[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: shared disk (DOS)

From: Peter Maydell
Subject: Re: shared disk (DOS)
Date: Mon, 12 Apr 2021 11:17:08 +0100

On Mon, 12 Apr 2021 at 10:57, Tomas By <tomas@basun.net> wrote:
> On Mon, 12 Apr 2021 11:16:21 +0200, Peter Maydell wrote:
> > No, that doesn't work. QEMU doesn't "see" filesystem formats at all.
> > To QEMU, a disk is simply a big bag of bytes. It can have anything
> > on it that the guest wants, and it's up to the guest how it interprets
> > it: QEMU just provides the functionality for "read some bytes"
> > and "write some bytes". Whether the disk has a FAT filesystem on it
> > or something else is up to the guest. (This is exactly the same way
> > that a real hard disk can have anything on it, and the disk itself
> > has no knowledge about the format.)
> Well... I have problems getting all this to fit together logically.
> Why then is there a difference (according to page linked earlier)
> between raw and ocr2/gfs?

According to which part of which page? There have been several links
mentioned in this thread...

> The point is not how qemu sees the bits, but how it handles concurrent
> access from multiple instances.
> Raw, as I understand it, just makes the first instance have exclusive
> access and treats the whole disk as one unit. I was hoping there would
> be some other system with more discriminiation.
> I don't see why this should not work. Qemu divides the disk into a
> bunch of pieces, no matter how, and the instances each have (more or
> less) random access.

In theory if the different guests all access entirely disjoint
sections of the file this could be made to work. But in that
case, why not just give them actually separate image files? That
would be 100 times easier and achieve exactly the same effect.

-- PMM

reply via email to

[Prev in Thread] Current Thread [Next in Thread]