Skip to contentSkip to conversation

drivers/nvme/host/core.c:5259 nvme_wait_freeze_timeout() warn: 'timeout' unsigned <= 0

PATCH Dan Carpenter 2 participants 3 messages May 21 NVMe

lore
Share & tools
lore

Dan Carpenter

author build bot
#0 lore raw

BCC: [email protected]
CC: [email protected]
CC: [email protected]
TO: Maurizio Lombardi <[email protected]>
CC: Keith Busch <[email protected]>
CC: Daniel Wagner <[email protected]>
CC: Mohamed Khalfella <[email protected]>
CC: Hannes Reinecke <[email protected]>
CC: Christoph Hellwig <[email protected]>

tree:   git://git.infradead.org/nvme.git nvme-7.2
head:   00d7b33351aac0ea55d17167561e12bbeca73138
commit: 23b6d2cbf75ff15647efbb7c0e5c03bd7ed1fe1a [2/7] nvme: remove redundant timeout argument from nvme_wait_freeze_timeout

:::::: branch date: 10 hours ago
:::::: commit date: 10 hours ago
config: s390-randconfig-r073-20260521 (https://download.01.org/0day-ci/archive/20260521/[email protected]/config)
compiler: clang version 18.1.8 (https://github.com/llvm/llvm-project 3b5b5c1ec4a3095ab096dd780e84d7ab81f3d7ff)
smatch: v0.5.0-9185-gbcc58b9c

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags

| Reported-by: kernel test robot <[email protected]>
| Reported-by: Dan Carpenter <[email protected]>
| Closes: https://lore.kernel.org/r/[email protected]/

smatch warnings:
drivers/nvme/host/core.c:5259 nvme_wait_freeze_timeout() warn: 'timeout' unsigned <= 0

vim +/timeout +5259 drivers/nvme/host/core.c

302ad8cc09339ea Keith Busch       2017-03-01  5248  
23b6d2cbf75ff15 Maurizio Lombardi 2026-05-14  5249  int nvme_wait_freeze_timeout(struct nvme_ctrl *ctrl)
302ad8cc09339ea Keith Busch       2017-03-01  5250  {
23b6d2cbf75ff15 Maurizio Lombardi 2026-05-14  5251  	unsigned long timeout = NVME_IO_TIMEOUT;
302ad8cc09339ea Keith Busch       2017-03-01  5252  	struct nvme_ns *ns;
be647e2c76b27f4 Keith Busch       2024-05-21  5253  	int srcu_idx;
302ad8cc09339ea Keith Busch       2017-03-01  5254  
be647e2c76b27f4 Keith Busch       2024-05-21  5255  	srcu_idx = srcu_read_lock(&ctrl->srcu);
6d1c69945ce63a9 Breno Leitao      2024-11-04  5256  	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
6d1c69945ce63a9 Breno Leitao      2024-11-04  5257  				 srcu_read_lock_held(&ctrl->srcu)) {
302ad8cc09339ea Keith Busch       2017-03-01  5258  		timeout = blk_mq_freeze_queue_wait_timeout(ns->queue, timeout);
302ad8cc09339ea Keith Busch       2017-03-01 @5259  		if (timeout <= 0)
302ad8cc09339ea Keith Busch       2017-03-01  5260  			break;
302ad8cc09339ea Keith Busch       2017-03-01  5261  	}
be647e2c76b27f4 Keith Busch       2024-05-21  5262  	srcu_read_unlock(&ctrl->srcu, srcu_idx);
7cf0d7c0f3c3b02 Sagi Grimberg     2020-07-30  5263  	return timeout;
302ad8cc09339ea Keith Busch       2017-03-01  5264  }
302ad8cc09339ea Keith Busch       2017-03-01  5265  EXPORT_SYMBOL_GPL(nvme_wait_freeze_timeout);
302ad8cc09339ea Keith Busch       2017-03-01  5266  

:::::: The code at line 5259 was first introduced by commit
:::::: 302ad8cc09339ea261eef58a8d5f4a116a8ffda5 nvme: Complete all stuck requests

:::::: TO: Keith Busch <[email protected]>
:::::: CC: Jens Axboe <[email protected]>

--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

#1 lore raw
Quoted 16 lines · Dan Carpenter
tree:   git://git.infradead.org/nvme.git nvme-7.2
head:   00d7b33351aac0ea55d17167561e12bbeca73138
commit: 23b6d2cbf75ff15647efbb7c0e5c03bd7ed1fe1a [2/7] nvme: remove redundant timeout argument from nvme_wait_freeze_timeout
config: s390-randconfig-r073-20260521 (https://download.01.org/0day-ci/archive/20260521/[email protected]/config)
compiler: clang version 18.1.8 (https://github.com/llvm/llvm-project 3b5b5c1ec4a3095ab096dd780e84d7ab81f3d7ff)
smatch: v0.5.0-9185-gbcc58b9c

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags

| Reported-by: kernel test robot <[email protected]>
| Reported-by: Dan Carpenter <[email protected]>
| Closes: https://lore.kernel.org/r/[email protected]/

smatch warnings:
drivers/nvme/host/core.c:5259 nvme_wait_freeze_timeout() warn: 'timeout' unsigned <= 0

Quoted 8 lines

vim +/timeout +5259 drivers/nvme/host/core.c

23b6d2cbf75ff15 Maurizio Lombardi 2026-05-14  5249  int nvme_wait_freeze_timeout(struct nvme_ctrl *ctrl)
302ad8cc09339ea Keith Busch       2017-03-01  5250  {
23b6d2cbf75ff15 Maurizio Lombardi 2026-05-14  5251  	unsigned long timeout = NVME_IO_TIMEOUT;
                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This is unsigned. NVME_IO_TIMEOUT is 30 * HZ. Can the &ctrl->namespaces list ever be empty?

I already had a patch for this, I am going to submit it today.

https://lore.kernel.org/linux-nvme/[email protected]/

Thanks.

Quoted 25 lines
302ad8cc09339ea Keith Busch       2017-03-01  5252  	struct nvme_ns *ns;
be647e2c76b27f4 Keith Busch       2024-05-21  5253  	int srcu_idx;
302ad8cc09339ea Keith Busch       2017-03-01  5254  
be647e2c76b27f4 Keith Busch       2024-05-21  5255  	srcu_idx = srcu_read_lock(&ctrl->srcu);
6d1c69945ce63a9 Breno Leitao      2024-11-04  5256  	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
6d1c69945ce63a9 Breno Leitao      2024-11-04  5257  				 srcu_read_lock_held(&ctrl->srcu)) {
302ad8cc09339ea Keith Busch       2017-03-01  5258  		timeout = blk_mq_freeze_queue_wait_timeout(ns->queue, timeout);
302ad8cc09339ea Keith Busch       2017-03-01 @5259  		if (timeout <= 0)

blk_mq_freeze_queue_wait_timeout() returns 0 on timeout and >= 1 on success.

302ad8cc09339ea Keith Busch       2017-03-01  5260  			break;
302ad8cc09339ea Keith Busch       2017-03-01  5261  	}
be647e2c76b27f4 Keith Busch       2024-05-21  5262  	srcu_read_unlock(&ctrl->srcu, srcu_idx);
7cf0d7c0f3c3b02 Sagi Grimberg     2020-07-30  5263  	return timeout;

This returns is nonsense. ;) >= 1 on success or empty. 0 on timeout.

302ad8cc09339ea Keith Busch 2017-03-01 5264 }

--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

#2 lore raw
In reply to Dan Carpenter
Quoted 8 lines · Dan Carpenter
tree:   git://git.infradead.org/nvme.git nvme-7.2
head:   00d7b33351aac0ea55d17167561e12bbeca73138
commit: 23b6d2cbf75ff15647efbb7c0e5c03bd7ed1fe1a [2/7] nvme: remove redundant timeout argument from nvme_wait_freeze_timeout
config: s390-randconfig-r073-20260521 (https://download.01.org/0day-ci/archive/20260521/[email protected]/config)
compiler: clang version 18.1.8 (https://github.com/llvm/llvm-project 3b5b5c1ec4a3095ab096dd780e84d7ab81f3d7ff)
smatch: v0.5.0-9185-gbcc58b9c
Quoted 6 lines
302ad8cc09339ea Keith Busch       2017-03-01  5260  			break;
302ad8cc09339ea Keith Busch       2017-03-01  5261  	}
be647e2c76b27f4 Keith Busch       2024-05-21  5262  	srcu_read_unlock(&ctrl->srcu, srcu_idx);
7cf0d7c0f3c3b02 Sagi Grimberg     2020-07-30  5263  	return timeout;

This returns is nonsense. ;) >= 1 on success or empty. 0 on timeout.

What about 0 on success, -ETIMEDOUT on timeout?

nvme: standardize nvme_wait_freeze_timeout return value

Currently, nvme_wait_freeze_timeout() returns the remaining timeout value (in jiffies) on success, and 0 on timeout.

Update the function to follow standard kernel conventions by returning 0 on success and -ETIMEDOUT on failure.

---
 drivers/nvme/host/core.c | 7 +++++--
 drivers/nvme/host/rdma.c | 2 +-
 drivers/nvme/host/tcp.c  | 2 +-
 3 files changed, 7 insertions(+), 4 deletions(-)
drivers/nvme/host/core.c +5 -2
index 26a65404279f..85d94d3b6a28 100644
@@ -5249,16 +5249,19 @@ int nvme_wait_freeze_timeout(struct nvme_ctrl *ctrl)
 	unsigned long timeout = ctrl->io_timeout;
 	struct nvme_ns *ns;
 	int srcu_idx;
+	int ret = 0;
 
 	srcu_idx = srcu_read_lock(&ctrl->srcu);
 	list_for_each_entry_srcu(ns, &ctrl->namespaces, list,
 				 srcu_read_lock_held(&ctrl->srcu)) {
 		timeout = blk_mq_freeze_queue_wait_timeout(ns->queue, timeout);
-		if (!timeout)
+		if (!timeout) {
+			ret = -ETIMEDOUT;
 			break;
+		}
 	}
 	srcu_read_unlock(&ctrl->srcu, srcu_idx);
-	return timeout;
+	return ret;
 }
 EXPORT_SYMBOL_GPL(nvme_wait_freeze_timeout);
 
drivers/nvme/host/rdma.c +1 -1
index bf73135c1439..3ac0a7e390e1 100644
@@ -888,7 +888,7 @@ static int nvme_rdma_configure_io_queues(struct nvme_rdma_ctrl *ctrl, bool new)
 	if (!new) {
 		nvme_start_freeze(&ctrl->ctrl);
 		nvme_unquiesce_io_queues(&ctrl->ctrl);
-		if (!nvme_wait_freeze_timeout(&ctrl->ctrl)) {
+		if (nvme_wait_freeze_timeout(&ctrl->ctrl)) {
 			/*
 			 * If we timed out waiting for freeze we are likely to
 			 * be stuck.  Fail the controller initialization just
drivers/nvme/host/tcp.c +1 -1
index 0552aa8a1150..52bd4478181e 100644
@@ -2208,7 +2208,7 @@ static int nvme_tcp_configure_io_queues(struct nvme_ctrl *ctrl, bool new)
 	if (!new) {
 		nvme_start_freeze(ctrl);
 		nvme_unquiesce_io_queues(ctrl);
-		if (!nvme_wait_freeze_timeout(ctrl)) {
+		if (nvme_wait_freeze_timeout(ctrl)) {
 			/*
 			 * If we timed out waiting for freeze we are likely to
 			 * be stuck.  Fail the controller initialization just

--
2.54.0

Maurizio