On 2/19/25 12:11 AM, Chaohai Chen wrote:
async_scan_lock is designed to protect the scanning_hosts list,
but there is no protection here.
Signed-off-by: Chaohai Chen <wdhh66@xxxxxxx>
---
drivers/scsi/scsi_scan.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/scsi/scsi_scan.c b/drivers/scsi/scsi_scan.c
index 087fcbfc9aaa..9a90e6ba5603 100644
--- a/drivers/scsi/scsi_scan.c
+++ b/drivers/scsi/scsi_scan.c
@@ -151,8 +151,12 @@ int scsi_complete_async_scans(void)
struct async_scan_data *data;
do {
- if (list_empty(&scanning_hosts))
+ spin_lock(&async_scan_lock);
+ if (list_empty(&scanning_hosts)) {
+ spin_unlock(&async_scan_lock);
return 0;
+ }
+ spin_unlock(&async_scan_lock);
/* If we can't get memory immediately, that's OK. Just
* sleep a little. Even if we never get memory, the async
* scans will finish eventually.
Has it been considered to use scoped_guard() as in the untested patch
below?
Thanks,
Bart.
diff --git a/drivers/scsi/scsi_scan.c b/drivers/scsi/scsi_scan.c
index 087fcbfc9aaa..efc90571ab47 100644
--- a/drivers/scsi/scsi_scan.c
+++ b/drivers/scsi/scsi_scan.c
@@ -151,8 +151,9 @@ int scsi_complete_async_scans(void)
struct async_scan_data *data;
do {
- if (list_empty(&scanning_hosts))
- return 0;
+ scoped_guard(spinlock, &async_scan_lock)
+ if (list_empty(&scanning_hosts))
+ return 0;
/* If we can't get memory immediately, that's OK. Just
* sleep a little. Even if we never get memory, the async
* scans will finish eventually.
