scsi: core: sysfs: Remove comments that conflict with the actual logic
authorJackie Liu <liuyun01@kylinos.cn>
Tue, 29 Mar 2022 02:12:51 +0000 (10:12 +0800)
committerMartin K. Petersen <martin.petersen@oracle.com>
Wed, 30 Mar 2022 03:54:26 +0000 (23:54 -0400)
Christoph Hellwig Says:
=======================
I think we should just handle the error properly and remove the comment.
There's no good reason to ignore bsg registration errors.

In fact, after commit 92c4b58b15c5 ("scsi: core: Register sysfs attributes
earlier"), we are already forced to return errno.

We discuss this issue in [1].

[1] https://lore.kernel.org/all/20211022010201.426746-1-liu.yun@linux.dev/

Link: https://lore.kernel.org/r/20220329021251.123805-1-liu.yun@linux.dev
Suggested-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jackie Liu <liuyun01@kylinos.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
drivers/scsi/scsi_sysfs.c

index 226a50944c005c55a1b2b2ce76b40cc4089247cd..dc6872e352bd4ea09b5cf7c9c945c2631c1af355 100644 (file)
@@ -1384,10 +1384,6 @@ int scsi_sysfs_add_sdev(struct scsi_device *sdev)
        if (IS_ENABLED(CONFIG_BLK_DEV_BSG)) {
                sdev->bsg_dev = scsi_bsg_register_queue(sdev);
                if (IS_ERR(sdev->bsg_dev)) {
-                       /*
-                        * We're treating error on bsg register as non-fatal, so
-                        * pretend nothing went wrong.
-                        */
                        error = PTR_ERR(sdev->bsg_dev);
                        sdev_printk(KERN_INFO, sdev,
                                    "Failed to register bsg queue, errno=%d\n",