blob: 298eac153284065a599459de91e4d183f381f65b [file] [log] [blame]
From bbd940b05d7be43a63aca9896b5014c4341b4ba8 Mon Sep 17 00:00:00 2001
From: Balbir Singh <>
Date: Wed, 18 Sep 2019 00:27:20 +0000
Subject: [PATCH] nvme-pci: Fix a race in controller removal
commit b224726de5e496dbf78147a66755c3d81e28bdd2 upstream.
User space programs like udevd may try to read to partitions at the
same time the driver detects a namespace is unusable, and may deadlock
if revalidate_disk() is called while such a process is waiting to
enter the frozen queue. On detecting a dead namespace, move the disk
revalidate after unblocking dispatchers that may be holding bd_butex.
changelog Suggested-by: Keith Busch <>
Signed-off-by: Balbir Singh <>
Reviewed-by: Keith Busch <>
Signed-off-by: Sagi Grimberg <>
Signed-off-by: Paul Gortmaker <>
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 963b4c6309b9..b174949d1fd2 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -103,10 +103,13 @@ static void nvme_set_queue_dying(struct nvme_ns *ns)
if (!ns->disk || test_and_set_bit(NVME_NS_DEAD, &ns->flags))
- revalidate_disk(ns->disk);
/* Forcibly unquiesce queues to avoid blocking dispatch */
+ /*
+ * Revalidate after unblocking dispatchers that may be holding bd_butex
+ */
+ revalidate_disk(ns->disk);
static void nvme_queue_scan(struct nvme_ctrl *ctrl)