How and when is a change of disk detected?

[Ruud]

I ran into this site that enables one to connect a PC drive to a Commodore 64. The original design uses the 44-pin version of the GM82C765B but I happen to have the 40-pin one. Creating a new schematic for the 40-pin version I found out that this IC lacks the pins DCHG and DCHGE. These are needed to detect a disk change. But having a look at the original schematics of the IBM FDC, I noticed that pin 34 of the FDD connector that signals a disk change, is not connected to anything on the card. So I started wondering, how is a disk change detected? I think I can answer this question myself because I think it has to do with the disk ID. But that leads to the second question: when does the OS check that? I cannot imagine that the OS checks the ID at every access. Or does it?

Many thanks in advance!

 

[modem7]

See the earlier thread at [here].

In that, the other thread referred to is at [here].

 

[Ruud]

I found those two already before my original post and did read them. But nothing about what actually happens. I even read the article about the 1.2 MB drive mentioning when omitting the /C parameter with DRIVPARM to force the PC to read the disk with every access. But then: what does a 5160 with a 360 KB drive that only loaded COMMAND.COM and nothing else? Or is that the default value: "read the disk every time"?

Also quite disturbing: I disassembled the software for this card but couldn't find anything checking bit 7 of the 2nd status register. I also noticed that the drive is read quite often. But the disassembly is not complete and I cannot pinpoint this to "checking for a disk change". But I start to I wonder if the software was written for the original 765 (or equivalent, thus before choosing this 44-pin version.

Thank you if you can enlighten me!

 

[krebizfan]

MS-DOS brute forced the issue. If the drive had been accessed in the last 2 seconds, the disk had not changed. Otherwise, reload the disk as if the disk was changed. It was possibly better than the CP/M method of requiring the user to press CTRL-C after the disk change.
One version of the story can be found at https://devblogs.microsoft.com/oldnewthing/20190924-00/?p=102915

 

[Ruud]

OK, that is a clear answer. That was probably meant in the first thread Modem7 mentioned but that was not clear to me until you explained this. It simply means there is no ingenious hard- or software trick but "just read the disk, stupid!".

Thank you!