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

Re: kernel

Roman Zippel wrote:

Why does something under Linux need to know about the simulated PTBL? That's
what I am missing. The host partitions (or partition image files) will appear
as /dev/sd[a-g]1 under Linux-m68k automagically. Linux-m68k recognizes the
simulated PTBL so everything will work out-of-the-box, transparently,

Damn, I somehow missed the connection to the Atari partition table support...

?? What Atari partition table support? The one in 2.6.26 Laurent mentioned yesterday? Or the one that is in the kernel for 15 years?

Please tell me just a single clear advantage I would have that this _adds_
to Linux?
as said above, the disk space is a partition for both the host and the system
running inside of ARAnyM. This is 1:1 mapping, it just makes sense and allows
formatting and mounting the disk space from both host and guest equally.
Adding an extra layer of partitions inside of ARAnyM would break the
possibility of easy mounting of the disk image from the host.

No, it wouldn't. You can easily mount filesystem images as complete block devices under Linux/Aranym

Nope. If you partition the disk (image) on Atari then you can't mount it easily on the host (unless you patch host operating system to recognize Atari PTBL). And vice-versa - if you partition a disk on the host then linux-m68k might have a problem mounting such partitions (until you patch it to recognize that host PTBL). That is a fact. Or are you going to argue about it?

there is no need for a synthetic partition table.

the synthetic PTBL hides the differences between various PTBL schemas and so is great for making the data transfer much easier.

Forcing it OTOH means it prevents me from creating my own partitions.

you can create your own partitions - just do it on the host side!

If you insist on playing with fdisk on linux-m68k then simply use the IDE disk drives but I believe that most people in most cases simply don't need the fdisk and will gladly skip it and mount the partitions directly.

So to answer your question, the simulated MBR and PTBL clearly adds the
possibility of simple and easy mounting of the partition from host which is
ideal for copying files between host and guest, for example.

My main point is still valid, it adds no value to Linux, that possibility was there already before.

My view is completely opposite - it adds the value of an easy sharing of data without headaches with PTBL formats. And last but not least - this is not Linux specific feature, it helps on all ARAnyM host platforms and most ARAnyM operating system so please don't insist on a specific value for Linux when it was designed for TOS without free disk driver/disk utilities.

I admit that ARAnyM in combination linux-host / linux-m68k guest is the most flexible one so when

 1) both kernels are compiled with all available PTBL formats, and ..
 2) their version is at least 2.6.26, and ..
 3) you know the special loop option for accessing partitions

then you can get almost the same flexibility as with the direct single partition access... Though if this is your main point then I'd say it's rather weak since its based on assumptions that are not true yet.


Reply to: