RDMA/bnxt_re: Fix hang during driver unload

Driver unload hits a hang during stress testing of load/unload.

stack trace snippet -

tasklet_kill at ffffffff9aabb8b2
bnxt_qplib_nq_stop_irq at ffffffffc0a805fb [bnxt_re]
bnxt_qplib_disable_nq at ffffffffc0a80c5b [bnxt_re]
bnxt_re_dev_uninit at ffffffffc0a67d15 [bnxt_re]
bnxt_re_remove_device at ffffffffc0a6af1d [bnxt_re]

tasklet_kill can hang if the tasklet is scheduled after it is disabled.

Modified the sequences to disable the interrupt first and synchronize
irq before disabling the tasklet.

Fixes: 1ac5a40479 ("RDMA/bnxt_re: Add bnxt_re RoCE driver")
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
Link: https://lore.kernel.org/r/1689322969-25402-3-git-send-email-selvin.xavier@broadcom.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
This commit is contained in:
Selvin Xavier
2023-07-14 01:22:49 -07:00
committed by Leon Romanovsky
parent b5bbc65512
commit 29900bf351
2 changed files with 9 additions and 10 deletions

View File

@@ -420,19 +420,19 @@ void bnxt_qplib_nq_stop_irq(struct bnxt_qplib_nq *nq, bool kill)
if (!nq->requested)
return;
tasklet_disable(&nq->nq_tasklet);
nq->requested = false;
/* Mask h/w interrupt */
bnxt_qplib_ring_nq_db(&nq->nq_db.dbinfo, nq->res->cctx, false);
/* Sync with last running IRQ handler */
synchronize_irq(nq->msix_vec);
if (kill)
tasklet_kill(&nq->nq_tasklet);
irq_set_affinity_hint(nq->msix_vec, NULL);
free_irq(nq->msix_vec, nq);
kfree(nq->name);
nq->name = NULL;
nq->requested = false;
if (kill)
tasklet_kill(&nq->nq_tasklet);
tasklet_disable(&nq->nq_tasklet);
}
void bnxt_qplib_disable_nq(struct bnxt_qplib_nq *nq)

View File

@@ -989,19 +989,18 @@ void bnxt_qplib_rcfw_stop_irq(struct bnxt_qplib_rcfw *rcfw, bool kill)
if (!creq->requested)
return;
tasklet_disable(&creq->creq_tasklet);
creq->requested = false;
/* Mask h/w interrupts */
bnxt_qplib_ring_nq_db(&creq->creq_db.dbinfo, rcfw->res->cctx, false);
/* Sync with last running IRQ-handler */
synchronize_irq(creq->msix_vec);
if (kill)
tasklet_kill(&creq->creq_tasklet);
free_irq(creq->msix_vec, rcfw);
kfree(creq->irq_name);
creq->irq_name = NULL;
creq->requested = false;
atomic_set(&rcfw->rcfw_intr_enabled, 0);
if (kill)
tasklet_kill(&creq->creq_tasklet);
tasklet_disable(&creq->creq_tasklet);
}
void bnxt_qplib_disable_rcfw_channel(struct bnxt_qplib_rcfw *rcfw)