From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C7BFD15AE3 for ; Wed, 9 Aug 2023 10:46:49 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 4A095C433C7; Wed, 9 Aug 2023 10:46:49 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1691578009; bh=86G+5AeVL5igzoEb2318b/GYvKv5NH01/jGTFknCEoE=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=1n8hHOX3rTZfi5U+43p4bLWBhCmTh9Nr5RkCGcH6yeDx8a0EKLxpyemGhhOT13kCF zl1X3ySkrzZTKx9xlhaEwxTu04pVEioFykY/n2/n1tIxB+lD2WB1IZLDeuLwBArfC8 er9mTfpK5CZ/uLuqvUFLp3KY+Qd83xXfzIpGC6AM= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Tomas Glozar , Jakub Sitnicki , John Fastabend , Paolo Abeni , Sasha Levin Subject: [PATCH 6.4 074/165] bpf: sockmap: Remove preempt_disable in sock_map_sk_acquire Date: Wed, 9 Aug 2023 12:40:05 +0200 Message-ID: <20230809103645.219963459@linuxfoundation.org> X-Mailer: git-send-email 2.41.0 In-Reply-To: <20230809103642.720851262@linuxfoundation.org> References: <20230809103642.720851262@linuxfoundation.org> User-Agent: quilt/0.67 Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Tomas Glozar [ Upstream commit 13d2618b48f15966d1adfe1ff6a1985f5eef40ba ] Disabling preemption in sock_map_sk_acquire conflicts with GFP_ATOMIC allocation later in sk_psock_init_link on PREEMPT_RT kernels, since GFP_ATOMIC might sleep on RT (see bpf: Make BPF and PREEMPT_RT co-exist patchset notes for details). This causes calling bpf_map_update_elem on BPF_MAP_TYPE_SOCKMAP maps to BUG (sleeping function called from invalid context) on RT kernels. preempt_disable was introduced together with lock_sk and rcu_read_lock in commit 99ba2b5aba24e ("bpf: sockhash, disallow bpf_tcp_close and update in parallel"), probably to match disabled migration of BPF programs, and is no longer necessary. Remove preempt_disable to fix BUG in sock_map_update_common on RT. Signed-off-by: Tomas Glozar Reviewed-by: Jakub Sitnicki Link: https://lore.kernel.org/all/20200224140131.461979697@linutronix.de/ Fixes: 99ba2b5aba24 ("bpf: sockhash, disallow bpf_tcp_close and update in parallel") Reviewed-by: John Fastabend Link: https://lore.kernel.org/r/20230728064411.305576-1-tglozar@redhat.com Signed-off-by: Paolo Abeni Signed-off-by: Sasha Levin --- net/core/sock_map.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/net/core/sock_map.c b/net/core/sock_map.c index 19538d6287144..08ab108206bf8 100644 --- a/net/core/sock_map.c +++ b/net/core/sock_map.c @@ -115,7 +115,6 @@ static void sock_map_sk_acquire(struct sock *sk) __acquires(&sk->sk_lock.slock) { lock_sock(sk); - preempt_disable(); rcu_read_lock(); } @@ -123,7 +122,6 @@ static void sock_map_sk_release(struct sock *sk) __releases(&sk->sk_lock.slock) { rcu_read_unlock(); - preempt_enable(); release_sock(sk); } -- 2.40.1