Browse Prior Art Database

Fast Two-Pass Write

IP.com Disclosure Number: IPCOM000035859D
Original Publication Date: 1989-Aug-01
Included in the Prior Art Database: 2005-Jan-28
Document File: 4 page(s) / 113K

Publishing Venue

IBM

Related People

Gregg, LE: AUTHOR [+3]

Abstract

The technique described herein performs the erase before writing, which is required with current optical disk drives, with minimal system involvement and in as unobtrusive a manner as possible.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 4

Fast Two-Pass Write

The technique described herein performs the erase before writing, which is required with current optical disk drives, with minimal system involvement and in as unobtrusive a manner as possible.

The technique starts by making the assumption that the entire disk is initially erased so it is ready for writing. In this condition, as read or write requests are received, it is then possible to satisfy them with one revolution of the disk. A simple flow for the host handling these requests is shown in Fig. 1.

In Fig. 1, as requests are received from an application program #1 they are immediately passed on to a task in the IOP #2. Meanwhile, the

(Image Omitted)

job which issued the request waits until the I/O request has been satisfied. When the IOP has finished its work, a response is returned to the host #3. The returned status of the request is checked and the waiting job in the host resumes #4.

Most file systems have a list of active files on the disk, and they also maintain a list of the free space on the disk. In this system the maintenance of these file "directories" is done in the host computer system. The IOP handles only specific read or write requests - without attaching any special significance to the disk address being referenced, or the content of the data being transferred. This system extends the traditional directories by adding another category for space which is "free, but not erased". When a file is deleted, the space is first moved to the "free, but not erased" area. Next, the newly freed space is erased, and then the space is added to the free space list.

Now consider the case when an existing file is to be erased. A simple flow for this processing (in the host) is shown in Fig. 2.

A request to erase a file is not a simple read or write, so further checking of the request type is done. For an "erase" the space occupied by the file is first moved to the free, but not erased list #5. Then the IOP is instructed to begin erasing these areas #6. Requests made to the IOP include a priority indication. Internally, the as free space #9. If the free but not erased space must be used, then that space is immediately zeroed #10 and then the new file is allocated. Of course, if a request is made to immediately erase some space (as in #10) then it is also necessary to cancel any low priority erase requests already made to the IOP. Otherwise, the space might erroneously be erased again later when the space is part of an active file. Obviously, if there is not enough space, then an insufficient space indication must be returned #12.

The flow in the IOP must be more complicated to handle both the normal requests from the host and the low-priority erase r...