RDMA/rxe: Use a dedicated and robust workqueue for RXE tasks
Subsystems (2)
PATCH 1/1 RDMA/rxe: Use a dedicated and robust workqueue for RXE tasks
Currently, the RXE driver uses the system-wide 'system_unbound_wq' for auxiliary tasks like ODP prefetching. This can lead to interference from other system services and lacks guaranteed forward progress under memory pressure.
Currently make all the tasks queue into the driver-specific 'rxe_wq'.
--- drivers/infiniband/sw/rxe/rxe_odp.c | 2 +- drivers/infiniband/sw/rxe/rxe_task.c | 10 +++++++++- drivers/infiniband/sw/rxe/rxe_task.h | 1 + 3 files changed, 11 insertions(+), 2 deletions(-)
drivers/infiniband/sw/rxe/rxe_odp.c +1 -1
@@ -545,7 +545,7 @@ static int rxe_ib_advise_mr_prefetch(struct ib_pd *ibpd, work->frags[i].mr = mr; } -queue_work (system_unbound_wq, &work->work); +rxe_queue_work (&work->work);return 0 ;
drivers/infiniband/sw/rxe/rxe_task.c +9 -1
@@ -10,7 +10,8 @@ static struct workqueue_struct *rxe_wq;int rxe_alloc_wq (void ) { - rxe_wq =alloc_workqueue ("rxe_wq" ,WQ_UNBOUND ,WQ_MAX_ACTIVE ); + rxe_wq =alloc_workqueue ("rxe_wq" ,WQ_UNBOUND |WQ_MEM_RECLAIM , +WQ_MAX_ACTIVE );if (!rxe_wq)return -ENOMEM ;
@@ -254,6 +255,13 @@ void rxe_sched_task(struct rxe_task *task)spin_unlock_irqrestore (&task->lock, flags); } +/* Helper to queue auxiliary tasks into rxe_wq. +*/ +void rxe_queue_work (struct work_struct *work) +{ +queue_work (rxe_wq, work); +} +/* rxe_disable/enable_task are only called from * rxe_modify_qp in process context. Task is moved * to the drained state by do_task.
drivers/infiniband/sw/rxe/rxe_task.h +1 -0
@@ -36,6 +36,7 @@ int rxe_alloc_wq(void);void rxe_destroy_wq (void ); +void rxe_queue_work (struct work_struct *work);/* * init rxe_task structure * qp => parameter to pass to func
Signature · 1 line
2.43.0
Quoted 33 lines · Zhu Yanjun
Currently, the RXE driver uses the system-wide '
system_unbound_wq' for auxiliary tasks like ODP prefetching. This can lead to interference from other system services and lacks guaranteed forward progress under memory pressure.Currently make all the tasks queue into the driver-specific '
rxe_wq'.Suggested-by Leon Romanovsky <[email protected]>Signed-off-by Zhu Yanjun <[email protected]>--- drivers/infiniband/sw/rxe/rxe_odp.c | 2 +- drivers/infiniband/sw/rxe/rxe_task.c | 10 +++++++++- drivers/infiniband/sw/rxe/rxe_task.h | 1 + 3 files changed, 11 insertions(+), 2 deletions(-)drivers/infiniband/sw/rxe/rxe_odp.c +1
-1@@ -545,7 +545,7 @@ static int rxe_ib_advise_mr_prefetch(struct ib_pd *ibpd, work->frags[i].mr = mr; } -queue_work (system_unbound_wq, &work->work); +rxe_queue_work (&work->work);return 0 ;drivers/infiniband/sw/rxe/rxe_task.c +0
-0
drivers/infiniband/sw/rxe/rxe_task.c static struct workqueue_struct *rxe_wq; · line 13
Quoted 5 lines
drivers/infiniband/sw/rxe/rxe_task.c +1
-1@@ -10,7 +10,8 @@ static struct workqueue_struct *rxe_wq;int rxe_alloc_wq (void ) { - rxe_wq =alloc_workqueue ("rxe_wq" ,WQ_UNBOUND ,WQ_MAX_ACTIVE ); + rxe_wq = alloc_workqueue("rxe_wq", WQ_UNBOUND | WQ_MEM_RECLAIM,
Why did you add WQ_MEM_RECLAIM flag? rxe_ib_advise_mr_prefetch() doesn't perform any memory reclaim.
Thanks
Quoted 35 lines · Leon Romanovsky
Quoted 34 lines · Zhu Yanjun
Currently, the RXE driver uses the system-wide '
system_unbound_wq' for auxiliary tasks like ODP prefetching. This can lead to interference from other system services and lacks guaranteed forward progress under memory pressure.Currently make all the tasks queue into the driver-specific '
rxe_wq'.Suggested-by Leon Romanovsky <[email protected]>Signed-off-by Zhu Yanjun <[email protected]>--- drivers/infiniband/sw/rxe/rxe_odp.c | 2 +- drivers/infiniband/sw/rxe/rxe_task.c | 10 +++++++++- drivers/infiniband/sw/rxe/rxe_task.h | 1 + 3 files changed, 11 insertions(+), 2 deletions(-)drivers/infiniband/sw/rxe/rxe_odp.c +1
-1@@ -545,7 +545,7 @@ static int rxe_ib_advise_mr_prefetch(struct ib_pd *ibpd, work->frags[i].mr = mr; } -queue_work (system_unbound_wq, &work->work); +rxe_queue_work (&work->work);return 0 ;drivers/infiniband/sw/rxe/rxe_task.c +0
-0@@ -10,7 +10,8 @@ static struct workqueue_struct *rxe_wq;
Quoted 6 lines
drivers/infiniband/sw/rxe/rxe_task.c static struct workqueue_struct *rxe_wq; · line 13
Quoted 4 lines
drivers/infiniband/sw/rxe/rxe_task.c +1
-1int rxe_alloc_wq (void ) { - rxe_wq =alloc_workqueue ("rxe_wq" ,WQ_UNBOUND ,WQ_MAX_ACTIVE ); + rxe_wq = alloc_workqueue("rxe_wq", WQ_UNBOUND | WQ_MEM_RECLAIM,Why did you add
WQ_MEM_RECLAIMflag?rxe_ib_advise_mr_prefetch()doesn't perform any memory reclaim.
You are correct that rxe_ib_advise_mr_prefetch() does not directly call memory reclaim functions.
However, the WQ_MEM_RECLAIM flag was added to prevent circular dependencies during
low-memory conditions.
Since rxe handles memory regions that may be part of the storage or network stack,
the workqueue must be able to make progress even when the system is under extreme
memory pressure. Without this flag, if the kernel attempts to reclaim memory and that
reclaim process depends on an RDMA operation being processed by this workqueue,
the system could deadlock because the workqueue might be unable to spawn a new
worker thread.
By setting WQ_MEM_RECLAIM, we ensure that a rescuer thread is pre-allocated,
guaranteeing that prefetch and MR-related tasks can complete and allow the
memory management subsystem to finish its reclaim cycle.
Zhu Yanjun
Quoted 2 lines
Thanks
Signature · 2 lines
Best Regards, Yanjun.Zhu
Quoted 63 lines · Zhu Yanjun
Quoted 29 lines · Leon Romanovsky
Quoted 28 lines · Zhu Yanjun
Currently, the RXE driver uses the system-wide '
system_unbound_wq' for auxiliary tasks like ODP prefetching. This can lead to interference from other system services and lacks guaranteed forward progress under memory pressure.Currently make all the tasks queue into the driver-specific '
rxe_wq'.Suggested-by Leon Romanovsky <[email protected]>Signed-off-by Zhu Yanjun <[email protected]>--- drivers/infiniband/sw/rxe/rxe_odp.c | 2 +- drivers/infiniband/sw/rxe/rxe_task.c | 10 +++++++++- drivers/infiniband/sw/rxe/rxe_task.h | 1 + 3 files changed, 11 insertions(+), 2 deletions(-)drivers/infiniband/sw/rxe/rxe_odp.c +1
-1@@ -545,7 +545,7 @@ static int rxe_ib_advise_mr_prefetch(struct ib_pd *ibpd, work->frags[i].mr = mr; } -queue_work (system_unbound_wq, &work->work); +rxe_queue_work (&work->work);return 0 ;drivers/infiniband/sw/rxe/rxe_task.c +0
-1--- a/drivers/infiniband/sw/rxe/rxe_task.cQuoted 8 lines
drivers/infiniband/sw/rxe/rxe_task.c static struct workqueue_struct *rxe_wq; · line 12
Quoted 6 lines
+++
b/drivers/infiniband/sw/rxe/rxe_task.cdrivers/infiniband/sw/rxe/rxe_task.c +1
-1@@ -10,7 +10,8 @@ static struct workqueue_struct *rxe_wq;int rxe_alloc_wq (void ) { - rxe_wq =alloc_workqueue ("rxe_wq" ,WQ_UNBOUND ,WQ_MAX_ACTIVE ); + rxe_wq = alloc_workqueue("rxe_wq", WQ_UNBOUND | WQ_MEM_RECLAIM,Why did you add
WQ_MEM_RECLAIMflag?rxe_ib_advise_mr_prefetch()doesn't perform any memory reclaim.You are correct that
rxe_ib_advise_mr_prefetch()does not directly call memory reclaim functions.However, the
WQ_MEM_RECLAIMflag was added to prevent circular dependencies duringlow-memory conditions.
Since rxe handles memory regions that may be part of the storage or network stack,
the workqueue must be able to make progress even when the system is under extreme
memory pressure. Without this flag, if the kernel attempts to reclaim memory and that
reclaim process depends on an RDMA operation being processed by this workqueue,
the system could deadlock because the workqueue might be unable to spawn a new
Quoted 7 lines
worker thread.
By setting
WQ_MEM_RECLAIM, we ensure that a rescuer thread is pre-allocated,guaranteeing that prefetch and MR-related tasks can complete and allow the
memory management subsystem to finish its reclaim cycle.
Zhu,
Please avoid relying on AI when answering ML-related questions. The response you received is broadly correct, but it is incorrect for RXE. You should set the WQ_MEM_RECLAIM flag only when the workqueue handlers free memory. RXE does the opposite in rxe_ib_advise_mr_prefetch().
Thanks
Quoted 12 lines
Zhu Yanjun
Quoted 2 lines
Thanks
Signature · 2 lines
Best Regards, Yanjun.Zhu
Quoted 66 lines · Leon Romanovsky
Quoted 65 lines · Zhu Yanjun
Quoted 29 lines · Leon Romanovsky
Quoted 28 lines · Zhu Yanjun
Currently, the RXE driver uses the system-wide '
system_unbound_wq' for auxiliary tasks like ODP prefetching. This can lead to interference from other system services and lacks guaranteed forward progress under memory pressure.Currently make all the tasks queue into the driver-specific '
rxe_wq'.Suggested-by Leon Romanovsky <[email protected]>Signed-off-by Zhu Yanjun <[email protected]>--- drivers/infiniband/sw/rxe/rxe_odp.c | 2 +- drivers/infiniband/sw/rxe/rxe_task.c | 10 +++++++++- drivers/infiniband/sw/rxe/rxe_task.h | 1 + 3 files changed, 11 insertions(+), 2 deletions(-)drivers/infiniband/sw/rxe/rxe_odp.c +1
-1@@ -545,7 +545,7 @@ static int rxe_ib_advise_mr_prefetch(struct ib_pd *ibpd, work->frags[i].mr = mr; } -queue_work (system_unbound_wq, &work->work); +rxe_queue_work (&work->work);return 0 ;drivers/infiniband/sw/rxe/rxe_task.c +0
-1--- a/drivers/infiniband/sw/rxe/rxe_task.cQuoted 8 lines
drivers/infiniband/sw/rxe/rxe_task.c static struct workqueue_struct *rxe_wq; · line 12
Quoted 6 lines
+++
b/drivers/infiniband/sw/rxe/rxe_task.cdrivers/infiniband/sw/rxe/rxe_task.c +1
-1@@ -10,7 +10,8 @@ static struct workqueue_struct *rxe_wq;int rxe_alloc_wq (void ) { - rxe_wq =alloc_workqueue ("rxe_wq" ,WQ_UNBOUND ,WQ_MAX_ACTIVE ); + rxe_wq = alloc_workqueue("rxe_wq", WQ_UNBOUND | WQ_MEM_RECLAIM,Why did you add
WQ_MEM_RECLAIMflag?rxe_ib_advise_mr_prefetch()doesn't perform any memory reclaim.You are correct that
rxe_ib_advise_mr_prefetch()does not directly call memory reclaim functions.However, the
WQ_MEM_RECLAIMflag was added to prevent circular dependencies duringlow-memory conditions.
Since rxe handles memory regions that may be part of the storage or network stack,
the workqueue must be able to make progress even when the system is under extreme
memory pressure. Without this flag, if the kernel attempts to reclaim memory and that
reclaim process depends on an RDMA operation being processed by this workqueue,
the system could deadlock because the workqueue might be unable to spawn a new
worker thread.
By setting
WQ_MEM_RECLAIM, we ensure that a rescuer thread is pre-allocated,
Quoted 8 lines
Quoted 3 lines
guaranteeing that prefetch and MR-related tasks can complete and allow the
memory management subsystem to finish its reclaim cycle.
Zhu,
Please avoid relying on AI when answering ML-related questions. The response you received is broadly correct, but it is incorrect for RXE. You should set the
WQ_MEM_RECLAIMflag only when the workqueue handlers
OK. Thanks a lot.
Zhu Yanjun
Quoted 13 lines
free memory. RXE does the opposite in
rxe_ib_advise_mr_prefetch().Thanks
Quoted 9 lines
Zhu Yanjun
Quoted 1 line
Thanks
Signature · 2 lines
Best Regards, Yanjun.Zhu