla;
info->ops = type->inner_ops;
+ /* At this time only meta and payload are supported, both are
+ * builtin. Therefore we can get away without a module
+ * reference.
+ */
+ rcu_read_unlock();
return 0;
-err_nla_parse:
+out_unlock:
+ rcu_read_unlock();
return err;
}
--
2.45.2
next prev parent reply other threads:[~2024-10-25 13:54 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-25 13:32 [PATCH nf-next 0/7] netfilter: nf_tables: avoid PROVE_RCU_LIST splats Florian Westphal
2024-10-25 13:32 ` [PATCH nf-next 1/7] netfilter: nf_tables: avoid false-positive lockdep splat on rule deletion Florian Westphal
2024-10-25 15:09 ` Matthieu Baerts
2024-10-26 9:35 ` Florian Westphal
2024-10-26 11:20 ` Matthieu Baerts
2024-10-25 13:32 ` [PATCH nf-next 2/7] netfilter: nf_tables: avoid false-positive lockdep splats with sets Florian Westphal
2024-10-25 13:32 ` [PATCH nf-next 3/7] netfilter: nf_tables: avoid false-positive lockdep splats with flowtables Florian Westphal
2024-10-25 13:32 ` [PATCH nf-next 4/7] netfilter: nf_tables: avoid false-positive lockdep splats in set walker Florian Westphal
2024-10-25 13:32 ` [PATCH nf-next 5/7] netfilter: nf_tables: avoid false-positive lockdep splats with basechain hook Florian Westphal
2024-10-25 14:40 ` Phil Sutter
2024-10-25 13:32 ` Florian Westphal [this message]
2024-10-26 9:39 ` [PATCH nf-next 6/7] netfilter: nf_tables: must hold rcu read lock while iterating expression type list Florian Westphal
2024-10-25 13:32 ` [PATCH nf-next 7/7] netfilter: nf_tables: must hold rcu read lock while iterating object " Florian Westphal
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20241025133230.22491-7-fw@strlen.de \
--to=fw@strlen.de \
--cc=netfilter-devel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.