This is a note to let you know that I've just added the patch titled RDMA/bnxt_re: Disable/kill tasklet only if it is enabled to the 6.4-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: rdma-bnxt_re-disable-kill-tasklet-only-if-it-is-enab.patch and it can be found in the queue-6.4 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. commit b4b5a32e709ed0e0fd42e4a33cd0eb431f06acb2 Author: Selvin Xavier <selvin.xavier@xxxxxxxxxxxx> Date: Thu May 18 23:48:11 2023 -0700 RDMA/bnxt_re: Disable/kill tasklet only if it is enabled [ Upstream commit ab112ee7899d6171da5acd77a7ed7ae103f488de ] When the ulp hook to start the IRQ fails because the rings are not available, tasklets are not enabled. In this case when the driver is unloaded, driver calls CREQ tasklet_kill. This causes an indefinite hang as the tasklet is not enabled. Driver shouldn't call tasklet_kill if it is not enabled. So using the creq->requested and nq->requested flags to identify if both tasklets/irqs are registered. Checking this flag while scheduling the tasklet from ISR. Also, added a cleanup for disabling tasklet, in case request_irq fails during start_irq. Check for return value for bnxt_qplib_rcfw_start_irq and in case the bnxt_qplib_rcfw_start_irq fails, return bnxt_re_start_irq without attempting to start NQ IRQs. Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") Link: https://lore.kernel.org/r/1684478897-12247-2-git-send-email-selvin.xavier@xxxxxxxxxxxx Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@xxxxxxxxxxxx> Signed-off-by: Selvin Xavier <selvin.xavier@xxxxxxxxxxxx> Signed-off-by: Jason Gunthorpe <jgg@xxxxxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c index 3073398a21834..cb4465b0792b9 100644 --- a/drivers/infiniband/hw/bnxt_re/main.c +++ b/drivers/infiniband/hw/bnxt_re/main.c @@ -283,15 +283,21 @@ static void bnxt_re_start_irq(void *handle, struct bnxt_msix_entry *ent) for (indx = 0; indx < rdev->num_msix; indx++) rdev->en_dev->msix_entries[indx].vector = ent[indx].vector; - bnxt_qplib_rcfw_start_irq(rcfw, msix_ent[BNXT_RE_AEQ_IDX].vector, - false); + rc = bnxt_qplib_rcfw_start_irq(rcfw, msix_ent[BNXT_RE_AEQ_IDX].vector, + false); + if (rc) { + ibdev_warn(&rdev->ibdev, "Failed to reinit CREQ\n"); + return; + } for (indx = BNXT_RE_NQ_IDX ; indx < rdev->num_msix; indx++) { nq = &rdev->nq[indx - 1]; rc = bnxt_qplib_nq_start_irq(nq, indx - 1, msix_ent[indx].vector, false); - if (rc) + if (rc) { ibdev_warn(&rdev->ibdev, "Failed to reinit NQ index %d\n", indx - 1); + return; + } } } diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index 8974f6235cfaa..b922181525035 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -399,6 +399,9 @@ static irqreturn_t bnxt_qplib_nq_irq(int irq, void *dev_instance) void bnxt_qplib_nq_stop_irq(struct bnxt_qplib_nq *nq, bool kill) { + if (!nq->requested) + return; + tasklet_disable(&nq->nq_tasklet); /* Mask h/w interrupt */ bnxt_qplib_ring_nq_db(&nq->nq_db.dbinfo, nq->res->cctx, false); @@ -406,11 +409,10 @@ void bnxt_qplib_nq_stop_irq(struct bnxt_qplib_nq *nq, bool kill) synchronize_irq(nq->msix_vec); if (kill) tasklet_kill(&nq->nq_tasklet); - if (nq->requested) { - irq_set_affinity_hint(nq->msix_vec, NULL); - free_irq(nq->msix_vec, nq); - nq->requested = false; - } + + irq_set_affinity_hint(nq->msix_vec, NULL); + free_irq(nq->msix_vec, nq); + nq->requested = false; } void bnxt_qplib_disable_nq(struct bnxt_qplib_nq *nq) @@ -449,8 +451,10 @@ int bnxt_qplib_nq_start_irq(struct bnxt_qplib_nq *nq, int nq_indx, snprintf(nq->name, sizeof(nq->name), "bnxt_qplib_nq-%d", nq_indx); rc = request_irq(nq->msix_vec, bnxt_qplib_nq_irq, 0, nq->name, nq); - if (rc) + if (rc) { + tasklet_disable(&nq->nq_tasklet); return rc; + } cpumask_clear(&nq->mask); cpumask_set_cpu(nq_indx, &nq->mask); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c index de90691031773..a668f877a7bf7 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c @@ -636,6 +636,10 @@ void bnxt_qplib_rcfw_stop_irq(struct bnxt_qplib_rcfw *rcfw, bool kill) struct bnxt_qplib_creq_ctx *creq; creq = &rcfw->creq; + + if (!creq->requested) + return; + tasklet_disable(&creq->creq_tasklet); /* Mask h/w interrupts */ bnxt_qplib_ring_nq_db(&creq->creq_db.dbinfo, rcfw->res->cctx, false); @@ -644,10 +648,8 @@ void bnxt_qplib_rcfw_stop_irq(struct bnxt_qplib_rcfw *rcfw, bool kill) if (kill) tasklet_kill(&creq->creq_tasklet); - if (creq->requested) { - free_irq(creq->msix_vec, rcfw); - creq->requested = false; - } + free_irq(creq->msix_vec, rcfw); + creq->requested = false; } void bnxt_qplib_disable_rcfw_channel(struct bnxt_qplib_rcfw *rcfw) @@ -693,8 +695,10 @@ int bnxt_qplib_rcfw_start_irq(struct bnxt_qplib_rcfw *rcfw, int msix_vector, tasklet_enable(&creq->creq_tasklet); rc = request_irq(creq->msix_vec, bnxt_qplib_creq_irq, 0, "bnxt_qplib_creq", rcfw); - if (rc) + if (rc) { + tasklet_disable(&creq->creq_tasklet); return rc; + } creq->requested = true; bnxt_qplib_ring_nq_db(&creq->creq_db.dbinfo, rcfw->res->cctx, true);