Re: Partitioning context

"Axel D=F6rfler"  <[email protected]> wrote:
> 
> "Ingo Weinhold" <bonefish-CFLBMwTPW48UNGrzBIF7/[email protected]> wrote:
> > "Axel D=3D3DF6rfler"  <[email protected]> wrote:
> > > - we should probably add an option to recognize files as a hard 
> > > disk 
> > > and be able to scan the partitions contained therein
> > I already thought about that some months ago, when we were 
> > designing 
> > the previous API. The issue I see is, where the devices are 
> > published. 
> > I would imagine it to work like this: There are syscalls to 
> > (un)register a file as virtual disk. On registration a device is 
> > published under /dev/disk/virtual/file/<filename>[uniqueness=3D
> > 3D5FID]/
> > raw. 
> > The rest can work completely transparent for the disk=3D3D5Fdevice=3D3D
> > 5Fmanager.
> 
> Hm, that might be the way to go, and a similar way could probably 
> also 
> be used for TTYs. So what's needed would be the ability to trigger 
> the 
> creation (and deletion) of /dev entries and the ability of attaching 
> data to those entries (which file it refers to, for example). AFAICT 
> that matches Thomas' proposed changes to the device management pretty 
> well :-)

Cool! :-))

> > So, probably an additional BPartition::IsLeaf() (better name=3D3D3F) 
> > would be 
> > a good idea. It would indicate whether the partition is of a type 
> > that 
> > doesn't allow to create child partitions using the parent's system 
> > in 
> > a 
> > non-nesting manner.
> 
> So that intel partitions in non-extended intel partitions would be 
> invalid=3D3F

No. IsLeaf() would only indicate, that if you want to have child 
partitions for this partition you'd need to use nesting. Perfectly 
legal (though quite likely a problem for other OSs).

> Do you think that this is a necessary restriction=3D3F As long as 
> we don't give the user the possibility to create those partitions 
> (too 
> easily), I don't see much of a problem.

Me neither. :-)

> The reader should support anything you throw at it, while the writer 
> should only output conformant layouts.

Yep.

> > BTW, DriveSetup users should be warned, when they are going to use 
> > nesting.
> 
> Might be a good idea, although a visual clue might be enough (i.e. 
> partition type: primary/logical (for intel style partioning), along 
> with a graphical representation (generic)).

In any case it doesn't harm to have a method to find out where the 
boundaries of a partitioning system are.

> The only problem with nesting partitions is, AFAICT missing support 
> in 
> other operating systems.

I guess so.

> > > both would be very convenient mostly for users of emulators (such 
> > > as 
> > > Bochs or Amithlon, the former accesses files as hard drivers, the 
> > > latter has the standard amiga partitioning system inside an Intel 
> > > partition), but I could imagine other useful applications as 
> > > well.
> > > What do you think=3D3D3D3F
> > Definitely a good idea.
> 
> Yep, it always annyoing not to be able to access those things 
> currently 
> :-)

BTW, if you really want to access such a thing, you can adjust the 
virtualdrive driver (src/tests/kits/storage/virtualdrive). The should 
be very easy to do.

CU, Ingo