summaryrefslogtreecommitdiffstats
path: root/drivers
diff options
context:
space:
mode:
authorSagi Grimberg <sagi@grimberg.me>2022-02-01 14:54:21 +0200
committerChristoph Hellwig <hch@lst.de>2022-02-02 09:19:07 +0100
commitb6bb1722f34bbdbabed27acdceaf585d300c5fd2 (patch)
treefeefdfcf880e3b2cc695749064b184fc80a5c329 /drivers
parentff9fc7ebf5c06de1ef72a69f9b1ab40af8b07f9e (diff)
downloadlinux-b6bb1722f34bbdbabed27acdceaf585d300c5fd2.tar.bz2
nvme-rdma: fix possible use-after-free in transport error_recovery work
While nvme_rdma_submit_async_event_work is checking the ctrl and queue state before preparing the AER command and scheduling io_work, in order to fully prevent a race where this check is not reliable the error recovery work must flush async_event_work before continuing to destroy the admin queue after setting the ctrl state to RESETTING such that there is no race .submit_async_event and the error recovery handler itself changing the ctrl state. Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Diffstat (limited to 'drivers')
-rw-r--r--drivers/nvme/host/rdma.c1
1 files changed, 1 insertions, 0 deletions
diff --git a/drivers/nvme/host/rdma.c b/drivers/nvme/host/rdma.c
index 850f84d204d0..9c55e4be8a39 100644
--- a/drivers/nvme/host/rdma.c
+++ b/drivers/nvme/host/rdma.c
@@ -1200,6 +1200,7 @@ static void nvme_rdma_error_recovery_work(struct work_struct *work)
struct nvme_rdma_ctrl, err_work);
nvme_stop_keep_alive(&ctrl->ctrl);
+ flush_work(&ctrl->ctrl.async_event_work);
nvme_rdma_teardown_io_queues(ctrl, false);
nvme_start_queues(&ctrl->ctrl);
nvme_rdma_teardown_admin_queue(ctrl, false);