FREEZER: mount seems not to work correctly with new dos_attach

[lydon42]

The new dos_attach MEGA65/mega65-core#628 shows that freezer does something wrong. If trying to mount an image to drive 1 after startup using freezer, a dir u9 following that will hang. The drive registers show that no image was mounted.

Inspection shows that the driveselect code does both use the hyppo call as well as wild poking to the registers. The poking is not needed, this is all done by dos_attach and the unfreeze. So get rid of the useless poking, and all should be fine?

Edit: the main goal is to check what exactly freezer does and how it interacts with HYPPO. It should do everything using HYPPO DOS Traps, so all internal states are updated correctly (just poking registers will not change the currenttask entries for example).

Todo:

• Check diskchooser
• Check makedisk

[lydon42]

diskchooser does not seem to do wild freeze_poke'ing, so that seem ok...

But makedisk does. makedisk should use dos_attach instead of poking into frozen memory.

[Schwefelholz]

Tried with development.20241022.13.76fac1b.

Mounting an existing disk image directly after startup using freezer and issuing dir u9 does not hang with this core.

Creating a new disk image in freezer, after pressing (1) EXTERNAL 1565 is successfully creating a new image, but instead of attaching it to drive 1 it is connected to drive 0. I would not have expected it that way...

[Pimau]

Same here, with mega65r6-20241022.13-develo-76fac1b.cor.

• DIR U9 working directly after boot, Freezer, mount existing image
• A newly created image after boot, Freezer, create is NOT attached to unit 9, but "overwrites" an already mounted on unit 8

[eilers]

Same here, with mega65r6-20241022.13-develo-76fac1b.cor.

• DIR U9 working directly after boot, Freezer, mount existing image
• A newly created image after boot, Freezer, create is NOT attached to unit 9, but "overwrites" an already mounted on unit 8

I can confirm this as well.