[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: SCSI I/O errors galore, PART II



Michael Schmitz wrote:


> > my voice, because my experience with this controller is that it
> > doesn't work too dang well.  With some drives it just plain can't
> > handle the higher speed.  But with some drives it can.  So it's hit
> 
> Seems in this case it can't. My external disk only runs at 5 MB (of
> course).
> 
> How to set MESH to 5 MB/s? I'd not object to providing a patch that makes
> it a driver option, if someone can point me to the correct combinattion of
> period and offset to limit the rate.

Eternal is different.  Since Apple wisely decided on a 25-pin scsi
cable scheme for external disks, they can only get up to 5MB/s
there.  [Sarcasm]  So they used a separate controller for that.  The
NCR 53Cwhatever.  Anyway, that controller is what drives the second
"slow" connector on the mobo's that have it, as well as the external
25pin connector.  Why Apple did it this way is beyond me and seems
so collosally stupid that there is just no way to explain it. 
Perhaps real 50 pin cables were deemed too expensive or something. 
A former author of the linux driver for the aic7xxx series of scsi
controllers once tried to explain exactly how NCR and apple managed
to hack the scsi standard to get a 25 pin cable to work with drives
that have a 50pin connector, but he got so upset about the whole
atrocity that I decided I should drop the subject.  (I have an old
scanner made for oldworld Apple's that has a 25pin connector and was
trying to get it to work with an aic7xxx controller on an x86
machine, which I eventually did.)  I still don't know how it's done,
but you can bet that it's sick and very wrong.

Suffice it to say that if your disk can't do over 5MB/s then you
have nothing to do or worry about.

a



Reply to: