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 D712130F85 for ; Mon, 9 Oct 2023 13:25:15 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.b="Ket3O47t" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 52D3CC433C7; Mon, 9 Oct 2023 13:25:15 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1696857915; bh=AMCEHO14sZUlR3NJmdxCq/Trduzr+wp2jHb92uz7/eY=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=Ket3O47tzPpyFcS2y+k5QAiFD1IY3AZB2iHMMFywVpJEcNubkpC107wNgShbcXPfb Cf9QdHPfOch+slaSkpnFDCKW7viC499je+Qeb0L5Yp1cejBnZSV8tHL20vHEjPcTev X2MIy2sivZ0nGa8InegqLKgyBbJRmb9PiBIhDwp0= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Trond Myklebust , Anna Schumaker , Sasha Levin Subject: [PATCH 5.15 07/75] NFSv4: Fix a state manager thread deadlock regression Date: Mon, 9 Oct 2023 15:01:29 +0200 Message-ID: <20231009130111.466670850@linuxfoundation.org> X-Mailer: git-send-email 2.42.0 In-Reply-To: <20231009130111.200710898@linuxfoundation.org> References: <20231009130111.200710898@linuxfoundation.org> User-Agent: quilt/0.67 X-stable: review X-Patchwork-Hint: ignore Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 5.15-stable review patch. If anyone has any objections, please let me know. ------------------ From: Trond Myklebust [ Upstream commit 956fd46f97d238032cb5fa4771cdaccc6e760f9a ] Commit 4dc73c679114 reintroduces the deadlock that was fixed by commit aeabb3c96186 ("NFSv4: Fix a NFSv4 state manager deadlock") because it prevents the setup of new threads to handle reboot recovery, while the older recovery thread is stuck returning delegations. Fixes: 4dc73c679114 ("NFSv4: keep state manager thread active if swap is enabled") Cc: stable@vger.kernel.org Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker Signed-off-by: Sasha Levin --- fs/nfs/nfs4proc.c | 4 +++- fs/nfs/nfs4state.c | 36 +++++++++++++++++++++++++----------- 2 files changed, 28 insertions(+), 12 deletions(-) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 32204c0b3d098..3275763b78bc8 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -10557,7 +10557,9 @@ static void nfs4_disable_swap(struct inode *inode) */ struct nfs_client *clp = NFS_SERVER(inode)->nfs_client; - nfs4_schedule_state_manager(clp); + set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state); + clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state); + wake_up_var(&clp->cl_state); } static const struct inode_operations nfs4_dir_inode_operations = { diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index d0183a7b01a49..7590d059eb78e 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1213,13 +1213,23 @@ void nfs4_schedule_state_manager(struct nfs_client *clp) { struct task_struct *task; char buf[INET6_ADDRSTRLEN + sizeof("-manager") + 1]; + struct rpc_clnt *clnt = clp->cl_rpcclient; + bool swapon = false; set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state); - if (test_and_set_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state) != 0) { - wake_up_var(&clp->cl_state); - return; + + if (atomic_read(&clnt->cl_swapper)) { + swapon = !test_and_set_bit(NFS4CLNT_MANAGER_AVAILABLE, + &clp->cl_state); + if (!swapon) { + wake_up_var(&clp->cl_state); + return; + } } - set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state); + + if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0) + return; + __module_get(THIS_MODULE); refcount_inc(&clp->cl_count); @@ -1236,8 +1246,9 @@ void nfs4_schedule_state_manager(struct nfs_client *clp) __func__, PTR_ERR(task)); if (!nfs_client_init_is_complete(clp)) nfs_mark_client_ready(clp, PTR_ERR(task)); + if (swapon) + clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state); nfs4_clear_state_manager_bit(clp); - clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state); nfs_put_client(clp); module_put(THIS_MODULE); } @@ -2726,22 +2737,25 @@ static int nfs4_run_state_manager(void *ptr) allow_signal(SIGKILL); again: - set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state); nfs4_state_manager(clp); - if (atomic_read(&cl->cl_swapper)) { + + if (test_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state) && + !test_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state)) { wait_var_event_interruptible(&clp->cl_state, test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state)); - if (atomic_read(&cl->cl_swapper) && - test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state)) + if (!atomic_read(&cl->cl_swapper)) + clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state); + if (refcount_read(&clp->cl_count) > 1 && !signalled() && + !test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state)) goto again; /* Either no longer a swapper, or were signalled */ + clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state); } - clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state); if (refcount_read(&clp->cl_count) > 1 && !signalled() && test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state) && - !test_and_set_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state)) + !test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state)) goto again; nfs_put_client(clp); -- 2.40.1