From: Junxiao Bi <junxiao.bi@oracle.com> To: Eric Ren <zren@suse.com>, ocfs2-devel@oss.oracle.com Cc: akpm@linux-foundation.org, mfasheh@versity.com, jlbec@evilplan.org, ghe@suse.com, jiangqi903@gmail.com, linux-kernel@vger.kernel.org Subject: Re: [PATCH 2/2] ocfs2: fix deadlocks when taking inode lock at vfs entry points Date: Mon, 16 Jan 2017 11:13:19 +0800 [thread overview] Message-ID: <d0585955-51fa-653d-94f3-51ca59e29453@oracle.com> (raw) In-Reply-To: <1d56b79c-1c2b-b988-7861-6b80df1d0a20@suse.com> On 01/16/2017 11:06 AM, Eric Ren wrote: > Hi Junxiao, > > On 01/16/2017 10:46 AM, Junxiao Bi wrote: >>>> If had_lock==true, it is a bug? I think we should BUG_ON for it, that >>>> can help us catch bug at the first time. >>> Good idea! But I'm not sure if "ocfs2_setattr" is always the first one >>> who takes the cluster lock. >>> It's harder for me to name all the possible paths;-/ >> The BUG_ON() can help catch the path where ocfs2_setattr is not the >> first one. > Yes, I understand. But, the problem is that the vfs entries calling > order is out of our control. > I don't want to place an assertion where I'm not 100% sure it's > absolutely right;-) If it is not the first one, is it another recursive locking bug? In this case, if you don't like BUG_ON(), you can dump the call trace and print some warning message. Thanks, Junxiao. > > Thanks, > Eric > >> >> Thanks, >> Junxiao. >> >>>> >>>>> + if (had_lock) >>>>> + arg_flags = OCFS2_META_LOCK_GETBH; >>>>> + status = ocfs2_inode_lock_full(inode, &bh, 1, arg_flags); >>>>> if (status < 0) { >>>>> if (status != -ENOENT) >>>>> mlog_errno(status); >>>>> goto bail_unlock_rw; >>>>> } >>>>> - inode_locked = 1; >>>>> + if (!had_lock) { >>>>> + ocfs2_add_holder(lockres, &oh); >>>>> + inode_locked = 1; >>>>> + } >>>>> if (size_change) { >>>>> status = inode_newsize_ok(inode, attr->ia_size); >>>>> @@ -1260,7 +1270,8 @@ int ocfs2_setattr(struct dentry *dentry, struct >>>>> iattr *attr) >>>>> bail_commit: >>>>> ocfs2_commit_trans(osb, handle); >>>>> bail_unlock: >>>>> - if (status) { >>>>> + if (status && inode_locked) { >>>>> + ocfs2_remove_holder(lockres, &oh); >>>>> ocfs2_inode_unlock(inode, 1); >>>>> inode_locked = 0; >>>>> } >>>>> @@ -1278,8 +1289,10 @@ int ocfs2_setattr(struct dentry *dentry, >>>>> struct iattr *attr) >>>>> if (status < 0) >>>>> mlog_errno(status); >>>>> } >>>>> - if (inode_locked) >>>>> + if (inode_locked) { >>>>> + ocfs2_remove_holder(lockres, &oh); >>>>> ocfs2_inode_unlock(inode, 1); >>>>> + } >>>>> brelse(bh); >>>>> return status; >>>>> @@ -1321,20 +1334,31 @@ int ocfs2_getattr(struct vfsmount *mnt, >>>>> int ocfs2_permission(struct inode *inode, int mask) >>>>> { >>>>> int ret; >>>>> + int has_locked; >>>>> + struct ocfs2_holder oh; >>>>> + struct ocfs2_lock_res *lockres; >>>>> if (mask & MAY_NOT_BLOCK) >>>>> return -ECHILD; >>>>> - ret = ocfs2_inode_lock(inode, NULL, 0); >>>>> - if (ret) { >>>>> - if (ret != -ENOENT) >>>>> - mlog_errno(ret); >>>>> - goto out; >>>>> + lockres = &OCFS2_I(inode)->ip_inode_lockres; >>>>> + has_locked = (ocfs2_is_locked_by_me(lockres) != NULL); >>>> The same thing as ocfs2_setattr. >>> OK. I will think over your suggestions! >>> >>> Thanks, >>> Eric >>> >>>> Thanks, >>>> Junxiao. >>>>> + if (!has_locked) { >>>>> + ret = ocfs2_inode_lock(inode, NULL, 0); >>>>> + if (ret) { >>>>> + if (ret != -ENOENT) >>>>> + mlog_errno(ret); >>>>> + goto out; >>>>> + } >>>>> + ocfs2_add_holder(lockres, &oh); >>>>> } >>>>> ret = generic_permission(inode, mask); >>>>> - ocfs2_inode_unlock(inode, 0); >>>>> + if (!has_locked) { >>>>> + ocfs2_remove_holder(lockres, &oh); >>>>> + ocfs2_inode_unlock(inode, 0); >>>>> + } >>>>> out: >>>>> return ret; >>>>> } >>>>> >> >
WARNING: multiple messages have this Message-ID (diff)
From: Junxiao Bi <junxiao.bi@oracle.com> To: Eric Ren <zren@suse.com>, ocfs2-devel@oss.oracle.com Cc: akpm@linux-foundation.org, mfasheh@versity.com, jlbec@evilplan.org, ghe@suse.com, jiangqi903@gmail.com, linux-kernel@vger.kernel.org Subject: [Ocfs2-devel] [PATCH 2/2] ocfs2: fix deadlocks when taking inode lock at vfs entry points Date: Mon, 16 Jan 2017 11:13:19 +0800 [thread overview] Message-ID: <d0585955-51fa-653d-94f3-51ca59e29453@oracle.com> (raw) In-Reply-To: <1d56b79c-1c2b-b988-7861-6b80df1d0a20@suse.com> On 01/16/2017 11:06 AM, Eric Ren wrote: > Hi Junxiao, > > On 01/16/2017 10:46 AM, Junxiao Bi wrote: >>>> If had_lock==true, it is a bug? I think we should BUG_ON for it, that >>>> can help us catch bug at the first time. >>> Good idea! But I'm not sure if "ocfs2_setattr" is always the first one >>> who takes the cluster lock. >>> It's harder for me to name all the possible paths;-/ >> The BUG_ON() can help catch the path where ocfs2_setattr is not the >> first one. > Yes, I understand. But, the problem is that the vfs entries calling > order is out of our control. > I don't want to place an assertion where I'm not 100% sure it's > absolutely right;-) If it is not the first one, is it another recursive locking bug? In this case, if you don't like BUG_ON(), you can dump the call trace and print some warning message. Thanks, Junxiao. > > Thanks, > Eric > >> >> Thanks, >> Junxiao. >> >>>> >>>>> + if (had_lock) >>>>> + arg_flags = OCFS2_META_LOCK_GETBH; >>>>> + status = ocfs2_inode_lock_full(inode, &bh, 1, arg_flags); >>>>> if (status < 0) { >>>>> if (status != -ENOENT) >>>>> mlog_errno(status); >>>>> goto bail_unlock_rw; >>>>> } >>>>> - inode_locked = 1; >>>>> + if (!had_lock) { >>>>> + ocfs2_add_holder(lockres, &oh); >>>>> + inode_locked = 1; >>>>> + } >>>>> if (size_change) { >>>>> status = inode_newsize_ok(inode, attr->ia_size); >>>>> @@ -1260,7 +1270,8 @@ int ocfs2_setattr(struct dentry *dentry, struct >>>>> iattr *attr) >>>>> bail_commit: >>>>> ocfs2_commit_trans(osb, handle); >>>>> bail_unlock: >>>>> - if (status) { >>>>> + if (status && inode_locked) { >>>>> + ocfs2_remove_holder(lockres, &oh); >>>>> ocfs2_inode_unlock(inode, 1); >>>>> inode_locked = 0; >>>>> } >>>>> @@ -1278,8 +1289,10 @@ int ocfs2_setattr(struct dentry *dentry, >>>>> struct iattr *attr) >>>>> if (status < 0) >>>>> mlog_errno(status); >>>>> } >>>>> - if (inode_locked) >>>>> + if (inode_locked) { >>>>> + ocfs2_remove_holder(lockres, &oh); >>>>> ocfs2_inode_unlock(inode, 1); >>>>> + } >>>>> brelse(bh); >>>>> return status; >>>>> @@ -1321,20 +1334,31 @@ int ocfs2_getattr(struct vfsmount *mnt, >>>>> int ocfs2_permission(struct inode *inode, int mask) >>>>> { >>>>> int ret; >>>>> + int has_locked; >>>>> + struct ocfs2_holder oh; >>>>> + struct ocfs2_lock_res *lockres; >>>>> if (mask & MAY_NOT_BLOCK) >>>>> return -ECHILD; >>>>> - ret = ocfs2_inode_lock(inode, NULL, 0); >>>>> - if (ret) { >>>>> - if (ret != -ENOENT) >>>>> - mlog_errno(ret); >>>>> - goto out; >>>>> + lockres = &OCFS2_I(inode)->ip_inode_lockres; >>>>> + has_locked = (ocfs2_is_locked_by_me(lockres) != NULL); >>>> The same thing as ocfs2_setattr. >>> OK. I will think over your suggestions! >>> >>> Thanks, >>> Eric >>> >>>> Thanks, >>>> Junxiao. >>>>> + if (!has_locked) { >>>>> + ret = ocfs2_inode_lock(inode, NULL, 0); >>>>> + if (ret) { >>>>> + if (ret != -ENOENT) >>>>> + mlog_errno(ret); >>>>> + goto out; >>>>> + } >>>>> + ocfs2_add_holder(lockres, &oh); >>>>> } >>>>> ret = generic_permission(inode, mask); >>>>> - ocfs2_inode_unlock(inode, 0); >>>>> + if (!has_locked) { >>>>> + ocfs2_remove_holder(lockres, &oh); >>>>> + ocfs2_inode_unlock(inode, 0); >>>>> + } >>>>> out: >>>>> return ret; >>>>> } >>>>> >> >
next prev parent reply other threads:[~2017-01-16 3:13 UTC|newest] Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top 2017-01-05 15:31 [PATCH 0/2] fix deadlock caused by recursive cluster locking Eric Ren 2017-01-05 15:31 ` [Ocfs2-devel] " Eric Ren 2017-01-05 15:31 ` [PATCH 1/2] ocfs2/dlmglue: prepare tracking logic to avoid recursive cluster lock Eric Ren 2017-01-05 15:31 ` [Ocfs2-devel] " Eric Ren 2017-01-06 6:07 ` Joseph Qi 2017-01-06 6:07 ` [Ocfs2-devel] " Joseph Qi 2017-01-06 7:03 ` Eric Ren 2017-01-06 7:03 ` [Ocfs2-devel] " Eric Ren 2017-01-06 7:24 ` Joseph Qi 2017-01-06 7:24 ` [Ocfs2-devel] " Joseph Qi 2017-01-06 8:04 ` Eric Ren 2017-01-06 8:04 ` [Ocfs2-devel] " Eric Ren 2017-01-13 3:59 ` Junxiao Bi 2017-01-13 3:59 ` [Ocfs2-devel] " Junxiao Bi 2017-01-13 6:12 ` Eric Ren 2017-01-13 6:12 ` [Ocfs2-devel] " Eric Ren 2017-01-16 2:42 ` Junxiao Bi 2017-01-16 2:42 ` [Ocfs2-devel] " Junxiao Bi 2017-01-16 3:31 ` Eric Ren 2017-01-16 3:31 ` [Ocfs2-devel] " Eric Ren 2017-01-05 15:31 ` [PATCH 2/2] ocfs2: fix deadlocks when taking inode lock at vfs entry points Eric Ren 2017-01-05 15:31 ` [Ocfs2-devel] " Eric Ren 2017-01-06 6:09 ` Joseph Qi 2017-01-06 6:09 ` [Ocfs2-devel] " Joseph Qi 2017-01-06 6:56 ` Eric Ren 2017-01-06 6:56 ` [Ocfs2-devel] " Eric Ren 2017-01-06 7:14 ` Joseph Qi 2017-01-06 7:14 ` [Ocfs2-devel] " Joseph Qi 2017-01-06 8:21 ` Eric Ren 2017-01-06 8:21 ` [Ocfs2-devel] " Eric Ren 2017-01-06 9:03 ` Joseph Qi 2017-01-06 9:03 ` [Ocfs2-devel] " Joseph Qi 2017-01-06 9:13 ` Eric Ren 2017-01-06 9:13 ` [Ocfs2-devel] " Eric Ren 2017-01-06 9:55 ` Joseph Qi 2017-01-06 9:55 ` [Ocfs2-devel] " Joseph Qi 2017-01-06 11:56 ` Eric Ren 2017-01-06 11:56 ` [Ocfs2-devel] " Eric Ren 2017-01-09 1:13 ` Joseph Qi 2017-01-09 1:13 ` [Ocfs2-devel] " Joseph Qi 2017-01-09 2:13 ` Eric Ren 2017-01-09 2:13 ` [Ocfs2-devel] " Eric Ren 2017-01-12 11:24 ` Eric Ren 2017-01-12 11:24 ` Eric Ren 2017-01-12 11:36 ` Joseph Qi 2017-01-12 11:36 ` Joseph Qi 2017-01-06 14:52 ` kbuild test robot 2017-01-06 14:52 ` [Ocfs2-devel] " kbuild test robot 2017-01-09 5:24 ` Eric Ren 2017-01-09 5:24 ` [Ocfs2-devel] " Eric Ren 2017-01-06 17:53 ` kbuild test robot 2017-01-06 17:53 ` [Ocfs2-devel] " kbuild test robot 2017-01-13 4:22 ` Junxiao Bi 2017-01-13 4:22 ` [Ocfs2-devel] " Junxiao Bi 2017-01-13 6:19 ` Eric Ren 2017-01-13 6:19 ` [Ocfs2-devel] " Eric Ren 2017-01-16 2:46 ` Junxiao Bi 2017-01-16 2:46 ` [Ocfs2-devel] " Junxiao Bi 2017-01-16 3:06 ` Eric Ren 2017-01-16 3:06 ` [Ocfs2-devel] " Eric Ren 2017-01-16 3:13 ` Junxiao Bi [this message] 2017-01-16 3:13 ` Junxiao Bi 2017-01-16 3:17 ` Eric Ren 2017-01-16 3:17 ` [Ocfs2-devel] " Eric Ren
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=d0585955-51fa-653d-94f3-51ca59e29453@oracle.com \ --to=junxiao.bi@oracle.com \ --cc=akpm@linux-foundation.org \ --cc=ghe@suse.com \ --cc=jiangqi903@gmail.com \ --cc=jlbec@evilplan.org \ --cc=linux-kernel@vger.kernel.org \ --cc=mfasheh@versity.com \ --cc=ocfs2-devel@oss.oracle.com \ --cc=zren@suse.com \ /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: linkBe 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.