diff options
| author | Jakub Kicinski <kuba@kernel.org> | 2026-02-26 10:20:47 -0800 |
|---|---|---|
| committer | Jakub Kicinski <kuba@kernel.org> | 2026-03-05 12:11:05 -0800 |
| commit | 0b1324cdd8de9f54f9daf689a4ae59783c333510 (patch) | |
| tree | 991952409fe93bd0ea73f8683b497e31bfe6952f /net/rds | |
| parent | ab99e1167293726f016360a5dbbbc10386fd7d66 (diff) | |
| parent | abacaf559950eec0d99d37ff6b92049409af5943 (diff) | |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Cross-merge networking fixes after downstream PR (net-7.0-rc3).
No conflicts.
Adjacent changes:
net/netfilter/nft_set_rbtree.c
fb7fb4016300 ("netfilter: nf_tables: clone set on flush only")
3aea466a4399 ("netfilter: nft_set_rbtree: don't disable bh when acquiring tree lock")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Diffstat (limited to 'net/rds')
| -rw-r--r-- | net/rds/tcp.c | 14 |
1 files changed, 10 insertions, 4 deletions
diff --git a/net/rds/tcp.c b/net/rds/tcp.c index 04f310255692..654e23d13e3d 100644 --- a/net/rds/tcp.c +++ b/net/rds/tcp.c @@ -490,18 +490,24 @@ bool rds_tcp_tune(struct socket *sock) struct rds_tcp_net *rtn; tcp_sock_set_nodelay(sock->sk); - lock_sock(sk); /* TCP timer functions might access net namespace even after * a process which created this net namespace terminated. */ if (!sk->sk_net_refcnt) { - if (!maybe_get_net(net)) { - release_sock(sk); + if (!maybe_get_net(net)) return false; - } + /* + * sk_net_refcnt_upgrade() must be called before lock_sock() + * because it does a GFP_KERNEL allocation, which can trigger + * fs_reclaim and create a circular lock dependency with the + * socket lock. The fields it modifies (sk_net_refcnt, + * ns_tracker) are not accessed by any concurrent code path + * at this point. + */ sk_net_refcnt_upgrade(sk); put_net(net); } + lock_sock(sk); rtn = net_generic(net, rds_tcp_netid); if (rtn->sndbuf_size > 0) { sk->sk_sndbuf = rtn->sndbuf_size; |