| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 304.1 | | NABETH::alan | Dr. File System's Home for Wayward Inodes. | Fri Mar 28 1997 11:39 | 10 |
| The Test Unit Ready code for mrd_ready(3mrd) on VMS was failing
(as expected), but not arranging for mrd_request_sense(3mrd)
to be called with the correct argument to get the Sense Data
for the failure. I've put a bandaid over it by explicitly calling
mrd_request_sense(3mrd) with SS$_DRVERR for the os_status.
The correct thing to do is have mrd_test_unit_ready(3mrd) return
an MRD_STATUS code that will indicate that it was a likely SCSI
error and go from there. An MRD_STATUS other than that one will
be fed into the dev_status_t.code and returned.
|
| 304.2 | Closing, Closing... | NABETH::alan | Dr. File System's Home for Wayward Inodes. | Mon Mar 31 1997 13:33 | 7 |
| The common startup code made use of mrd_request_sense(3mrd) in a
similar way to mrd_ready(3mrd). This has been fixed, with some
extra clean-up to make it easier for API users use
mrd_request_sense(3mrd).
The NT version of the API probably needs to make the change
as well, but Digital UNIX and OpenVMS are done.
|
| 304.3 | | NABETH::alan | Dr. File System's Home for Wayward Inodes. | Tue Apr 15 1997 17:43 | 11 |
| While testing the most recent changes to Request Sense, I
tried something a bit different. I set the MRU_READY_DELAY
to be very short and after opening the port door, let it
sit until the loop timed out. Then I started a "show slot"
and part way through closed the door.
Not surprisingly, one of the Read Element Status commands
"failed" with Port Accessed. I have to wonder it it is
appropriate for the TL820 to report this as a full up
Unit Attention. It isn't an error, it is a status report.
I'd as soon not see it at all.
|