* [PATCH v2] iwcm: don't hold the irq disabled lock on iw_rem_ref
@ 2019-09-04 1:10 Sagi Grimberg
2019-09-04 1:39 ` Mark Bloch
0 siblings, 1 reply; 2+ messages in thread
From: Sagi Grimberg @ 2019-09-04 1:10 UTC (permalink / raw)
To: linux-rdma; +Cc: Jason Gunthorpe
This may be the final put on a qp and result in freeing
resourcesand should not be done with interrupts disabled.
Produce the following warning:
--
[ 317.026048] WARNING: CPU: 1 PID: 443 at kernel/smp.c:425 smp_call_function_many+0xa0/0x260
[ 317.026131] Call Trace:
[ 317.026159] ? load_new_mm_cr3+0xe0/0xe0
[ 317.026161] on_each_cpu+0x28/0x50
[ 317.026183] __purge_vmap_area_lazy+0x72/0x150
[ 317.026200] free_vmap_area_noflush+0x7a/0x90
[ 317.026202] remove_vm_area+0x6f/0x80
[ 317.026203] __vunmap+0x71/0x210
[ 317.026211] siw_free_qp+0x8d/0x130 [siw]
[ 317.026217] destroy_cm_id+0xc3/0x200 [iw_cm]
[ 317.026222] rdma_destroy_id+0x224/0x2b0 [rdma_cm]
[ 317.026226] nvme_rdma_reset_ctrl_work+0x2c/0x70 [nvme_rdma]
[ 317.026235] process_one_work+0x1f4/0x3e0
[ 317.026249] worker_thread+0x221/0x3e0
[ 317.026252] ? process_one_work+0x3e0/0x3e0
[ 317.026256] kthread+0x117/0x130
[ 317.026264] ? kthread_create_worker_on_cpu+0x70/0x70
[ 317.026275] ret_from_fork+0x35/0x40
--
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
---
changes from v1:
- don't release the lock before qp pointer is cleared.
drivers/infiniband/core/iwcm.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
index 72141c5b7c95..ad6fd5019285 100644
--- a/drivers/infiniband/core/iwcm.c
+++ b/drivers/infiniband/core/iwcm.c
@@ -427,8 +427,10 @@ static void destroy_cm_id(struct iw_cm_id *cm_id)
break;
}
if (cm_id_priv->qp) {
- cm_id_priv->id.device->ops.iw_rem_ref(cm_id_priv->qp);
cm_id_priv->qp = NULL;
+ spin_unlock_irqrestore(&cm_id_priv->lock, flags);
+ cm_id_priv->id.device->ops.iw_rem_ref(cm_id_priv->qp);
+ spin_lock_irqsave(&cm_id_priv->lock, flags);
}
spin_unlock_irqrestore(&cm_id_priv->lock, flags);
--
2.17.1
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH v2] iwcm: don't hold the irq disabled lock on iw_rem_ref
2019-09-04 1:10 [PATCH v2] iwcm: don't hold the irq disabled lock on iw_rem_ref Sagi Grimberg
@ 2019-09-04 1:39 ` Mark Bloch
0 siblings, 0 replies; 2+ messages in thread
From: Mark Bloch @ 2019-09-04 1:39 UTC (permalink / raw)
To: Sagi Grimberg, linux-rdma; +Cc: Jason Gunthorpe
On 9/3/2019 18:10, Sagi Grimberg wrote:
> This may be the final put on a qp and result in freeing
> resourcesand should not be done with interrupts disabled.
>
> Produce the following warning:
> --
> [ 317.026048] WARNING: CPU: 1 PID: 443 at kernel/smp.c:425 smp_call_function_many+0xa0/0x260
> [ 317.026131] Call Trace:
> [ 317.026159] ? load_new_mm_cr3+0xe0/0xe0
> [ 317.026161] on_each_cpu+0x28/0x50
> [ 317.026183] __purge_vmap_area_lazy+0x72/0x150
> [ 317.026200] free_vmap_area_noflush+0x7a/0x90
> [ 317.026202] remove_vm_area+0x6f/0x80
> [ 317.026203] __vunmap+0x71/0x210
> [ 317.026211] siw_free_qp+0x8d/0x130 [siw]
> [ 317.026217] destroy_cm_id+0xc3/0x200 [iw_cm]
> [ 317.026222] rdma_destroy_id+0x224/0x2b0 [rdma_cm]
> [ 317.026226] nvme_rdma_reset_ctrl_work+0x2c/0x70 [nvme_rdma]
> [ 317.026235] process_one_work+0x1f4/0x3e0
> [ 317.026249] worker_thread+0x221/0x3e0
> [ 317.026252] ? process_one_work+0x3e0/0x3e0
> [ 317.026256] kthread+0x117/0x130
> [ 317.026264] ? kthread_create_worker_on_cpu+0x70/0x70
> [ 317.026275] ret_from_fork+0x35/0x40
> --
>
> Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
> ---
> changes from v1:
> - don't release the lock before qp pointer is cleared.
>
> drivers/infiniband/core/iwcm.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
> index 72141c5b7c95..ad6fd5019285 100644
> --- a/drivers/infiniband/core/iwcm.c
> +++ b/drivers/infiniband/core/iwcm.c
> @@ -427,8 +427,10 @@ static void destroy_cm_id(struct iw_cm_id *cm_id)
> break;
> }
> if (cm_id_priv->qp) {
> - cm_id_priv->id.device->ops.iw_rem_ref(cm_id_priv->qp);
> cm_id_priv->qp = NULL;
> + spin_unlock_irqrestore(&cm_id_priv->lock, flags);
> + cm_id_priv->id.device->ops.iw_rem_ref(cm_id_priv->qp);
You are calling it with NULL now :)
I was thinking more about something like this:
diff --git a/drivers/infiniband/core/iwcm.c b/drivers/infiniband/core/iwcm.c
index 72141c5b7c95..c5d9a1ebace6 100644
--- a/drivers/infiniband/core/iwcm.c
+++ b/drivers/infiniband/core/iwcm.c
@@ -373,6 +373,7 @@ static void destroy_cm_id(struct iw_cm_id *cm_id)
{
struct iwcm_id_private *cm_id_priv;
unsigned long flags;
+ struct ib_qp *qp;
cm_id_priv = container_of(cm_id, struct iwcm_id_private, id);
/*
@@ -426,12 +427,14 @@ static void destroy_cm_id(struct iw_cm_id *cm_id)
BUG();
break;
}
- if (cm_id_priv->qp) {
- cm_id_priv->id.device->ops.iw_rem_ref(cm_id_priv->qp);
- cm_id_priv->qp = NULL;
- }
+
+ qp = cm_id_priv->qp;
+ cm_id_priv->qp = NULL;
spin_unlock_irqrestore(&cm_id_priv->lock, flags);
+ if (qp)
+ cm_id_priv->id.device->ops.iw_rem_ref(qp);
+
if (cm_id->mapped) {
iwpm_remove_mapinfo(&cm_id->local_addr, &cm_id->m_local_addr);
iwpm_remove_mapping(&cm_id->local_addr, RDMA_NL_IWCM);
> + spin_lock_irqsave(&cm_id_priv->lock, flags);
> }
> spin_unlock_irqrestore(&cm_id_priv->lock, flags);
>
>
^ permalink raw reply related [flat|nested] 2+ messages in thread
end of thread, other threads:[~2019-09-04 1:39 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-09-04 1:10 [PATCH v2] iwcm: don't hold the irq disabled lock on iw_rem_ref Sagi Grimberg
2019-09-04 1:39 ` Mark Bloch
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).