Re: New file system IDs

[Julian Reschke-FJR010 <Julian_Reschke-FJR010@email.mot.com>]

felsch@tu-harburg.de%INTERNET wrote:
> 
> Hi!
> 
> While working on a new port of GNU file-utils I've add the following
> patch to
> MiNT lib PL48.
> By default statfs() set the statfs.f_type to 0 and only Minixfs support
> a
> special Dcnt() which generate a f_type != 0. With this patch the new
> MiNT FATFS and the modified ramfs.xfs also returning there real f_type.
> The second thing I've change is the struct 'mfsinfo'. MinixFS do not
> returned
> the current blocksize, because it use always 1024. For new MiNT/ramfs I
> need
> a posibility to return the current blocksize via the Dcntl(0x104). Now
> I've
> use the first of the reserved longs in the mfsinfo structure. This would
> be
> backward-compatible because MinixFS doen't touch the reserved compones
> of the
> structure.
> 
> The new port of file-utils 3.16 use this patch and produce the following
> output on my system:
> 
>         Drive    Type   1024-blocks  Used Available Capacity Mounted on
>         a:       minix       1421       1     1420      0%   u:/
>         c:       tos        30650   25274     5376     82%   u:/
>         d:       minix2     30071   18028    12043     60%   u:/
>         e:       fat16     102328   44000    58328     43%   u:/
>         f:       minix2    117392   94415    22977     80%   u:/
>         g:       vfat16    119780     512   119268      0%   u:/
>         h:       ramfs       8630       0     8630      0%   u:/
> 
> I thing it's nice to see, which kind of file system each partition is
> used!
> Any comments? If not, Yves, please add this patch to the lib.

Yes, indeed.

MagiC also has a Dcntl which returns an 8-char filesystem *name.

Generally, we should define a more powerful mechanism, which does
not rely on an enumeration value.

I will try to make a proposal in the next few days.


Regards, jr