tmaster@ucscb.UCSC.EDU (The Master) (03/09/91)
I'm curious as to the safety (contention wise) of having a system with 2 apple scsi controllers (Rev C cards) with hard drives. This system would be on apple ///'s, in which I have my own drivers. The drivers make smartport calls. Do the cards check and handle contention if the scsi bus is already in use? If not, does anyone have some documented code which will handle this (Since it is using device drivers, I can put this code in..) My hope is to have 2 apple ///'s and 2 scsi drives all hooked on the same bus (I can make one scsi drive read only to one computer and the other read only to the other computer through the device drivers to control problems with overwrites.) With this setup, either computer can therefor access all the information on the two drives (therefor making it easier to copy from one system to another). Also, what about DMA scsi cards? Are they different for handling the above situation? The Master tmaster@ucscb.ucsc.edu
toddpw@nntp-server.caltech.edu (Todd P. Whitesel) (03/10/91)
tmaster@ucscb.UCSC.EDU (The Master) writes: > I'm curious as to the safety (contention wise) of having a system with >2 apple scsi controllers (Rev C cards) with hard drives. > This system would be on apple ///'s, in which I have my own drivers. >The drivers make smartport calls. Do the cards check and handle contention >if the scsi bus is already in use? If not, does anyone have some >documented code which will handle this (Since it is using device >drivers, I can put this code in..) I don't know about the Rev. C card, but I do know for a fact that the DMA SCSI firmware is 100% ANSI compliant (and then some, because certain Apple SCSI devices aren't 100% ANSI compliant) -- this includes full bus arbitration, something the Mac still doesn't do, regardless of what Incider's hard drive articles claim. There's one caveat about the DMA SCSI, though: it has built in termination that can't be removed unless you want to snip off three resistor packs and void your nifty 1 year warranty. The Rev. C didn't have any termination of its own. Todd Whitesel toddpw @ tybalt.caltech.edu
gwyn@smoke.brl.mil (Doug Gwyn) (03/11/91)
In article <1991Mar9.161552.22392@nntp-server.caltech.edu> toddpw@nntp-server.caltech.edu (Todd P. Whitesel) writes: >There's one caveat about the DMA SCSI, though: it has built in termination that >can't be removed unless you want to snip off three resistor packs and void your >nifty 1 year warranty. The Rev. C didn't have any termination of its own. Why would one WANT to remove the bus terminating resistors? Lack of proper termination was one of the main problems with the previous (non-DMA) Apple SCSI card.
toddpw@nntp-server.caltech.edu (Todd P. Whitesel) (03/11/91)
gwyn@smoke.brl.mil (Doug Gwyn) writes: >Why would one WANT to remove the bus terminating resistors? So you can get more than 2 of them on a bus together. A properly terminated SCSI bus has terminators on each end, and nowhere in between. With 2 DMA SCSI's that means you have 1 computer on each end of the bus and all the devices in between them with internal termination removed and no external terminators. Try to add more DMA SCSI's and you will run into problems because of the extra termination. >Lack of proper termination was one of the main problems with >the previous (non-DMA) Apple SCSI card. External terminators fixed that. Apple should have included one free with the Rev. C SCSI, but it doesn't really matter now. Todd Whitesel toddpw @ tybalt.caltech.edu
gwyn@smoke.brl.mil (Doug Gwyn) (03/12/91)
In article <1991Mar11.045531.3866@nntp-server.caltech.edu> toddpw@nntp-server.caltech.edu (Todd P. Whitesel) writes: >>Why would one WANT to remove the bus terminating resistors? >So you can get more than 2 of them on a bus together. A properly terminated >SCSI bus has terminators on each end, and nowhere in between. With 2 DMA SCSI's >that means you have 1 computer on each end of the bus and all the devices in >between them with internal termination removed and no external terminators. >Try to add more DMA SCSI's and you will run into problems because of the extra >termination. >>Lack of proper termination was one of the main problems with >>the previous (non-DMA) Apple SCSI card. >External terminators fixed that. Apple should have included one free with the >Rev. C SCSI, but it doesn't really matter now. No, in fact proper termination does require terminators precisely at the two ends of the SCSI bus and nowhere in between. Thus, the terminating resistors DO belong (permanently) on the Apple SCSI Card. Putting them one "system cable" length from the card for the old (non-DMA) SCSI card was electrically inferior to having them on the card. The scenario of more than two SCSI computer interfaces (bus controllers) strikes me as unrealistic. I have heard of only a few installations to have even two controllers, and Apple's operating systems do not properly interlock their shared access to device contents (for example, local caching is not coordinated among the systems). It is terminators at each peripheral that may need to be installed or removed; Apple's use of external terminators makes that convenient but expensive. Some people leave the internal termination installed in the last drive on the SCSI bus (farthest from the interface card), to avoid having to purchase an external terminator. If you seldom reconfigure the SCSI chain, this is adequate. For frequent reconfiguration, external termination is the only way to go. In all realistic cases, the SCSI interface card should retain its built-in termination.