[PATCH v2 1/4] fs/afs/vlocation: Remove deprecated create_singlethread_workqueue

[PATCH v2 1/4] fs/afs/vlocation: Remove deprecated create_singlethread_workqueue

[Bhaktipriya Shridhar]

The workqueue "afs_vlocation_update_worker" queues a single work item
&afs_vlocation_update and hence it doesn't require execution ordering.
Hence, alloc_workqueue has been used to replace the deprecated
create_singlethread_workqueue instance.

Since the workqueue is being used on a memory reclaim path, WQ_MEM_RECLAIM
flag has been set to ensure forward progress under memory pressure.

Since there are fixed number of work items, explicit concurrency
limit is unnecessary here.

Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
---
 Changes in v2:
	-No change.

 fs/afs/vlocation.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/afs/vlocation.c b/fs/afs/vlocation.c
index 5297678..45a8639 100644
--- a/fs/afs/vlocation.c
+++ b/fs/afs/vlocation.c
@@ -594,8 +594,8 @@ static void afs_vlocation_reaper(struct work_struct *work)
  */
 int __init afs_vlocation_update_init(void)
 {
-	afs_vlocation_update_worker =
-		create_singlethread_workqueue("kafs_vlupdated");
+	afs_vlocation_update_worker = alloc_workqueue("kafs_vlupdated",
+						      WQ_MEM_RECLAIM, 0);
 	return afs_vlocation_update_worker ? 0 : -ENOMEM;
 }

--
2.1.4

[PATCH v2 2/4] fs/afs/rxrpc: Remove deprecated create_singlethread_workqueue

[Bhaktipriya Shridhar]

The workqueue "afs_async_calls" queues a single work item viz
&call->async_work and hence it doesn't require execution ordering.
Hence, alloc_workqueue has been used to replace the deprecated
create_singlethread_workqueue instance.

The WQ_MEM_RECLAIM flag has been set to ensure forward progress under
memory pressure because the workqueue is being used on a memory reclaim
path.

Since there are fixed number of work items, explicit concurrency
limit is unnecessary here.

Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
---
 Changes in v2:
	-No change

 fs/afs/rxrpc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/afs/rxrpc.c b/fs/afs/rxrpc.c
index 63cd9f9..4f9e9b3 100644
--- a/fs/afs/rxrpc.c
+++ b/fs/afs/rxrpc.c
@@ -85,7 +85,7 @@ int afs_open_socket(void)

 	skb_queue_head_init(&afs_incoming_calls);

-	afs_async_calls = create_singlethread_workqueue("kafsd");
+	afs_async_calls = alloc_workqueue("kafsd", WQ_MEM_RECLAIM, 0);
 	if (!afs_async_calls) {
 		_leave(" = -ENOMEM [wq]");
 		return -ENOMEM;
--
2.1.4

[PATCH v2 3/4] fs/afs/callback: Remove deprecated create_singlethread_workqueue

[Bhaktipriya Shridhar]

The workqueue "afs_callback_update_worker" queues multiple work items
viz  &vnode->cb_broken_work, &server->cb_break_work which require strict
execution ordering. Hence, an ordered dedicated workqueue has been used.

Since the workqueue is being used on a memory reclaim path, WQ_MEM_RECLAIM
has been set to ensure forward progress under memory pressure.

Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
---
 Changes in v2:
	-No change

 fs/afs/callback.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/afs/callback.c b/fs/afs/callback.c
index 7ef637d..1e9d2f8 100644
--- a/fs/afs/callback.c
+++ b/fs/afs/callback.c
@@ -461,8 +461,8 @@ static void afs_callback_updater(struct work_struct *work)
  */
 int __init afs_callback_update_init(void)
 {
-	afs_callback_update_worker =
-		create_singlethread_workqueue("kafs_callbackd");
+	afs_callback_update_worker = alloc_ordered_workqueue("kafs_callbackd",
+							     WQ_MEM_RECLAIM);
 	return afs_callback_update_worker ? 0 : -ENOMEM;
 }

--
2.1.4

[PATCH v2 4/4] fs/afs/flock: Remove deprecated create_singlethread_workqueue

[Bhaktipriya Shridhar]

The workqueue "afs_lock_manager" queues a single work item
&vnode->lock_work, and hence it doesn't require execution ordering.
Hence, alloc_workqueue has been used to replace the deprecated
create_singlethread_workqueue instance.

The WQ_MEM_RECLAIM flag has been set to ensure forward progress under
memory pressure because the workqueue is being used on a memory reclaim
path.

Since there are fixed number of work items, explicit concurrency
limit is unnecessary here.

Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
---
 Changes in v2:
	-No Change

 fs/afs/flock.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/afs/flock.c b/fs/afs/flock.c
index d91a9c9..3191dff 100644
--- a/fs/afs/flock.c
+++ b/fs/afs/flock.c
@@ -36,8 +36,8 @@ static int afs_init_lock_manager(void)
 	if (!afs_lock_manager) {
 		mutex_lock(&afs_lock_manager_mutex);
 		if (!afs_lock_manager) {
-			afs_lock_manager =
-				create_singlethread_workqueue("kafs_lockd");
+			afs_lock_manager = alloc_workqueue("kafs_lockd",
+							   WQ_MEM_RECLAIM, 0);
 			if (!afs_lock_manager)
 				ret = -ENOMEM;
 		}
--
2.1.4

Re: [PATCH v2 2/4] fs/afs/rxrpc: Remove deprecated create_singlethread_workqueue

[David Howells]

Bhaktipriya Shridhar <bhaktipriya96@gmail.com> wrote:

> The workqueue "afs_async_calls" queues a single work item viz
> &call->async_work and hence it doesn't require execution ordering.
> ...
> Since there are fixed number of work items, explicit concurrency
> limit is unnecessary here.

There can be more than one single call and there's no fixed number.

> The workqueue "afs_lock_manager" queues a single work item
> &vnode->lock_work, and hence it doesn't require execution ordering.
> ...
> Since there are fixed number of work items, explicit concurrency
> limit is unnecessary here.

There may be more than one single vnode and there's no fixed number.

David

Re: [PATCH v2 2/4] fs/afs/rxrpc: Remove deprecated create_singlethread_workqueue

[Tejun Heo]

On Tue, Aug 30, 2016 at 11:15:43PM +0100, David Howells wrote:
> Bhaktipriya Shridhar <bhaktipriya96@gmail.com> wrote:
> 
> > The workqueue "afs_async_calls" queues a single work item viz
> > &call->async_work and hence it doesn't require execution ordering.
> > ...
> > Since there are fixed number of work items, explicit concurrency
> > limit is unnecessary here.
> 
> There can be more than one single call and there's no fixed number.
> 
> > The workqueue "afs_lock_manager" queues a single work item
> > &vnode->lock_work, and hence it doesn't require execution ordering.
> > ...
> > Since there are fixed number of work items, explicit concurrency
> > limit is unnecessary here.
> 
> There may be more than one single vnode and there's no fixed number.

Ah, okay, the work items are dynamically allocated.  Do they need
strict execution ordering or can they be executed concurrently?

Thanks.

-- 
tejun

Re: [PATCH v2 2/4] fs/afs/rxrpc: Remove deprecated create_singlethread_workqueue

[David Howells]

Tejun Heo <tj@kernel.org> wrote:

> Ah, okay, the work items are dynamically allocated.  Do they need
> strict execution ordering or can they be executed concurrently?

Concurrently is fine.

David

Re: [PATCH v2 2/4] fs/afs/rxrpc: Remove deprecated create_singlethread_workqueue

[Tejun Heo]

Hello, David.

On Wed, Aug 31, 2016 at 09:24:40PM +0100, David Howells wrote:
> Tejun Heo <tj@kernel.org> wrote:
> 
> > Ah, okay, the work items are dynamically allocated.  Do they need
> > strict execution ordering or can they be executed concurrently?
> 
> Concurrently is fine.

I see.  Bhaktpriya, can you please update the patch description?  The
code itself should be fine, I think.

Thanks for the clarification.

-- 
tejun

Re: [PATCH v2 2/4] fs/afs/rxrpc: Remove deprecated create_singlethread_workqueue

[Bhaktipriya Shridhar]

On Thu, Sep 1, 2016 at 2:47 AM, Tejun Heo <tj@kernel.org> wrote:
> On Wed, Aug 31, 2016 at 09:24:40PM +0100, David Howells wrote:
>> Concurrently is fine.
>
> I see.  Bhaktpriya, can you please update the patch description?  The
> code itself should be fine, I think.
>
Sure. Will update it in v2.

Thanks,
Bhaktipriya