On Sun, 5 Feb 2006, Matthew Wilcox wrote:
> Seems to me it'd be much easier just to initialise sdev to
> ERR_PTR(-ENODEV) at the top of the function. That way, __scsi_add_device
> doesn't need to know what the return codes from scsi_probe_and_add_lun
> mean. Something like this (untested):
> @@ -1278,12 +1277,8 @@ struct scsi_device *__scsi_add_device(st
>
> get_device(&starget->dev);
> mutex_lock(&shost->scan_mutex);
> - if (scsi_host_scan_allowed(shost)) {
> - res = scsi_probe_and_add_lun(starget, lun, NULL, &sdev, 1,
> - hostdata);
> - if (res != SCSI_SCAN_LUN_PRESENT)
> - sdev = ERR_PTR(-ENODEV);
> - }
> + if (scsi_host_scan_allowed(shost))
> + scsi_probe_and_add_lun(starget, lun, NULL, &sdev, 1, hostdata);
> mutex_unlock(&shost->scan_mutex);
> scsi_target_reap(starget);
> put_device(&starget->dev);
This assumes that scsi_probe_and_add_lun doesn't assign anything to the
&sdev pointer if it returns anything other than SCSI_SCAN_LUN_PRESENT.
Since that assumption is true for the current code, this version of the
patch will work as well as mine.
Alan Stern
-
: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
