Message-ID: <fb748c596936651a6b42781653f0ca29dc45fb5b.camel@kernel.org>
Date: Sun, 18 Feb 2024 07:16:34 -0500
From: Jeff Layton <jlayton@...nel.org>
To: Steven Rostedt <rostedt@...dmis.org>, Masami Hiramatsu
 <mhiramat@...nel.org>,  Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
 Chuck Lever <chuck.lever@...cle.com>, Alexander Viro
 <viro@...iv.linux.org.uk>, Christian Brauner <brauner@...nel.org>, Jan Kara
 <jack@...e.cz>, Eric Van Hensbergen <ericvh@...nel.org>, Latchesar Ionkov
 <lucho@...kov.net>, Dominique Martinet <asmadeus@...ewreck.org>, Christian
 Schoenebeck <linux_oss@...debyte.com>, David Howells <dhowells@...hat.com>,
 Marc Dionne <marc.dionne@...istor.com>, Xiubo Li <xiubli@...hat.com>, Ilya
 Dryomov <idryomov@...il.com>, Alexander Aring <aahringo@...hat.com>, David
 Teigland <teigland@...hat.com>, Andreas Gruenbacher <agruenba@...hat.com>,
 Neil Brown <neilb@...e.de>, Olga Kornievskaia <kolga@...app.com>, Dai Ngo
 <Dai.Ngo@...cle.com>, Tom Talpey <tom@...pey.com>, Trond Myklebust
 <trond.myklebust@...merspace.com>, Anna Schumaker <anna@...nel.org>, Mark
 Fasheh <mark@...heh.com>, Joel Becker <jlbec@...lplan.org>, Joseph Qi
 <joseph.qi@...ux.alibaba.com>, Steve French <sfrench@...ba.org>, Paulo
 Alcantara <pc@...guebit.com>, Ronnie Sahlberg <ronniesahlberg@...il.com>,
 Shyam Prasad N <sprasad@...rosoft.com>, Namjae Jeon
 <linkinjeon@...nel.org>, Sergey Senozhatsky <senozhatsky@...omium.org>,
 Miklos Szeredi <miklos@...redi.hu>
Cc: linux-kernel@...r.kernel.org, linux-trace-kernel@...r.kernel.org, 
 linux-fsdevel@...r.kernel.org, v9fs@...ts.linux.dev, 
 linux-afs@...ts.infradead.org, ceph-devel@...r.kernel.org,
 gfs2@...ts.linux.dev,  linux-nfs@...r.kernel.org,
 ocfs2-devel@...ts.linux.dev,  linux-cifs@...r.kernel.org
Subject: Re: [PATCH v3 25/47] filelock: convert __locks_insert_block,
 conflict and deadlock checks to use file_lock_core

On Wed, 2024-01-31 at 18:02 -0500, Jeff Layton wrote:
> Have both __locks_insert_block and the deadlock and conflict checking
> functions take a struct file_lock_core pointer instead of a struct
> file_lock one. Also, change posix_locks_deadlock to return bool.
> 
> Signed-off-by: Jeff Layton <jlayton@...nel.org>
> ---
>  fs/locks.c | 132 +++++++++++++++++++++++++++++++++----------------------------
>  1 file changed, 72 insertions(+), 60 deletions(-)
> 
> diff --git a/fs/locks.c b/fs/locks.c
> index 1e8b943bd7f9..0dc1c9da858c 100644
> --- a/fs/locks.c
> +++ b/fs/locks.c
> @@ -757,39 +757,41 @@ EXPORT_SYMBOL(locks_delete_block);
>   * waiters, and add beneath any waiter that blocks the new waiter.
>   * Thus wakeups don't happen until needed.
>   */
> -static void __locks_insert_block(struct file_lock *blocker,
> -				 struct file_lock *waiter,
> -				 bool conflict(struct file_lock *,
> -					       struct file_lock *))
> +static void __locks_insert_block(struct file_lock *blocker_fl,
> +				 struct file_lock *waiter_fl,
> +				 bool conflict(struct file_lock_core *,
> +					       struct file_lock_core *))
>  {
> -	struct file_lock *fl;
> -	BUG_ON(!list_empty(&waiter->c.flc_blocked_member));
> +	struct file_lock_core *blocker = &blocker_fl->c;
> +	struct file_lock_core *waiter = &waiter_fl->c;
> +	struct file_lock_core *flc;
>  
> +	BUG_ON(!list_empty(&waiter->flc_blocked_member));
>  new_blocker:
> -	list_for_each_entry(fl, &blocker->c.flc_blocked_requests,
> -			    c.flc_blocked_member)
> -		if (conflict(fl, waiter)) {
> -			blocker =  fl;
> +	list_for_each_entry(flc, &blocker->flc_blocked_requests, flc_blocked_member)
> +		if (conflict(flc, waiter)) {
> +			blocker =  flc;
>  			goto new_blocker;
>  		}
> -	waiter->c.flc_blocker = blocker;
> -	list_add_tail(&waiter->c.flc_blocked_member,
> -		      &blocker->c.flc_blocked_requests);
> -	if ((blocker->c.flc_flags & (FL_POSIX|FL_OFDLCK)) == FL_POSIX)
> -		locks_insert_global_blocked(&waiter->c);
> +	waiter->flc_blocker = file_lock(blocker);
> +	list_add_tail(&waiter->flc_blocked_member,
> +		      &blocker->flc_blocked_requests);
>  
> -	/* The requests in waiter->fl_blocked are known to conflict with
> +	if ((blocker->flc_flags & (FL_POSIX|FL_OFDLCK)) == (FL_POSIX|FL_OFDLCK))

Christian,

There is a bug in the above delta. That should read:

    if ((blocker->flc_flags & (FL_POSIX|FL_OFDLCK)) == FL_POSIX)

I suspect that is the cause of the performance regression noted by the
KTR.

I believe the bug is fairly harmless -- it's just putting OFD locks into
the global hash when it doesn't need to, which probably slows down
deadlock checking. I'm going to spin up a patch and test it today, but I
wanted to give you a heads up.

I'll send the patch later today or tomorrow.
 
> +		locks_insert_global_blocked(waiter);
> +
> +	/* The requests in waiter->flc_blocked are known to conflict with
>  	 * waiter, but might not conflict with blocker, or the requests
>  	 * and lock which block it.  So they all need to be woken.
>  	 */
> -	__locks_wake_up_blocks(&waiter->c);
> +	__locks_wake_up_blocks(waiter);
>  }
>  
>  /* Must be called with flc_lock held. */
>  static void locks_insert_block(struct file_lock *blocker,
>  			       struct file_lock *waiter,
> -			       bool conflict(struct file_lock *,
> -					     struct file_lock *))
> +			       bool conflict(struct file_lock_core *,
> +					     struct file_lock_core *))
>  {
>  	spin_lock(&blocked_lock_lock);
>  	__locks_insert_block(blocker, waiter, conflict);
> @@ -846,12 +848,12 @@ locks_delete_lock_ctx(struct file_lock *fl, struct list_head *dispose)
>  /* Determine if lock sys_fl blocks lock caller_fl. Common functionality
>   * checks for shared/exclusive status of overlapping locks.
>   */
> -static bool locks_conflict(struct file_lock *caller_fl,
> -			   struct file_lock *sys_fl)
> +static bool locks_conflict(struct file_lock_core *caller_flc,
> +			   struct file_lock_core *sys_flc)
>  {
> -	if (lock_is_write(sys_fl))
> +	if (sys_flc->flc_type == F_WRLCK)
>  		return true;
> -	if (lock_is_write(caller_fl))
> +	if (caller_flc->flc_type == F_WRLCK)
>  		return true;
>  	return false;
>  }
> @@ -859,20 +861,23 @@ static bool locks_conflict(struct file_lock *caller_fl,
>  /* Determine if lock sys_fl blocks lock caller_fl. POSIX specific
>   * checking before calling the locks_conflict().
>   */
> -static bool posix_locks_conflict(struct file_lock *caller_fl,
> -				 struct file_lock *sys_fl)
> +static bool posix_locks_conflict(struct file_lock_core *caller_flc,
> +				 struct file_lock_core *sys_flc)
>  {
> +	struct file_lock *caller_fl = file_lock(caller_flc);
> +	struct file_lock *sys_fl = file_lock(sys_flc);
> +
>  	/* POSIX locks owned by the same process do not conflict with
>  	 * each other.
>  	 */
> -	if (posix_same_owner(&caller_fl->c, &sys_fl->c))
> +	if (posix_same_owner(caller_flc, sys_flc))
>  		return false;
>  
>  	/* Check whether they overlap */
>  	if (!locks_overlap(caller_fl, sys_fl))
>  		return false;
>  
> -	return locks_conflict(caller_fl, sys_fl);
> +	return locks_conflict(caller_flc, sys_flc);
>  }
>  
>  /* Determine if lock sys_fl blocks lock caller_fl. Used on xx_GETLK
> @@ -881,28 +886,31 @@ static bool posix_locks_conflict(struct file_lock *caller_fl,
>  static bool posix_test_locks_conflict(struct file_lock *caller_fl,
>  				      struct file_lock *sys_fl)
>  {
> +	struct file_lock_core *caller = &caller_fl->c;
> +	struct file_lock_core *sys = &sys_fl->c;
> +
>  	/* F_UNLCK checks any locks on the same fd. */
>  	if (lock_is_unlock(caller_fl)) {
> -		if (!posix_same_owner(&caller_fl->c, &sys_fl->c))
> +		if (!posix_same_owner(caller, sys))
>  			return false;
>  		return locks_overlap(caller_fl, sys_fl);
>  	}
> -	return posix_locks_conflict(caller_fl, sys_fl);
> +	return posix_locks_conflict(caller, sys);
>  }
>  
>  /* Determine if lock sys_fl blocks lock caller_fl. FLOCK specific
>   * checking before calling the locks_conflict().
>   */
> -static bool flock_locks_conflict(struct file_lock *caller_fl,
> -				 struct file_lock *sys_fl)
> +static bool flock_locks_conflict(struct file_lock_core *caller_flc,
> +				 struct file_lock_core *sys_flc)
>  {
>  	/* FLOCK locks referring to the same filp do not conflict with
>  	 * each other.
>  	 */
> -	if (caller_fl->c.flc_file == sys_fl->c.flc_file)
> +	if (caller_flc->flc_file == sys_flc->flc_file)
>  		return false;
>  
> -	return locks_conflict(caller_fl, sys_fl);
> +	return locks_conflict(caller_flc, sys_flc);
>  }
>  
>  void
> @@ -980,25 +988,27 @@ EXPORT_SYMBOL(posix_test_lock);
>  
>  #define MAX_DEADLK_ITERATIONS 10
>  
> -/* Find a lock that the owner of the given block_fl is blocking on. */
> -static struct file_lock *what_owner_is_waiting_for(struct file_lock *block_fl)
> +/* Find a lock that the owner of the given @blocker is blocking on. */
> +static struct file_lock_core *what_owner_is_waiting_for(struct file_lock_core *blocker)
>  {
> -	struct file_lock *fl;
> +	struct file_lock_core *flc;
>  
> -	hash_for_each_possible(blocked_hash, fl, c.flc_link, posix_owner_key(&block_fl->c)) {
> -		if (posix_same_owner(&fl->c, &block_fl->c)) {
> -			while (fl->c.flc_blocker)
> -				fl = fl->c.flc_blocker;
> -			return fl;
> +	hash_for_each_possible(blocked_hash, flc, flc_link, posix_owner_key(blocker)) {
> +		if (posix_same_owner(flc, blocker)) {
> +			while (flc->flc_blocker)
> +				flc = &flc->flc_blocker->c;
> +			return flc;
>  		}
>  	}
>  	return NULL;
>  }
>  
>  /* Must be called with the blocked_lock_lock held! */
> -static int posix_locks_deadlock(struct file_lock *caller_fl,
> -				struct file_lock *block_fl)
> +static bool posix_locks_deadlock(struct file_lock *caller_fl,
> +				 struct file_lock *block_fl)
>  {
> +	struct file_lock_core *caller = &caller_fl->c;
> +	struct file_lock_core *blocker = &block_fl->c;
>  	int i = 0;
>  
>  	lockdep_assert_held(&blocked_lock_lock);
> @@ -1007,16 +1017,16 @@ static int posix_locks_deadlock(struct file_lock *caller_fl,
>  	 * This deadlock detector can't reasonably detect deadlocks with
>  	 * FL_OFDLCK locks, since they aren't owned by a process, per-se.
>  	 */
> -	if (caller_fl->c.flc_flags & FL_OFDLCK)
> -		return 0;
> +	if (caller->flc_flags & FL_OFDLCK)
> +		return false;
>  
> -	while ((block_fl = what_owner_is_waiting_for(block_fl))) {
> +	while ((blocker = what_owner_is_waiting_for(blocker))) {
>  		if (i++ > MAX_DEADLK_ITERATIONS)
> -			return 0;
> -		if (posix_same_owner(&caller_fl->c, &block_fl->c))
> -			return 1;
> +			return false;
> +		if (posix_same_owner(caller, blocker))
> +			return true;
>  	}
> -	return 0;
> +	return false;
>  }
>  
>  /* Try to create a FLOCK lock on filp. We always insert new FLOCK locks
> @@ -1071,7 +1081,7 @@ static int flock_lock_inode(struct inode *inode, struct file_lock *request)
>  
>  find_conflict:
>  	list_for_each_entry(fl, &ctx->flc_flock, c.flc_list) {
> -		if (!flock_locks_conflict(request, fl))
> +		if (!flock_locks_conflict(&request->c, &fl->c))
>  			continue;
>  		error = -EAGAIN;
>  		if (!(request->c.flc_flags & FL_SLEEP))
> @@ -1140,7 +1150,7 @@ static int posix_lock_inode(struct inode *inode, struct file_lock *request,
>  	 */
>  	if (request->c.flc_type != F_UNLCK) {
>  		list_for_each_entry(fl, &ctx->flc_posix, c.flc_list) {
> -			if (!posix_locks_conflict(request, fl))
> +			if (!posix_locks_conflict(&request->c, &fl->c))
>  				continue;
>  			if (fl->fl_lmops && fl->fl_lmops->lm_lock_expirable
>  				&& (*fl->fl_lmops->lm_lock_expirable)(fl)) {
> @@ -1442,23 +1452,25 @@ static void time_out_leases(struct inode *inode, struct list_head *dispose)
>  	}
>  }
>  
> -static bool leases_conflict(struct file_lock *lease, struct file_lock *breaker)
> +static bool leases_conflict(struct file_lock_core *lc, struct file_lock_core *bc)
>  {
>  	bool rc;
> +	struct file_lock *lease = file_lock(lc);
> +	struct file_lock *breaker = file_lock(bc);
>  
>  	if (lease->fl_lmops->lm_breaker_owns_lease
>  			&& lease->fl_lmops->lm_breaker_owns_lease(lease))
>  		return false;
> -	if ((breaker->c.flc_flags & FL_LAYOUT) != (lease->c.flc_flags & FL_LAYOUT)) {
> +	if ((bc->flc_flags & FL_LAYOUT) != (lc->flc_flags & FL_LAYOUT)) {
>  		rc = false;
>  		goto trace;
>  	}
> -	if ((breaker->c.flc_flags & FL_DELEG) && (lease->c.flc_flags & FL_LEASE)) {
> +	if ((bc->flc_flags & FL_DELEG) && (lc->flc_flags & FL_LEASE)) {
>  		rc = false;
>  		goto trace;
>  	}
>  
> -	rc = locks_conflict(breaker, lease);
> +	rc = locks_conflict(bc, lc);
>  trace:
>  	trace_leases_conflict(rc, lease, breaker);
>  	return rc;
> @@ -1468,12 +1480,12 @@ static bool
>  any_leases_conflict(struct inode *inode, struct file_lock *breaker)
>  {
>  	struct file_lock_context *ctx = inode->i_flctx;
> -	struct file_lock *fl;
> +	struct file_lock_core *flc;
>  
>  	lockdep_assert_held(&ctx->flc_lock);
>  
> -	list_for_each_entry(fl, &ctx->flc_lease, c.flc_list) {
> -		if (leases_conflict(fl, breaker))
> +	list_for_each_entry(flc, &ctx->flc_lease, flc_list) {
> +		if (leases_conflict(flc, &breaker->c))
>  			return true;
>  	}
>  	return false;
> @@ -1529,7 +1541,7 @@ int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
>  	}
>  
>  	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, c.flc_list) {
> -		if (!leases_conflict(fl, new_fl))
> +		if (!leases_conflict(&fl->c, &new_fl->c))
>  			continue;
>  		if (want_write) {
>  			if (fl->c.flc_flags & FL_UNLOCK_PENDING)
> 

-- 
Jeff Layton <jlayton@...nel.org>

Powered by blists - more mailing lists