wilker@gauss.math.purdue.edu (Clarence Wilkerson) (01/23/91)
I believe that the 8meg limit is set in the coding of the BDOS, and inforced in the BIOS in the sense that certain scratch ram in the BIOS data area has a preset size. I think several replacement BDOS clones, e.g. DOSPLUS, offer larger maximum logical disk size. There would be some perfomance penalties for using larger sizes, namely more directory entries to check or larger logical block sizes. Clarence Wilkerson
rlb@skyler.mavd.honeywell.com (Randal L. Barnes) (01/24/91)
In article <D4J8V2w163w@cynic.wimsey.bc.ca>, pevans@cynic.wimsey.bc.ca (Phillip Evans) writes: > Where did/does the 8 meg limit on logical disk size in CP/M come from? It has to do with the fact that CP/M uses a 16 bit logical block pointer, Since CP/M also uses 128 byte logical sectors (sector blocking/deblocking routines change the physical sector size on the disk but the logical sector size remains at 128), the maximum logical disk size is: 65536 * 128 = 8192K My Molecular Series 9 system has drive B set up as an 8 meg drive. Doing a STAT B:DSK: yields the following: B: Drive Characteristics 65536: 128 Byte Record Capacity 8192: Kilobyte Drive Capacity 1024: 32 Byte Directory Entries 0: Checked Directory Entries 256: Records/ Extent 32: Records/ Block 256: Sectors/ Track 0: Reserved Tracks -- Randal L. Barnes rlb@skyler.mavd.honeywell.com Tea, Earl Grey, HOT! Honeywell Military Avionics Division 612/542-5021
wilker@descartes.math.purdue.edu (Clarence Wilkerson) (01/24/91)
It seems to me like RLB's reply is a good explanation of why the maximum size of file would be 8 megs, if we can access sectors of the file by a 16 bit integer. Why does it dictate the max. logical size of the disk, however? Clarence
syswtr@iowasp.physics.uiowa.edu (01/25/91)
> pevans@cynic.wimsey.bc.ca (Phillip Evans) writes: >> Where did/does the 8 meg limit on logical disk size in CP/M come from? NOte that this restriction is removed in CPM3 and MPM2. The file size restriction is limited by the 16 bit record pointer in MPM2 (I haven't combed the CPM3 manuals that closely), but the logical device size is limited by the cluster size. Device size is expressed in terms of the number of clusters, so the maximum device is 32767*16384 (or 65535*16384) as the largest allocation cluster is 16K bytes. When you get these large devices, however, MPM2 signifigantly slows down when doing any directory operations as it scans the (potentially) large directory. A little work with caching the directory can make a world of difference... Willy
fzsitvay@techbook.com (Frank Zsitvay) (01/25/91)
In article <4314@mentor.cc.purdue.edu> wilker@gauss.math.purdue.edu.UUCP (Clarence Wilkerson) writes: >I believe that the 8meg limit is set in the coding of >the BDOS, and inforced in the BIOS in the sense that >certain scratch ram in the BIOS data area has a preset >size. > I think several replacement BDOS clones, e.g. >DOSPLUS, offer larger maximum logical disk size. > There would be some perfomance penalties for >using larger sizes, namely more directory entries >to check or larger logical block sizes. >Clarence Wilkerson well, not really. it is limited by the directory structure that is at the heart of all cp/m disk operations. cp/m only allows for a maximum of 65536 128 byte blocks. that is the maximum size of a file, or a whole disk. there is a built in mechanism in cp/m to allow the creation of partitions on a single physical disk. dosplus is essentially a different operating system than cp/m, and it allows a larger total disk space through a somewhat revamped directory structure. zcpr3 and zsystem do it through other techniques. -- fzsitvay@techbook.COM - but don't quote me on that.... American Oil Company motto - Bend over, We'll pump!!!
fzsitvay@techbook.com (Frank Zsitvay) (01/26/91)
In article <4402@mentor.cc.purdue.edu> wilker@descartes.math.purdue.edu (Clarence Wilkerson) writes: >It seems to me like RLB's reply is a good explanation of why >the maximum size of file would be 8 megs, if we can access >sectors of the file by a 16 bit integer. Why does it dictate >the max. logical size of the disk, however? >Clarence well, it's like this - that same 16 bit integer that is used to address sectors in a file is also used to address blocks in a disk. therefore, the maximum available blocks on a disk is 65536, or 8192K bytes total. to access larger disks you would need a larger block pointer, something on the order of 24 bits or 32 bits. however, there is too much overhead associated with maintaining two pointers, and large ones at that, in an 8 bit system. it is just easier to accept the 16 bit pointers and work with them. -- fzsitvay@techbook.COM - but don't quote me on that.... American Oil Company motto - Bend over, We'll pump!!!
wilker@gauss.math.purdue.edu (Clarence Wilkerson) (01/29/91)
But the size of a disk block can be large, i.e. 1k, 4k even 16k. Even 1k x 65535 = 64 megs. Of course it's hard to imagine doing a DIR to a disk with 64k 1k files....... Clarence
john@attcan.UUCP (John Benfield) (02/14/91)
In article <1991Jan25.012406.13752@techbook.com>, fzsitvay@techbook.com (Frank Zsitvay) writes: > In article <4314@mentor.cc.purdue.edu> wilker@gauss.math.purdue.edu.UUCP (Clarence Wilkerson) writes: > >I believe that the 8meg limit is set in the coding of > >the BDOS, and inforced in the BIOS in the sense that > > well, not really. it is limited by the directory structure that is at > the heart of all cp/m disk operations. cp/m only allows for a maximum of > 65536 128 byte blocks. that is the maximum size of a file, or a whole > disk. You can actually do some major diddling with the BDOS if you want and modify the directory block size. It's a major hack, but I did it on my Cromemco system to allow me to make use of a kludged hard disk I hotwired in. It got a bit flaky with anything that was explicity accessing directory records but if you're running a 'single application' box and you need to up the file size, it's a good bet. ______Opinions stated are my own. Transcripts available by request______ === =--==== AT&T Canada Inc. John Benfield =----==== 3650 Victoria Park Ave. Network Support Analyst (MIS) =----==== Suite 800 ==--===== Willowdale, Ontario attmail : ~jbenfield ======= M2H-3P7 email : uunet!attcan!john === (416) 756-5221 Compu$erve: 72137,722 ____Eagles may soar, but weasels don't get sucked into jet engines._____