jim@b11.UUCP (Jim Levie ) (01/30/88)
I suppose that this is primarily directed at the CATS folks, but somebody else may also know the answers. First of all where does the limit of 17 blocks (sectors) per track come from? Is it a hardware consideration in the ST506 circuitry or a software limit in the driver? I tried an Atasi 3046 with the 2090, which works at 17 sectors per track, but not above. The Atasi has over 10K bytes per track which should allow 20 sectors. Since one has to to create a mount list entry for RESn: before running prep why does prep need the number of heads, sectors, low and high cylinder, etc? Seems to me that the only question prep should really need to ask concerns precomp cylinder and parking. Also the mountlist entries allow the specification of memory type, but prep doesn't. I would assume that prep sets the buffer memory type to public, is this correct? Why will format guru the system if one accidentally asks to format more cylinders than the disk has? Seems to me that, at least for ST506 devices, the hardware should detect an attempt to seek past the last cylinder and the driver should gracefully handle it. I'm not sure if this has already been answered or not. When an ST506 disk is formatted by AmigaDOS do bad blocks get mapped out, and if so how? The possibilities that occured to me are that format could just chain out the bad blocks or there could be a replacement table. While the disk was formatting it seems occasionally to retry, I assume that this was bad block detection. Prep asks about manual bad block entries, but I don't know how one would have the information without some type of 'bad block' utility. A little experimentation with an embedded SCSI disk drive (Rhodime 652) leads me to belive that it is not possible to use SCSI adapters with the current hddisk driver. A SCSI adapter will need the drive geometry to be set up before accessing the drive and I don't think that the current version of prep or hddisk.device are capable of doing this. Will this restriction be removed when the Fast File System is released, or have I missed something? The 2090 manual is fine for installing and setting up the controller and drive, however there isn't any information about the Amiga Interface to the controller in the manual (csr's etc). Any chance that this info is going to be released? -- =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ Jim Levie REMTECH Inc Huntsville, Al The opinions expressed above are just that. Ph. (205) 536-8581 email: uunet!ingr!b11!jim
grr@cbmvax.UUCP (George Robbins) (01/31/88)
In article <229@b11.UUCP> jim@b11.UUCP (Jim Levie ) writes: > I suppose that this is primarily directed at the CATS folks, but somebody > else may also know the answers. > > First of all where does the limit of 17 blocks (sectors) per track come from? > Is it a hardware consideration in the ST506 circuitry or a software limit in > the driver? I tried an Atasi 3046 with the 2090, which works at 17 sectors > per track, but not above. The Atasi has over 10K bytes per track which should > allow 20 sectors. The 17 sector "limit" just reflects the number of sectors you get with the standard ST506 clock rate, rotation speed and inter-block gaps. To go beyond these parameter requires some kind of special controller. For example "RLL" controller/drives change to data encoding format. Adaptec has sold controllers that vary the clock rate to put different numbers of sectors on differend radial zones of the disk. Typically the drives with "SCSI" interfaces use one of these or other techniques to increase capacity, however the details are hidden between the drive and SCSI controller, allowing the system to see a single standardized interface. > The 2090 manual is fine for installing and setting up the controller and > drive, however there isn't any information about the Amiga Interface to the > controller in the manual (csr's etc). Any chance that this info is going to > be released? There's nothing secret about it. I guess nobody's had the time or energy to get all the random pieces together and write an intelligible document. -- George Robbins - now working for, uucp: {uunet|ihnp4|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@uunet.uu.net Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
gunther@cbmvax.UUCP (Ed Gunther QA) (02/04/88)
In article <229@b11.UUCP> jim@b11.UUCP (Jim Levie ) writes: > Since one has to to create a mount list entry for RESn: before running prep > why does prep need the number of heads, sectors, low and high cylinder, etc? > Seems to me that the only question prep should really need to ask concerns > precomp cylinder and parking. Also the mountlist entries allow the > specification of memory type, but prep doesn't. I would assume that prep > sets the buffer memory type to public, is this correct? > First, do realize that all the information in the mountlist is not necessarily supported. Also, what version of prep do you have? As far as I know the latest version is 33.19. It does not ask for detailed drive information if you are using a supported drive, only if you specify drive type 0, namely user defined. Yes, the buffer memory type specification has no effect on where the buffers actually get located. The hddisk.device uses whatever is next. You can get the buffers in fast ram by running slowmemlast before binddrivers. > Why will format guru the system if one accidentally asks to format more > cylinders than the disk has? Seems to me that, at least for ST506 devices, > the hardware should detect an attempt to seek past the last cylinder and > the driver should gracefully handle it. > I have done this, strictly by accident of course, but I do not recall guruing the system. What does happen is the format locks up the CLI in which is it running, which I'll agree is not great either. > I'm not sure if this has already been answered or not. When an ST506 disk > is formatted by AmigaDOS do bad blocks get mapped out, and if so how? The > possibilities that occured to me are that format could just chain out the bad > blocks or there could be a replacement table. While the disk was formatting > it seems occasionally to retry, I assume that this was bad block detection. > Prep asks about manual bad block entries, but I don't know how one would > have the information without some type of 'bad block' utility. > The AmigaDOS format exercises each block something like five times during the format (so like this is why it takes so long compared to the formatting of as similar drive on a PC) and maps out any bad blocks and assigns replacements up to 126 bad blocks. The occasional retries of which you speak are undoubtably bad block retries as the system makes sure a block really is bad. Yes, an entire drive bad block test utility might be useful and forthcoming one of these days, but every drive is tested by the manufacturer for defects and has a list of defect information on the drive. This information, namely the cylinder, head, and offset of these known defects is precisely the information which prep asks for when entering bad block data. A "good" bad block utility would allow one to specify the physical setup of a drive, i.e. number of head, cylinders, ect., and allow one to determine the block number of a defect given the cylinder, head, and offset as well as the inverse, the cylinder, head, and offset from the block number. I have written some lousy routines that do this for a four head drive which may be real easy to fixup. Dare I ask if anyone is interested in these? Let me know. I'd rather send them by mail so I don't have to be too embarrassed. > A little experimentation with an embedded SCSI disk drive (Rhodime 652) leads > me to belive that it is not possible to use SCSI adapters with the current > hddisk driver. I'm not sure I'm following this one. Is this just a scsi drive or a scsi to ST-506 adapter hooked up to an ST-506 drive or what. Send me some mail on this, okay? > > The 2090 manual is fine for installing and setting up the controller and > drive, however there isn't any information about the Amiga Interface to the > controller in the manual (csr's etc). Any chance that this info is going to > be released? > Find out from CATS about this. There is information about the hardware in the C-A A500/A2000 Technical Reference Manual. As far as the software interface goes, there must be something developers are getting about this, but I haven't seen it. > > =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ > Jim Levie REMTECH Inc Huntsville, Al > The opinions expressed above are just that. > Ph. (205) 536-8581 email: uunet!ingr!b11!jim -- (((((((((()((((((((())(((((((()))((((((())))(((((()))))((((())))))(((())))))) __ Ed Gunther COMMODORE AMIGA TEST ENGINEERING / /\ "The LAST line of defense, if "THEY" let us!" < == /__\ UUCP ...{allegra,ihnp4,rutgers}!cbmvax!gunther \__ / \ PHONE 215-431-9356 (life or death only please!) "Disclaimer, heck no! NOooo PROBLEM. You have myyyy word on it." (((((((((()((((((((())(((((((()))((((((())))(((((()))))((((())))))(((()))))))
page@swan.ulowell.edu (Bob Page) (02/04/88)
gunther@cbmvax.UUCP (Ed Gunther QA) wrote:
>all the information in the mountlist is not necessarily supported.
What the hell does this mean? Steve? Andy?
I HOPE you meant "the hddisk.device doesn't use all the possible fields."
..Bob
--
Bob Page, U of Lowell CS Dept. page@swan.ulowell.edu ulowell!page
"I don't know such stuff. I just do eyes." -- from 'Blade Runner'cs178abu@sdcc8.ucsd.EDU (John Schultz) (02/04/88)
I've just ordered a 2090 and am wondering what type of drive has the least problems: ST506 or SCSI? With the current revision of the 2090, what works best; what will work best with the (alledged) 2090A/Dos 1.3? How 'bout that fat agnes? Wize it there? John
gunther@cbmvax.UUCP (Ed Gunther QA) (02/05/88)
In article <784@sdcc8.ucsd.EDU> cs178abu@sdcc8.ucsd.edu.UUCP (John Schultz) writes: > > I've just ordered a 2090 and am wondering what type of drive has > the least problems: ST506 or SCSI? With the current revision of the > 2090, what works best; what will work best with the (alledged) > 2090A/Dos 1.3? We are an equal opportunity drive employer here at Commodore with the A2090, ST-506 or SCSI drives should work equally well. ST-506 drives tend to be fairly cheap because they are used extensively in PC's. However, SCSI drives can be faster, depending on the drive, than ST-506, but, these will tend to cost more. Certain drives, both ST-506 and SCSI, have been thoroughly tested for use with the A2090, these are indicated at the front of the A2090 user's guide. If you have a specific question about a certain manufacturer or drive, send me email. I will answer you (I prrromise). > > John -- (((((((((()((((((((())(((((((()))((((((())))(((((()))))((((())))))(((())))))) __ Ed Gunther COMMODORE AMIGA TEST ENGINEERING / /\ "The LAST line of defense, if "THEY" let us!" < == /__\ UUCP ...{allegra,ihnp4,rutgers}!cbmvax!gunther \__ / \ PHONE 215-431-9356 (life or death only please!) "Disclaimer, heck no! NOooo PROBLEM. You have myyyy word on it." (((((((((()((((((((())(((((((()))((((((())))(((((()))))((((())))))(((()))))))