diff mbox series

[01/17] pmcraid: add missing scsi_device_put() in pmcraid_eh_target_reset_handler()

Message ID 20231016092430.55557-2-hare@suse.de (mailing list archive)
State Superseded
Headers show
Series scsi: EH rework prep patches, part 2 | expand

Commit Message

Hannes Reinecke Oct. 16, 2023, 9:24 a.m. UTC 
When breaking out of a shost_for_each_device() loop one need to do
an explicit scsi_device_put(). And while at it convert to use
shost_priv() instead of a direct reference to ->hostdata.

Fixes: c2a14ab3b9b3 ("scsi: pmcraid: Select device in pmcraid_eh_target_reset_handler()")
Signed-off-by: Hannes Reinecke <hare@suse.de>
---
 drivers/scsi/pmcraid.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

Comments

Christoph Hellwig Oct. 16, 2023, 1:29 p.m. UTC | #1 
On Mon, Oct 16, 2023 at 11:24:14AM +0200, Hannes Reinecke wrote:
> When breaking out of a shost_for_each_device() loop one need to do
> an explicit scsi_device_put(). And while at it convert to use
> shost_priv() instead of a direct reference to ->hostdata.

While both part of the patch looks good, mixing these totally unrelated
bits doesn't seem like a good idea.
Hannes Reinecke Oct. 16, 2023, 1:37 p.m. UTC | #2 
On 10/16/23 15:29, Christoph Hellwig wrote:
> On Mon, Oct 16, 2023 at 11:24:14AM +0200, Hannes Reinecke wrote:
>> When breaking out of a shost_for_each_device() loop one need to do
>> an explicit scsi_device_put(). And while at it convert to use
>> shost_priv() instead of a direct reference to ->hostdata.
> 
> While both part of the patch looks good, mixing these totally unrelated
> bits doesn't seem like a good idea.
> 
I can easily send this as a separate patch, seeing that it's a bugfix to 
the previous merge.

Cheers,

Hannes
diff mbox series

Patch

diff --git a/drivers/scsi/pmcraid.c b/drivers/scsi/pmcraid.c
index a831b34c08a4..d946fb014474 100644
--- a/drivers/scsi/pmcraid.c
+++ b/drivers/scsi/pmcraid.c
@@ -2702,8 +2702,7 @@  static int pmcraid_reset_device(
 	unsigned long lock_flags;
 	u32 ioasc;
 
-	pinstance =
-		(struct pmcraid_instance *)scsi_dev->host->hostdata;
+	pinstance = shost_priv(scsi_dev->host);
 	res = scsi_dev->hostdata;
 
 	if (!res) {
@@ -3026,8 +3025,7 @@  static int pmcraid_eh_device_reset_handler(struct scsi_cmnd *scmd)
 static int pmcraid_eh_bus_reset_handler(struct scsi_cmnd *scmd)
 {
 	struct Scsi_Host *host = scmd->device->host;
-	struct pmcraid_instance *pinstance =
-		(struct pmcraid_instance *)host->hostdata;
+	struct pmcraid_instance *pinstance = shost_priv(host);
 	struct pmcraid_resource_entry *res = NULL;
 	struct pmcraid_resource_entry *temp;
 	struct scsi_device *sdev = NULL;
@@ -3066,6 +3064,7 @@  static int pmcraid_eh_target_reset_handler(struct scsi_cmnd *scmd)
 {
 	struct Scsi_Host *shost = scmd->device->host;
 	struct scsi_device *scsi_dev = NULL, *tmp;
+	int ret;
 
 	shost_for_each_device(tmp, shost) {
 		if ((tmp->channel == scmd->device->channel) &&
@@ -3078,9 +3077,11 @@  static int pmcraid_eh_target_reset_handler(struct scsi_cmnd *scmd)
 		return FAILED;
 	sdev_printk(KERN_INFO, scsi_dev,
 		    "Doing target reset due to an I/O command timeout.\n");
-	return pmcraid_reset_device(scsi_dev,
-				    PMCRAID_INTERNAL_TIMEOUT,
-				    RESET_DEVICE_TARGET);
+	ret = pmcraid_reset_device(scsi_dev,
+				   PMCRAID_INTERNAL_TIMEOUT,
+				   RESET_DEVICE_TARGET);
+	scsi_device_put(scsi_dev);
+	return ret;
 }
 
 /**