From patchwork Wed Jan 12 04:29:04 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Xiubo Li X-Patchwork-Id: 12710921 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 77CACC433EF for ; Wed, 12 Jan 2022 04:29:20 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1350696AbiALE3T (ORCPT ); Tue, 11 Jan 2022 23:29:19 -0500 Received: from us-smtp-delivery-124.mimecast.com ([170.10.129.124]:47592 "EHLO us-smtp-delivery-124.mimecast.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231799AbiALE3S (ORCPT ); Tue, 11 Jan 2022 23:29:18 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1641961757; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding; bh=i4oNaLzVFZJz612kIu4HzMBn4l/zssbMXysVEH5dcMI=; b=M+YSIRbZB/OBOsg3k1NadvgX2F3UOcm81LyElnuQx3zyhj8WHXMAQu8IugPlkz6aw04ji2 7JaqmlGMWFq1zsSCvkMgoyXkPMZN3gWqJGMHCgWHXOyMtTDn5jBhvosups9PZxSi26fg6C CavUERqew7nDLLbBwX6Z0/1cy52gBpc= Received: from mimecast-mx01.redhat.com (mimecast-mx01.redhat.com [209.132.183.4]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id us-mta-317-1Ddu3zIVOfSMOUKMqiq-lQ-1; Tue, 11 Jan 2022 23:29:14 -0500 X-MC-Unique: 1Ddu3zIVOfSMOUKMqiq-lQ-1 Received: from smtp.corp.redhat.com (int-mx07.intmail.prod.int.phx2.redhat.com [10.5.11.22]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mimecast-mx01.redhat.com (Postfix) with ESMTPS id 9F8DD10151E0; Wed, 12 Jan 2022 04:29:13 +0000 (UTC) Received: from lxbceph1.gsslab.pek2.redhat.com (unknown [10.72.47.117]) by smtp.corp.redhat.com (Postfix) with ESMTP id 79C401037F44; Wed, 12 Jan 2022 04:29:11 +0000 (UTC) From: xiubli@redhat.com To: jlayton@kernel.org Cc: idryomov@gmail.com, vshankar@redhat.com, ceph-devel@vger.kernel.org, Xiubo Li Subject: [PATCH] ceph: put the requests/sessions when it fails to alloc memory Date: Wed, 12 Jan 2022 12:29:04 +0800 Message-Id: <20220112042904.8557-1-xiubli@redhat.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.84 on 10.5.11.22 Precedence: bulk List-ID: X-Mailing-List: ceph-devel@vger.kernel.org From: Xiubo Li When failing to allocate the sessions memory we should make sure the req1 and req2 and the sessions get put. And also in case the max_sessions decreased so when kreallocate the new memory some sessions maybe missed being put. And if the max_sessions is 0 krealloc will return ZERO_SIZE_PTR, which will lead to a distinct access fault. URL: https://tracker.ceph.com/issues/53819 Signed-off-by: Xiubo Li Fixes: e1a4541ec0b9 ("ceph: flush the mdlog before waiting on unsafe reqs") Reported-by: Niklas Hambuechen Reviewed-by: Venky Shankar --- fs/ceph/caps.c | 55 +++++++++++++++++++++++++++++++++----------------- 1 file changed, 37 insertions(+), 18 deletions(-) diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 944b18b4e217..5c2719f66f62 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2276,6 +2276,7 @@ static int unsafe_request_wait(struct inode *inode) struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_mds_request *req1 = NULL, *req2 = NULL; + unsigned int max_sessions; int ret, err = 0; spin_lock(&ci->i_unsafe_lock); @@ -2293,37 +2294,45 @@ static int unsafe_request_wait(struct inode *inode) } spin_unlock(&ci->i_unsafe_lock); + /* + * The mdsc->max_sessions is unlikely to be changed + * mostly, here we will retry it by reallocating the + * sessions arrary memory to get rid of the mdsc->mutex + * lock. + */ +retry: + max_sessions = mdsc->max_sessions; + /* * Trigger to flush the journal logs in all the relevant MDSes * manually, or in the worst case we must wait at most 5 seconds * to wait the journal logs to be flushed by the MDSes periodically. */ - if (req1 || req2) { + if ((req1 || req2) && likely(max_sessions)) { struct ceph_mds_session **sessions = NULL; struct ceph_mds_session *s; struct ceph_mds_request *req; - unsigned int max; int i; - /* - * The mdsc->max_sessions is unlikely to be changed - * mostly, here we will retry it by reallocating the - * sessions arrary memory to get rid of the mdsc->mutex - * lock. - */ -retry: - max = mdsc->max_sessions; - sessions = krealloc(sessions, max * sizeof(s), __GFP_ZERO); - if (!sessions) - return -ENOMEM; + sessions = kzalloc(max_sessions * sizeof(s), GFP_KERNEL); + if (!sessions) { + err = -ENOMEM; + goto out; + } spin_lock(&ci->i_unsafe_lock); if (req1) { list_for_each_entry(req, &ci->i_unsafe_dirops, r_unsafe_dir_item) { s = req->r_session; - if (unlikely(s->s_mds >= max)) { + if (unlikely(s->s_mds >= max_sessions)) { spin_unlock(&ci->i_unsafe_lock); + for (i = 0; i < max_sessions; i++) { + s = sessions[i]; + if (s) + ceph_put_mds_session(s); + } + kfree(sessions); goto retry; } if (!sessions[s->s_mds]) { @@ -2336,8 +2345,14 @@ static int unsafe_request_wait(struct inode *inode) list_for_each_entry(req, &ci->i_unsafe_iops, r_unsafe_target_item) { s = req->r_session; - if (unlikely(s->s_mds >= max)) { + if (unlikely(s->s_mds >= max_sessions)) { spin_unlock(&ci->i_unsafe_lock); + for (i = 0; i < max_sessions; i++) { + s = sessions[i]; + if (s) + ceph_put_mds_session(s); + } + kfree(sessions); goto retry; } if (!sessions[s->s_mds]) { @@ -2358,7 +2373,7 @@ static int unsafe_request_wait(struct inode *inode) spin_unlock(&ci->i_ceph_lock); /* send flush mdlog request to MDSes */ - for (i = 0; i < max; i++) { + for (i = 0; i < max_sessions; i++) { s = sessions[i]; if (s) { send_flush_mdlog(s); @@ -2375,15 +2390,19 @@ static int unsafe_request_wait(struct inode *inode) ceph_timeout_jiffies(req1->r_timeout)); if (ret) err = -EIO; - ceph_mdsc_put_request(req1); } if (req2) { ret = !wait_for_completion_timeout(&req2->r_safe_completion, ceph_timeout_jiffies(req2->r_timeout)); if (ret) err = -EIO; - ceph_mdsc_put_request(req2); } + +out: + if (req1) + ceph_mdsc_put_request(req1); + if (req2) + ceph_mdsc_put_request(req2); return err; }