FDC discussion

Post by **SarahWalker** » Thu 12 Nov, 2015 9:32 pm

Discussion moved from viewtopic.php?f=2&t=344.

Attached first patch for separating out FDD emulation, and providing support for motor speed control. Not been tested thoroughly, but seems to work for all the main cases using .IMG format (.FDI still needs updating). Battler, you should be able to implement the 3-mode / drive control line stuff on top of this, let me know if there are any issues.

Post by **SarahWalker** » Tue 17 Nov, 2015 9:42 pm

Added function to swap drives.

Battler, how are you getting on with this?

Battler · Post by **Battler** » Wed 18 Nov, 2015 5:32 pm

Downloaded the latest mainline PCem source code now, will now get the patch too and apply it, then start the work.

Battler · Post by **Battler** » Wed 18 Nov, 2015 10:30 pm

I can't apply the v2 patch cleanly to the latest PCem code, possibly because of FDC fixes that were made in the mean time.

Post by **SarahWalker** » Thu 19 Nov, 2015 6:33 pm

See attached.

Battler · Post by **Battler** » Sun 22 Nov, 2015 2:21 pm

I have a slight problem with data rate 1. On normal drives, this means 300 kbps, but on type 1 3-mode 3.5" drives, it means 500 kbps @ 360 rpm (while data rate 0 becomes specifically 500 kbps @ 300 rpm). However currently, PCem gives me no way to specify bit rate or bitcell period per drive, but only globally, which makes it a bit awkward since the two drives can be of different types.

Edit: And I also made the drives aware of the data rate since real drives would be, as the information is passed on the DRATE pins.

Battler · Post by **Battler** » Wed 25 Nov, 2015 4:20 pm

Just wanted to say I'm still waiting for some input on this, I can't continue until then.

Post by **SarahWalker** » Wed 25 Nov, 2015 6:45 pm

As I understand it, the implementation should probably look like this :

- Provide the DRATE pins to fdd.c. This should be used for calculating fdd_getrpm() for 3-mode drives.
- In fdc.c, select the bit rate based on the drive type (which I believe is programmed into the SuperIO control registers? At least on the FDC37c669 datasheet I'm looking at). You will probably need to modify fdc_update_rate() to take the current drive number, and call it when the drive select changes.

Post by **SarahWalker** » Sun 29 Nov, 2015 1:01 pm

Added generic sector-based image handling, this should make it easier to support formats with non-standard layouts.

Battler · Post by **Battler** » Sun 06 Dec, 2015 2:59 pm

The patch is attacted. It includes all my floppy work, but I also replaced the emulation of the FDC37C665 Super I/O chip with mine as I tried to add the floppy-related things into it, and I quite didn't like the idea of every single register being processed regardless of which you write to. The floppy parts have both 3-Mode support (the RPM, DENSEL, etc. stuff) and XDF images working.

Battler · Post by **Battler** » Sun 06 Dec, 2015 5:38 pm

New patch attached, this time made with hg diff.

Post by **SarahWalker** » Tue 08 Dec, 2015 9:07 pm

Can you also attach fdd.* and disc_sector.*? They are missing from the second patch.

Battler · Post by **Battler** » Tue 08 Dec, 2015 10:20 pm

Yes, they are attached.

Battler · Post by **Battler** » Tue 08 Dec, 2015 10:21 pm

Another post for fdd.h because I can't attach more than 3 files in a single post.

Battler · Post by **Battler** » Wed 09 Dec, 2015 10:53 pm

A .RAR file with patch v2 is attached. Archive includes patch file, as well as disc_sector.* and fdd.*. This patch removes the fdd_rate_compare stuff that was unneeded, adds the floppy density "hole", reduces the two 3-Mode types to one, introduces the usage of the type variable of fdd rather than drive_types[] in disc.c, makes changing drive type in settings require a hard reset, and makes the READ and WRITE commands check for the N (sector size code) part of sector ID too, and fail if it mismatches. The FDD-FDC rate compare that was unneeded, was replaced with a "drive can read this medium" based on the density "hole" and drive type flag.

Post by **SarahWalker** » Sat 12 Dec, 2015 4:43 pm

Attached v3 patch. This patch has the following changes :

- Revert to the pre-existing fdc37c665 code.
- Formatting fixes
- Fixes to 'read track' command
- Added gap between sectors to disc_sector.c
- Bit of generic cleanup

Post by **SarahWalker** » Sat 12 Dec, 2015 8:40 pm

Committed in rev 413.

te_lanus · Post by **te_lanus** » Fri 25 Dec, 2015 4:32 pm

Wonder if PCem will one day support some of the other Disk Image formats that are available today. IE

FDI: Formatted Disk Image
td0: Teledisk
imd: ImageDisk

Post by **SarahWalker** » Sat 26 Dec, 2015 9:33 am

FDI is supported in v10...

Battler · Post by **Battler** » Sun 27 Dec, 2015 9:24 pm

I can try to write some code for reading TD0. :p

Battler · Post by **Battler** » Mon 04 Jan, 2016 9:43 pm

Going to upload a floppy fix patch soon. Among other things, it's going to fix a pretty bad bug I accidentally left in, namely the sector size for non-XDF floppies is obtained from the wrong IMG array entry.

FDC discussion

FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion

Re: FDC discussion