rogerj@batcomputer.tn.cornell.edu (Roger Jagoda) (09/20/89)
Well people, it's here! I've got my 60 copies of 1.0. Guess what, it comes with a free 1.0 ROM chip AND a disposable (can you do this?) wrist grounding strap so you can upgrade the ROM yerself! Oh...goodie... Something else I see, a WARNING the ROM chip MUST be installed before the 1.0 disk can be inserted. Now does this mean that 0.9 disks are incompatible, that the ROM chip is just for the new passwd protection, or both? How absolute is this requirement? Just as a side note, I don't see SyBase SQL on the front lid of the OD as included. Did this not make it into the distribution. If this is indeed the case, MNAY people are going to be GREATLY disappointed as it was a selling feature to be included with the bundled software. I'll let you know more after I install the ROM...now, where's that grounding strap...:-) Roger Jagoda Cornell University FQOJ@CORNELLA.CIT.CORNELL.EDU (607) 255-8960
cs021042@brunix (Ronald Antony) (09/20/89)
What is this with the ROM? Is this the kind of ROM with a serial number for each machine? Or what sort of "new passwd protection" do you mean? Does a new ROM imply that the 0.8 or 0.9 OS can't be ran anymore after the ROM is installed? I'd like to do so, at least for some time to compare certain things (I guess it is interesting to compare the different versions side by side..) Ronald
eht@f.word.cs.cmu.edu (Eric Thayer) (09/20/89)
In article <8887@batcomputer.tn.cornell.edu> rogerj@tcgould.tn.cornell.edu (Roger Jagoda) writes: > >Just as a side note, I don't see SyBase SQL on the front lid of the >OD as included. Did this not make it into the distribution. If this >is indeed the case, MNAY people are going to be GREATLY disappointed >as it was a selling feature to be included with the bundled software. > Look harder, it's there. (/usr/sybase) -- Eric H. Thayer School of Computer Science, Carnegie Mellon (412) 268-7679 5000 Forbes Ave, Pittsburgh, PA 15213
feldman@umd5.umd.edu (Mark Feldman) (09/20/89)
In article <15676@brunix.UUCP> cs021042@cslab8e.UUCP (Ronald Antony) writes: >What is this with the ROM? Is this the kind of ROM with a serial number for >each machine? Or what sort of "new passwd protection" do you mean? >Does a new ROM imply that the 0.8 or 0.9 OS can't be ran anymore after the >ROM is installed? I'd like to do so, at least for some time to compare >certain things (I guess it is interesting to compare the different versions >side by side..) >Ronald The ROM contains the ROM monitor -- a piece of software that has been changed for release 1.0. The ROM monitor tests the system at power-up, starts the boot process, and provides for low level (very low level) debugging. Some of the changes to the ROM monitor are bug fixes and others are enhancements. The ROM monitor now supports a hardware password (stored in non-volatile memory with the other configuration information) which can be used to prevent random people (those without the password) from booting in single-user mode or booting from other than the configured boot device (OD, SCSI, net). (Thanks, NeXT!) The format of the OD disk label has changed between 0.9 and 1.0. The new disk label format, revision 3, can only be read using the new ROM. The new ROM can also read old 0.8 and 0.9 disks. The older ROM code, upon seeing a 1.0 disk, with what is unknown to it as the new disk label, will attempt to be helpful and "fix" the label, and thereby proceed to make its contents unreadable. Of course, you can still use the disk as a blank, but you've lost all of the nice, new, improved 1.0 software. As a side effect of changing the ROM, the Ethernet address of the cube changes, too. The cube's Ethernet address, which must be unique, is stored in the ROM and when you change ROMs, you change Ethernet addresses. The only times that Ethernet addresses matter are when you are doing network booting based on the Ethernet address and when a software package uses the Ethernet address in a copy protection scheme (yuk!). The new ROM monitor and the consequences of changing the Ethernet address of a cube that is using the net for configuration/booting are discussed in the on-line documentation. Now, if only the UPS truck would bring us our 1.0 disks... Mark
cs141043@brunix (Ronald Antony) (09/21/89)
Some questions to the following: >As a side effect of changing the ROM, the Ethernet address of the cube >changes, too. The cube's Ethernet address, which must be unique, is stored >in the ROM and when you change ROMs, you change Ethernet addresses. The >only times that Ethernet addresses matter are when you are doing network >booting based on the Ethernet address and when a software package uses the >Ethernet address in a copy protection scheme (yuk!). How did it work before? How could one network if the adresses were not different? How can one network different computers, if it is not sure that e.g. some SUN has not the same address? Or asked in a different way, is there a) some organisation verifiing that there is each address used only once, or b) some way of changing the address short of burning a new (ep)rom with a different address? What kind of address is this ethernet address? Something similar to the 32bit integers of the internet addresses? Maybe these are stupid questions, but well, I can't know everything :-) Ronald
labc-1aa@e260-4d.berkeley.edu (Bob Heiney) (09/21/89)
Ethernet addresses are not the same thing as InterNet addresses. They are regulated by someone (I don't know who), but each company has its own set of numbers assigned to it by the regulators. Therefore, there is no confusion about who is who. InterNet addresses are a hardware independent way of identifying machines, and are part of the DARPA InterNet Protocol. Two network protocols can run on the same Ethernet network (e.g. TCP/IP and DECnet). Basically, the InterNet or DECnet address resolves to an Ethernet address when it is determined that a packet is going to be sent from one machine to another using the Ethernet network technology. EtherNet address look like this: 08-00-2B-02-F0-36 I.e. 6 bytes. I believe the two most significant bytes identify the hardware vendor. Hope I didn't obsure the matter further, Bob labc-1aa@WEB.Berkeley.edu